Parafusin, an exocytic-sensitive phosphoprotein, is the primary acceptor for the glucosylphosphotransferase in Paramecium tetraurelia and rat liver

Evolutionary implications of localization of the signaling scaffold protein parafusin to both cilia and the nucleus

Parafusin (PFUS), a 63 kDa protein first discovered in the eukaryote Paramecium and known for its role in apicomplexan exocytosis, provides a model for the common origin of cellular systems employing scaffold proteins for targeting and signaling. PFUS is closely related to eubacterial rather than archeal phosphoglucomutases (PGM) - as we proved by comparison of their 88 sequences - but has no PGM activity. Immunofluorescence microscopy analysis with a PFUS-specific peptide antibody showed presence of this protein around the base region of primary cilia in a variety of mammalian cell types, including mouse embryonic (MEFs) and human foreskin fibroblasts (hFFs), human carcinoma stem cells (NT-2 cells), and human retinal pigment epithelial (RPE) cells. Further, PFUS localized to the nucleus of fibroblasts, and prominently to nucleoli of MEFs. Localization studies were confirmed by Western blot analysis, showing that the PFUS antibody specifically recognizes a single protein of ca. 63 kDa in both cytoplasmic and nuclear fractions. Finally, immunofluorescence microscopy analysis showed that PFUS localized to nuclei and cilia in Paramecium. These results support the suggestion that PFUS plays a role in signaling between nucleus and cilia, and that the cilium and the nucleus both evolved around the time of eukaryotic emergence. We hypothesize that near the beginnings of eukaryotic cell evolution, scaffold proteins such as PFUS arose as peripheral membrane protein identifiers for cytoplasmic membrane trafficking and were employed similarly during the subsequent evolution of exocytic, nuclear transport, and ciliogenic mechanisms.

Role of the parafusin orthologue, PRP1, in microneme exocytosis and cell invasion in Toxoplasma gondii

The association of PRP1, a Paramecium parafusin orthologue, with Toxoplasma gondii micronemes, now confirmed by immunoelectron microscopy, has here been studied in relation to exocytosis and cell invasion. PRP1 becomes labelled in vivo by inorganic 32P and is dephosphorylated when ethanol is used to stimulate Ca2+-dependent exocytosis of the micronemes. The ethanol Ca2+-stimulated exocytosis is accompanied by translocation of PRP1 and microneme content protein (MIC3) from the apical end of the parasite. Immunoblotting showed that PRP1 is redistributed inside the parasite, while microneme content is secreted. To study whether similar changes occur during cell invasion, quantitative microscopy was performed during secretion, invasion and exit (egress) from the host cell. Time-course experiments showed that fluorescence intensities of PRP1 and MIC3 immediately after invasion were reduced 10-fold compared to preinvasion levels, indicating that PRP1 translocation and microneme secretion accompanies invasion. MIC3 regained fluorescence intensity and apical distribution after 15 min, while PRP1 recovered after 1 h. Intensity of both proteins then increased throughout the parasite division period until host cell lysis, suggesting the need to secrete microneme proteins to egress. These studies suggest that PRP1 associated with the secretory vesicle scaffold serves an important role in Ca2+-regulated exocytosis and cell invasion.

Complex glycosylation of Skp1 in Dictyostelium: implications for the modification of other eukaryotic cytoplasmic and nuclear proteins

Recently, complex O-glycosylation of the cytoplasmic/nuclear protein Skp1 has been characterized in the eukaryotic microorganism Dictyostelium. Skp1's glycosylation is mediated by the sequential action of a prolyl hydroxylase and five conventional sugar nucleotide-dependent glycosyltransferase activities that reside in the cytoplasm rather than the secretory compartment. The Skp1-HyPro GlcNAcTransferase, which adds the first sugar, appears to be related to a lineage of enzymes that originated in the prokaryotic cytoplasm and initiates mucin-type O-linked glycosylation in the lumen of the eukaryotic Golgi apparatus. GlcNAc is extended by a bifunctional glycosyltransferase that mediates the ordered addition of beta1,3-linked Gal and alpha1,2-linked Fuc. The architecture of this enzyme resembles that of certain two-domain prokaryotic glycosyltransferases. The catalytic domains are related to those of a large family of prokaryotic and eukaryotic, cytoplasmic, membrane-bound, inverting glycosyltransferases that modify glycolipids and polysaccharides prior to their translocation across membranes toward the secretory pathway or the cell exterior. The existence of these enzymes in the eukaryotic cytoplasm away from membranes and their ability to modify protein acceptors expose a new set of cytoplasmic and nuclear proteins to potential prolyl hydroxylation and complex O-linked glycosylation.

