Identification, sequence, and expression of an invertebrate caveolin gene family from the nematode Caenorhabditis elegans. Implications for the molecular evolution of mammalian caveolin genes.

Caveolae are vesicular organelles that represent an appendage of the plasma membrane. Caveolin, a 21-24-kDa integral membrane protein, is a principal component of caveolae membranes in vivo. Caveolin has been proposed to function as a plasma membrane scaffolding protein to organize and concentrate signaling molecules within caveolae, including heterotrimeric G proteins (α and βγ subunits). In this regard, caveolin interacts directly with Gα subunits and can functionally regulate their activity. To date, three cDNAs encoding four subtypes of caveolin have been described in vertebrates. However, evidence for the existence of caveolin proteins in less complex organisms has been lacking. Here, we report the identification, cDNA sequence and genomic organization of the first invertebrate caveolin gene, Cavce (for caveolin from Caenorhabditis elegans). The Cavce gene, located on chromosome IV, consists of two exons interrupted by a 125-nucleotide intron sequence. The region of Cavce that is strictly homologous to mammalian caveolins is encoded by a single exon in Cavce. This suggests that mammalian caveolins may have evolved from the second exon of Cavce. Cavce is roughly equally related to all three known mammalian caveolins and, thus, could represent a common ancestor. Remarkably, the invertebrate Cavce protein behaves like mammalian caveolins: (i) Cavce forms a high molecular mass oligomer, (ii) assumes a cytoplasmic membrane orientation, and (iii) interacts with G proteins. A 20-residue peptide encoding the predicted G protein binding region of Cavce possesses “GDP dissociation inhibitor-like activity” with the same potency as described earlier for mammalian caveolin-1. Thus, caveolin appears to be structurally and functionally conserved from worms to man. In addition, we find that there are at least two caveolin-related genes expressed in C. elegans, defining an invertebrate caveolin gene family. These results establish the nematode C. elegans as an invertebrate model system to study caveolae and caveolin in vivo.

H-ras (G12V) and v-abl); down-regulation of caveolin protein coincides with the disappearance of caveolae from these transformed cells (28). Also, recombinant overexpression of caveolin in caveolin-deficient cells can lead to the formation of caveolae in lymphocytes (29) and insect Sf21 cells (30).
However, there are certain cell lines which morphologically contain caveolae, but fail to express caveolin (31). This finding has suggested that other caveolin-related proteins may exist that are immunologically distinct from caveolin. In support of this notion, two novel caveolin-related proteins have recently been identified and cloned. These proteins, termed caveolin-2 and caveolin-3, are the products of separate caveolin genes (27,32,33). Thus, caveolin (retermed caveolin-1) is the first member of a multigene family (27).
Caveolins 1, 2, and 3 are structurally homologous proteins but are immunologically distinct molecules; they have different but overlapping tissue distributions (27,(32)(33)(34). For example, the expression of caveolin-3 is absolutely muscle-specific (skeletal and cardiac muscle cells) (32,34). Caveolin-1 is not expressed in these striated muscle tissues, but smooth muscle cells co-express caveolins 1 and 3 (34). Furthermore, caveolin-1 and -2 are co-expressed in adipocytes and share the same overlapping tissue distribution (27). Thus, a given mammalian cell, such as smooth muscle cells or fibroblasts, may co-express up to three or four immunologically distinct caveolin protein products.
Due to the complexity inherent in studying mammalian systems, we have chosen to search for caveolin genes in model invertebrate organisms. Here, we report the identification of the first invertebrate caveolin gene, Cav ce , in the nematode Caenorhabditis elegans. Cav ce is roughly equally related to all three known mammalian caveolins and, thus, could represent a common ancestor. The identification and sequencing of Cav ce represents a starting point for reverse genetics experiments designed to isolate animals mutated in a caveolin gene.

EXPERIMENTAL PROCEDURES
Materials-The cDNAs for mammalian caveolins 1, 2, and 3 were as we described previously (27,31,32). The C. elegans EST clone yk74b2 was obtained from Dr. Yuji Kohara (Gene Library Laboratory, National Institute of Genetics, Japan). Rabbit polyclonal IgG and mouse monoclonal IgG (mAb 2297) directed against caveolin-1 were the generous gift of Dr. John R. Glenney (Transduction Laboratories, Lexington, KY). Anti-Myc IgG (9E10) and anti-FLAG (M2) IgG were from Santa Cruz Biotechnologies and IBI, respectively. Other reagents were from the following commercial suppliers: prestained protein markers, fetal bovine serum, and other cell culture reagents (Life Technologies, Inc.); and Slow-Fade anti-fade reagent (Molecular Probes, Inc.). Peptide synthesis was performed by the Biopolymers Facility at the Massachusetts Institute of Technology.
