Protist

Protist, Vol. 151, 161–169, August 2000 © Urban & Fischer Verlag http://www.urbanfischer.de/journals/protist

ORIGINAL PAPER

Paramecium GPI Proteins: Variability of Expression and Localization Yvonne Capdeville1 Centre de Génétique Moléculaire, Centre National de la Recherche Scientifique, associé à l’Université Pierre et Marie Curie, F-91198 Gif-sur-Yvette, France Submitted December 29, 1999; Accepted: April 17, 2000 Monitoring Editor: Michael Melkonian

In Paramecium primaurelia, the two major classes of cell surface proteins, the surface antigen (SAg) and the surface GPI proteins (SGPs), are linked to the plasma membrane through a glycosylphosphatidylinositol (GPI) anchor. In the present study, we have characterized the expression of the SGPs in several geographical strains of P. primaurelia and P. tetraurelia at different temperatures, 23 °C and 32 °C. The identification of the expressed SGPs was performed on purified cilia, by establishing the SGP SDS-PAGE profiles under four different conditions: with or without their anchoring lipid, cleaved with a Bacillus thuringiensis phosphatidylinositol-specific phospholipase C (PI-PLC), and either in a reduced or in an unreduced state. This screening revealed the existence of specific sets of ciliary SGPs, as a function of temperature and the geographical origin of the strains. The SGPs the most abundant at 23 °C and 32 °C displayed a rapid turnover. We also looked for the presence of PI-PLC releasable proteins in purified cortices. In addition to the SAg and SGPs, the cortical fraction was shown to contain other PI-PLC releasable proteins, not found in the ciliary fraction, thus localized exclusively in the interciliary region.

Introduction In a wide variety of eukaryotic cells, many membrane proteins are linked to the outer leaflet of the plasma membrane via a glycosylphosphatidylinositol (GPI) anchor, the core structure of which (ethanolamine - phosphate - 6Manα1 - 2Manα1 6Manα1-4GlucNH2α1 - 6myoinositol - phospholipid) has been highly conserved throughout evolution (Ferguson and Williams 1988; Low 1989). This mode of anchoring is used particularly in parasitic protozoa to attach not only proteins but also protein-free glycoconjugates to the plasma membrane (McConville and Ferguson 1993). In the free-living ciliated protozoon Paramecium primaurelia, all the major cell surface proteins have 1 fax 33-1- 69823150 e-mail [email protected]

been shown to be GPI-anchored. A GPI-anchor was first demonstrated for the surface antigen (SAg), also referred to as the immobilization antigen (Azzouz and Capdeville 1992; Capdeville et al. 1987). The SAg, a high molecular weight polypeptide coating the whole cell surface, belongs to a multigene family, the expression and antigenic variation of which is controlled by environmental factors (Beale 1954; Capdeville et al. 1993; Caron and Meyer Abbreviations: GPI: glycosylphosphatidylinositol; GPIPLC: glycosylphosphatidylinositol-specific phospholipase C; PI-PLC: phosphatidylinositol-specific phospholipase C; SAg: surface antigen; sSAg: soluble form of surface antigen; mSAg: membrane-bound form of surface antigen; SGP: surface GPI protein; mSGP: membrane-bound form of surface GPI protein; sSGP: soluble form of surface GPI protein 1434-4610/00/151/02-161 $ 12.00/0

162

Y. Capdeville

1989). More recently, by combining biosynthetic labeling with metabolites specific for the GPI anchor and a treatment of purified cilia with a phosphatidylinositol-specific phospholipase C (PI-PLC), we have identified another class of major GPI proteins consisting of a set of three GPI-anchored polypeptides with apparent molecular weights in the range of 30 35 kDa (Capdeville and Benwakrim 1996). These proteins, named surface GPI proteins (SGPs), actually correspond to the “40K” proteins, previously described in purified cilia of P. tetraurelia strain 51, and assumed to be intrinsic membrane proteins (Adoutte et al. 1980; Ramanathan et al. 1981). The GPI anchors of both SAg and SGPs are sensitive to bacterial PI-PLCs and to an endogenous glycosylphosphatidylinositol-specific phospholipase C (GPI-PLC) -like hydrolase and share common features (Capdeville et al. 1986). Their lipid moiety is composed of a ceramide, whose amide-linked fatty acid component varies as a function of the growth temperature (Benwakrim et al. 1998), while their core glycan is probably modified. Indeed, the final GPI precursor, identified in a P. primaurelia cell-free system by Azzouz et al. (1995), bears an additional phosphate mannose. As a preliminary study suggested that the expression of the SGPs was also temperature-dependent (Capdeville and Benwakrim 1994), we have further analyzed this phenomenon, by studying the expression of these proteins at 23 °C and 32 °C, in several strains of both P. primaurelia and P. tetraurelia.