A parafusin-related Toxoplasma protein in Ca2+-regulated secretory organelles

We cloned a gene, PRPI, of Toxoplasma gondii encoding a 637-amino-acids protein having a calculated mass of 70 kDa. The sequence showed high homology to parafusin, a protein that in Paramecium tetraurelia participates in Ca2+-regulated exocytosis and is a paralog of phosphoglucomutase. We show that Toxoplasma gondii homogenate and an expressed recombinant PRP1 fusion protein cross-react with a specific peptide-derived antibody to parafusin in Western blots. Antibodies to the recombinant PRP1 showed cross-reaction with parafusin and recognized PRP1, as bands at M, 63 x 10(3) and 68 x 10(3), respectively. PRP1 is labeled when Toxoplasma gondii cells are incubated with inorganic 32P and appears as the major band on autoradiograms of SDS-PAGE gels. The localization of PRP1 was examined in secretory organelles of Toxoplasma gondii by deconvolution light microscopy followed by three dimensional reconstruction using pairwise combinations of specific antibodies. PRP1 localized to the apical third of the cell. It co-localized with micronemes, the only secretory organelle the secretion of which is Ca2+ dependent. Quantification of the co-localized stain suggests that only mature micronemes ready for exocytosis have PRP1. These findings suggest that PRP1, parafusin and other members of the phosphoglucomutase superfamily have a conserved role in Ca2+-regulated exocytic processes.

A non-Golgi alpha 1,2-fucosyltransferase that modifies Skp1 in the cytoplasm of Dictyostelium

Skp1 is a subunit of the SCF-E3 ubiquitin ligase that targets cell cycle and other regulatory factors for degradation. In Dictyostelium, Skp1 is modified by a pentasaccharide containing the type 1 blood group H trisaccharide at its core. To address how the third sugar, fucose alpha1,2-linked to galactose, is attached, a proteomics strategy was applied to determine the primary structure of FT85, previously shown to copurify with the GDP-Fuc:Skp1 alpha 1,2-fucosyltransferase. Tryptic-generated peptides of FT85 were sequenced de novo using Q-TOF tandem mass spectrometry. Degenerate primers were used to amplify FT85 genomic DNA, which was further extended by a novel linker polymerase chain reaction method to yield an intronless open reading frame of 768 amino acids. Disruption of the FT85 gene by homologous recombination resulted in viable cells, which had altered light scattering properties as revealed by flow cytometry. FT85 was necessary and sufficient for Skp1 fucosylation, based on biochemical analysis of FT85 mutant cells and Escherichia coli that express FT85 recombinantly. FT85 lacks sequence motifs that characterize all other known alpha 1,2-fucosyltransferases and lacks the signal-anchor sequence that targets them to the secretory pathway. The C-terminal region of FT85 harbors motifs found in inverting Family 2 glycosyltransferase domains, and its expression in FT85 mutant cells restores fucosyltransferase activity toward a simple disaccharide substrate. Whereas most prokaryote and eukaryote Family 2 glycosyltransferases are membrane-bound and oriented toward the cytoplasm where they glycosylate lipid-linked or polysaccharide precursors prior to membrane translocation, the soluble, eukaryotic Skp1-fucosyltransferase modifies a protein that resides in the cytoplasm and nucleus.

A comparative hybridization analysis of yeast DNA with Paramecium parafusin- and different phosphoglucomutase-specific probes

Molecular probes designed for the parafusin (PFUS), the Paramecium exocytic-sensitive phospho glyco protein, gave distinct hybridization patterns in Saccharomyces cerevisiae genomic DNA when compared with different phosphoglucomutase specific probes. These include two probes identical to segments of yeast phosphoglucomutase (PGM) genes 1 and 2. Neither of the PGM probes revealed the 7.4 and 5.9 kb fragments in Bgl II-cut yeast DNA digest detected with the 1.6 kb cloned PFUS cDNA and oligonucleotide constructed to the PFUS region (insertion 3 – I-3) not found in other species. PCR amplification with PFUS-specific primers generated yeast DNA-species of the predicted molecular size which hybridized to the I-3 probe. A search of the yeast genome database produced an unassigned nucleotide sequence that showed 55% identity to parafusin gene and 37% identity to PGM2 (the major isoform of yeast phosphoglucomutase) within the amplified region.Key words: parafusin, phosphoglucomutase, yeast, hybridization, PCR.