Recombinant Expression of Cav ce -Epitope-tagged forms of the Cav ce cDNA were subcloned into the MCS (KpnI/XbaI) of the vector pCB7 for expression in COS-7 cells. A Myc epitope was incorporated into the N terminus (MEQKLISEEDLNGG-Cav ce ) and a FLAG epitope tag was incorporated into the C terminus (Cav ce -GGDYKDDDDK); GG was placed as a spacer between the epitopes and the Cav ce coding sequence, as performed previously for caveolins 1, 2, and 3 (11,27,32,33,35,36). Using this scheme, both singly tagged and doubly tagged versions of Cav ce were produced. Constructs were transiently transfected into COS-7 cells, as described previously. Expression was detected using mAb 9E10 that recognizes the Myc epitope or mAb M2 that recognizes the FLAG epitope.
Velocity Gradient Centrifugation-The molecular mass of Cav ce was estimated as described previously for mammalian caveolins 1, 2, and 3 (19,27,32). Briefly, samples were dissociated in Mes-buffered saline containing 60 mM octyl glucoside. Solubilized material was loaded atop a 5-40% linear sucrose gradient and centrifuged at 50,000 rpm (340,000 ϫ g) for 10 h in an SW60 rotor (Beckman Instruments). Gradient fractions were collected from above and subjected to immunoblot analysis. Molecular mass standards for velocity gradient centrifugation were as described previously (19,27,32). Note that caveolin-2 is a dimer of ϳ40 kDa, and it correctly migrates between the 29-kDa and 66-kDa molecular mass standards using this velocity gradient system (27).
Immunofluorescence-All reactions were performed at room temperature. Transfected COS-7 cells were briefly washed three times with PBS and fixed for 20 min in PBS containing 4% paraformaldehyde. Fixed cells were rinsed with PBS and treated with 25 mM NH 4 Cl in PBS for 10 min to quench free aldehyde groups. Cells were then permeabilized with 0.1% Triton X-100 for 10 min and washed with PBS (three times, 10 min each). For double-labeling, the cells were then successively incubated with PBS, 3% bovine serum albumin containing: (i) 50 g/ml each of normal goat and donkey IgGs, (ii) a 1:400 dilution of mAb 9E10 and 40 g/ml anti-caveolin-1 polyclonal IgG, and (iii) lissamine rhodamine B sulfonyl chloride-conjugated goat anti-mouse antibody (5 g/ml) and fluorescein isothiocyanate-conjugated donkey anti-rabbit antibody (5 g/ml). Alternatively, a 1:400 dilution of anti-FLAG IgG (mAb M2) was utilized in step (ii). The first incubation was 30 min, while primary and secondary antibody reactions were 60 min each. Cells were washed three time with PBS between incubations. Slides were mounted with Slow-Fade anti-fade reagent and observed under a Bio-Rad MR600 confocal fluorescence microscope.
Cell Fractionation-Transfected COS-7 cells grown to confluence in 100-mm dishes were used to prepare caveolin-enriched membrane fractions, essentially as we have described previously (24, 26, 31, 36 -38). However, two specific modifications were introduced to allow the purification of caveolin-rich domains without the use of detergent (18,34). Triton X-100 was replaced with sodium carbonate buffer, and a sonication step was introduced to more finely disrupt cellular membranes (18,34). After two washes with ice-cold PBS, cells (two confluent 100-mm dishes) were scraped into 2 ml of 500 mM sodium carbonate, pH 11.0. Homogenization was carried out sequentially in the following order using: (i) a loose fitting Dounce homogenizer (10 strokes); (ii) a Polytron tissue grinder (three 10 s bursts; Kinematica GmbH, Brinkmann Instruments, Westbury, NY); and (iii) a sonicator (three 20-s bursts; Branson Sonifier 250, Branson Ultrasonic Corp., Danbury, CT). The homogenate was then adjusted to 45% sucrose by addition of 2 ml of 90% sucrose prepared in MBS (25 mM Mes, pH 6.5, 0.15 M NaCl) and placed at the bottom of an ultracentrifuge tube. A 5-35% discontinuous sucrose gradient was formed above (4 ml of 5% sucrose, 4 ml of 35% sucrose; both in MBS containing 250 mM sodium carbonate) and centrifuged at 39,000 rpm for 16 -20 h in an SW41 rotor (Beckman Instruments). A light-scattering band confined to the 5-35% sucrose interface was observed that contained caveolin, but excluded most other cellular proteins.
