Sterol Carrier Protein 2 Regulates Proximal Tubule Size in the Xenopus Pronephric Kidney by Modulating Lipid Rafts

Abstract

The kidney is a homeostatic organ required for waste excretion and reabsorption of water, salts and other macromolecules. To this end, a complex series of developmental steps ensures the formation of a correctly patterned and properly proportioned organ. While previous studies have mainly focused on the individual signaling pathways, the formation of higher order receptor complexes in lipid rafts is an equally important aspect. These membrane platforms are characterized by differences in local lipid and protein compositions. Indeed, the cells in the Xenopus pronephric kidney were positive for the lipid raft markers ganglioside GM1 and Caveolin-1. To specifically interfere with lipid raft function in vivo, we focused on the Sterol Carrier Protein 2 (scp2), a multifunctional protein that is an important player in remodeling lipid raft composition. In Xenopus, scp2 mRNA was strongly expressed in differentiated epithelial structures of the pronephric kidney. Knockdown of scp2 did not interfere with the patterning of the kidney along its proximo-distal axis, but dramatically decreased the size of the kidney, in particular the proximal tubules. This phenotype was accompanied by a reduction of lipid rafts, but was independent of the peroxisomal or transcriptional activities of scp2. Finally, disrupting lipid micro-domains by inhibiting cholesterol synthesis using Mevinolin phenocopied the defects seen in scp2 morphants. Together these data underscore the importance for localized signaling platforms in the proper formation of the Xenopus kidney.

INTRODUCTION

The kidney is an essential organ required for the excretion of waste products and the recovery of water, salts and organic compounds (Smith, 1953; Saxén, 1987; Vize et al., 2003). Three distinct kidney forms, the pronephros, the mesonephros and the metanephros are found in vertebrates. They are characterized by increasing levels of complexity with the pronephros being the most primitive and only active form in larval stages of aquatic animals. The metanephros is the adult kidney of higher vertebrates; the exceptions are birds, fish and amphibians, which do not progress past the mesonephric kidney. Importantly, the three kidney types are evolutionarily interdependent, as the formation of the pronephros is a prerequisite for the formation of the meso- and metanephric kidney (Bouchard et al., 2002). Moreover, the nephron, the functional unit of each kidney form, is very stereotypical. It is divided into functionally distinct segments characterized by a repertoire of transporters, channels and signaling molecules (Zhou and Vize, 2004; Raciti et al., 2008) and its formation is regulated by evolutionarily conserved gene regulatory networks (Heller and Brandli, 1997; Carroll et al., 1999; Reggiani et al., 2007; White et al., 2010; Marcotte et al., 2013). Interestingly, many aspects in kidney development are still poorly understood; these include the determination of total number of cells present in the nephron, the regulation of the size of individual cells, but also cellular specification of kidney epithelial cells such as the formation of the proximal tubular brush border or the contributions of lipid rafts (Wessely et al., 2013).

Lipid rafts are 10–200 nm small, heterogeneous, highly dynamic, cholesterol- and sphingolipid-rich domains within the lipid bilayer (Pike, 2006) and are thought to function by compartmentalizing multiple cellular processes (Pike, 2005). In fact, they potentiate the efficiency of growth factor receptor signaling by local clustering of receptors and downstream signaling components. Lipid rafts have been primarily studied in cell culture, where biochemical approaches elucidated their lipid and protein composition as well as their impact on cellular function (Simons and Ikonen, 1997; Pike, 2005; Jacobson et al., 2007). More recently, the in vivo contributions of lipid rafts have started to emerge. Since cholesterol is one of the major lipid raft components, removing cholesterol (using e.g. Methyl-β-cyclodextrin) or inhibiting cholesterol synthesis (using e.g. Mevinolin) disrupts signaling events mediated by lipid rafts (Klein et al., 1995; Taraboulos et al., 1995). They cause dramatic defects during embryonic development affecting signaling pathways like sonic hedgehog and Wnt (Cooper et al., 2003; Tadjuidje and Hollemann, 2006; Anderson et al., 2011; Reis et al., 2012). Similarly, mice lacking protein components of lipid rafts such as Caveolin-1 display a wide range of embryological defects (Drab et al., 2001; Razani et al., 2001; Zhao et al., 2002).

Another aspect of lipid rafts is that they are very dynamic structures and their lipid and protein composition must be constantly adjusted. One of the proteins involved in this process is the Sterol Carrier Protein 2 (scp2). This multifunctional protein binds to a subset of lipids such as cholesterol and sphingomyelin and transports them between cell membranes (Milis et al., 2006; Filipp and Sattler, 2007). In addition to its role in the cytoplasm, where it regulates lipid raft composition, scp2 is also present in peroxisomes, where it contributes to the import of fatty acids and β-oxidation (Mukherji et al., 2002; Atshaves et al., 2003; Atshaves et al., 2007a; Atshaves et al., 2007c). Interestingly, even though scp2 has been knocked out in mice, its in vivo role in lipid rafts has not yet been investigated (Seedorf et al., 1998; Fuchs et al., 2001; Atshaves et al., 2007b).

Here we study the connection between scp2, lipid rafts and kidney development using Xenopus laevis as a model (Wessely and Tran, 2011). We demonstrate that scp2 mRNA is expressed in the entire pronephros and that interfering with its function caused a dramatic reduction in kidney size. Moreover, this effect of scp2 knockdown was a result of interfering with lipid rafts and not due to its transcriptional or peroxisomal activities because it could be mimicked by disruption of lipid rafts using the cholesterol synthesis inhibitor Mevinolin.

