Traffic 2004; 5: 493–502 Blackwell Munksgaard

Copyright

#

Blackwell Munksgaard 2004

doi: 10.1111/j.1600-0854.2004.00194.x

Novel Secretory Vesicle Proteins Essential for Membrane Fusion Display Extracellular-Matrix Domains Marine Froissard, Anne-Marie Keller, Jean-Claude Dedieu and Jean Cohen* Centre de Ge´ne´tique Mole´ulaire, Centre National de la Recherche Scientifique, Avenue de la Terrasse, 91198 Gif-sur-Yvette cedex, France * Corresponding author: Jean Cohen, [email protected] Exocytotic mutants can be obtained in Paramecium that affect the organization of the fusion machinery, visible by electron microscopy. The site of action of the genes in the plasma membrane, cytosol or secretory compartment can easily be determined in such mutants. Functional complementation cloning of exocytotic mutants specifically affected in the secretory compartment, nd2-1 and nd169-1, reported here, and the previously studied nd7-1, led to the discovery of a set of novel proteins that display PSI and EGF domains, normally found in extracellular matrix proteins and involved in transmembrane signaling. The structure of one of these proteins, Nd2p, and of the product of a paralog found in the genome Nd22p, corresponds to that of type I membrane receptors, generally involved in protein and vesicle sorting. Our characterization suggests that the proteins we have identified are required to indicate the presence of a mature secretory vesicle to the plasma membrane, to prepare the machinery for fusion. We propose to name this novel subclass of receptors VEMIF, for Vesicular Extracellular-Matrix-like proteins Involved in preparing membrane Fusion.

signals in the cytosolic part of the sequence of the transmembrane receptors. In targeting to lysosomes, for instance, the cytoplasmic tail of the mannose-6-phosphate receptor, a type I membrane protein whose lumenal domain is involved in cargo sorting, interacts with specific coat proteins, the GGAs, that identify them as destined to lysosomes (2). However, in other membrane traffic pathways, such as that culminating in regulated exocytosis, little is known about potential equivalents to sorting receptors involved in plasma membrane recognition. In this work, we took advantage of the properties of the exocytotic pathway of Paramecium, in particular the possibility to easily identify gene products associated with the secretory compartment, to address this question.

Received 7 January 2004, revised and accepted for publication 25 March 2004

Paramecium secretes defensive organelles called trichocysts (Figure 1a) through a pathway of regulated exocytosis (3–5). The main advantage of the model is that exocytotic membrane fusion is naturally blocked at a late stage of connection between both plasma and trichocyst membranes, so that this step can be studied independently of all previous membrane fusion steps in the secretory pathway. Interestingly, specialized microdomains have been identified cytologically at the contacts between trichocysts and plasma membrane and can be easily visualized by freeze fracture and transmission electron microscopy, respectively, in the form of a rosette of intramembranous particles and electron-dense connecting material between membranes (6). The stage at which trichocyst exocytosis is naturally blocked in the absence of trigger is precisely situated between microdomain assembly and membrane fusion. Numerous mutants, called nd for nondischarge (7), have been obtained, which are only deficient for the last step(s) of membrane fusion, including connecting material and rosette assembly (3,5).

A key event in membrane fusion is the trans association of N-ethylmaleimide attachment protein receptors (SNAREs), integral proteins of the membranes to be fused (1). However, other molecules are required to precisely orchestrate all intracellular membrane fusion events to avoid wrong timing and illegitimate cargo delivery. Each cargo has a specific destination and is sorted to budding vesicles either by intrinsic properties, such as differential solubility, or by interaction with trans-membrane receptors, generating differentiated vesicles with specific contents. In turn, these vesicles are sorted within the intracellular traffic to avoid illegitimate membrane fusion. The recognition of vesicles by the transport/fusion machinery is carried out by specific

Among the ND genes previously cloned and sequenced, the ND7 gene (8) displays a cysteine-rich motif, which turns out to be a PSI motif, evoking the occurrence of disulfide bond-stabilized structures found in plexins, semaphorins and integrins of the extracellular-matrix (9). The site of action of the ND7 gene, which was determined to be the trichocyst (vesicular) compartment (10), indicates that Nd7p is most likely associated with or included in the secretory vesicle. Proteins with extracellular matrix-like domains within vesicular compartments were considered an exception until the recent description in several species of such proteins, e.g. transmembrane microneme proteins in Toxoplasma (11) and transmembrane proteins of prevacuolar compartments in plants (12,13). These proteins

Key words: EGF, microdomains, Paramecium, PSI domains, trichocyst, vesicular sorting

493

Froissard et al.

motif and have a structure of type I membrane receptors, with a unique C-terminal transmembrane domain followed by a short acidic cytoplasmic tail. The way in which such proteins of the secretory compartment can help to organize exocytotic microdomains is discussed. The involvement of these proteins in preparing membrane fusion, which distinguishes them from general sorting receptors, as well as their probable general occurrence, have led us to name this novel functional class VEMIF, for Vesicular Extracellular-Matrix-like proteins Involved in preparing membrane Fusion.