Functional diversity of the phosphoglucomutase superfamily: structural implications

Three-dimensional structural models of three members of the phosphoglucomutase (PGM) superfamily, parafusin, phosphoglucomutase-related protein and sarcoplasmic reticulum phosphoglucomutase, were constructed by homology modeling based on the known crystal structure of rabbit muscle phosphoglucomutase. Parafusin, phosphoglucomutase-related protein and sarcoplasmic reticulum phosphoglucomutase each have 50% or more identity with rabbit muscle phosphoglucomutase at the amino acid level and all are reported to exhibit no or minor phosphoglucomutase activity. There are four major insertions and two deletions in the parafusin sequence relative to PGM, all of which are located in surface-exposed loops connecting secondary structural elements. The remaining amino acid substitutions are distributed throughout the sequence and are not predicted to alter the polypeptide fold. Parafusin contains a putative protein kinase C site located on a surface loop in domain II that is not present in the homologs. Although the general domain structure and the active site of rabbit muscle phosphoglucomutase are preserved in the model of phosphoglucomutase-related protein, a major structural difference is likely to occur in domain 1 due to the absence of 55 amino acid residues in PGM-RP. This deletion predicts the loss of three alpha-helices and one beta-strand from an anti-parallel beta-sheet in this domain as compared with the rabbit muscle phosphoglucomutase.

Parafusin is a membrane and vesicle associated protein that cycles at exocytosis

In the unicellular eukaryote Paramecium tetraurelia, stimulation of exocytosis leads to Ca2+ activation of an alpha Glc-1-phosphodiesterase that dephosphoglucosylates the phosphoglycoprotein parafusin (PFUS). This process fails in exo mutant nd9 and also in the absence of Ca2+ influx upon stimulation suggesting that PFUS dephosphoglucosylation may be causally related to exocytosis. To further corroborate the hypothesis that PFUS is involved in the molecular events in exocytosis, we used laser confocal scanning microscopy and a PFUS specific peptide antibody to perform localization studies of PFUS in wild type (wt) and mutant Paramecium. In unstimulated wt cells, PFUS was associated both with the exocytic site of the cell membrane and with the membrane of the dense core secretory vesicles. Localization at these two sites was shown to be independent of each other since the exocytosis mutant (exo-) tam8, in which docking of its vesicles is blocked, still showed cell membrane staining. Immunofluorescence and immunoblotting of isolated intact secretory vesicles also revealed PFUS association. Upon stimulation of exocytosis, PFUS dissociated from both the dense core secretory vesicles and the cell membrane in a Ca(2+)-dependent manner. During recovery of exocytic capacity, PFUS reassociated with the newly formed secretory vesicles in the cytoplasm prior to their docking at the exocytic sites. Immunoblot analysis of PFUS during this time showed no changes in levels of the protein. Stimulation of exocytosis in wt in Mg2+ buffer or in the exo- temperature sensitive mutant (nd9) at the non-permissive temperature did not lead to dissociation of the PFUS. We conclude that PFUS is a novel critical component whose cycling probably participates in the molecular exocytic fusion machinery in these cells.

In vivo analysis of the major exocytosis-sensitive phosphoprotein in Tetrahymena

Phosphoglucomutase (PGM) is a ubiquitous highly conserved enzyme involved in carbohydrate metabolism. A number of recently discovered PGM-like proteins in a variety of organisms have been proposed to function in processes other than metabolism. In addition, sequence analysis suggests that several of these may lack PGM enzymatic activity. The best studied PGM-like protein is parafusin, a major phosphoprotein in the ciliate Paramecium tetraurelia that undergoes rapid and massive dephosphorylation when cells undergo synchronous exocytosis of their dense-core secretory granules. Indirect genetic and biochemical evidence also supports a role in regulated exocytotic membrane fusion. To examine this matter directly, we have identified and cloned the parafusin homologue in Tetrahymena thermophila, a ciliate in which protein function can be studied in vivo. The unique T. thermophila gene, called PGM1, encodes a protein that is closely related to parafusin by sequence and by characteristic post-translational modifications. Comparison of deduced protein sequences, taking advantage of the known atomic structure of rabbit muscle PGM, suggests that both ciliate enzymes and all other PGM-like proteins have PGM activity. We evaluated the activity and function of PGM1 through gene disruption. Surprisingly, DeltaPGM1 cells displayed no detectable defect in exocytosis, but showed a dramatic decrease in PGM activity. Both our results, and reinterpretation of previous data, suggest that any potential role for PGM-like proteins in regulated exocytosis is unlikely to precede membrane fusion.