MATERIAL & METHODS

Embryo Manipulations

Xenopus embryos were obtained by in vitro fertilization, maintained in 0.1x modified Barth medium (Sive et al., 2000) and staged according to Nieuwkoop and Faber (1994). Antisense morpholino oligomers (MOs) were obtained from GeneTools. The sequences of the MOs used in this study were 5’-AGC CAT GTT CCA CAG CAG CAG GTA T-3’ (scp2-MO1) and 5’-CCC CAA GGC CAA TAT TGT GCT GCA G-3’ (scp2-MO2). MOs were diluted to a concentration of 1 mM. To target both scp2 transcripts the two MOs were mixed at a 1:1 ratio (scp2-MO1+2). The pCS2-scp(P1)-GFP and pCS2-scp(P2)-GFP constructs were generated by PCR and sub-cloned into pCS2-AcGFP (White et al., 2010). For synthetic mRNA the plasmids were linearized with NotI and transcribed with SP6 RNA polymerase using the mMessage mMachine (Life Technologies). For all injections a total of 8 nl of morpholino oligomer solution was injected radially at the 2- to 4-cell stage into Xenopus embryos. For the GFP reporter assays these injections were followed by two injections of 2 ng of synthetic mRNA into two animal blastomeres at the 8-cell stage. Rescue experiments were performed by injecting embryos with GFP-scp2ΔL DNA into one blastomere at the 2-cell stage followed by injection of scp2-MO-1+2 into all 4 blastomeres at the 4-cell stage. Embryos were analyzed by 3G8 staining at stage 40 comparing the GFP-scp2ΔL-injected side to the contralateral side.

To interfere with cholesterol synthesis Xenopus embryos were cultured until stage 32 and treated with 125 µM Mevinolin (#M2147, Sigma). Embryos were cultured until sibling controls reached stage 39/40 and processed for subsequent analyses.

Global metabolic profiling was performed by Metabolon, using six independent sets of uninjected control and scp2-MO1+2-injected whole embryos at stage 38.

Whole Mount In Situ Hybridization

The procedure for the in situ hybridizations and the analysis by paraplast sectioning has been previously described (Belo et al., 1997; Tran et al., 2007). To generate antisense probes plasmids were linearized and transcribed as follows: pSK-β1-Na/K-ATPase - EcoRI/T7 (Tran et al., 2007), pCMV-SPORT6-Basigin - SalI/T7 (Clone ID: 6631233), pSK-Catalase - EcoRI/T7 (Clone ID: XL156c02), pSK-Ncc - EcoRI/T7 (Tran et al., 2007), pSK-Nkcc2 - SmaI/T7 (Tran et al., 2007), pCMV-SPORT6-Nphs1 -EcoRI/T7 (Tran et al., 2007), pSK-Pmp70 - EcoRI/T7 (Clone ID: XL044i24), pSK-scp2(P1+P2) - EcoRI/T7 (Clone ID: Xl217c18), pSK-scp2(P1) - XbaI/T7, pCMV-SPORT6-Sglt1-K- SalI/T7 (Zhou and Vize, 2004).

Histology and Immunohistochemistry

For histological staining, dissected mesonephroi were fixed in Bouin’s Fixative, cleared in 70% ethanol, embedded in paraplast, sectioned at 7 µm, dewaxed, and stained with Hematoxylin and Eosin. For immunohistochemistry embryos were fixed in Dent’s fixative (Methanol:DMSO=4:1). For whole mount immunostaining, embryos were incubated overnight with the 3G8 and 4A6 monoclonal antibodies (Vize et al., 1995) followed by incubation with an anti-mouse Alexa-555 secondary antibody or a Horseradish peroxidase-coupled anti-mouse secondary antibody, which was developed using the ImmPACT DAB kit (Vector Laboratories). For the analysis of yolk platelet proteins or Caveolin-1 embryos were first processed for 3G8 antibody staining on whole mounts (using anti-mouse Alexa-555 as secondary antibody). Xenopus embryos were then embedded in paraplast, sectioned at 25 µm and probed with antibodies recognizing Vitellogenin (a kind gift of Dr. M. Kirschner), Caveolin-1 (#610407, BD or #3267, Cell Signaling), Clathrin (#4796, Cell Signaling), Rab5 (#3547, Cell Signaling) and Rab7 (#9367, Cell Signaling). Primary antibodies were visualized either by staining with anti-rabbit Alexa-647 (in the case of Vitellogenin) or by anti mouse/rabbit HRP followed by Tyramide Signal Amplification (Perkin Elmer) using Cy5 or Cy3. In some of the immunostainings, 3G8 was replaced by Erythrina Cristagalli Lectin (ECL, Vector Laboratories). In contrast to 3G8 immunostaining, which needs to be performed in whole mounts, ECL also labels proximal tubules (as initially reported by Dr. P. Vize on Xenbase.org), but is more versatile. Vitellogenin labeling intensity and number of Caveolin-1-positive foci were quantified by the ImageJ software using only the 3G8- or ECL-positive tubules of 4–5 sections per embryo and multiple embryos of at least three independent fertilizations for each condition. Co-localization analyses were performed using the co-localization module of the Leica Analysis Software (LAS AF). To determine the overall numbers of cells present in the proximal tubules, we followed the protocol reported in Romaker et al. (2012).

To visualize the GM1 present in lipid rafts in whole mounts, embryos were fixed in MEMFA (Sive et al., 2000), incubated with Horseradish peroxidase-conjugated Cholera Toxin Subunit B (CT-B, 1:1000, #C34780, Invitrogen) and developed using Tyramide Signal Amplification (Perkin Elmer). Cryostat sections were imaged using FITC-labeled CT-B (1:1000, #C1655, Sigma). Peroxisomal peroxidase activity was visualized by 3,3'-diaminobenzidine (DAB) staining following the procedure described by Krysko et al. (2010).