PI-PLC releasable GPI proteins were recovered from two cellular fractions, cilia and cortex, which contains the non-ciliary plasma membrane, also called the pellicular or somatic plasma membrane (Allen 1971; Hufnagel 1969; Kéryer et al. 1990; Stelly et al. 1991). In addition to the SAg and SGPs detected in both fractions, we identified a new array of GPI-anchored proteins present only in the cortex fractions, i.e. restricted to the interciliary domains of the plasma membrane.

Results A protocol combining both biosynthetic labeling with metabolites specific for the GPI anchor and a PI-PLC treatment of purified cilia to specifically solubilize GPI-anchored molecules was used to identify the SGPs expressed by cells grown at either 23 °C or 32 °C, in different geographical strains of P. primaurelia and P. tetraurelia. Cells were labeled by feeding with [35S]- or [33P]-labeled bacteria during several divisions, or with [3H]ethanolamine for a short time, allowing only the molecules with a rapid turnover to be well labeled. The cilia were purified, treated with B. thuringiensis PI-PLC, then centrifuged. The pellet contained the membrane-bound form of the SGPs (mSGP), with the anchoring lipid; the soluble form (sSGP), lacking the lipid but retaining the ethanolamine-glycosylinositolphosphate part of the GPI anchor was recovered in the super-

Figure 1. Identification of the ciliary SGPs in P. primaurelia strain 513 grown at 32 °C. Purified [35S]- or [33P]- or [3H]ethanolamine-labeled cilia were divided into two equal parts, treated with or without B. thuringiensis PI-PLC, then centrifuged. Each ciliary pellet and supernatant was analyzed by SDS-PAGE under non-reducing conditions. In all three panels (A, B, C): lanes P-: control ciliary pellet; lanes S-: control ciliary supernatant; lanes P+: PI-PLCtreated ciliary pellet; lanes S+: PI-PLC-treated ciliary supernatant. Large arrows: membrane-bound form of SAg; large arrowheads: soluble form of SAg; small arrows: membrane-bound form of SGPs; small arrowheads: soluble form of SGPs; T: tubulins (identified by Western blotting). Molecular weight standards are indicated in kDa.

Paramecium GPI Proteins

natant. Both fractions were subjected to SDS-PAGE under reducing- and non-reducing conditions and the electrophoregram visualized by autoradiography or fluorography. Figure 1 shows, for P. primaurelia strain 513 grown at 32 °C, the identification of the unreduced SGPs in both forms (membrane-bound or soluble), after labeling with [35S]- or [33P]-labeled bacteria, or with [3H]ethanolamine. Only the SAg and SGPs could be labeled with all three radiolabeled metabolites. For the three types of labeling, the pellets contained, in addition to a high molecular weight band corresponding to the membranebound form of the SAg (mSAg), a poorly labeled band at 30 kDa, and an intensely labeled band at 34 kDa. Their labeling was stronger in the control than in the PI-PLC-treated samples. The PI-PLCtreated supernatants yielded the soluble form of the SAg (sSAg) band, migrating more slowly than the mSAg band, and a sSGP band at 51 kDa, resulting from solubilization of the 34 kDa mSGP. The 34 kDa mSGP and 51 kDa sSGP bands thus correspond to a SGP which is well expressed at 32 °C. The fact that this SGP was strongly labeled with [3H]ethanolamine indicates a rapid turnover. Occasionally (as in panel A), small amounts of mSAg, mSGP and tubulins were recovered in the supernatants. In the control and the PI-PLC-treated supernatants from [33P]-labeled cilia, bands observed below the 18 kDa marker may correspond to shed glycophospholipids.

163

We noticed an important shift in the electrophoretic mobility between the unreduced mSGP and sSGP. After reduction (patterns not shown), no such difference was observed: the reduced mSGP was located at 43 kDa while the corresponding reduced sSGP was located at 44 kDa. The difference of migration is characteristic of unreduced SGPs (the mSGPs migrating more rapidly) and thus provides a signature of SGPs.