Results

Figure 1: Trichocyst localization of the Nd7-GFP fusion protein in Paramecium. a. Living Paramecium cell transformed with the ND7-GFP fusion gene viewed in phase contrast microscopy showing the trichocysts as black ‘carrots’ anchored at the cell periphery, under the plasma membrane (arrow). b. Same cell, viewed in epifluorescence using a bright field objective and GFP selective filters, in which tricho- cysts appear fluorescent (arrow), along with cytoplasmic vesicles. Bar ¼ 10 mm.

found in vesicles are thought to be sorting receptors for the relevant cargos. However, in the nd7 mutant, apparently no sorting defect occurred, since normally structured trichocysts develop and reach their correct location. The defect lies only in a distal event, the assembly of the fusion machinery. This suggested a role in cross talk between a PSI domain on the lumenal side of a vesicle with cytosolic and plasma membrane components to prepare membrane fusion. As PSI domains are generally involved in protein/protein interactions, we looked for putative interacting proteins by studying other ND genes for which the site of action is also the secretory compartment. We show here that the ND2 and ND169 genes cloned by functional complementation, and the ND22 gene found in a genome project (14), similar to ND2, display extracellularmatrix protein motifs, PSI repeats or EGF domains. In addition, the ND2 and ND22 genes present a kelch b-propeller 494

nd mutations affecting the trichocyst compartment In order to focus on proteins likely to be localized in the secretory (trichocyst) compartment, we considered the nd mutants in our stock collection (http://Paramecium. cgm.cnrs-gif.fr/souches_public/) defective in the trichocyst itself, as determined by experiments of transfer of cytoplasm from strain to strain (15). The known mutants in this category are nd2-1, nd7-1, nd126-1 and nd169-1 as determined by Lefort-Tran et al. and Pouphile et al. (10,16), for nd7-1 and by Bonnemain et al. (17), for nd126-1 and nd 169-1. The site of action of nd2-1 (formerly called ndB) is also known (15). To be sure that the nd2-1 strain we kept in stock is the same strain, we confirmed the site of action of the mutation by repeating cytoplasm transfers: a sample of cytoplasm, containing a few trichocysts, transferred from the nd2-1 mutant into the reference tam38-1 strain lacking functional trichocysts, failed to provide exocytosis, whereas transfer of competent trichocysts, free in the cytoplasm, from the tam6-1 mutant into nd2-1 provided efficient exocytosis of the injected trichocysts. To confirm this localization for at least one of the gene products, we transformed the nd7 mutant with an ND7GFP fusion gene. Although variations could be observed between cells, green fluorescent protein (GFP)-labeled trichocysts could be visualized in vivo along with other vesicular structures, using an intensifying camera (Figure 1). Although other investigations will be required to determine the nature of the other labeled vesicles and their possible role in the exocytotic pathway, this experiment clearly establishes that the conclusions drawn from physiologic experiments are consistent with the localization of the fusion protein in situ. Cloning ND2 and ND169 by functional complementation Using the procedure developed by Keller & Cohen (18), the cloning of the ND126 gene failed, but the ND2 and ND169 genes could be respectively isolated in the p77o7 (7386 bp insert) and p89m7 (9491 bp insert) plasmids. After sequencing the inserts, and in order to localize the relevant rescuing genes, we amplified several polymerase chain reaction Traffic 2004; 5: 493–502

Extracellular-Matrix-Like Proteins in Membrane Fusion

(PCR) products covering parts of the inserts, according to the localization of putative Open Reading Frames (ORFs) in the sequence, and tested them by transformation for their ability to rescue the mutants. The PCR fragments rescuing the genes ND2 and ND169 were of 3 kb and 1.5 kb, respectively. Sequence analysis revealed no predicted introns in the ND2 gene and three putative small introns in the ND169 gene (Figure 2).

By Southern analysis under stringent conditions, both genes were found to be unique in the genome (not shown), and, by Northern analysis, the sizes of the transcripts were determined to be 2.4 kb for ND2 and a doublet at 1.9 and 2.3 kb for ND169 (data not shown). The size of the ND169 messenger is larger than the size of the rescuing DNA fragment. We therefore analyzed the ND169 cDNA and looked for introns, 50 and 30 ends. This revealed that the ND169 open reading frame is longer than previously expected due to the presence of two additional introns, changing the frames in the 30 part of the gene and encompassing the previously predicted stop codon. The deduced transcript is compatible with a size of 1.9 kb. It is interesting to note that, in our complementation experiments, a 30 truncated version of the gene could rescue the mutant, indicating that the C-terminus of the protein is not essential, at least when the rest of the protein is present at high copy number. The sequences of the ND2 and ND169 genes allow us to predict translation products of 775 and 583 amino acids, respectively. The sequence of the corresponding genes in the four nd2 mutants and in the unique nd169 mutant revealed mutations (Figure 2), confirming that the genes cloned were the genuine ones, not multicopy suppressors. In silencing experiments using a method of microinjection into the macronucleus of PCR-amplified DNA corresponding to the coding sequence (19), we observed that cells acquired an nd phenotype in both cases, showing that both ND2 and ND169 genes are only involved in the final step of trichocyst exocytosis.

Figure 2: The ND2, ND22 and ND169 novel genes and their products. a. Schematic representation of the, ND2, ND22 and ND169 genes indicating the positions of the mutations and of the oligonucleotides, described in Table 1, used for PCR amplification. Full and dotted lines schematize PCR products used for overexpression and silencing experiments, respectively. Vertical bars indicate the positions of the mutations found by sequencing the gene versions obtained in the corresponding mutants. As the nd169-1 mutant was obtained in a different geographic stock (Stock 169), two silent polymorphic sites (in gray) were observed in addition to the frameshift insertion corresponding to the nd169-1 mutation. b. Schematic representation of the Nd2p, Nd22p, Nd169p and Nd7p proteins. PS: peptide signal, TM: transmembrane domain, C: cytoplasmic tail. kelch, EGF and PSI domains are schematized by different boxes. Also indicated are the mutations as well as the predicted length of putative mutated peptides in case of frameshift.