Immunofluorescence analyses of mouse kidneys (embryonic day E16.5, postnatal day P5 and adult) were performed as described for the Xenopus embryos. The only differences were that 10 µm paraplast sections were used. Proximal tubules were visualized using Lotus tetragonolobus agglutinin (LTA, Vector Laboratories), collecting ducts were counterstained with Dolichos Biflorus Agglutinin (DBA, Vector Laboratories) or an antibody against Carbonic Anhydrase 2 (#SC25596, Santa Cruz Biotechnologies).

Transmission Electron Microscopy

Embryos were fixed in 2.5% glutaradehyde/4% paraformaldehyde in 0.2 M cacodylate buffer overnight at 4°C and post-fixed in 1% Osmium Tetroxide for 1 hour at 4°C. After washing in Maleate buffer they were incubated in 1% uranyl acetate for 1 hour, washed again and dehydrated into propylene oxide. Samples were infiltrated with Eponate and allowed to polymerize for 24 hrs. Ultra thin sections of 85 nm were cut with a diamond knife, stained with uranyl acetate and lead citrate, and then observed with a Philips CM12 electron microscope operated at 60 kV.

Cell Culture Experiments

HEK-293T cells were maintained in Ham's F-12/DMEM supplemented with 10% FBS and 1% Pen-Strep (5000 U/ml). GFP-scp2 fusion proteins were generated by PCR by inserting the GFP from the pAcGFP1-C1 vector into pCS2-pro-scp2, pCS2-scp2 and pCS2-scp2ΔL. Cells were transfected using Lipofectamine 2000 (Invitrogen). Western blot analyses of cell lysates were conducted after 48 hrs using an anti-GFP antibody (#AB3080, Millipore) and anti-α-Actin antibody (#A4700, Sigma). Immunofluorescence analyses were performed using the anti-PMP70 antibody (#P0497, Sigma) followed by an Alexa-555 coupled anti-rabbit secondary antibody.

Statistical Analysis

All experiments were repeated at least three times. In the case of Xenopus experiments the repetitions used independent fertilizations from different females. Data were analyzed by Student’s t-Test.

RESULTS

Lipid Rafts in the Pronephric Kidney of Xenopus

It has recently been shown that lipid rafts are present during early Xenopus development (Reis et al., 2012). However, little is known about their occurrence and function during kidney development. In an attempt to address this we performed immunofluorescence analysis on stage 42 Xenopus embryos assaying for the distribution of the ganglioside GM1, a lipid highly enriched in lipid rafts (Simons and Ikonen, 1997). We incubated embryos with the HRP- or FITC-conjugated B subunit of cholera toxin (CT-B), which specifically binds to GM1 (Holmgren, 1973; Fishman, 1982). Embryos were co-labeled with the 3G8 antibody (Vize et al., 1995) or Erythrina Cristagalli Lectin [ECL, (Romaker et al., 2014)] to visualize the kidney proximal tubules. Pronephroi were dissected and imaged as whole mounts using 2-photon microscopy. As shown in Fig. 1A, CT-B staining was present throughout the kidney with more abundant signal in the proximal tubules. This was confirmed using cryostat sections comparing the 3G8-positive proximal tubules to the structures labeled with the distal tubule and duct marker 4A6 (Vize et al., 1995) (Fig. 1E–H’). Examining earlier developmental stages revealed that this enrichment was present throughout kidney formation but was particularly evident at stage 42 (Fig. 1C–E’). At the cellular level, CT-B staining was present throughout the membranes, but was enriched apically (Fig. 1F,F’). In fact, it may be a result of the formation of the proximal tubular brush border. These consist of an unusual lipid rafts specialized to concentrate transporters, enzymes and channels (Parkin et al., 2001; Danielsen and Hansen, 2003).

Fig. 1.

Lipid Rafts in the Pronephric Kidney. (A,B) Whole mount immunostaining of Xenopus pronephroi using the B-subunit of Cholera Toxin (CT-B) or an antibody against Caveolin-1 at stage 42. Proximal tubules were counterstained using the 3G8 antibody (A) or ECL (B). pt, proximal tubules; dt, distal tubules. (C–H’) Cryostat sections of stage 35, 39 and 42 Xenopus embryos stained with CT-B and either 3G8 (C–F’) or 4A6 (G–H’). (I–N’) Paraplast sections of stage 35, 39 and 42 Xenopus embryos stained with Caveolin-1 and 3G8 (I–L’) or 4A6 antibodies (M–N’). The gray-scale panels show CT-B and Caveolin-1 staining only, while the color panels show the merged images. (O–S’) Paraplast sections of stage 40 Xenopus embryos stained with anti-Caveolin-1 in combination with ECL (O–P’), Clathrin (Q,Q’), Rab5 (R,R’) and Rab7 (S,S’). In panel P and P’ co-localization of Caveolin-1 and ECL staining in the apical domain was analyzed using the co-localization module of the LAS AF software and is indicated by white pixels. The panels F,F’,H,H’,L,L’,N,N’,O’,P’,Q’,R’ and S’ are close-ups of the areas indicated by the white boxes.