SGP Sets Vary According to the Growth Temperature Comparative analyses of the SGPs in the unreduced pellets and supernatants from the control and the PI-PLC-treated cilia, of [35S]-labeled cells grown at either 23 °C or 32 °C, were performed for P. primaurelia strains 513 (Fig. 2A) and P. tetraurelia strain d4-2 (Fig. 2B). Table 1 recapitulates the apparent molecular weights of the mSGPs and sSGPs, resolved at both temperatures under non-reducing and reducing conditions. For both strains (Fig. 2), the unreduced pellets and PI-PLC-treated supernatants contained three major SGP bands at 23 °C and only one major band at 32 °C, comigrating with one of the bands seen at 23 °C. In strain 513, the 37 kDa mSGP band and the 43 kDa and 40 kDa sSGP bands present at 23 °C in the pellets and the PI-PLC-treated supernatant, respectively, were absent at 32 °C. In strain d4-2, among the three mSGP bands (at 38 kDa, 37 kDa

Figure 2. Comparison of the electrophoretic patterns of the ciliary SGPs expressed at different temperatures. After growth at either 23 °C or 32 °C, cilia were purified from [35S]-labeled cells of P. primaurelia strain 513 (A) and of P. tetraurelia strain d4-2 (B), and were treated with or without B. thuringiensis PI-PLC, then centrifuged. The control and PI-PLC-treated ciliary pellets and supernatants were analyzed by SDS-PAGE under non-reducing conditions. For abbreviations and symbols, see the legend of Figure 1. The electrophoretic patterns of whole cilia from cells grown at both temperatures were identical to those of the pellets (results not shown).

164

Y. Capdeville

Table 1. Apparent molecular weights of the mSGPs and sSGPs from P. primaurelia strain 513 and P. tetraurelia strain d4-2, grown at either 23 °C or 32 °C. STRAIN

23 °C (NR) –––––––––––––––––– mSGP sSGP

32 °C (NR) –––––––––––––––––– mSGP sSGP

513

37 [1] 34 [2] 30 [3]

51 [2’] 43 [3’] 40 [1’]

34 (30)

(38) [1] 37 [2] 34 [3]

68 52 [3’] 48

d4-2

51

52 34

23 °C (R) –––––––––––––––––– mSGP sSGP

32 °C (R) –––––––––––––––––––– mSGP sSGP

43 39 35

44 41

43

44

47 43 39

52 47

47 43

52 47

The apparent molecular weights (in kDa) of the unreduced (NR) and reduced (R) membrane-bound (mSGP) and soluble (sSGP) forms of the ciliary SGPs were determined according to Weber and Osborn (1969). Barely visible bands are indicated in brackets. A correspondence between the nonreduced membrane-bound and soluble forms of each SGP has been established in P. primaurelia strain 513 grown at 23 °C, on the basis of [33P] and [3H]ethanolamine labeling (Capdeville and Benwakrim 1996). [1], [2] and [3] designate the unreduced mSGPs and [1’], [2’] and [3’] designate the corresponding unreduced sSGPs.

and 34 kDa) and the three sSGPs (at 68 kDa, 52 kDa and 48 kDa) observed at 23 °C, only the 34 kDa mSGP band and the 52 kDa sSGP band were detected at 32 °C. A correspondence between the membrane-bound and the soluble form of each unreduced SGP could be established for strain 513 (as indicated in Table 1) but not for strain d4-2, except for one SGP. The presence of the 34 kDa mSGP band in the pellets and of the 52 kDa sSGP band in the PI-PLC-treated supernatant at 32 °C, suggests that these bands correspond to the membranebound and soluble forms, respectively, of the same unreduced SGP. After reduction, the electrophoretic patterns (not shown) of the PI-PLC-treated ciliary samples differed notably: at both temperatures, the reduced mSGPs migrated more slowly while the corresponding reduced sSGPs migrated more rapidly (Table 1, R). The SDS-PAGE patterns obtained after labeling with [3H]ethanolamine of both strains grown at either 23 °C or 32 °C showed that the mSGP and sSGP bands, detected at both temperatures, were the most intensely labeled (results not shown). The same comparative analyses on the expression of the SGPs at 23 °C and at 32 °C were also performed for P. primaurelia strain 61. The electrophoretic patterns at both temperatures were similar to those of strain 513 (results not shown). Altogether, these results suggest that at 32 °C, paramecia express a subset of the repertoire expressed at 23 °C, and that the SGPs expressed at both temperatures display a rapid turnover.

SGP Sets Differ According to the Strains Analyses of SGP expression in strains 156 and 168 of P. primaurelia and in strain 32 of P. tetraurelia showed that, at a given temperature, the SDS-PAGE profiles of the expressed SGPs were distinct. Several differences between strains 156 and 168 grown at 23 °C were observed (Fig. 3). Several mSGP bands differed in labeling intensities: only the highest mSGP band was strongly labeled in both strains. Strain 156 displayed both higher and lower sSGP bands, while strain 168 showed a major sSGP band with the same migration as the lower sSGP

Figure 3. Comparison of the ciliary SGP profiles from different geographical strains grown at 23 °C. Purified cilia from [35S]-labeled cells of P. primaurelia strains 156, 168 and iso-4 (carrying the 156 allele of the G SAg in a 168 genetic background) were treated with PI-PLC and then centrifuged. The PI-PLC-treated ciliary pellets and supernatants were analyzed by SDS-PAGE under non-reducing conditions. The middle part of the gel including the SGP bands in their membrane-bound and soluble forms is shown. Abbreviations and symbols, as in Figure 1.