Traffic 2004; 5: 493–502

The cloning of a paralogue of ND2 and its identification as an ND gene: ND22 By examination of the data of a Paramecium genome pilot project (14,20, http://Paramecium.cgm.cnrs-gif.fr/ ), we identified a clone (R12C02r) not identical to, but homologous with, the 50 region of the ND2 gene. The full gene was cloned using an inverse PCR strategy. The sequence of the DNA fragment obtained revealed a 2460 bp ORF without introns, corresponding to an 820 aa protein very similar to Nd2p. The unique remarkable difference lies in a 160 aa divergent segment with an insertion of 40 aa in Nd22p. Both sequences share 70% sequence identity at the DNA level and 66% identity (80% similarity) at the protein level in the regions that can be aligned. To see whether this gene has a function in trichocyst exocytosis, we silenced it using a DNA fragment covering part of the ORF (Figure 2), according to the principle defined by Ruiz et al. (19), and Galvani & Sperling (21). The clones derived from injected cells displayed an nd phenotype, without any other apparent effects on the cells. To be sure that the effect was specific to the target gene, and not a cosilencing of ND2 due to the presence of portions of sequence sharing more than 85% identity at the DNA level with ND2 (apparent threshold to obtain 495

Froissard et al.

cosilencing), we used the ca. 500 bp corresponding to the only divergent portion of the gene as a PCR fragment amplified with oligos L-M (Table 1, Figure 2) in silencing experiments. In this case, an nd phenotype could also be obtained. The silenced phenotype was transient, as could be expected from the small size of the injected DNA fragment, which is barely maintained in the macronucleus. Hence, the new gene was demonstrated to be an ND gene and called ND22. In order to see whether ND22 could replace ND2, we tried to complement the nd2-1 and the nd2-2 mutants with a PCR fragment containing the ND22 ORF plus 50 and 30 flanking segments. No exocytotic rescue occurred, although Northern analysis of transformed clones revealed the presence of large amounts of the corresponding messenger (Figure 3). Therefore, both ND2 and ND22 genes appear to be essential for exocytosis and have independent functions. Ultrastructural defects provoked by gene silencing In order to see whether the nd phenotype induced by silencing also concerned rosette assembly, we analyzed the ultrastructural defects in ND2 and ND22 silenced cells, compared to cells carrying the mutations nd2-1 and nd2-2. As shown in Figure 4, rosettes fail to assemble in silenced cells, as in the mutants. The silencing of ND2 and ND22 thus affects exocytosis through direct effects on the mechanism of rosette assembly. Conserved domains in Nd2p, Nd22p and Nd169p Although BLAST analyses of the genes ND2, ND22 and ND169 did not reveal true orthologues in other species,

they made it possible for us to identify strong similarities with proteins containing conserved domains (Figure 5): two PSI (plexin-semaphorin-integrin) repeats are found in Nd169p (similar to the PSI domain already found in Nd7p), six kelch repeats and five EGF (Epidermal Growth Factor) motifs in each of the proteins Nd2p and Nd22p. All three polypeptides contain a signal peptide at the N-terminus and Nd2p and Nd22p have a transmembrane domain at the C-terminus, characteristic of the organization of type I integral membrane proteins, and a short cytoplasmic tail (Figure 5). PSI domains, found here in Nd7p and Nd169p, are always present in extracellular matrix proteins and are probably responsible for dimerization of integrins and semaphorins. This domain is characterized firstly by the presence of eight cysteines, essential for the secondary structure, implicated in disulfide bonds between C1/C6, C2/C4, C3/ C8, C5/C7 and, secondly, by the presence of conserved aromatic residues (Figure 5c). All known molecules with these repeats are signaling receptors (for review (9)). Kelch motifs, found here in Nd2p and Nd22p, first described in the Drosophila kelch gene (22), are characterized by the conservation of four hydrophobic residues followed by a glycine pair, separated from two characteristically spaced aromatic residues (Figure 5a). This consensus pattern leads to the formation of a b-propeller structure. kelch domains are present in various proteins with distinct cellular functions (for review (23)). The presence of these repeats seems to have a structural rather than a catalytic role. EGF repeats, found here in Nd2p and Nd22p, diagnostic of extracellular matrix proteins, are recognizable by the conserved cysteine residues (Figure 5b) (for review (24,25)), which, as for PSI domains, permit the formation of disulfide bonds between C1/C3, C2/C4 and C5/C6 to stabilize a secondary structure.

Figure 3: Overexpression of the ND22 gene in the nd2-1 mutant. Northern blot of total RNA isolated from three independent nd2-1 cell lines transformed by the ND22 gene in which the transgene is overexpressed, compared to an uninjected control (note the variability in expression between injected clones reflecting the variable amount of injected DNA). The ND22 DNA was used as a probe to reveal the level of expression, as well as ND7 DNA as a control. Exocytotic capacity of the nd2-1 mutant could not be rescued by overexpression of the ND22 gene.

496

The cytoplasmic tails of type I membrane receptors are involved in interaction with cytosolic proteins to form a coat and acquire specificity for transport and fusion machineries with cognate targets (26). The cytoplasmic tails of Nd2p and Nd22p end with a short acidic stretch followed by two isoleucines (Figure 5d), which could be related to the DXXLL motif responsible for interaction with specific adaptor proteins for some type I membrane receptors (26). This or other characteristics of these tails could well represent the specific signal for trichocyst identification by the site of fusion at the plasma membrane.