As the focus of our study was not on the formation of brush borders, but on lipid rafts and their involvement in signaling, we next examined the distribution of Caveolin-1. This structural protein is concentrated in caveolar lipid rafts, but absent in brush borders (Parton and Simons, 2007). In contrast to CT-B, Caveolin-1 was present in a punctate pattern throughout the entire kidney without any obvious proximal-distal specificity (Fig. 1B, K–N’). Moreover, the pattern could be detected from stage 35 onward (Fig. 1I–J’). However, in contrast to the early stages, the increased brush border-associated CT-B staining at stage 42 (Fig. 1E–F’) is not accompanied by enrichment in Caveolin-1 expression (Fig. 1K–L’).

Co-localization studies revealed that Caveolin-1-positive foci were present at the cell membrane, but were also found intracellularly (Fig. 1O–P’). It has been reported that Caveolin-1-positive lipid rafts undergo Clathrin/Rab5-independent endocytosis forming stable transport vesicles that fuse to so-called caveosomes (Parton and Simons, 2007). To confirm this we performed co-localization studies with Clathrin, and the two small G-proteins Rab5 and Rab7 that stain early and late endosomes, respectively. As shown in Fig.1Q–S’ each of those endosomal markers show a speckled appearance similar to Caveolin-1; yet co-localization analyses did not detect an overlap with Caveolin-1.

One puzzling aspect of the Caveolin-1 pattern in the developing pronephric kidney was that its distribution is different from the one described for adult mouse kidneys (Zhuang et al., 2011; Paunescu et al., 2013). Instead of a punctate expression, these studies did not detect any Caveolin-1 staining in proximal tubules and basolateral enrichment in collecting ducts. As an identical Caveolin-1 pattern in the Xenopus pronephros was observed with two different antibodies (compare Fig. 1O,O’ and Supplementary Figure S1A,A’), we reasoned that this reflected differences in lipid rafts during kidney development and adult kidney homeostasis. To confirm this, we examined Caveolin-1 staining in developing (i.e. embryonic day E16.5 and postnatal day P5) and adults mouse kidneys. Indeed, proximal tubules displayed a punctate staining pattern at E16.5, but lost this expression by P5 (Supplementary Fig. S2A–H and data not shown). Conversely, collecting ducts initially exhibited rather sparse Caveolin-1 staining, but acquired strong, baso-laterally enriched staining (Supplementary Fig. S2I–P). Finally, as in Xenopus, the Caveolin-1-positive foci in E16.5 mouse proximal tubules did not co-stain with markers of the endosomal compartment (Supplementary Fig. S3).

Together, these experiments suggest that the cells of the pronephros contain Caveolin-1-positive rafts, which seem to be actively recycling during kidney development.

Identification and Characterization of Xenopus Scp2

We next wanted to address the functional relevance of lipid rafts in pronephros development. While depleting cholesterol or interfering with its biosynthesis is a well-established method to study lipid rafts (Hooper, 1999), we instead decided to focus on Sterol Carrier Protein 2 (scp2), a multifunctional protein (Fig. 2A) that is known to be involved in the remodeling of the lipid composition of rafts (Gallegos et al., 2001; Wirtz, 2006).

Fig. 2.

Expression of Scp2 mRNA during Xenopus Development. (A) Scheme depicting the multiple described activities of scp2 protein. (B) Schematic of the two different antisense probes used. (C–H) Whole mount in situ hybridization of 4-cell stage, stage 28 and stage 35 embryos using the scp2(P1+P2) or the scp2(P1) probe. Inset in (D) and (G) shows a frontal view highlighting the hatching gland expression. (I–J’) In situ hybridization on paraplast sections of stage 45 embryos using both scp2 mRNA probes. Panels (I’) and (J’) show close-ups of the pronephric kidney area. no, notochord; nt, neural tube; so, somites; pn, pronephros; en, endoderm. (K–M) Hematoxylin and Eosin (H&E) staining (K) and in situ hybridization on paraplast sections using scp2(P1+P2) or scp2(P1) RNA probes (L,M) of an adolescent Xenopus mesonephros.

Identification and characterization of the Xenopus homologue demonstrated that Xenopus scp2 is an evolutionarily conserved protein that is localized to the cytoplasm and peroxisomes (Supplementary Fig. S4 and S5). Most importantly, it was expressed in the pronephric kidney. Whole mount in situ hybridization analyses detected full-length scp2 mRNA [scp2(P1+P2)] maternally in the animal region (Fig. 2B,C); zygotic expression was rather low and uniform until tailbud stages, when staining in the hatching gland and pronephros were first detected (Fig. 2D). At stage 35, additional expression domains in the head, branchial arches, optic and otic vesicles, neural tube, notochord and somites appeared (Fig. 2E and data not shown). At stage 45, scp2 mRNA was found in all endodermal organs and the pronephros, but little expression was now detected in neural and muscle tissue (Fig. 2I,I’). Finally, the kidney expression of scp2 was not only restricted to the pronephric kidney, but could also be observed in the mesonephros of adolescent Xenopus (Fig. 2K,L).

In designing a knockdown strategy for scp2 using antisense morpholino oligomers (MOs) we needed to consider its unusual genomic organization. In humans, SCP2 protein is encoded by two transcripts, SCPX and PRO-SCP2 (Ohba et al., 1994; Ohba et al., 1995; Ferdinandusse et al., 2006; Atshaves et al., 2007b). Indeed, a similar genomic organization was identified in Xenopus laevis and Xenopus tropicalis (Supplementary Fig. S4 and data not shown). To prove that both transcripts are indeed present in Xenopus, we first performed in silico analyses examining the Xenopus scp2 EST sequences available in the public databases. The comparison revealed three classes, clones sequenced from the 3’ end, clones sequenced from the 5’ end that correspond to the full-length transcript and finally clones sequenced from the 5’ end that mapped to about 1,100 nucleotides downstream (Supplementary Fig. S6). This class is in agreement with a transcript initiated from an alternative promoter similar to the pro-scp2 transcript in humans.