Paramecium GPI Proteins

band present in strain 156, suggesting that this sSGP could result from solubilization of the mSGP displaying the highest apparent molecular weight in both strains. We also observed in strain 168 a minor band below the major sSGP band. The SGP profiles of strains 156 and 168 also differed from those of strain 513, because strain 513 grown at 23 °C demonstrated three sSGP bands in the unreduced supernatant (Fig. 2A). Concerning strain 32 of P. tetraurelia, grown at 23 °C, the SGP SDS-PAGE profiles (not shown) differed from those of P. tetraurelia strain d4-2: only one sSGP band was observed in the unreduced supernatant of strain 32 versus three sSGP bands for strain d4-2 (Fig. 2B). The SGP SDS-PAGE profiles of strains 156 and 168 were compared with that of the isogenic strain iso4 which carries the 156 allele of the G SAg in a 168 genetic background. As shown in Figure 3, the iso4 strain retained the SGP profile of the 168 strain. These results show that cells can display differences in the expressed SGP set according to the geographical origin of strains, and that the loci coding for the G SAg and the SGPs are not linked.

Presence of the SAg, SGPs and Other GPI Proteins in the Cortex Fractions We screened GPI proteins specifically released by PI-PLC in two different preparations of cortex: the “5th wash cortex” and the “Percoll gradient cortex”

165

(see protocols). Both types of preparations gave identical electrophoretical patterns of PI-PLC releasable GPI proteins (results not shown). Figure 4 shows autoradiographs of SDS-PAGEs of the pellets and the supernatants from the control and the PI-PLC-treated cortex (“5th wash cortex”), issued from strain 513 grown at 32 °C and [35S]-labeled, under non-reducing (NR) and reducing (R) conditions. The profiles of the unreduced PI-PLC-treated supernatants revealed new bands, in addition to the sSAg band (large arrowhead) and the sSGP band (small arrowhead). These additional bands (several minor bands located between the sSAg band and the 105 kDa marker, and a major band located at 29 kDa) correspond to the unreduced soluble form of specific cortical GPI proteins. After reduction, the bands corresponding to the reduced sSAg and sSGP were easily identified. In contrast, the specific cortical GPI proteins of high apparent molecular weights in the unreduced soluble form were not easy to identify when reduced. The band located at 29 kDa was replaced by two new strongly labeled bands with lower apparent molecular weights. The same set consisting of a 29 kDa band and a doublet of low apparent molecular weight bands was consistently observed for cortex fractions prepared from strain 513, grown at 23 °C, as well as for strains 61 and P. tetraurelia d4-2, grown at either 23 °C or 32 °C. In contrast, there was some interstrain variability (in number and electrophoretic mobility) for the cortical bands with high apparent molecular weights (results not shown).

Discussion

Figure 4. PI-PLC releasable GPI proteins from purified cortices of P. primaurelia strain 513, grown at 32 °C. The [35S]-labeled “5th wash cortex” was treated with or without PI-PLC. After centrifugation, the cortical samples were analyzed by SDS-PAGE under non-reducing (NR) and reducing (R) conditions. Lane P-: control cortical pellet; lane S-: control cortical supernatant; lane P+: PI-PLC-treated cortical pellet; lane S+: PI-PLC-treated cortical supernatant. Symbols, as in Figure 1.

A systematic screening for PI-PLC releasable GPI proteins performed on purified cellular fractions: cilia and cortex, in different strains of P. primaurelia and P. tetraurelia, grown at either 23 °C or 32 °C, allowed us to further characterize the expression and the localization of the SGPs and to detect the existence of new GPI proteins, localized exclusively in the interciliary region.

Regulation of SGP Expression Capdeville and Benwakrim (1996) estimated the abundance and the turnover of the ciliary GPI proteins and showed that within a given ciliary SGP set, each SGP could be distinguished by its turnover and level of expression. This distinct behaviour

166

Y. Capdeville

leads us to postulate that each SGP may be coded by a particular gene. The expressed SGP set differed according to the growth temperature and the geographical origin of strains. This is reminiscent of the properties of SAgs and leads us to wonder to what extent a parallel can be drawn between these two classes of major GPIanchored surface proteins. Do external factors other than the temperature, such as pH or salt concentration, modulate the expression of the SGPs? Do the various geographical strains differ by the composition of the SGP repertoire, or the regulation of their expression, or both? Although the analogy between SAgs and SGPs is not complete, in so far as paramecia express only a single SAg of the repertoire but several SGPs, the number of characteristics shared by these surface proteins suggests that the SGPs, like the SAgs, are coded by a multigene family whose expression is controlled by environmental factors.