Protein phosphatase and kinase activities possibly involved in exocytosis regulation in Paramecium tetraurelia

In Paramecium tetraurelia cells synchronous exocytosis induced by aminoethyldextran (AED) is accompanied by an equally rapid dephosphorylation of a 63 kDa phosphoprotein (PP63) within 80 ms. In vivo, rephosphorylation occurs within a few seconds after AED triggering. In homogenates (P)P63 can be solubilized in all three phosphorylation states (phosphorylated, dephosphorylated and rephosphorylated) and thus tested in vitro. By using chelators of different divalent cations, de- and rephosphorylation of PP63 and P63 respectively can be achieved by an endogenous protein phosphatase/kinase system. Dephosphorylation occurs in the presence of EDTA, whereas in the presence of EGTA this was concealed by phosphorylation by endogenous kinase(s), thus indicating that phosphorylation of P63 is calcium-independent. Results obtained with protein phosphatase inhibitors (okadaic acid, calyculin A) allowed us to exclude a protein serine/threonine phosphatase of type I (with selective sensitivity in Paramecium). Protein phosphatase 2C is also less likely to be a candidate because of its requirement for high Mg2+ concentrations. According to previous evidence a protein serine/threonine phosphatase of type 2B (calcineurin; CaN) is possibly involved. We have now found that bovine brain CaN dephosphorylates PP63 in vitro. Taking into account the specific requirements of this phosphatase in vitro, with p-nitrophenyl phosphate as a substrate, we have isolated a cytosolic phosphatase of similar characteristics by combined preparative gel electrophoresis and affinity-column chromatography. In Paramecium this phosphatase also dephosphorylates PP63 in vitro (after 32P labelling in vivo). Using various combinations of ion exchange, affinity and hydrophobic interaction chromatography we have also isolated three different protein kinases from the soluble fraction, i.e. a cAMP-dependent protein kinase (PKA), a cGMP-dependent protein kinase (PKG) and a casein kinase. Among the kinases tested, PKA cannot phosphorylate P63, whereas either PKG or the casein kinase phosphorylate P63 in vitro. On the basis of these findings we propose that a protein phosphatase/kinase system is involved in the regulation of exocytosis in P. tetraurelia cells.

A 63 kDa phosphoprotein undergoing rapid dephosphorylation during exocytosis in Paramecium cells shares biochemical characteristics with phosphoglucomutase

We have enriched phosphoglucomutase (PGM; EC 5.4.2.2) approximately 20-fold from Paramecium tetraurelia cells by combined fractional precipitation with (NH4)2SO4, gel filtration and anion-exchange chromatography yielding two PGM peaks. Several parameters affecting PGM enzymic activity, molecular mass and pI were determined. Phosphorylation studies were done with isolated endogenous protein kinases. Like the 63 kDa phosphoprotein PP63, which is dephosphorylated within 80 ms during synchronous trichocyst exocytosis [Höhne-Zell, Knoll, Riedel-Gras, Hofer and Plattner (1992) Biochem. J. 286, 843-849], PGM has a molecular mass of 63 kDa and forms of identical pI. Since mammalian PGM activity depends on the presence of glucose 1,6-bisphosphate (Glc-1,6-P2) (which is lost during anion-exchange chromatography), we analysed this aspect with Paramecium PGM. In this case PGM activity was shown not to be lost, due to p-nitrophenyl phosphate-detectable phosphatase(s) (which we have separated from PGM), but also due to loss of Glc-1,6-P2. Like PGM from various vertebrate species, PGM activity from Paramecium can be fully re-established by addition of Glc-1,6-P2 at 10 nM, and it is also stimulated by bivalent cations and insensitive to chelating or thiol reagents. The PGM which we have isolated can be phosphorylated by endogenous cyclic-GMP-dependent protein kinase or by endogenous casein kinase. This results in three phosphorylated bands of identical molecular mass and pI values, as we have shown to occur with PP63 after phosphorylation in vivo (forms with pI 6.05, 5.95, 5.85). In ELISA, antibodies raised against PGM from rabbit skeletal muscle were reactive not only with original PGM but also with PGM fractions from Paramecium. Therefore, PGM and PP63 seem to be identical with regard to widely different parameters, i.e. co-elution by chromatography, molecular mass, phosphorylation by the two protein kinases tested, pI values of isoforms, and immuno-binding. Recent claims that PP63 ('parafusin') would not be identical with PGM specifically in Paramecium are critically evaluated. Since some glycolytic enzymes are discussed as being associated with the Ca(2+)-release channel in muscle sarcoplasmic reticulum, and since sub-plasmalemmal Ca2+ stores in Paramecium closely resemble sarcoplasmic reticulum, a possible function of PP63/PGM in exocytosis regulation is discussed, particularly since dephosphorylation strictly parallels exocytosis.