Next, we examined whether the two transcripts differed in their expression patterns by in situ hybridizations. The initial antisense probe [scp2(P1+P2)] corresponded to the entire mRNA and recognized transcripts from both promoters (Fig. 2B). A probe encompassing only the sequence found in the longer transcript, scp2(P1), detected the same expression domains as scp2(P1+P2) (Fig. 2F–H, J,J’ and M). Unfortunately, the reverse experiment visualizing only the short transcript was impossible due to the extensive sequence overlap. However, the fact that the staining using the scp2(P1) probe was generally much weaker than the scp2(P1+P2) suggested that both transcripts were present in the pronephros. This interpretation could also be confirmed by RT-PCR analyses using transcript-specific primers (data not shown).

Based on these data, we conclude that Xenopus scp2 is transcribed from two promoters and is expressed in the pronephric kidney.

Knockdown of scp2 Interferes with Proximal Tubule Development

With both promoters active in early kidney development, we designed two antisense morpholino oligomers (MOs) to block translation of both transcripts (scp2-MO1 and scp2-MO2) (Fig. 3A). The efficacy of both MOs were tested using GFP-fusion protein constructs containing the respective MO binding sites [scp2(P1)-GFP and scp2(P2)-GFP]. In contrast to uninjected controls, Xenopus embryos injected at the 8-cell stage with 2 ng of either scp2(P1)-GFP or scp2(P2)-GFP mRNA displayed green fluorescence at the gastrula stage (Fig. 3B,C,E); this expression was lost when the embryos were co-injected at the 2- to 4-cell stage with either 3.2 pmol scp2-MO1 or scp2-MO2 (Fig. 3D,F).

Fig. 3.

Phenotype of Xenopus Embryos Lacking Scp2. (A) Sequence alignment of the two scp2 pseudo-alleles in the regions targeted by either scp2-MO1 or scp2-MO2. The position of the MOs and the start codons (P1 and P2) are indicated. Non-conserved nucleotides are highlighted in red. (B–F) Xenopus embryos were injected animally with the scp2-MO1 and scp2-MO2 at the 2-4 cell stage followed by two diametral injections of synthetic mRNA encoding either scp2(P1)-GFP or scp2(P2)-GFP at the 8-cell stage, while control embryos were only injected with the scp2(P1)-GFP or scp2(P2)-GFP mRNA. Embryos were analyzed at stage 10.5 by florescence microscopy. Representative images are shown and the percentage of GFP-positive embryos is indicated in the upper right corner of each panel. (G–R) 3G8 or 4A6 immunostaining at stage 40 (G–N) or β1-Na/K-ATPase whole mount in situ hybridization at stage 39 (O–R) of Xenopus embryos injected with scp2-MO1, scp2-MO2 or a combination of both MOs (scp2-MO1+2). (S) Quantification of the 3G8 whole mount immunostaining. The embryos were categorized in three groups representing increasing severity of the phenotype. The amount of MOs injected and the number of embryos analyzed is indicated. (T–V) 3G8 immunostaining of stage 40 controls or embryos injected with scp2-MO1+2 in the presence or absence of GFP-scp2ΔL DNA. Panels show representative images of four independent experiments. (W) Bar diagram of the number of proximal tubular cells in embryos injected on one side with scp2-MO1+2 and comparing the injected and uninjected side at stages 40 and 42. The number of embryos analyzed is indicated in the individual bars. Data were analyzed by paired Student’s t-Test; the three asterisks indicate a significance of p<0.001. (X) Bar diagram of the measured intensity of the vitellogenin immunostainings (see panel Z, Z’). Data were analyzed by Student’s t-Test and asterisk represent a significance of p<0.05. (Y–Z’) Uninjected and scp2-MO1+2-injected embryos at stage 40 were analyzed by transmission electron microscopy (Y,Y’) or immunofluorescence staining of proximal tubules using 3G8 (red) and anti-vitellogenin antibodies (green) (Z,Z’). Nuclei were counterstained with DAPI. Red arrowheads in (Y,Y’) indicate individual electron-dense yolk platelets.

Next, we tested whether loss-of-scp2 affected pronephric development. To this end, Xenopus embryos were injected at the 2- to 4-cell stage with 3.2 pmol scp2-MO1, 3.2 pmol scp2-MO2 or a mixture of both MOs (each at a concentration of 3.2 pmol; scp2-MO1+2), cultured until stage 40 and processed for 3G8/4A6 immunohistochemistry. Embryos lacking scp2 showed a dramatic reduction of the 3G8-positive area and a drastic loss of 4A6 staining (Fig. 3G–N,S and Supplementary Fig. S7A). Since both 3G8 and 4A6 label terminally differentiating renal epithelial cells, we also performed whole mount in situ hybridization using a range of pronephric marker genes. These, however, did not show defects in overall kidney formation (β1-Na/K-ATPase, Fig. 3O–R) or proximal-distal patterning (Nphs1, Sglt1K, Nkcc2, Ncc, Supplementary Fig. S7B–E’). The phenotypes were observed in the injections of the individual MOs (with scp2-MO1 being more effective than scp2-MO2), but were more pronounced upon co-injection of both MOs (Fig. 3G–N,S and Supplementary Fig. S7A). The cooperativity supported our interpretation that the scp2-MOs had a gene-specific phenotype. Importantly, injection of a standard control MO did not result in any of these defects [data not shown and (Tran et al., 2007; Agrawal et al., 2009)]. We also performed rescue experiments of the scp2 knockdown phenotype by co-injection of constructs resistant to scp2-MO1+2 activity. Unfortunately, scp2 mRNA injected embryos did not survive past gastrula stage due to dissociation of epithelial cell layers followed by apoptosis (data not shown). To circumvent this early lethality, we injected embryos with plasmid DNA instead. Even though this still impaired survival, a percentage of embryos survived until stage 40. Most importantly, those embryos exhibited a partially restored size of the 3G8-positive proximal tubule (Fig. 3T–V).