Localization of the SGPs The finding of SGPs in the cortical fraction indicates that these molecules are also present in the non-ciliary plasma membrane domains. Because in the fractionation methods used for isolating the cortex, only the interciliary plasma membrane was accessible to PI-PLC (since the alveoli retain their in vivo topological orientation (Stelly et al. 1991)), we could not determine whether the SGPs were also linked to the alveolar membranes. There is some evidence that this is not the case, since no labeling of the alveolar membranes was observed by Flötenmeyer et al. (1999), using antisera obtained by immunizing rabbits with P. tetraurelia ethanolic extracts. These antisera were assumed to contain antibodies against all the surface proteins, i.e. the SAg and SGPs in their soluble forms, since incubation of living cells in ethanolic solution permits the endogenous Paramecium GPI-PLC to act (Capdeville et al. 1986). This result contrasts with the alveolar membrane labeling observed (Capdeville et al. 1993) with antisera prepared against whole ciliary membranes (Eisenbach et al. 1983). However, in the latter case, we cannot exclude the possibility that the labeling was due to antibodies directed against the membrane lipids linked to the anchoring lipid of the GPI proteins, owing to the strong interactions between the GPI anchor lipid and the membrane lipids (Benwakrim et al. 1998). Therefore, a definitive conclusion regarding the presence of the SGPs in alveolar vesicles demands further experiments with specific antibodies raised against purified SGPs in the lipidfree soluble form.

Non-ciliary GPI Proteins and Membrane Domains In the course of screening for SGPs in the cortical fraction, we detected other PI-PLC releasable GPI proteins, besides the SAg and the SGPs, which were not found in the purified cilia. These specific cortical GPI proteins identified in their soluble form comprised several minor proteins of high apparent molecular weights and one major protein of apparent molecular weight of 29 kDa under non-reducing conditions. Under reducing conditions, the disappearance of the 29 kDa band and the appearance of two new bands of equal intensity and located in a lower molecular weight range, suggest that the 29 kDa GPI protein is composed of two subunits, although the apparent molecular mass of the two subunits exceeds 29 kDa. This discrepancy might be due to the GPI anchorage itself, which influences the migration of GPI proteins on SDS gels. In contrast to the SAg and the SGPs which are evenly distributed over the entire cell surface, the cortical GPI proteins are localized only in the nonciliary domains of the plasma membrane. Segregation of GPI proteins into different domains of the plasma membrane has not been previously reported in other protozoa, whether parasitic or free-living. This finding is reminiscent of that described in the fetal porcine intestinal epithelium where a GPI-anchored transferrin-like protein is present in flat or invaginated domains, but never on microvilli structures (Danielsen and van Deurs 1995). This suggests that the localization of GPI proteins can be regulated in a specific way, possibly through lipid-lipid interactions, to form particular microdomains. Such microdomains containing glycolipids and probably transmembrane proteins (Ahmed et al. 1997; Simons and Ikonen 1997) might be involved in signal transduction pathways (Brown 1993; Brown and London 1998; Hoessli and Robinson 1998). They could possibly be implicated in mechanisms operating to adjust both the metabolism and swimming behavior of paramecia to environmental conditions. The segregation of membrane proteins into different domains of the plasma membrane in Paramecium has been observed for ion channels, which are specific for either ciliary or non-ciliary plasma membranes (Kung and Saimi 1982); for mechanoreceptors (Machemer and Ogura 1979) and for chemoreceptors (van Houten 1990), which are located on the cell body but not on the cilia. Some of the specific cortical GPI proteins could correspond to the folate receptor (shown to be GPI-anchored in mammals (Rothberg et al. 1990)) and to the glutamate receptor. Van Houten et al. (1999, in Mol Biol Cell 10:

Paramecium GPI Proteins

197a) has found that paramecia expressing antisense transcripts for a human protein termed PIG-A (phosphatidylinositol glycan – class A), one of the first enzymes in the GPI anchor synthesis pathway (Miyata et al. 1993), are specifically defective in chemoresponse to folate and glutamate.