Immunoblotting of Gradient Fractions-Gradient fractions were separated by SDS-PAGE (10% acrylamide) and transferred to nitrocellulose. After transfer, nitrocellulose sheets were stained with Ponceau S to visualize protein bands and subjected to immunoblotting with anticaveolin-1 IgG (mAb 2297; 1:400) or with 9E10 ascites (1:500) to visualize epitope-tagged Cav ce . For immunoblotting, incubation conditions were as described by the manufacturer (Promega; Amersham Corp.), except we supplemented our blocking solution with both 1% bovine serum albumin and 1% non-fat dry milk (Carnation). The amount of caveolin-1 and Cav ce that remains associated with caveolin-enriched fractions was estimated by immunoblotting, as described elsewhere (37). Quantitation was performed with a Molecular Dynamics computing densitometer. To ensure that these estimates were made in the linear range, we used multiple autoradiographic exposures and monitored their linearity using the densitometer, essentially as described previously (39).
GTP Hydrolysis and GTP␥S Binding Assays-Trimeric G i2 purified from bovine spleen was provided by T. Asano (40). Trimeric G o purified from bovine brain was provided by T. Katada (41). Steady-state GTP hydrolysis activity was examined as we described previously (42). Briefly, the assay was performed for 20 min at 37°C in the presence of 20 M Mg 2ϩ with 10 nM G protein. The GTP␥S binding assay was performed as we described elsewhere (42) by incubating 10 nM G o with GTP in the presence of 20 M Mg 2ϩ for 2 min at 37°C. The Cav cederived polypeptide contained the sequence DCVWRLNHTVFTAVR-LFIYR, corresponding to amino acids 132-151 of Cav ce .
Cloning and Analysis of the Genomic Copies of Murine Caveolins 2 and 3-Probes corresponding to the cDNAs of both caveolin-2 and caveolin-3 were labeled with digoxigenin, according to the manufacter's instructions (Boehringer Mannheim). These labeled probes were used to screen a previously described -2 murine genomic library (43,44) (generous gift of H. Wu and R. Jaenisch, Whitehead Institute). A total of 100,000 plaque-forming units were screened; three independent genomic clones were isolated for caveolins 2 and 3.

RESULTS AND DISCUSSION
A Caveolin Homologue from the Nematode C. elegans: Identification, Sequence, and Genomic Organization-In order to identify an invertebrate caveolin gene, we searched existing data bases of model invertebrate organisms using the protein sequences of mammalian caveolins 1, 2, and 3. Using this approach, we identified a short EST sequence potentially encoding a caveolin homologue in the nematode C. elegans.
This EST clone (yk74b2, accession no. D71360) was obtained from Dr. Yuji Kohara (C. elegans Genome Project) for further study. Analysis and sequencing of clone yk74b2 revealed that it contains an open reading frame encoding a caveolin-related protein of 235 amino acids (708 bases). Fig. 1 shows the deduced protein sequence. For simplicity, we have termed this protein caveolin ce (or Cav ce ; for caveolin from C. elegans). Cav ce contains a 151-amino acid N-terminal domain, a 32-amino acid membrane spanning segment, and a 52-amino acid C-terminal domain, with a predicted molecular mass of 26,293 Da. Cav ce is substantially longer than the known mammalian caveolins, which are ϳ147-178-amino acid in length.
