Flotillins are integral membrane proteins that are enriched in lipid rafts, which also contain caveolins and glycosylphosphatidylinositol (GPI) -anchored proteins (Bickel et al., 1997; Galbiati et al., 1998; Lang et al., 1998). Lipid rafts are segregated cell membrane domains that are thought to be very important in a variety of signal transduction events (Bickel, 2002). For example, lipid modification of Wnt proteins is crucial for their function, and lipidation of Drosophila Wnt-1 localizes this protein to specialized lipid raft domains (Willert et al., 2003; Zhai et al., 2004). Flotillin-1 has been implicated in axonal regeneration, insulin signaling, membrane scaffolding and cell polarity, phagocytosis, immune cell activation and a variety of signal transduction events in several adult cell types (e.g., Schulte et al., 1997; Baumann et al., 2000; Dermine et al., 2001; Salzer and Prohaska, 2001; Stuermer et al., 2001; Solomon et al., 2002; Rajendran et al., 2003; Slaughter et al., 2003; Neumann-Giesen et al., 2004). In neurons and astrocytes, Flotillins (also called Reggie in rat and goldfish) closely associate with fyn kinases and the GPI-linked proteins Thy1 and F3 (Stuermer et al., 2001) and are enriched at synapses (Kokubo et al., 2003). Thus, Flotillins are generally thought to be important components of the plasma membrane domains involved in the transduction of specialized signaling events.
Numerous signaling events regulate crucial developmental processes such as axis determination, germ layer formation, tissue inductions, and organogenesis. However, the normal role of Flotillins during embryonic development has not been explored in any animal; we find only one study reporting that Flotillin-1 levels are reduced in dihydroxyacetonephosphate acyltransferase mutant mice that display eye defects (Rodemer et al., 2003). Indeed, there is a paucity of descriptions of Flotillin expression patterns during development. In the initial description of Drosophila flotillin-1, whole-mount staining was performed only during germ-band shortening (stages 12–16) to demonstrate transcripts in the brain and ventral nerve cord (Galbiati et al., 1998). In an initial characterization of an antibody directed against Flotillin (called Neurolin), immunostaining was performed only in day 5 goldfish larvae to demonstrate staining on retinal ganglion cells and their axons (Paschke et al., 1992). As Flotillins are likely to have important roles in developmental signaling processes, herein we provide the first report of the sequence and detailed expression pattern of Xenopus flotillin1. Xenopus laevis has a partially duplicated genome (Kobel and Du Pasquier, 1986) and, thereby, carries two paralogous copies of flotillin1 (flotillin1A and flotillin1B). In addition, we identified a polymorphic version of flotillin1A (flotillin1A′). We show that both paralogues are expressed maternally. Zygotic expression begins during gastrulation in the embryonic ectoderm and becomes enhanced in the neural plate. As the neural tube closes and differentiates, expression becomes greatly enhanced in the dorsal neural tube and in peripheral sensory structures, including the olfactory pit, retina, lens, otocyst, and cranial ganglia. At late tail bud stages, dorsal primary neurons of the spinal cord intensely express this gene. This report is the first comprehensive developmental description in any animal of the expression pattern of flotillin1, a gene whose protein product is likely to play important roles in multiple specialized signal transduction events.
RESULTS AND DISCUSSION
Proteins that are characteristic for caveolae and lipid rafts have been isolated independently from various species. Reggie-1/-2 were identified in goldfish and rat where they are expressed predominantly in the nervous system (Schulte et al., 1997; Lang et al., 1998). In goldfish, Reggie-1/-2 are proposed to play roles during regeneration of retinal ganglion cell axons (Schulte et al., 1997). Membrane-associated proteins isolated from mouse and Drosophila that are highly homologous to Reggie-1/-2 were named Flotillin-2/-1, respectively (Bickel et al., 1997; Galbiati et al., 1998). We isolated a large number of cDNA clones with homology to flotillin1 from oocyte, cleavage, and tadpole stage Xenopus laevis cDNA libraries. These clones fell into three distinct groups of ∼1.7-kb cDNAs that encode highly similar 429 amino acid proteins (Fig. 1A). Flotillin1A (GenBank accession no. AF545659) differs from Flotillin1A′ (GenBank accession no. AF545660) by one amino acid (aa; V→A223; 99.8% identity); thus, they likely represent polymorphic versions of the same gene. Flotillin1B (GenBank accession no. AF545661) differs from Flotillin1A by 31 aa (92.8% identity) and, thus, likely is the paralogous copy expected due to the partial duplication of the Xenopus laevis genome (Kobel and Du Pasquier, 1986). In zebrafish and goldfish, which also underwent genome duplication, there are two copies of flotillin1 (reggie2) and two copies of flotillin2 (reggie1) (Malaga-Trillo et al., 2002). Xenopus Flotillin1A shares high amino acid identity with human (80.8%), mouse (81.1%), rat (80.8%), goldfish (78.3%), zebrafish (79.1%), and Drosophila (61.3%) Flotillin1 (Fig. 1B). Xenopus Flotillin1A shares only 47% and Flotillin1B shares only 49% amino acid identity to mouse Flotillin2, in accord with mouse Flotillin1 and mouse Flotillin2 sharing only 47% amino acid identity. Furthermore, a Xenopus cDNA, annotated as flotillin2 (GenBank accession no. AAH43770), shares only 42% amino acid identity to Xenopus Flotillin1A. These comparisons confirm the identity of the three groups of genes we isolated as all belonging to the Xenopus Flotillin1 subfamily.