The posttranslational modification of phosphoglucomutase is regulated by galactose induction and glucose repression in Saccharomyces cerevisiae

The enzyme phosphoglucomutase functions at a key point in carbohydrate metabolism. In this paper, we show that the synthesis of the major isoform of yeast phosphoglucomutase, encoded by the GAL5 (PGM2) gene, is regulated in a manner that is distinct from that previously described for other enzymes involved in galactose metabolism in the yeast Saccharomyces cerevisiae. Accumulation of this isoform increased four- to sixfold when the culture experienced either glucose depletion or heat shock. However, heat shock induction did not occur unless the cells were under glucose repression. This nonadditive increase in expression suggests that the regulatory mechanisms controlling the heat shock induction and glucose repression of the GAL5 gene are functionally related. We previously demonstrated that phosphoglucomutase is modified by a posttranslational Glc-phosphorylation reaction. We now show that this posttranslational modification, like phosphoglucomutase expression itself, is also regulated by galactose induction and glucose repression. Finally, no evidence was found to indicate that the Glc-phosphorylation of phosphoglucomutase alters its enzymatic activity under the conditions examined.

Cloning and sequencing of parafusin, a calcium-dependent exocytosis-related phosphoglycoprotein

A cDNA for parafusin, an evolutionarily conserved phosphoglycoprotein involved in exocytosis, has been cloned and sequenced from a unicellular eukaryote, Paramecium tetraurelia. A Paramecium cDNA library was screened with an oligonucleotide probe synthesized to an internal amino acid sequence of isolated parafusin. The insert was 3 kb long with an open reading frame of 1.75 kb. Data base searches of the deduced amino acid sequence showed that Paramecium parafusin had a 50.7% sequence identity to rabbit muscle phosphoglucomutase, although no detectable phosphoglucomutase activity has been detected in isolated parafusin. The deduced parafusin amino acid sequence had four inserts and two deletions, which might confer on the protein specific functions in signal transduction events related to exocytosis. Furthermore, searches for potential phosphorylation sites showed the presence of a protein kinase C site (KDFSFR) specific to parafusin. Southern blot analysis with probes specific for parafusin and phosphoglucomutase suggested that these proteins were products of different genes. We propose that parafusin and phosphoglucomutase are members of a superfamily that conserve homologies important for the tertiary structure of the molecules.

Isolation and expression of a rat liver cDNA encoding phosphoglucomutase

A cDNA encoding phosphoglucomutase (PGM) has been isolated from a rat liver cDNA library following screening with a polymerase chain reaction product. The cDNA was found to contain a 53-base-pair (bp) 5' untranslated region (5' UTR), a single start codon and consensus initiation sequence, an open reading frame (ORF) of 1686 bp, and a 3' untranslated tail. A comparison to the rabbit and human muscle PGM cDNAs [Whitehouse et al., Proc. Natl. Acad. Sci. USA 89 (1992) 411-415] showed 90% identity of rat cDNA to both, while a comparison to the deduced amino acid sequences showed 97 and 96% identity, respectively. Northern blot analyses determined that PGM was encoded by a single mRNA in rat liver, of approximately 2.2 kb. Following transfection of COS-7 cells with a plasmid containing the entire PGM ORF, indirect immunofluorescence analyses using a PGM-specific monoclonal antibody determined that approximately 5% of the cells displayed 50-100 times greater fluorescence than that seen in the remainder of the cells or in mock transfects. The enhanced production of PGM was also demonstrated by Western blotting and by enzymatic activity assays.