Using the C. elegans Blast server, we compared the deduced protein sequence of Cav ce with the sequences of catalogued genomic clones prepared by the C. elegans Genome Project. Our searches indicate that the Cav ce gene is included within the ϳ40-kb insert of cosmid T13F2 which maps to chromosome IV, more specifically within the major sperm protein region. This direct nucleotide sequence comparison between the Cav ce cDNA and the corresponding genomic sequence of cosmid T13F2 allowed us to deduce the intron-exon organization of the Cav ce gene. As illustrated in Fig. 2, the Cav ce gene consists of two exons interrupted by a 125-base intron sequence. The 5Ј and 3Ј DNA sequences flanking the intron correspond to the accepted consensus sequences for donor and acceptor splice sites (Fig. 2B). Fig. 3 shows an alignment of the Cav ce protein sequence with the known protein sequences of mammalian caveolins. The region of highest homology is restricted to Cav ce residues 100 -235. This 135-amino acid region is 67% similar and 37% identical to the mammalian caveolins 1 and 3; 65% similarity and 32% identity were obtained by comparision with caveolin-2. Thus, Cav ce is roughly equally related to all three known mammalian caveolins and could represent a common ancestor. Interestingly, Cav ce residues 100 -235 are restricted to the second exon of Cav ce that encodes residues 93-235. Thus, the region of Cav ce that is strictly homologous to mammalian caveolins is encoded by a single exon in Cav ce . This suggests that mammalian caveolins may have evolved from this second exon of Cav ce .
This Cav ce region (residues 93-235) is 142 amino acids in length, approximating the size of mammalian caveolins. Also, this Cav ce region is predicted to contain the most important features of mammalian caveolins, the N-terminal oligomerization domain and G protein-interacting sequences, the membrane-spanning segment, and the C terminus.
Perhaps the first exon encoded by Cav ce evolved into a separate mammalian gene product. Interestingly, searches with the deduced protein sequence of the first coding exon revealed that this region of the Cav ce protein is most closely related to a region of the mammalian MARCKS protein (myristoylated alanine-rich C-kinase substrate) and a MARCKS-related protein termed MAC-MARCKS.  pound sign (#); a C-terminal cysteine residue that is a potential site for palmitoylation is indicated by an asterisk.

Recombinant Expression and Characterization of the Cav ce
Protein Product-To study the properties of the Cav ce protein, we constructed an epitope-tagged form of Cav ce for expression in mammalian cells. A Myc epitope tag was placed at its extreme N terminus and a FLAG epitope tag was placed at its extreme C terminus. Note that epitope tagging in this fashion does not affect the behavior of mammalian caveolins 1, 2, and 3, or their subcellular localization within caveolae membranes (11,27,32,33,35,36). Fig. 4A shows the recombinant expression of the epitopetagged Cav ce protein in COS-7 cells. Expression yielded a protein product of ϳ32-33 kDa. Similar results were obtained using singly tagged versions of Cav ce containing exclusively either the Myc or the FLAG epitope. Thus, as with mammalian caveolins (27,32), Cav ce migrates several kilodaltons higher than is predicted from its amino acid sequence.
Caveolin-1 forms a ϳ350-kDa homo-oligomer containing ϳ14 -16 caveolin monomers per oligomer (19,20). These homooligomers are thought to function as building blocks in the construction of caveolae membranes. Similarly, caveolin-3 forms homo-oligomers of the same size as caveolin-1 (32). In contrast, caveolin-2 exists as a homodimeric complex (27). In caveolin-1, the oligomerization domain has been localized to a 41-amino acid membrane proximal region of the cytoplasmic N-terminal domain using a battery of glutathione S-transferase-caveolin fusion proteins expressed in Escherichia coli (19). This region is highly conserved in all mammalian caveolin subtypes and invertebrate Cav ce .
Thus, we next investigated the oligomeric state of Cav ce . For this pupose, we employed an established velocity gradient system developed previously to study the oligomeric state of mammalian caveolins 1, 2, and 3 (19,27,32). Fig. 4B shows that Cav ce behaved as a high molecular mass complex, migrating between the 200-and 443-kDa molecular mass standards (peak fractions 7 and 8). The migration of caveolin-1 is shown for comparison (peak fraction 6). As expected, Cav ce migrated with a slightly higher molecular mass than caveolin-1, as Cav ce has a higher monomeric molecular mass than caveolin-1.