Methods Reagents: [35S] Sulphate (Sulfur-35; SA: 0.07–1.5 GBq/mmol or 2-40Ci/mmol), 1-[3H] ethan-1-ol-2amine hydrochloride (SA: 0.56–1.48 TBq/mmol or 15–40 Ci/mmol), [33P] phosphoric acid (SA: 92.5–129.5 TBq/mmol or 2500-35000Ci/mmol) were from Amersham (Les Ulis, France). Leupeptin and phenylmethylsulfonylfluoride (PMSF) were from Sigma (St Quentin Fallavier, France), and Percoll from Pharmacia (Saclay, France). Bacillus thuringiensis phosphatidylinositol-specific phospholipase C (PI-PLC) was from Immunotech (Marseille, France). Strains: Different geographical strains of P. primaurelia and P. tetraurelia (Sonneborn 1975) were used: P. primaurelia strains 513 (Chantilly, France), 61 (Massachusetts, USA), 156 (New Haven, USA), 168 (Sendai, Japan) and iso-4 which is a genetically constructed strain, homozygous for the 156 allele of the G SAg in a 168 genetic background (Capdeville et al. 1978); P. tetraurelia strain 32 (Maryland, USA) and strain d4-2 deriving from strain 51 (Indiana, USA). Cell culture: Cells were routinely grown at 23 °C, according to the standard techniques (Sonneborn 1970) in phosphate-buffered hay-grass infusion for P. primaurelia cells or Wheat Grass Powder infusion (Pines International, Lawrence, KS) for P. tetraurelia cells. Both media were inoculated the day before use, with Enterobacter aerogenes and supplemented with b-sitosterol (0.8 mg/ml for the haygrass medium, 0.4 mg/ml for the Wheat Grass Powder medium). In the case of P. primaurelia strains, cultures were checked by immobilization tests, performed with homologous antisera to identify the SAg expressed and to ascertain the phenotypic homogeneity of the population: expression of the G and D SAgs at medium and high temperatures, respectively (Capdeville 1971). Before labeling at 32 °C, cells were grown at 32 °C for about ten days. Practically, paramecia were transferred from 23 °C to 36 °C for 6 to 12 hours to induce the change in the SAg expression more rapidly. They were then transferred to 32 °C. Daily isolations of single cells for three days and mass cultures were performed at this temperature before initiating the labeling.

167

Biosynthetic labeling: Exponentially grown paramecia were pelleted at 250 g in a GGT (Giovanni Giaccardo Torino) centrifuge (Jouan, Paris, France) and washed twice with natural mineral water (Volvic, Puy de Dôme/France). Biosynthetic labeling experiments were performed, as previously described (Capdeville and Benwakrim 1996). Washed paramecia were incubated in mineral water (supplemented with β-sitosterol), containing [35S]- or [33P]-labeled bacteria, for two days at 23 °C and for one day at 32 °C, or with [3H]ethanolamine (1 mCi) for a few hours. [35S]-Labeled bacteria were obtained by growth of Enterobacter aerogenes in minimum medium, sulfate-limited (0.05 mM) and supplemented with [35S]sulfate (1 mCi). [33P]-Labeled bacteria were obtained by growth in Neidhardt medium (Neidhardt et al. 1974), phosphate-limited (0.9 mM) and supplemented with [33P]phosphoric acid (1 mCi). At both temperatures of growth, paramecia incorporated 75 to 95% of the [35S] or [33P] radioactivity contained in [35S]- or in [33P]-labeled bacteria, while they incorporated only 3 to 8% of the [3H]ethanolamine radioactivity. Purification of cilia and cortices: Cilia were purified from labeled cells as previously described (Capdeville and Benwakrim 1996). After purification, pelleted cilia were washed in Tris-buffered saline, pH 7.4 (TBS). 1 to 2% of the [35S] radioactivity contained in the paramecia was recovered in cilia. Cortices were purified from labeled whole cells or deciliated cells, according to Kéryer et al. (1990), with slight modifications. Pelleted whole cells or deciliated cells were resuspended in two volumes of cold 20 mM Tris-maleate buffer (pH 7.8), 3 mM EDTA, containing 0.25 M sucrose (homogenization medium, HM), in the presence of 0.01% leupeptin and 0.25 mM PMSF (homogenization medium inhibitors, HMI). The resuspended pellets were transferred to a Potter homogenizer (reference no 3431E15, size A, Thomas, Philadelphia, PA) equipped with a teflon pestle (0.15 mm clearance), and left to stand on ice for 10–20 min. They were then subjected to 60–100 hand strokes; the extent of cell breakage was monitored by phase contrast microscopy, until about 95% of the cells had been broken, yielding cortex fragments very heterogeneous in size but mostly detached from gullets. After dilution in 30 volumes of cold HMI, the homogenate was centrifuged at 350 g for 5 min in a Beckman J2-MC centrifuge. The pellet derived from deciliated cells was washed again three times under the same conditions and washed with TBS and centrifuged at 4,500 g to yield the “5th wash cortex”. The pellet derived from whole cells was washed once more under the same conditions and subjected to fractionation