Next, we decided to more accurately assess the size of the proximal tubules in the absence of scp2. To this end, we followed the approach described in Romaker et al. (2012) and quantified the total number of cells present in the proximal tubules. Uninjected control embryos increased their proximal tubules from approximately 50 cells at stage 34 to about 230 cells at stage 40, and 320 cells at stage 42 (Fig. 3W) (Romaker et al., 2014). However, scp2 morphants showed little growth and the proximal tubules consisted in average of 110 cells both at stage 40 and 42. This reduction of proximal tubular size caused by scp2 knockdown was very reminiscent of a recent study on G-protein α_s function in the pronephric kidney (Zhang et al., 2013). There we demonstrated that the defect in proximal tubule size was accompanied by an impairment of yolk degradation that normally occurs during embryonic development. The same was observed in scp2 morphants. Transmission electron microscopy revealed more electron-dense yolk platelets in the proximal tubules of scp2-MO1+2-injected embryos compared to uninjected control embryos (Fig. 3Y,Y’). This was confirmed by immunofluorescence using an antibody against the major yolk protein, vitellogenin (Jorgensen et al., 2009) (Fig. 3X,Z,Z’).

Together, these data show that loss-of-scp2 interferes with the formation of a correctly proportioned pronephric kidney by impairing the proximal tubular growth and yolk platelet degradation.

Scp2 and Lipid Rafts

Scp2 is involved in the regulation of lipid raft composition in a variety of cell lines (Zhou et al., 2004; Atshaves et al., 2007a). To directly evaluate if the loss-of-scp2 causes a defect in lipid rafts in vivo, we performed immunofluorescence analysis for lipid rafts using the two different Caveolin-1 antibodies. Scp2-MO1+2-injected embryos displayed a reduction in the number and intensity of Caveolin-1-positive foci in the proximal tubules (labeled with ECL) when compared to uninjected controls (Fig. 4A–B’ and Supplementary Fig. S1A–B’). Quantification of several independent embryos from multiple experiments demonstrated that this effect was statistically significant (Fig. 4E). Moreover, other domains with scp2 expression such as the distal pronephric tubules, neural tube and notochord exhibited a reduced number of Caveolin-1-positive foci in the scp2 morphants (Supplementary Fig. S8).

Fig. 4.

Scp2 and Lipid Rafts. (A–D’) Immunofluorescence analysis of uninjected controls and scp2-MO1+2-injected embryos at stage 40 using Caveolin-1 (A–B’) or Rab5 (C–D’) antibodies with the gray-scale image showing Caveolin-1 or Rab5 staining alone and the color panels showing merged images. ECL was used to identify proximal tubules (green), DAPI (blue) to visualize nuclei. (E) Quantification of the number of Caveolin-1-positive foci identified in (A,B). The number of embryos analyzed is indicated in the individual bars. Data were analyzed by Student’s t-Test and the three asterisks indicate a significance of p<0.001. (F) Metabolomics analysis of uninjected and scp2-MO1+2-injected embryos at stage 38 showing Box-and-Whisker Plots for cholesterol, 7-dehydrocholesterol and palmitoyl sphingomyelin.

Scp2 has been implicated in the function of lipid rafts, but not in the regulation of other endocytic pathways (Atshaves et al., 2003; Zhou et al., 2004). To validate the specificity of the defect, we performed immunofluorescence analysis for Rab5 and Clathrin comparing the proximal tubules of uninjected controls and scp2-MO1+2-injected embryos. Indeed, no significant changes in the staining patterns could be detected (Fig. 4C–D’ and Supplementary Figure S1C–D’).

Mechanistically, scp2 transports cholesterol and other lipids between different membrane compartments, but has rather little effect on overall cholesterol levels (Seedorf et al., 1998). In an attempt to validate this in vivo, we performed a metabolomics analysis comparing uninjected and scp2-MO1+2-injected whole Xenopus embryos at stage 38. As shown in Fig. 4F, the levels of neither cholesterol, nor its metabolic precursor, 7-dehydrocholesterol, were significantly different between the two conditions. Furthermore, palmitoyl sphingomyelin, another lipid bound by scp2 and enriched in lipid rafts (Atshaves et al., 2007a), was not altered as well.

Based on these data we conclude that loss-of-scp2 causes a decrease in Caveolin-1-positive lipid raft, which in turn could be responsible for the reduction in proximal tubule size.

Non-Raft Functions of Scp2

To further substantiate this interpretation, we next examined whether the non-raft functions of scp2 (Fig. 2A) contribute to the proximal tubular phenotype. Scp2 has been implicated in gene transcription by acting as a positive co-factor for the expression of Basigin (also known as CD147) in a glioblastoma cell line (Ko and Puglielli, 2007). While the Xenopus homologue of Basigin was strongly expressed in the pronephric kidney from stage 28 onwards (Supplementary Fig. S9), its expression levels were not downregulated upon microinjection of scp2-MO1+2 (Supplementary Fig. S7I,I’).