Subcellular Distribution and Membrane Topology of Cav ce -To determine if caveolin-1 and Cav ce co-fractionate when Cav ce is expressed in mammalian cells, we transiently expressed Cav ce in COS-7 cells and subjected them to subcellular fractionation. A protocol involving homogenization in sodium carbonate followed by equilibrium sucrose density centrifugation was used to separate membranes enriched in caveolin-1 from the bulk of cellular membranes and cytosolic proteins (18,34). In this fractionation scheme, immunoblotting with anti-caveolin-1 IgG can be used to track the position of caveolae-derived membrane domains (24, 26, 31, 36 -38). Fig. 5 shows that ϳ90 -95% of caveolin-1 and Cav ce co-fractionate and are localized to the same low density fractions (gradient fractions 5 and 6).
Immunostaining of COS-7 cells expressing epitope-tagged Cav ce revealed punctate fluorescence or micropatches along the surface of the cell and within the perinuclear region (Fig. 6,  left). A similar staining pattern has been observed for the distribution of mammalian caveolins (9, 11, 12, 27, 31-33, 35, 36, 47). In addition, double-labeling of cells co-transfected with Cav ce and caveolin-1 revealed significant co-localization of these distinct caveolin proteins (Fig. 6). This is consistent with results demonstrating their co-fractionation during subcellular fractionation (Fig. 5).
The transmembrane domain of mammalian caveolins is thought to form a hairpin loop within the membrane, allowing

FIG. 4. Recombinant expression of Cav ce in mammalian cells.
A, expression of epitope-tagged Cav ce in COS-7 cells yielded a protein product of ϳ32-33 kDa. B, velocity gradient analysis of Cav ce . COS-7 cells expressing epitoped-tagged Cav ce were solubilized and loaded atop a 5-40% sucrose gradient, as described previously for mammalian caveolin proteins. After centrifugation, fractions were collected and subjected to SDS-PAGE/Western blot analysis. Note that Cav ce migrates mainly in fractions 7 and 8. The migration of caveolin-1 is shown for comparison; caveolin-1 migrates mainly in fraction 6. Arrows mark the positions of molecular mass standards. In A and B, expression of Cav ce was detected with the mAb 9E10 that recognizes the Myc epitope. Similar results were obtained using mAb M2 that recognizes the FLAG epitope (not shown). Caveolin-1 was detected with mAb 2297. both the N-terminal domain and the C-terminal domain to remain entirely cytosolic (35,36,48). In accordance with this topology, caveolin is inaccessible to biotinylation probes that efficiently label other plasma membrane proteins within caveolae (19). This unusual membrane topology has been confirmed using several independent approaches, including epitope tagging (35,36). In this regard, immunolocalization of an N-terminal epitope tag or a C-terminal epitope tag attached to caveolin-1 requires detergent permeabilization to allow specific IgG access to the cytoplasm (35,36).
Here, we have performed the analogous experiment with Cav ce . As illustrated in Fig. 7, immunolocalization of Cav ce using the N-terminal Myc tag or the C-terminal FLAG tag required detergent permeabilization; no immunostaining was observed if the permeabilization step was omitted. Thus, it appears that Cav ce assumes the same unusual cytoplasmic membrane topology as mammalian caveolins, with both Nterminal and C-terminal domains facing the cytoplasm.
A Cav ce -derived Peptide Suppresses the GTPase Activity and GTP␥S Binding of Purified Heterotrimeric G Proteins-Mam-malian caveolins functionally interact directly with G protein ␣-subunits (18,21,24,27,32). This binding activity is encoded by a 20-amino acid membrane proximal region within caveolins 1, 2, and 3. Peptides encoding this region differentially affect the GTPase activity of G proteins, depending on the caveolin subtype. For example, the caveolin-1-derived peptide acts as a FIG. 6. Localization of Cav ce and caveolin-1 within a single cell. COS-7 cells were co-transfected with epitope-tagged Cav ce and untagged mammalian caveolin-1. Cav ce expression was detected with the mouse mAb M2 that recognizes the FLAG tag; caveolin-1 expression was detected using rabbit polyclonal IgG directed against caveolin-1. Control experiments using singly transfected populations of cells confirmed the specificity of these antibodies; no cross-reaction was observed (not shown). Transfected cells expressing both Cav ce and caveolin-1 were selected for imaging by laser confocal fluorescence microscopy. Primary antibodies were detected using distinctly tagged fluorescent secondary antibodies (rhodamine-conjugated for Cav ce (left) and fluorescein-conjugated for caveolin-1 (right)). Virtually identical results were obtained when mAb 9E10 that recognizes the Myc epitope was used to detect epitope-tagged Cav ce (not shown). GDI, while the same region of caveolin-2 acts as a GTPaseactivating protein (GAP) (24,27). The caveolin-3-derived peptide expresses both GDI and GAP activities: (i) at nanomolar concentrations, the caveolin-3-derived peptide stimulates their GTPase activity; and (ii) at micromolar concentrations, it suppresses their GTPase activity (32). We have suggested such GAP and GDI activities could function in concert to recruit and sequester G proteins in the GDP-liganded conformation within caveolae (27,32). In this two-step mechanism, GAP activity would first actively place the G protein in the inactive GDPbound state, and GDI activity would then hold the G protein in the inactive GDP-bound state by preventing GDP/GTP exchange (27,32).