168

Y. Capdeville

on a Percoll gradient according to Stelly et al. (1991). The cortex band was washed in 5 volumes of cold 20 mM Tris-maleate buffer (pH 7.8) and centrifuged at 4,500 g for 10 min at 4 °C in a Beckman J2-MC centrifuge. PI-PLC treatment: Purified cilia or cortices were resuspended in TBS, and then divided into two equal aliquots. One aliquot was treated with 2 U/ml of Bacillus thuringiensis PI-PLC for 30 minutes at 37 °C. The other, used as a control, was incubated without PI-PLC for 30 minutes at 37 °C. After incubation, the two samples were centrifuged at 23,000 g for 20 minutes at 4 °C in a Sigma 2 K15 centrifuge. The control and the PI-PLC-treated supernatants were centrifuged again while the pellets were washed once in TBS and solubilized with 5% SDS. All the fractions were analyzed by SDS-PAGE. For each fraction, supernatant and solubilized pellet, the same volume of the control and the PI-PLC-treated samples were loaded on to gels. Protein electrophoresis and autoradiographv: SDS-PAGE was performed according to Laemmli (1970) on 5–15% gradient gels. b-mercaptoethanol was omitted from the sample buffer for electrophoresis of unreduced samples. Radiolabeled proteins were visualized as previously described (Capdeville and Benwakrim 1996). The exposure times of films were one day for [35S]-labeled proteins, three days for [33P]-labeled proteins and one month for [3H]ethanolamine-labeled proteins.

Acknowledgements I thank very much Dr J. Beisson and Dr J. Cohen for helpful comments and critical reading of the manuscript and Dr R. Karess for improvements to the English. I also thank Dr L. Sperling and C. Sassier for assistance in image engineering. This work was supported by the Centre National de la Recherche Scientifique (CNRS, France).

References Adoutte A, Ramanathan R, Lewis RR, Ling KY, Kung C, Nelson DL (1980) Biochemical studies of the excitable membrane of Paramecium tetraurelia. J Cell Biol 84: 717–738 Ahmed SN, Brown DA, London E (1997) On the origin of sphingolipid-cholesterol rich detergent-insoluble domains in cell membrane: Physiological concentrations of cholesterol and sphingolipid induce formation of a detergent-insoluble, liquid-ordered lipid phase in model membranes. Biochemistry 36: 10944–10953

Allen RD (1971) Fine structure of membranous and microfibrillar systems in the cortex of Paramecium caudatum. J Cell Biol 49: 1–20 Azzouz N, Capdeville Y (1992) Structural comparisons between the soluble and the GPI-anchored forms of the Paramecium temperature-specific 156G surface antigen. Biol Cell 75: 217–223 Azzouz N, Striepen B, Gerold P, Capdeville Y, Schwarz RT (1995) Glycosylinositol-phosphoceramide in the free-living protozoan Paramecium primaurelia: modification of core glycans by mannosyl phosphate. EMBO J 14: 4422–4433 Beale GH (1954) The antigens. In Salt G (ed) The Genetics of Paramecium aurelia. Cambridge, University Press, pp 77–123 Benwakrim A, Trémolières A, Labarre J, Capdeville Y (1998) The lipid moiety of the GPI-anchor of the major plasma membrane proteins in Paramecium primaurelia is a ceramide: variation of the amide-linked fatty acid composition as a function of growth temperature. Protist 149: 39–50 Brown DA (1993) The tyrosine kinase connection: how GPI-anchored proteins activate T cells. Current Opin Immunol 5: 349–354 Brown DA, London E (1998) Functions of lipid rafts in biological membranes. Annu Rev Cell Dev Biol 14: 111–136 Capdeville Y (1971) Allelic modulation in Paramecium aurelia heterozygotes. Mol Gen Genet 112: 306–316 Capdeville Y, Benwakrim A (1994) Temperature-dependent expression of ciliary GPI proteins in Paramecium. Brazil J Med Biol Res 27: 415–420 Capdeville Y, Benwakrim A (1996) The major ciliary membrane proteins in Paramecium primaurelia are all glycosylphosphatidylinositol-anchored. Eur J Cell Biol 70: 339–346 Capdeville Y, Cardoso de Almeida ML, Deregnaucourt C (1987) The membrane-anchor of Paramecium temperature-specific surface antigens is a glycosylinositol phospholipid. Biochem Biophys Res Commun 147: 1219–1225 Capdeville Y, Vierny C, Keller AM (1978) Regulation of surface antigens expression in Paramecium primaurelia. Genetic and physiological factors involved in allelic exclusion. Mol Gen Genet 161: 23–29 Capdeville Y, Baltz T, Deregnaucourt C, Keller AM (1986) Immunological evidence of a common structure between Paramecium surface antigens and Trypanosoma variant surface glycoproteins. Exp Cell Res 167: 75–86 Capdeville Y, Charret R, Antony C, Delorme J, Nahon P, Adoutte A (1993) Ciliary and Plasma Membrane Proteins in Paramecium: Description, Localization, and Intracellular Transit. In Plattner H (ed) Mem-