Carbohydrate cycling in signal transduction: parafusin, a phosphoglycoprotein and possible Ca(2+)-dependent transducer molecule in exocytosis in Paramecium

Parafusin, a cytosolic phosphoglycoprotein of M(r) 63,000, is dephosphorylated and rephosphorylated rapidly in a Ca(2+)-dependent manner upon stimulation of exocytosis in vivo in wild-type (wt) Paramecium. In contrast, the temperature-sensitive exocytosis mutant nd9, grown at the nonpermissive temperature (27 degrees C), does not exocytose or dephosphorylate parafusin upon stimulation in the presence of Ca2+; grown at the permissive temperature (18 degrees C), nd9 cells show a wt phenotype. Parafusin contains two types of phosphorylation sites: one where glucose 1-phosphate is added by an alpha-glucose-1-phosphate phosphotransferase and removed by a phosphodiesterase and one where phosphate from ATP is added directly to a serine residue by a protein kinase and removed by a phosphatase. We show here that, in cell fractions from wt Paramecium, both reactions can be carried out in vitro by using uridine(5'-[beta-[35S]thio])diphospho(1)-glucose (UDP[beta 35S]-Glc) and [gamma-32P]ATP, respectively. The characteristics of these pathways are different. Specifically, in the presence of Ca2+, the amount of UDP[beta 35S]-Glc label in parafusin is reduced. In contrast, identical labeling experiments with [gamma-32P]ATP show that Ca2+ enhances labeling of parafusin. Mg2+ had no appreciable effect on either labeling. Removal of the UDP[beta 35S]-Glc label on parafusin in the presence of Ca2+ correlates with the in vivo dephosphorylation seen upon exocytosis. Incubations with UDP[beta 35S]-Glc were then performed with homogenates and nd9 cell fractions grown at 27 degrees C under the ionic conditions used for wt cells. These labelings were not affected by Ca2+, in contrast to results from wt cells but in accord with those obtained earlier with nd9-27 mutant cells in vivo. Factors responsible for both dephosphorylation and Ca2+ sensitivity were found in the high-speed pellet (P2) in wt cells, suggesting that the putative phosphodiesterase is in this fraction and that the defect in the mutant nd9-27 residues in the Ca2+ activation of the phosphodiesterase. We conclude that the in vivo dephosphorylation of parafusin that occurs upon exocytosis is a dephosphoglucosylation due to removal of the alpha-glucose 1-phosphate and more generally that carbohydrates on cytoplasmic glycoproteins may be cyclically added and/or removed in response to extracellular stimuli.

A cortical phosphoprotein ('PP63') sensitive to exocytosis triggering in Paramecium cells. Immunolocalization and quenched-flow correlation of time course of dephosphorylation with membrane fusion

We had previously shown that a phosphoprotein of 63 kDa ('PP63') is rapidly and selectively dephosphorylated during synchronous (less than or equal to 1 s) trichocyst exocytosis in Paramecium cells and then rephosphorylated within less than or equal to 1 min [Zieseniss & Plattner (1985) J. Cell Biol. 101, 2028-2035]. Using a new quenched-flow device, we now find a strict correlation between PP63 dephosphorylation and the process of membrane fusion, both occurring within 80 ms. Uptake of 32P over 90 min, followed by exocytosis and rephosphorylation for 1 min, results in a rather selective phosphorylation of the dephosphorylated form, P63, to PP63. Solubilization by repeated freezing and thawing allows isolations of P63 and PP63. On isoelectric focusing autoradiograms they have pI values of 6.05, 5.95 (major spots), 5.85 and 5.75. All spots are sensitive to alkaline, but not to acidic, hydrolysis (except for the pI-6.05 spot). On two-dimensional-gel autoradiograms the most prominent spot, of pI 5.95, is most extensively de- and re-phosphorylated. This spot, from de- and re-phosphorylated samples, was used to produce monospecific antibodies. A cortical localization of PP63 was revealed by producing Western blots from isolated cell-surface fragments ('cortices') and by immunofluorescence labelling. We assume that both P63 and PP63 are attached to cortical structures, e.g. around trichocysts, though they are partly soluble. This localization and the strict correlation of PP63 dephosphorylation with exocytotic membrane fusion suggests a role in fusion regulation.