The predicted G protein binding region of Cav ce is indicated in Fig. 1. We synthesized a peptide encoding this 20-amino acid region of Cav ce to evaluate its effect on the functional properties of heterotrimeric G proteins. Fig. 8 shows the effect of this Cav ce -derived peptide on the GTPase activity and GTP␥S binding of the purified trimeric G o protein. Note that the steadystate GTP hydrolysis activity of G o was dose-dependently inhibited by this peptide with an IC 50 value of 4 M. A similar inhibitory effect was also observed for trimeric G i2 , yielding an IC 50 value of 2 M. For both G o and G i2 , complete inhibition was observed at 10 M. Thus, the Cav ce -derived peptide negatively affects both G o and G i2 GTPase activity in a similar fashion. In addition, at a concentration of 10 M, the Cav ce -derived peptide completely abolished GDP/GTP exchange as measured by GTP␥S binding of purified G o . Thus, it appears that this Cav cederived peptide possesses "GDI-like activity" with the same potency as described earlier for mammalian caveolin-1 (24). A Based on the cDNA sequence of caveolin-3, the expected size of PCR products for lanes 1-4 is 153, 306, 456, and 118 bp, respectively. Identical results were obtained with the caveolin-3 cDNA and a genomic clone. C, proposed genomic organization of Cav ce and known mammalian caveolin genes. The structure of the chicken caveolin-1 gene was published previously by Glenney and Soppet (9); we confirmed the same genomic organization after cloning the murine caveolin-1 gene (data not shown). Note that in many cases Cav ce -2 is more similar to mammalian caveolin-2.
homologue of the mammalian G o protein has recently been identified in C. elegans (49).
Genomic Organization of Murine Caveolin-2 and Caveolin-3 Genes-Independently of our work on Cav ce , we cloned the genomic copies of the murine caveolin-2 and caveolin-3 genes. To our surprise, caveolins 2 and 3 appear to be intronless single exon genes. Using PCR, we attempted to determine their intron-exon organization. In both cases, the cDNAs for caveolins 2 and 3 were used as positive controls. Fig. 9 shows that the same size PCR products were obtained when either of the cDNAs or the genomic clones were used as the template. These results indicate that no introns are present within the caveolin-2 and caveolin-3 genes. These result provide further support for the hypothesis that mammalian caveolins evolved from a single exon, related to the second exon of the Cav ce gene. As the caveolin-1 gene contains three exons (9), it is possible that caveolin-1 evolved from caveolin-3. Caveolins 1 and 3 are the most closely related; the protein sequences of caveolins 1 and 3 are ϳ85% similar and ϳ65% identical. These results have interesting implications for understanding the molecular evolution of the mammalian caveolin gene family.
Are There Multiple Caveolin Genes in C. elegans?-Given the existence of multiple mammalian caveolin genes, we continued to search for other caveolin-related genes in C. elegans. Additional data base searches revealed a genomic sequence (accession no. Z77655; cosmid C56A3) that appeared similar to mammalian caveolins and Cav ce . This gene is located on chromosome V and is, thus, distinct from the Cav ce gene.