Discussion In this study, we identified genes of the Paramecium membrane-fusion mutants nd2-1 and nd169-1, as well as Traffic 2004; 5: 493–502

Extracellular-Matrix-Like Proteins in Membrane Fusion

e Figure 4: Effect of ND2 and ND22 silencing on rosette structure. a–d. Examples of wild type rosettes (a and b) and nd mutant sites displaying one or two rosette particles (black arrowheads) per site (c and d). Bar: 100 nm. e. Histogram plotting the percentage of exocytotic sites according to the number of rosette particles they contain. The wild type strain, taken as a control for normal exocytosis, displays 98% of 58 sites with five particles or more, the number required for trichocyst exocytosis. Wild type cells, in which the ND2 or ND22 genes have been silenced, display 97% of 100 sites or 96% of 83 sites with fewer than five particles per rosette, respectively. Similar values are observed for the two mutants nd2-1 and nd2-2, in which 100% of 62 sites and 96% of 73 sites, respectively, contain fewer than five particles.

the gene ND22 by sequence homology to ND2, encoding novel proteins that display PSI and EGF motifs normally found in the extracellular matrix, a characteristic already found for the ND7 gene product. These gene products are predicted to be associated with the secretory compartment, according to determination of the site of action of the mutation by trichocyst transfer from cell to cell. To confirm this localization, GFP fusion genes have been used: nd2 mutant cells overexpressing the corresponding wild type gene fused to the GFP gene, present a wild type phenotype and display abundant GFP fusion mRNA and protein as judged by Northern and Western experiments, but do not provide any fluorescence under the microscope (data not shown). In contrast, the nd7 mutant transformed with the ND7-GFP fusion gene displays both a rescued exocytotic phenotype and fluorescence that can be detected in vesicles and in trichocysts using an intensifying camera. Traffic 2004; 5: 493–502

The ND2 and ND22 genes display the structure of type I membrane receptors, with a single C-terminal transmembrane domain followed by a short cytoplasmic tail. The type I membrane receptor structure is widespread in intracellular membrane proteins. However, the occurrence of extracellular-matrix domains in such proteins has only recently been described in plant vesicles destined to fuse with the vacuole (12,13), and in Toxoplasma gondii micronemes (11), which are secretory organelles anchored at the plasma membrane much like Paramecium trichocysts. So far, these receptors have been described as sorting receptors or ‘escorting’ proteins to target their vesicle. In Paramecium, mutations in these proteins prevent exocytotic membrane fusion and, more precisely, the formation of the particle arrays within the plasma membrane, the rosettes, which define microdomains where fusion takes place. This indicates that, in addition to any role in sorting, the Ndp 497

Froissard et al.

Figure 5: Alignment of repeated regions of the proteins using CLUSTALW and BOXSHADE programs, which shades amino acids which are identical (black) or conserved (gray) in a majority of the sequences. a. Alignment of the kelch repeats of Nd2p and Nd22p, compared to two kelch repeats of Drosophila melanogaster host cell factor. The signature of this motif is the series of four hydrophobic residues and two glycines. Each repeat forms four b strands (arrows) arranged in a six-blade b-propeller structure. b. Alignment of EGF repeats of Nd2p and Nd22p, compared to two EGF repeats of Mus musculus tenascin X. The six cysteines conserved are necessary for the secondary structure of the domain by forming disulfide bonds (C1/C3, C2/C4, C5/C6). The position of glycines and aromatic residues are also conserved. c. Alignment of three PSI domains, one from Nd7p and two from Nd169p, compared to two PSI domains of Mus musculus plexin A2. PSI repeats are characterized by the presence of eight cysteines engaged in disulfide bonds (C1/C6, C2/C4, C3/C8, C5/C7) and the conservation of aromatic residues as shown in the consensus line under the alignment. d. Alignments of the cytoplasmic tails of Nd2p and Nd22p under which the acidic stretch is underlined with asterisks. Consensus lines: h ¼ hydrophobic amino acid, A ¼ aromatic amino acid. Accession numbers: Drosophila melanogaster HCF AAK28427, Mus musculus Tenascin X NP_112453, Mus musculus Plexin A2 NP_032908.

proteins described here play a role in membrane fusion by cross talk between the vesicle and the plasma membrane to prepare the fusion machinery. We thus propose to highlight this peculiarity by naming this novel membrane receptor subclass VEMIF, for Vesicular Extracellular-Matrix-like proteins Involved in preparing membrane Fusion.

Similarities between Paramecium VEMIFs and extracellular matrix proteins Both Nd7p and Nd169p proteins have PSI motifs, the only conserved domains in their sequences, and do not seem 498

to present a transmembrane domain (there is only weak probability of a transmembrane domain in the N-terminus of Nd7p). Nd2p and Nd22p proteins, very similar in structure, but nonredundant in function, display a type I membrane receptor structure with a large N-terminal (lumenal) part with a kelch-type b-propeller and five EGF repeats. To date, no proteins with the same domain organization as Nd2p and Nd22p have been identified in other organisms. However, many extracellular-matrix proteins can be identified in the databases with PSI, EGF and kelch domains, whether on the same or on different molecules. For example, attractin contains kelch, PSI and EGF repeats (27). The integrins, Traffic 2004; 5: 493–502