Paramecium GPI Proteins

brane Traffic in Protozoa, Vol 2A. JAI Press LTD, Hampton Hill, UK, pp 181–226 Caron F, Meyer E (1989) Molecular basis of surface antigen variation in paramecia. Annu Rev Microbiol 43: 23–42 Danielsen EM, van Deurs B (1995) A transferrin-like GPI-linked iron-binding protein in detergent-insoluble noncaveolar microdomains at the apical surface of fetal intestinal epithelial cells. J Cell Biol 131: 939–950 Eisenbach L, Ramanathan R, Nelson DL (1983) Biochemical studies of the excitable membrane of Paramecium tetraurelia. IX. Antibodies against ciliary membrane proteins. J Cell Biol 97: 1412–1420 Ferguson MAJ, Williams AF (1988) Cell surface anchoring of proteins via glycosyl-phosphatidylinositol structures. Annu Rev Biochem 57: 285–320 Flötenmeyer M, Momayezi M, Plattner H (1999) Immunolabeling analysis of biosynthetic and degradative pathways of cell surface components (glycocalyx) in Paramecium cells. Eur J Cell Biol 78: 67–77 Hoessli DC, Robinson PJ (1998) GPI-anchors and cell membrane: a special relationship. Trends in Cell Biol 8: 87–89

169

Machemer H, Ogura A (1979) Ionic conductance of membranes in ciliated and deciliated Paramecium. J Physiol (Lond) 296: 49–60 McConville MJ, Ferguson MAJ (1993) The structure, biosynthesis and function of glycosylated phosphatidylinositols in the parasitic protozoa and higher eukaryotes. Biochem J 294: 305–324 Miyata T, Takeda J, Iida Y, Yamada N, Inoue N, Takahashi M, Maeda K, Kitani T, Kinoshita T (1993) The clonoing of PIG-A, a component in the early step of GPI anchor biosynthesis. Science 259: 1318–1320 Neidhardt EC, Bloch PL, Smith DF (1974) Culture medium for enterobacteria. J Bacteriol 119: 736–747 Ramanathan R, Adoutte A, Dute RR (1981) Biochemical studies of the excitable membrane of Paramecium tetraurelia: Effects of proteases on the ciliary membrane. Biochim Biophys Acta 641: 349–365 Rothberg KG, Ying Y, Kolhouse JF, Kamen BA, Anderson RGW (1990) Glycophospholipid-linked folate receptor internalizes folate without entering the clathrincoated pit endocytic pathway. J Cell Biol 110: 637–649 Simons K, Ikonen E (1997) Functional rafts in membranes. Nature 387: 569–572

Hufnagel L (1969) Cortical ultrastructure of Paramecium aurelia. Studies on isolated pellicles. J. Cell Biol. 40: 779–801

Sonneborn TM (1970) Methods in Paramecium Research. In Prescott DM (ed) Methods in Cell Biology. Academic Press, New York, pp 241–339

Kéryer G, Adoutte A, Ng SF, Cohen J, Garreau de Labouesse N, Rossignol M, Stelly N, Beisson J (1990) Purification of the surface membrane-cytoskeleton (cortex) of Paramecium and identification of its protein constituents. Europ J Protistol 25: 209–225

Sonneborn TM (1975) The Paramecium aurelia complex of fourteen sibling species. Trans Amer Microsc Soc 94: 155–178

Kung C, Saimi Y (1982) The physiological basis of taxes in Paramecium. Annu Rev Physiol 44: 519–534

Stelly N, Mauger JP, Claret M, Adoutte A (1991) Cortical alveoli of Paramecium: a vast submembranous calcium storage compartment. J Cell Biol 113: 103–112

Laemmli UK (1970) Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227: 680–685

van Houten J (1990) Chemosensory Transduction in Paramecium. In Armitage JP, Lackie JM (eds) Biology of the Chemotactic Response. Soc Gen Microbiol Symp Vol. 46, Cambridge University Press, Cambridge, pp 297–321

Low MG (1989) The glycosylphosphatidylinositol anchor of membrane proteins. Biochim Biophys Acta 988: 427–454

Weber K, Osborn M (1969) The reliability of molecular weight determinations by dodecyl sulfate-polyacrylamide gel electrophoresis. J Biol Chem 244: 4406–4412