To determine if this gene is expressed, we used a PCR-based approach to clone the cDNA for this gene from a C. elegans cDNA library. This approach yielded an ϳ1-kb product of the appropriate size. DNA sequencing of multiple clones and direct sequencing of PCR products revealed an open reading frame encoding a protein of 351 amino acids (Fig. 10A), with a calculated molecular mass of 40,822.1 Da. For simplicity, we have termed this second novel caveolin-related protein Cav ce -2 and retermed Cav ce as Cav ce -1. Cav ce -2 contains a predicted 261amino acid N-terminal domain, a 29-amino acid membrane spanning segment, and a 61-amino acid C-terminal domain. Fig. 10B shows an alignment of the Cav ce -2 protein with the protein sequences of Cav ce -1 and known mammalian caveolin genes. The most homologous region of Cav ce -2 is a 138-amino acid stretch from residues 214 -351. This homologous region includes the putative N-terminal oligomerization and G protein binding domains, the membrane spanning region and the entire C-terminal domain. This 138-amino acid region of Cav ce -2 is ϳ44% similar and ϳ24% identical to the mammalian caveolins 1, 2, and 3. Perhaps surprisingly, Cav ce -2 is only ϳ28% identical to Cav ce -1. Thus, Cav ce -2 is roughly equally related to all three known mammalian caveolins and Cav ce -1. Also, analysis of the Cav ce -2 protein sequence using the Prosite data base reveals potential sites for phosphorylation and SH-3 domain binding ( Fig. 10A; see legend).
To ensure that both Cav ce -1 and Cav ce -2 are expressed, we performed Northern analysis using purified poly(A) ϩ RNA (Fig. 11). Hybridization with the Cav ce -1 cDNA probe revealed an mRNA species of ϳ1.4 kb, while hybridization with the Cav ce -2 cDNA probe demonstrated a 2.4-kb species. Little or no FIG. 11. Northern blot analysis of the expression of Cav ce -1 and Cav ce -2. Each lane contains ϳ5 g of poly(A) ϩ RNA prepared from eggs or mixed stages of C. elegans. Blots were first probed with the Cav ce -1 cDNA, stripped, and reprobed with the Cav ce -2 cDNA. Cav ce -1 and Cav ce -2 have different transcript sizes: ϳ1.4 kilobases for Cav ce -1 and ϳ2.4 kilobases for Cav ce -2. Note that the message for Cav ce -1 is more abundant in eggs, while the message for Cav ce -2 is roughly equal in eggs and mixed stages.
FIG. 12. Summary of known caveolin family members. The overall structure of the four known mammalian caveolin gene products (caveolins 1␣, 1␤, 2, and 3) is shown and compared with the structure of invertebrate Cav ce -1 and -2. Caveolin-1 exists as two isoforms (␣ and ␤) differing in their translation initiation sites; the ␤-isoform lacks the first 31 amino acids (36). All four mammalian caveolin products contain the invariant sequence FEDVIAEP within their hydrophillic N-terminal domains (27,32); this peptide sequence is FEDIFGEA in Cav ce -1 and FFEVFNEP in Cav ce -2. Both mammalian caveolins and invertebrate caveolins contain a characteristic and unusually long membrane spanning segment (TM) of ϳ29 -33 amino acids. Overall amino acid length, percent similarity, and identity to caveolin-1 and GDI or GAP activities are summarized at the right. *From Cav ce -1 residues 100 -235; @from Cav ce -2 residues 214 -351. cross-hybridization was observed even at lower stringency. It is also interesting to note that the message for Cav ce -1 was more abundant in eggs than mixed stages, suggesting developmental regulation of its expression.
Conclusions-The small soil nematode, C. elegans, has become an established model organism for studying developmental processes and signal transduction using genetic approaches. Its genome is relatively small, roughly 1 ⁄30 that of the human (50). During development, 131 of the 1090 total somatic cells undergo apoptosis (51). The mature adult contains only 302 neurons, which represent ϳ 1 ⁄3 of the total cells of the organism. C. elegans expresses many signaling molecules that have known mammalian counterparts, such as heterotrimeric G proteins, Ras proteins, and receptor-tyrosine kinases (49,52). Many of these signaling molecules have been shown to function as key regulators in C. elegans development. Moreover, the C. elegans system provides an opportunity to perform reversegenetic analysis in a moderately complex organism through (i) transposon insertion and excision mutagenesis (53) and (ii) the construction of transgenic animals (54).
Here, we have described the identification and characterization of two C. elegans homologues of mammalian caveolins, termed Cav ce -1 and Cav ce -2 (see Fig. 12 for a comparison). This critical first step will allow us to take advantage of the C. elegans system to study the proposed functions of caveolae and caveolin using the established in vivo genetic approaches outlined above. Thus, it is likely that multiple caveolin genes exist in other simple model organisms such as Drosophila melanogaster and Saccharomyces cerevisiae.