Finally, scp2 is also found in peroxisomes. Here, it functions as a fatty acyl-CoA carrier and solubilizing protein that facilitates the degradation of long and branched fatty acids (Seedorf et al., 1998; Mukherji et al., 2002). Thus, we performed two functional analyses to identify, whether scp2 morphants exhibit peroxisomal defects. To visualize peroxisomal catalase activity, embryos were stained with 3,3'-diaminobenzidine (DAB) (Fahimi and Baumgart, 1999; Krysko et al., 2010). But no obvious differences could be detected between uninjected controls and scp2-MO1+2-injected embryos (Supplementary Fig. S7F,F’). Secondly, as peroxisomes are an important organelle in fatty acid degradation, we surveyed our metabolomics data for changes in fatty acid levels. Indeed, several long chain fatty acids were significantly reduced in embryos lacking scp2, while medium chain fatty acids, which are primarily degraded in the mitochondria (Wanders, 2004), did not show such changes (Supplementary Table S1). In line with these assays whole mount in situ hybridization for Catalase did not show changes, while the peroxisomal fatty acid transporter, Pmp70, was greatly reduced (Supplementary Fig. S7G–H’). However, this effect was not only observed in the pronephros, but throughout the entire embryo.

Together these data demonstrate that scp2 morphants also show changes in peroxisomal-mediated fatty acid degradation, but not in transcriptional regulation of Basigin.

Reduction in Lipid Rafts Causes Reduced Proximal Tubule Size

The data so far do not confirm that the lipid raft defects are causative for the decrease in proximal tubular size. Thus, we decided to address this question by interfering with lipid raft function using Mevinolin, an inhibitor of HMG-CoA reductase, the rate-limiting enzyme in cholesterol biosynthesis (Taraboulos et al., 1995). To assess whether the phenotype resembles the one of scp2 morphants, embryos were treated with 125 µM Mevinolin from stage 32 onwards and analyzed at late tailbud stages. As expected, Caveolin-1 immunofluorescence analysis demonstrated that the number of lipid rafts in proximal tubules labeled with ECL was greatly reduced (Fig. 5A–C). Importantly, Mevinolin-treated embryos - like the scp2-MO1+2-injected counterparts - displayed reduced proximal tubular size as measured by 3G8 immunostaining followed by cell counting (Fig. 5D–E’). They also showed reduced intensity of 4A6 staining in the pronephric duct, yet this effect was less prominent than in scp2-MO1+2-injected embryos (Fig. 5F,F’). Additionally, Mevinolin-treated embryos exhibited ectopic 4A6 staining in the proximal tubular area (indicated by red arrowheads in Fig. 5F’). This phenotype was occasionally observed in scp2 morphants (data not shown), but was a hallmark of embryos with hyperactivated G-protein α_s function (Zhang et al., 2013). Finally, - like in the case of scp2 morphants - the kidney phenotype in Mevinolin-treated embryos was restricted to later events in pronephros development. Whole mount in situ hybridizations using β1-Na/K-ATPase and a panel of segment-specific marker genes showed changes in the architecture but not patterning of the kidney between control and Mevinolin-treated embryos (Fig. 5G,G’ and Supplementary Fig. S10).

Fig. 5.

Proximal Tubule Phenotype upon Inhibition of Cholesterol Synthesis. (A–B’) Immunofluorescence analysis of untreated controls and embryos treated with 125 µM Mevinolin at stage 40 using anti-Caveolin-1 (red) antibody and ECL (green) with the gray-scale image showing Caveolin-1 staining alone and the color panels showing merged images. DAPI (blue) was used to visualize nuclei. (C) Quantification of the number of Caveolin-1-positive foci identified in (A,B). The number of embryos analyzed is indicated in the individual bars. Data were analyzed by Student’s t-Test and asterisk represent a significance of p<0.05. (D) Bar diagram of the number of proximal tubular cells in untreated controls and Mevinolin-treated embryos at stages 40. The number of embryos analyzed is indicated in the individual bars. Data were analyzed by Student’s t-Test and the three asterisks indicate a significance of p<0.001. (E–I’) Untreated controls and Mevinolin-treated embryos were processed for 3G8 and 4A6 immunohistochemistry at stage 40 (E–F’) or whole mount in situ hybridization with β1-Na/K ATPase (G,G’), Pmp70 (H,H’) and Catalase (I,I’) at stage 39. (J,J’) Model for the role of lipid rafts in proximal tubule elongation in the wild-type situation (J) or in the absence of scp2 or upon treatment with Mevinolin (J’). See discussion for details.

Importantly, the Mevinolin experiments argue against a peroxisomal contribution, since the stunted proximal tubular growth by Mevinolin was independent of any detectable peroxisomal defects. Whole mount in situ hybridizations of Mevinolin-treated embryos did not show any changes in the expression of Pmp70 or Catalase (Fig. 5H–I’).

Together these data indicate that interfering with lipid rafts by inhibiting cholesterol synthesis (i.e. Mevinolin) or by knocking down scp2 results in similar proximal tubular growth phenotypes.