flotillin1 expression extends from oocyte to tadpole stages. By polymerase chain reaction (PCR) analysis, transcripts are highly expressed at all these stages (Fig. 2A). By in situ hybridization analysis, flotillin1 transcripts are expressed throughout the cytoplasm at the earliest stages of oogenesis (Fig. 3A), and they remain abundant through cleavage stages (Fig. 3B). Although by in situ hybridization transcripts are detected primarily in the animal hemisphere of the late oocyte and cleavage embryo (Fig. 3A,B), PCR analysis demonstrates that flotillin1A transcripts are present in both animal and vegetal cells (Fig. 2B). During gastrula stages, flotillin1A transcripts are detected throughout the embryonic ectoderm, with a slight enhancement on the dorsal side (Fig. 3C). This expression becomes more notably enhanced in the neural ectoderm during neural plate formation (Fig. 3D–F). During neural tube closure (stage [st.] 16–20), ectodermal expression is greatly enhanced in the neural tube, and there is additional staining in the paraxial mesoderm (Fig. 3G). At early tail bud stages (st. 21–26), flotillin1 is additionally expressed in the branchial arches (Fig. 3H,I) and in the overlying placodal ectoderm (Fig. 3J).
At later tail bud stages (st. 28–34), flotillin1A expression is additionally identified in several placodal/neural crest derivatives, including the olfactory pit, lens, otocyst, and cranial ganglia (Fig. 3K). These structures were confirmed with tissue sections (Fig. 3N–Q) and a neural-specific marker (n-tubulin, Fig. 3M, Q′; Good et al., 1989; Moody et al., 1996). In addition, branchial arch mesenchyme (neural crest and/or somitomeric mesoderm) and the dorsal tips of somites also are stained (Fig. 3K,N,R,T). At these stages, neural tube staining is confined to the dorsal domains, from forebrain to spinal cord (Fig. 3O,Q,R,T), and retina staining is confined to the ciliary margins (Fig. 3O). Beginning around stages 32–34 and continuing into tadpole stages, the primary neurons in the dorsal spinal cord are highly expressive (Fig. 3R,T,X). Based on the dorsal/dorsolateral and peripheral positions of the cell bodies, these cells are most likely Rohon-Beard and/or dorsolateral commissural interneurons (Nordlander, 1984; Moody, 1989; Roberts and Sillar, 1990) and are most likely not to be other commissural interneurons, descending interneurons, ascending interneurons or reticulospinal neurons (Soffe et al., 1984; Nordlander et al., 1985; Roberts and Alford, 1986; Roberts et al., 1987; van Mier and ten Donkelaar, 1989). However, positive identification of these flotillin1-positive cells will require characterization by neurotransmitter phenotype, axonal projections, and electrophysiological properties (e.g., Clarke et al., 1984; Gallagher and Moody, 1987; Roberts et al., 1987, 1988). At tadpole stages (st. 39–42), flotillin1 expression remains highly expressed in the olfactory pit, cranial ganglia, hindbrain, and dorsal primary spinal neurons but is significantly reduced in lens, retina, and otocyst (Fig. 3V–Z). This reduced staining may indicate down-regulation of flotillin1 transcription as these structures differentiate, but we also cannot rule out reduced penetration of the in situ probe in these older embryos.