Extracellular-Matrix-Like Proteins in Membrane Fusion

involved in cell adhesion and signal transduction, are dimeric proteins that contain EGF, PSI and b-propeller domains, although not kelch repeats. The b-propeller is present on the a subunit and the PSI and EGF domains on the b subunit. The formation of structurally elaborate heterodimers gives the active conformation necessary for integrin function. Since, in many cases, domains can act similarly, whether they are carried by one or by several molecules, we can extrapolate that Paramecium VEMIFs could function much like extracellular-matrix proteins, through protein/protein interactions via their conserved domains. This suggests the occurrence of protein/protein interactions on the lumenal side of the trichocyst membrane. Possible occurrence of VEMIFs in other species Unless they have been experimentally identified, extracellular-matrix-like proteins associated with a vesicular compartment are difficult to recognize from their sequence, among the tremendous number of true extracellular matrix proteins. However, a few examples of potential VEMIFs do exist. In plants, the BP-80 family of proteins first discovered in pea [AtELP in Arabidopsis thaliana (28), or PV72 in pumpkin (29)] are type I membrane proteins with EGF domains present in vesicles destined to fuse with the lytic vacuole (12,13,30). In T. gondii, and in other apicomplexan parasites, proteins similar in organization have been described: TgMIC in T. gondii and PfTRAP in Plasmodium falciparum for example (11). Candidate VEMIFs also exist in animals: the milk fat globule EGF8 (MFG-E8) protein, in addition to its role in the interaction with integrins at the cell surface, has a function in secreted vesicles where it seems to regulate the number of secreted exosomes (31). The list is not exhaustive, but we can infer from these observations that VEMIFs are widespread proteins in membrane traffic regulation. Tentative model for the role of VEMIFs in membrane fusion The main role of extracellular-matrix proteins is in cell adhesion and in ligand binding and signal transduction. The underlying mechanisms involve lateral protein/protein interactions. Another role of extracellular matrix proteins is to prepare membrane fusion. Cell-to-cell membrane fusion in fertilization (32) and myoblast fusion in myogenesis (33) both require extracellular-matrix proteins of partner membranes for recognition and membrane fusion. After allowing cell-to-cell adhesion, these proteins recruit transmembrane proteins, such as the tetraspanin CD9, to the contact site to promote membrane fusion. In the case of VEMIFs, a role in intracellular membrane fusion cannot involve trans interactions between their extracellular-matrix-like domains, since they do not face each other in the membranes to be fused. This role is rather played by SNAREs on the cytoplasmic faces of the membranes. However, the role of recruiting fusion proteins at the right place could well be assumed by VEMIFs. Traffic 2004; 5: 493–502

Indeed, they can display lateral protein/protein interactions, which, in turn, can promote the organization of microdomains where fusion proteins are recruited. The recruited proteins could just as well be SNAREs as small proteins with four transmembrane segments such as tetraspanin (34) or proteolipid subunits of the V-ATPase, also candidates to be involved in membrane fusion (35). In Paramecium, the inactivation of VEMIF function, either by mutation or by gene silencing of the ND genes, prevents assembly of the membrane fusion site in the trans partner, the plasma membrane. In contrast, the secretory partner, the trichocyst, is perfectly positioned in nd mutants, indicating that the mutated proteins, even Nd21p and Nd2-4p, which lack transmembrane domains and cytoplasmic tails, do not impair sorting of trichocyst proteins or escorting of the trichocyst to the plasma membrane. The only affected function is the preparation of membrane fusion. To understand how this type of protein can transmit a message from the secretory compartment to the plasma membrane, further work is needed on the structure of the cytoplasmic tail of Nd2p and Nd22p as well as proteins with which they interact. These proteins should be involved in connecting trichocyst and plasma membranes and triggering assembly of rosette particles. Nd9p, a Paramecium protein with Armadillo-like repeats (36), also found in b-catenin, is a good candidate to carry out this function.

Materials and Methods Strains and culture conditions The wild type strain used in this study was Paramecium tetraurelia stock d4-2, derived from stock 51 (37). The mutations nd2-1 and nd2-2 were isolated after nitrosoguanidine treatment F. Ruiz (7; unpublished). The mutations nd2-3 and nd2-4 have been recently isolated in the laboratory (Keller, Yanagi, Froissard, Koll and Cohen, unpublished). The mutations nd126-1 and nd169-1 come from wild type stocks and were transferred to stock 51 background (38). Cells were grown at 27 °C in grass infusion (Wheat Grass Powder, Pines International, Lawrence, KS), bacterized with Klebsiella pneumoniae the day before use, and supplemented with 0.4 mg/mL b-sitosterol (39). Monitoring exocytosis To monitor trichocyst exocytosis, a sample of living cells is triggered with a saturated solution of the fixative picric acid and examined under dark-field light microscopy with a 10
objective. Wild type cells are surrounded with a crown of discharged trichocysts, whereas mutants or silenced cells appear naked. Microinjection experiments Microinjections were made under an inverted Nikon phase-contrast microscope, using a Narishige micro499

Froissard et al. Table 1: Oligonucleotides used in this study Primer

50 to 30 Sequence

Use

A B C D E F G H I J K L M N O P Q R S T U V

GTTATTTGTCAGGTAATGTTGC GTTGGAATTAGTTGTCTTGGC TAGATACTCAATCAATGACC CTTGCAGGTAGATATTTACC CTGAAGTAGGCTTCTTTAGC TTTTGATAGCTAGGTTTGGG ATTTCATCGACTTCTCCTTC CTCTGCAAAACCGAATAGATG CAAAGTCAACAATCGGATGCA CAACATTACCTGACAAATAACC CATAATCCTTCAACAGATCC TGTCTAAGGAACAGAATCTGG AACCTTTGGAGGTACCATCA AATTCCAACTCCTACTCTCC TGCAAAATGTATGAGGTTGAC TTCACTTAGATCCTGTTGTTG ATGCCAGAGTTATCAGAAGATG TACATAAGCACTAGTTTTACCAC GTTAGTTGTATAGGTTGTAG ATTGGCATCACATAGCAGG GGCCACGCGTCGACTAGTACTTTTTTTTTTTTVN GGCCACGCGTCGACTAGTAC