DISCUSSION

Lipid rafts are instrumental in compartmentalizing diverse cellular processes such as signaling, endocytosis and mechanosensing in the cellular plasma membrane (Simons and Ikonen, 1997; Hooper, 1999; Vogel and Sheetz, 2006; Parton and Simons, 2007). The importance for lipid rafts has been well established in cell culture. Surprisingly, their in vivo relevance is generally accepted, but thorough analyses are rare. This, to a large part, is due to the fact that many techniques used to study rafts in vitro (e.g. density gradients) are not applicable to the mixed population of different cell types found in living organisms. Most in vivo data are based on removing or lowering the core lipid components of rafts such as cholesterol and sphingolipids using chemicals (Hooper, 1999). Unfortunately, the effect of these compounds is obviously not restricted to the lipid microdomains, but interferes with all cell membranes and the interpretation of the results can therefore be difficult. Thus, interfering with proteins contributing to lipid rafts formation or maintenance - such as the one performed in this study - has emerged as the most promising alternative (Drab et al., 2001; Razani et al., 2001; Zhao et al., 2002; Ludwig et al., 2010).

Cell culture studies have implicated scp2 as a lipid raft modulator (Starodub et al., 2000; Milis et al., 2006; Atshaves et al., 2007a). In this study we now demonstrate that these findings are in vivo relevant and scp2 is an important protein for lipid raft composition and function. Moreover, it is independent of other membrane-associated pathways such as Clathrin-mediated endocytosis. This interpretation is based on the following three facts: (1) Xenopus embryos lacking scp2 protein show a reduced number of Caveolin-1 positive lipid rafts, but no defects in the classical endosomal compartment (Fig. 4A–E). (2) The overall levels of cholesterol and palmitoyl sphingomyelin are not altered in scp2 morphants (Fig. 4F). (3) The pronephric kidney phenotype of embryos lacking scp2 mimics the one of embryos treated with the cholesterol synthesis blocker Mevinolin (compare Figs. 4 and 5).

The instrumental role of scp2 as a lipid carrier for raft formation is also underscored by gain-of-function analyses. Embryos microinjected with a variety of scp2 mRNA or DNA exhibited poor to no survival until late stages of development. Thus, we did not examine the effect of ectopic scp2 expression on the pronephric kidney. However, the mRNA overexpression phenotype was quite revealing. Animal cap cells from injected embryos displayed weaker cell-to-cell contacts and dropped out of the ectodermal cell layer into the blastocoel and underwent apoptosis as early as blastula stage (data not shown). Even though we did not thoroughly test it, we hypothesized that this phenotype was caused by changes in the cell membrane lipid composition, which thereby modified the biophysical characteristics of the cells (e.g. surface tension or membrane fluidity).

One puzzling aspect of our study is that the scp2 knockdown phenotype in Xenopus is more severe than the one in the mutant mice (Seedorf et al., 1998; Atshaves et al., 2007b). Scp2-deficient mice display defects in lipid metabolism, but are viable and do not exhibit developmental deficiencies. However, these mice have not been studied in respect to lipid raft biology and the relatively weak phenotype may be explained by compensation by other lipid carriers. The best candidate is Caveolin-1, which like scp2 transports cholesterol and other lipids from the intracellular sites to the plasma membrane and is involved in lipid raft formation (Smart et al., 1996). Importantly and in line with the possibility of redundancy, even Caveolin-1 mutant mice only show a relatively mild phenotype considering its presumed central role in lipid raft biogenesis and function (Drab et al., 2001; Razani et al., 2001; Zhao et al., 2002).

Finally, the pronephric phenotype observed by interfering with lipid rafts (i.e. shortened proximal tubules) is highly reminiscent to two recent reports by our group (Zhang et al., 2013; Romaker et al., 2014). Indeed, we believe that all three phenotypes are interconnected. Zhang et al. (2013) described that changes in cAMP levels cause increased endocytosis and decreased exocytosis, while Romaker et al. (2014) identified the role of Insulin and Igf2 in promoting proximal tubule growth. Insulin receptor signaling is well characterized (Pike, 2005; Taniguchi et al., 2006). One particular feature is that the localization of the Insulin/Igf receptors to rafts is essential for active signaling (Gustavsson et al., 1999; Vainio et al., 2002; Karlsson et al., 2004). The lipid rafts are thought to potentiate the signaling efficiency by local clustering of the receptors and downstream signaling components (Pike, 2005). Thus, the simplest scenario for the data presented here (Fig. 5J,J’) is that Insulin/Igf2 trigger proliferation by binding to Insulin/Igf receptors clustered in lipid rafts. In the absence of scp2 and the reduction of lipid raft numbers, this process is impaired causing reduced proximal tubular growth. One alternative, not necessarily exclusive hypothesis is that the secretion of Insulin/Igf2 from the proximal tubules is reduced. Indeed, lipid rafts are the place of intense vesicle trafficking and exocytosis (Salaun et al., 2004). This would also be in line with the data presented in Zhang et al. (2013), which demonstrate that inhibition of secretion by Golgicide A causes shortened proximal tubules. In the future, it will be important to experimentally address these two possibilities by directly assaying ligand secretion and receptor activation using e.g. GFP-tagged versions of Insulin or Igf2. But besides these open questions, it is very intriguing that three quite different experimental approaches converged on a single process.

In summary, this study exemplifies the advantages of Xenopus as a model to bridge cell culture and in vivo experiments. The accessibility of the developing frog embryo to chemical compounds (Adams and Levin, 2006; Kalin et al., 2009) and microinjections allowed us to directly compare a classical cell biological approach with a modern loss-of-function study. This type of approach will undoubtedly allow us to translate observations seen in cell culture to a relevant in vivo situation and better understand the basic molecular mechanisms involving lipid rafts.

Highlights.

• Visualization of lipid rafts in the Xenopus pronephric kidney

• First time characterization of the Xenopus homologue of Sterol Carrier Protein 2

• Interfering with lipid rafts disrupts pronephric size control mechanisms

• Sterol Carrier Protein 2 is a very valuable tool to study lipid raft function

• Use of metabolomics to study developmental processes