The expression patterns using flotillin1A and 1A′ probes are identical. The expression of flotillin1B is identical to flotillin1A up to tail bud stages and then becomes more restricted. flotillin1B expression is confined to smaller domains in the dorsal neural tube, is reduced in the branchial arches, and is not detected in the somites (Fig. 3L,S,U,W). This more restricted expression pattern suggests that the two flotillin1 paralogues may be under the control of different regulatory regions in the genome. Where different paralogues have been identified in Xenopus laevis, similar small differences in expression patterns have been reported (e.g., for FoxD5: Solter et al., 1999; Fetka et al., 2000; Sullivan et al., 2001).
Previous reports demonstrate that flotillins/reggies are associated with the nervous system. For example, in adult mouse and human, flotillin-1 is expressed at highest concentrations in the brain and heart muscle (Bickel et al., 1997; Edgar and Polak, 2001). In the late gastrulating fly, flotillin1 is highly expressed in the growing nervous system (Galbiati et al., 1998), and in the 5-day goldfish larva flotillin1 expression (reggie2) has been reported in the developing retina (Paschke et al., 1992). Our data concur that there is enhanced, albeit not restricted, expression of flotillin1 in the nervous system and provide the first comprehensive temporal and spatial description in any animal of the dynamic expression patterns of both paralogous copies of this gene.
It has not yet been fully resolved whether Flotillins reside in caveolae, in noncaveolar lipid rafts, or in both (Bickel, 2002); however, we show that Xenopus flotillin1 is expressed in patterns distinct from the Xenopus caveolins 1–3. Whereas flotillin1 is highly enriched in neural structures with lower levels of expression in non-neural ectoderm and some mesodermal tissues, Xenopus cav-1 and cav-2 are highly expressed in the notochord, fat body, and lungs, and cav-3 is expressed in punctate structures in the non-neural epidermis, heart, and skeletal muscles (Razani et al., 2002). These different patterns argue against a strict colocalized expression of Flotillin1 and caveolar-associated proteins.
Although Flotillins are involved in membrane scaffolding, phagocytosis, and/or signal transduction in several differentiated cell types, we know very little about their developmental functions. Flotillin proteins are expressed during retinal axon outgrowth and regeneration (Schulte et al., 1997), are involved in actin reorganization during neurite outgrowth (Haglund et al., 2004), and when expressed in Cos cells, induce the formation of a neuronal morphology (Hazarika et al., 1999). Because Xenopus is a particularly well-suited vertebrate in which to test the developmental function of novel genes, this description of the early and progressively restricted expression of flotillin1 before and during nervous system development provides important groundwork for future functional studies.
A dorsal 16-cell blastomere cDNA library was constructed and screened for pools of maternal mRNAs that alter dorsal axis formation. A NotI–SalI fragment of one partial clone was used to screen an oocyte library (kindly provided by Bruce Blumberg, University of California, Irvine), and an EcoRI–XhoI fragment of a cDNA from the oocyte library was used to screen a stage 42 library (kindly provided by Michael King, University of Indiana) according to standard methods. Twelve full-length cDNAs were fully sequenced (both strands). These represented three unique but highly similar genes submitted to GenBank as Xenopus flotillin1a (accession no. AF545659), flotillin1b (accession no. AF545660), and flotillin 1c (accession no. AF545661). After sequence comparison, these were renamed flotillin1A (former 1a), flotillin1A′ (former 1b), and flotillin1B (former 1c) to indicate polymorphic versions vs. paralogous copies.
Total RNA from whole embryos at different developmental stages or from animal and vegetal halves dissected from 32-cell embryos was extracted (Gentra RNA Isolation Kit). First-strand cDNAs were synthesized by using Superscript Reverse Transcriptase (Invitrogen). PCR was performed by using the Taq Amplification Kit (Epicenter) under standard conditions and Xenopus flotillin1A primer sequences (F, 5′-ACATGGGGATCAGTGTGGTT-3′; R, 5′-GTTTGGTCTTGGCCACCTGT-3′).
Whole-mount in situ hybridization was carried out by standard procedures (Sive et al., 2000). A probe for the neural-specific n-tubulin (Good et al., 1989) was used in stage-matched embryos to confirm neural identity of flotillin1-positive tissues. After whole-mount analysis, several specimens were embedded in paraffin and serially sectioned at 10 μm for histological analysis using differential interference contrast optics. Some stage-matched embryos were fixed, serially sectioned with a cryostat, and stained with an anti–neural-specific Tubulin antibody, as previously described (Moody et al., 1996).
We thank Lianhua Yang and Himani Datta Majumdar for histological preparations, Bruce Blumberg (UC Irvine) for the gift of the oocyte cDNA library, Michael King (Indiana University) for the gift of the stage 42 cDNA library, and Peter Good (NIH-NHGRI) for the n-tubulin plasmid.