Reverse PCR for ND22 Reverse PCR for ND22 Reverse PCR for ND22 PCR of ND2 for complementation PCR of ND2 for complementation PCR of ND2 for silencing PCR of ND2 for silencing PCR of ND22 for complementation PCR of ND22 for complementation PCR of ND22 for silencing and probe PCR of ND22 for silencing PCR of ND22 for silencing (short fragment) PCR of ND22 for silencing (short fragment) PCR of ND22 for probe PCR of ND169 for complementation PCR of ND169 for complementation PCR of ND169 for silencing PCR of ND169 for silencing PCR for ND169 30 -RACE Nested PCR for ND169 30 -RACE Reverse transcription of ND169 mRNA PCR for ND169 30 -RACE

manipulation device, and an Eppendorf air-pressure microinjector. Transfer of cytoplasm to determine the site of action of the mutations was as described by Lefort-Tran et al. (10). DNA microinjections were performed after filtration on Millex-GV 0.22 mm (Millipore, Billerica, MA), precipitation and dissolution in water at a final concentration between 10 and 30 mg/mL. When DNA was microinjected into wildtype cells, they were pretreated with a solution of the vital secretagogue aminoethyldextran (40) at 0.01% to trigger trichocyst discharge and avoid disturbance during the microinjection procedure. Gene cloning The complementation cloning of the ND2 and ND169 genes was performed by DNA microinjection into the macronucleus of the nd2-1 and nd169-1 mutants, respectively, using the sib selection procedure and the indexed library described by Keller & Cohen (18). RNA extractions and Northern blots Total RNA was prepared essentially according to the method of Chomczynski & Sacchi (41) using the Trizol reagent (Gibco BRL, Cergy-Pontoise, France), except that the cells were lysed by vortexing in the presence of glass beads. Total RNA was fractionated on formaldehyde/ 1.25% agarose gels, and transferred to positively charged nylon membranes (Ambion, Austin, TX) by capillarity. Hybridization was carried out at 60 °C in 6
SSC, 2
Denhardt’s solution 0.1% SDS (42). The membranes were then washed at the same temperature with decreas500

ing concentrations of SSC, in the presence of 0.1% SDS as follows: 2
SSC for 30 min and 0.2
SSC for 30–45 min (42). Images were obtained by using a Phosphorimager (Molecular Dynamics, Sunnyvale, CA). Hybridization was quantified by ImageQuant Software (Molecular Dynamics).

PCR amplifications and RT-PCR Inverse PCR was used to clone the ND22 gene. One microgram of EcoRV digested macronuclear DNA (18) was circularized using 400 mm of T4 DNA ligase (New England Biolabs, Beverly, MA) in 100 mL, incubated overnight at room temperature (RT) and precipitated. Nested PCR was performed with a first amplification on 500 ng of circularized DNA using the Expand Long Template PCR system (Roche Diagnostics, Mannheim, Germany) with primers A and B (Table 1, Figure 1) and a second round of amplification on 1 mL of 1/50 dilution of the first round product, after primer removal using the QIAquick PCR Purification Kit (Qiagen, Hilden, Germany), using primers A and C (Table 1, Figure 1). The reactions were performed with one cycle of denaturation (1 min, 92°), and 30 cycles of denaturation (30 s, 92°), annealing (45 s, 50°) and extension (7 min, 72°), with a final extension (10 min, 72°). DNA used for overexpression and silencing was obtained by PCR amplification using the Expand Long Template PCR system (Roche Diagnostics, Mannheim, Germany) as described previously (36) with primers D–G for ND2, H–N for ND22 and O–R for ND169 (Table 1, Figure 2). Amplifications were performed as described above, except an annealing temperature of 54 °C. Traffic 2004; 5: 493–502

Extracellular-Matrix-Like Proteins in Membrane Fusion

The 30 extremity of the ND169 mRNA was determined using the 30 -RACE System for rapid amplification of cDNA ends (Invitrogen Life Technol., Carlsbad, CA) with the oligonucleotides S-V (Table 1).

DNA sequencing and sequence analyses Random in vitro transposition was performed on the p77o7 insert using the Genome Priming System (Biolabs) to provide sequencing templates with internal primer sites. The transposition reaction was done with the Transprimer-1 (KanR) using the manufacturer’s instructions. The clones obtained were sequenced with an ABI 310 sequencer (Perkin Elmer, Foster City, CA) using the BigDye Primer Cycle Sequencing Ready Reaction Kit (Perkin Elmer). The ND22 PCR product was sequenced using primer walking. The p89m7 insert was sequenced by MWG Biotech. Both insert sequences were verified on both strands by primer walking. The ND2, ND22 and ND169 sequences have been deposited in the EMBL/Genebank/DDBJ databases under the accession numbers AJ437480, AJ566620 and AJ582224, respectively. The characterization and annotation of DNA and protein sequences were first performed with DNA Strider (43), then with Artemis (44), using the ciliate genetic code where UAA and UAG encode glutamine instead of stop. Homology searches were performed with the BLAST program (45). Secondary structures and motifs contained in the proteins were detected using InterProScan (46), SMART (47), Pfam (48), SignalP (49), and TMHMMPred (50).

Preparation of radioactive probes The ND2 template for 32P labeling was a PCR fragment used for overexpression, and the ND22 template was a PCR fragment obtained with primer J and primer N: (Table 1, Figure 2) on macronuclear DNA. Probes were synthesized by a32PdATP incorporation using a Random Primers Labeling System (Gibco-BRL, Cergy-Pontoise, France), according to the supplier’s protocol.

Freeze-fracture electron microscopy Freeze fracture of mutant and silenced cells was performed on small pools of cells as described by Froissard et al. (51). To prevent unwanted stimulation during the fixation step, cells were incubated in medium supplemented with 20 mM MgCl2 for 3 min before transfer into 1% glutaraldehyde, 10 mM Na-phosphate buffer pH 7.0 for 1 h at room temperature and two washes in the same buffer without glutaraldehyde. Cells were placed onto hydrophilic 10-mm-thick round copper plates, then rapidly covered by another plate. The rest of the treatment, including freeze fracture and electron microscopy, was as previously described by Bonnemain et al. (17). Traffic 2004; 5: 493–502

ND7-GFP fusion gene A sequence corresponding to the GFP with a codon usage adapted to the one of Paramecium (E. Meyer and J. Cohen, unpublished) was inserted into the ND7 gene at the unique SpeI site, close to the C-terminus of the encoded protein. The construction was microinjected into the macronucleus of nd7-1 mutant cells and grown for a few days. The efficiency of transformation was assessed by the rescue of trichocyst exocytosis capacity. Rescued clones were examined under a Zeiss epifluorescence microscope equipped with a Roper Coolsnap-CF intensifying camera, using GFP filters.

Acknowledgments We thank France Koll, Delphine Gogendeau and Linda Sperling for critical reading of the manuscript. Financial support from the Association pour la Recherche contre le Cancer to M.F. and from the Microbiology Program of the Ministe´re de la Recherche are gratefully acknowledged.

References 1. Jahn R, Sudhof TC. Membrane fusion and exocytosis. Annu Rev Biochem 1999;68:863–911. 2. Dellı´Angelica EC, Payne GS. Intracellular cycling of lysosomal enzyme receptors: cytoplasmic tailsı´tales. Cell 2001;106:395–398. 3. Vayssie L, Skouri F, Sperling L, Cohen J. Molecular genetics of regulated secretion in Paramecium. Biochimie 2000;82:269–288. 4. Plattner H. My favorite cell – Paramecium. Bio Essays 2002;24: 649–658. 5. Plattner H, Kissmehl R. Dense-core secretory vesicle docking and exocytotic membrane fusion in Paramecium cells. Biochim Biophys Acta 2003;1641:185–193. 6. Beisson J, Lefort-Tran M, Pouphile M, Rossignol M, Satir B. Genetic analysis of membrane differentiation in Paramecium. Freeze-fracture study of the trichocyst cycle in wild-type and mutant strains. J Cell Biol 1976;69:126–143. 7. Pollack S. Mutations affecting the trichocysts in Paramecium aurelia. I. Morphology and description of the mutants. J Protozool 1974;21: 352–362. 8. Skouri F, Cohen J. Genetic approach to regulated exocytosis using functional complementation in Paramecium: identification of the ND7 gene required for membrane fusion. Mol Biol Cell 1997;8:1063–1071. 9. Bork P, Doerks T, Springer TA, Snel B. Domains in plexins: links to integrins and transcription factors. Trends Biochem Sci 1999;24:261–263. 10. Lefort-Tran M, Aufderheide K, Pouphile M, Rossignol M, Beisson J. Control of exocytotic processes: cytological and physiological studies of trichocyst mutants in Paramecium tetraurelia. J Cell Biol 1981;88:301–311. 11. Meissner M, Reiss M, Viebig N, Carruthers VB, Toursel C, Tomavo S, Ajioka JW, Soldati D. A family of transmembrane microneme proteins of Toxoplasma gondii contain EGF-like domains and function as escorters. J Cell Sci 2002;115:563–574. 12. Watanabe E, Shimida T, Kuroyanagi M, Nishimura M, Hara-Nishimura I. Calcium-mediated association of a putative vacuolar sorting receptor PV72 with a propeptide of 2S albumin. J Biol Chem 2002;277: 8708–8715. 13. Shimida T, Watanabe E, Tamura K, Hayashi Y, Nishimura M, HaraNishimura I. A vacuolar sorting receptor PV72 on the membrane of vesicles that accumulate precursors of seed storage proteins (PAC vesicles). Plant Cell Physiol 2002;43:1086–1095.

501

Froissard et al. 14. Sperling L, Dessen P, Zagulski M, Pearlman RE, Migdalski A, Gromadka R, Froissard M, Keller AM, Cohen J. Random sequencing of Paramecium somatic DNA. Eukaryot Cell 2002;1:341–352. 15. Aufderheide K. The effective sites of some mutations affecting exocytosis in Paramecium tetraurelia. Mol Gen Genet 1978;165:199–205. 16. Pouphile M, Lefort-Tran M, Plattner H, Rossignol M, Beisson J. Genetic dissection of the morphogenesis of exocytosis sites in Paramecium. Biol Cell 1986;56:151–162. 17. Bonnemain H, Gulik-Krzywicki T, Grandchamp C, Cohen J. Interactions between genes involved in exocytotic membrane fusion in Paramecium. Genetics 1992;130:461–470. 18. Keller AM, Cohen J. An indexed genomic library for Paramecium complementation cloning. J Eukaryot Microbiol 2000;47:1–6. 19. Ruiz F, Vayssie L, Klotz C, Sperling L, Madeddu L. Homology-dependent gene silencing in Paramecium. Mol Biol Cell 1998;9:931–943. 20. Dessen P, Zagulski M, Gromadka R, Plattner H, Kissmehl R, Meyer E, Betermier M, Schultz JE, Linder JU, Pearlman RE, Kung C, Forney J, Satir BH, Van Houten JL, Keller A-M, . Paramecium genome survey: a pilot project. Trends Genet 2001;17:306–308. 21. Galvani A, Sperling L. Transgene-mediated post-transcriptional gene silencing is inhibited by 30 non-coding sequences in Paramecium. Nucl Acids Res 2001;29:4387–4394. 22. Xue F, Cooley L. kelch encodes a component of intercellular bridges in Drosophila egg chambers. Cell 1993;72:681–693. 23. Adams J, Kelso R, Cooley L. The kelch repeat superfamily of proteins: propellers of cell function. Trends Cell Biol 2000;10:17–24. 24. Appella E, Weber IT, Blasi F. Structure and function of epidermal growth factor-like regions in proteins. FEBS Lett 1988;231:1–4. 25. Engel J. EGF-like domains in extracellular matrix proteins: localized signals for growth and differentiation? FEBS Lett 1989;251:1–7. 26. Bonifacino JS, Traub LM. Signals for sorting of transmembrane proteins to endosomes and lysosomes. Annu Rev Biochem 2003;72:395–447. 27. He L, Gunn TM, Bouley DM, Lu X-Y, Watson SJ, Schlossman SE, Duke-Cohan JS, Barsh GS. A biochemical function for attractin in agouti-induced pigmentation and obesity. Nat Genet 2001;27:40–47. 28. Ahmed SU, Bar-Peled M, Raikhel NV. Cloning and subcellular location of an Arabidopsis receptor-like protein that shares common features with protein-sorting receptors of eukaryotic cells. Plant Physiol 1997;114:325–336. 29. Shimida T, Kuroyanagi M, Nishimura M, Hara-Nishimura I. A pumpkin 72-kDa membrane protein of precursor-accumulating vesicles has characteristics of a vacuolar sorting receptor. Plant Cell Physiol 1997;38: 1414–1420. 30. Paris N, Rogers SW, Jiang L, Kirsch T, Beevers L, Phillips TE, Rogers JC. Molecular cloning and further characterization of a probable plant vacuolar sorting receptor. Plant Physiol 1997;115:29–39. 31. Oshima K, Aoki N, Kato T, Kitajima K, Matsuda T. Secretion of a peripheral membrane protein, MFG-E8, as a complex with membrane vesicles. Eur J Biochem 2002;269:1209–1218. 32. Cuasnc PS, Ellerman DA, Cohen DJ, Busso D, Morgenfeld MM, Da Ros VG. Molecular mechanisms involved in mammalian gamete fusion. Arch Med Res 2001;32:614–618.

502

33. Schwander M, Leu M, Stumm M, Dorchies OM, Ruegg UT, Schitttny J, Mu¨ller U. b1 integrins regulate myoblast fusion and sarcomere assembly. Dev Cell 2003;4:673–685. 34. Hemler ME. Specific tetraspanin functions. J Cell Biol 2001;155: 1103–1107. 35. Peters C, Bayer M, Bu¨hler S, Andersen J, Mann M. Trans-complex formation by proteolipid channel in the terminal phase of membrane fusion. Nature 2001;409:581–588. 36. Froissard M, Keller AM, Cohen J. ND9P, a novel protein with armadillolike repeats involved in exocytosis: physiological studies using allelic mutants in Paramecium. Genetics 2001;157:611–620. 37. Sonneborn TM. Paramecium aurelia. In: Handbook of Genetics. King R. C. Ed, New York: Plenum, 1974: 469–594. 38. Nyberg D. Genetic analysis of trichocyst discharge of the wild stocks of Paramecium tetraurelia. J Protozool 1978;25:107–112. 39. Sonneborn TM. Methods in Paramecium research. Meth Cell Physiol 1970;4:242–335. 40. Plattner H, Matt H, Kersken H, Haacke B, Sturzl R. Synchronous exocytosis in Paramecium cells. I. A novel approach. Exp Cell Res 1984;151:6–13. 41. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate- phenol-chloroform extraction. Anal Biochem 1987;162:156–159. 42. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: A laboratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 1989. 43. Marck C. ‘DNA Strider’: a ‘C’ program for the fast analysis of DNA and protein sequences on the Apple Macintosh family of computers. Nucl Acids Res 1988;16:1829–1836. 44. Rutherford K, Parkhill J, Crook J, Horsnell T, Rice P, Rajandream M-A, Barrell B. Artemis: sequence visualisation and annotation. Bioinformatics 2000;16:944–945. 45. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucl Acids Res 1997;25: 3389–3402. 46. Apweiler R. Protein sequence databases [Process Citation]. Adv Protein Chem 2000;54:31–71. 47. Schultz J, Copley RR, Doerks T, Ponting CP, Bork P. SMART: a webbased tool for the study of genetically mobile domains. Nucl Acids Res 2000;28:231–234. 48. Batemen A, Birney E, Durbin R, Eddy SR, Howe KL, Sonnhammer EL. The Pfam protein families database. Nucl Acids Res 2000;28: 263–266. 49. Nielsen H, Brunak S, von Heijne G. Machine learning approaches for the prediction of signal peptides and other protein sorting signals. Protein Eng 1999;12:3–9. 50. Moller S, Croning MDR, Apweiler R. Evaluation of methods for the prediction of membrane spanning regions. Bioinformatics 2001;17: 646–653. 51. Froissard M, Kissmehl R, Dedieu JC, Gulik-Krzywicki T, Plattner H, Cohen J. NSF is required to organize functional exocytotic microdomains in Paramecium. Genetics 2002;161:643–650.

Traffic 2004; 5: 493–502