Report

Requirement of Prorenin Receptor and Vacuolar H+-ATPase–Mediated Acidification for Wnt Signaling

See allHide authors and affiliations

Science  22 Jan 2010:
Vol. 327, Issue 5964, pp. 459-463
DOI: 10.1126/science.1179802

Abstract

Wnt/β-catenin signaling is important in stem cell biology, embryonic development, and disease, including cancer. However, the mechanism of Wnt signal transmission, notably how the receptors are activated, remains incompletely understood. We found that the prorenin receptor (PRR) is a component of the Wnt receptor complex. PRR functions in a renin-independent manner as an adaptor between Wnt receptors and the vacuolar H+–adenosine triphosphatase (V-ATPase) complex. Moreover, PRR and V-ATPase were required to mediate Wnt signaling during antero-posterior patterning of Xenopus early central nervous system development. The results reveal an unsuspected role for the prorenin receptor, V-ATPase activity, and acidification during Wnt/β-catenin signaling.

Wnt/β-catenin signaling is implicated in stem cell biology and human disease, including cancer, and has important roles during embryonic development, such as axis formation and patterning of the central nervous system (14). Wnt binding to its receptors, low-density lipoprotein receptor-related protein 6 (LRP6) and frizzled (Fz), induces receptor aggregation in signalosomes (5) and phosphorylation of LRP6 by the kinases casein kinase 1γ (CK1γ) (6) and glycogen synthase kinase 3 (GSK3) (7). This requires the action of dishevelled (Dvl) and leads to recruitment of the negative regulator axin, thus stabilizing β-catenin (5, 8, 9).

To identify previously undescribed Wnt pathway components regulating Wnt receptors, we carried out a genome-wide small inhibitory RNA (siRNA) screen and identified the prorenin receptor gene (PRR) (10). In brief, human embryonic kidney (HEK293T) cells were transfected individually with siRNA pools targeting about 18,500 human genes, stimulated with Wnt3a, and analyzed for transcription of a Wnt-responsive luciferase reporter (11). The PRR is a single spanning transmembrane protein located at the plasma membrane that transmits renin and prorenin signals (1216). Hypomorphic PRR mutation causes mental retardation and epilepsy in humans (17). PRR has a short cytoplasmic domain mediating renin signal transduction but contains no obvious motifs (18).

Three independent siRNAs targeting PRR inhibited luciferase reporter activity stimulated by Wnt3a but not by β-catenin, attesting to the specificity of the effect (Fig. 1A and fig. S1). Although Wnt1 or Wnt3a signaling was inhibited by PRR siRNA, signaling induced by downstream components of the Wnt pathway, including constitutively active LRP6 (LRP6ΔE1-4), Dvl, or β-catenin, remained unaffected (Fig. 1B). This indicates that PRR is required for Wnt/β-catenin signaling at the level of or upstream of the coreceptor LRP6, consistent with it being a transmembrane protein. Overexpressed PRR did not activate Wnt/β-catenin signaling by itself, but a C-terminally truncated construct (PRRΔC) synergized with Wnt3a in reporter activation (Fig. 1C), in causing duplication of the Xenopus embryonic axis (Fig. 1D), as well as in increasing transcription of the direct Wnt response gene siamois in Xenopus animal cap assays (Fig. 1E).

Fig. 1

Regulation of Wnt/β-catenin signaling by PRR at the level of or upstream of LRP6. (A to C) Wnt luciferase reporter assays in HEK293T cells stimulated with Wnt3a-conditioned medium or by transfection with the indicated constructs, in the presence of the indicated siRNAs. Co, reporter only; siPRR, siRNA pool. Error bars indicate SDs; N = 2 in (A) and (B) and N = 3 in (C). (D) Axis duplication assay by injection of the indicated Xenopus mRNAs into the ventral blastomeres of Xenopus embryos at the four-cell stage. (E) Reverse transcription polymerase chain reaction (RT-PCR) analysis of animal caps from Xenopus embryos injected with the indicated mRNAs. Animal caps were excised from blastula embryos and cultivated until stage 10. Sia, siamois; H4, histone H4.

To analyze whether PRR is required for Wnt signaling in vivo, we studied its role in Xenopus embryos, where the gene shows weak expression in most tissues and prominent expression in the central nervous system (fig. S2A). Tadpoles that developed from cleavage-stage embryos injected with PRR antisense morpholino (Mo) oligonucleotides had small heads, shortened tails, and defects in melanocyte and eye pigmentation (Fig. 2, A and B). This phenotype appeared to be specific because it was efficiently rescued by coinjection of human PRR mRNA (Fig. 2, A and B). A very similar phenotype was observed for zebrafish PRR mutant embryos (19). Phenotypic rescue was also obtained by constructs with deletion in the intracellular domain (ΔC) but not those with deletions in the extracellular or transmembrane domain (Fig. 2B; see Fig. 3B for constructs), suggesting that PRR does not directly transduce a cytoplasmic signal in Wnt signaling.

Fig. 2

Requirement of PRR for Wnt/β-catenin signaling and antero-posterior neural patterning in Xenopus. (A and B) Tailbud stage Xenopus embryos injected in all blastomeres of the animal hemisphere at the four-cell stage with PRR antisense Mo oligonucleotides in the absence or presence of wild-type or mutated human PRR mRNAs. (For PRR mutants, see Fig. 3B.) CoMo, control Mo; Co, PPL mRNA. (C) Wnt luciferase reporter assay of whole embryos at the indicated stages injected in all blastomeres at the animal pole with PRR Mo and/or Wnt3a mRNA. Luciferase activity in uninjected embryos was set to 1. Error bars indicate SDs; N = 3. (D) Whole-mount in situ hybridization of neurula-stage embryos injected in the animal blastomeres with the indicated Mo plus β-galactosidase mRNA lineage tracer (red or light blue), showing reduced expression (arrowheads) of otx2 (65%, n = 26) and en2 (96%, n = 25), but not of Krox20 (0%, n = 24). (E and F) Quantitative PCR (qPCR) analysis of indicated genes in neurulae (E) or animal caps cultured until neurula stage (F). Eight-cell-stage embryos were injected into indicated blastomeres (E) or animal 4 blastomeres (F) with PRRMo and the following mRNA and DNA: chordin mRNA, 250 pg; PRR mRNA, 200 pg; Wnt3a DNA, 50 pg; and β-catenin DNA, 100 pg. Error bars show SDs of biological triplicates. Gene expression in CoMo or uninjected explants was set to 1.

Fig. 3

Binding of PRR to LRP6, Fz8, and components of the vacuolar H+-ATPase. (A) CoIP from lysates of HEK293T cells transfected with the indicated constructs. (B) Schematic representation of PRR constructs used. SP, signal peptide; TM, transmembrane domain; C, cytoplasmic domain; and TMC, cytoplasmic and transmembrane domains. (C) Wnt luciferase reporter assay in HEK293T cells transfected with control or PRR siRNA and stimulated with Wnt3a-conditioned medium in absence or presence of the indicated transfected Flag-tagged PRR constructs. Error bars indicate SDs; N = 3. (D) CoIP of endogenous PRR with endogenous LRP6 from lysates of HEK293T cells in the presence or absence of Wnt3a. (E) CoIP of V5-PRR with Flag-tagged ATP6V0C from lysates of transfected HEK293T cells. Flag-FLRT3 serves as negative control. (F) CoIP of endogenous ATP6V0D1 and endogenous ATP6V0C with endogenous PRR from HEK293T lysates using antibody against PRR (anti-PRR). Anti-PLSCR1 serves as negative control.

In Wnt reporter assays in Xenopus embryos, PRR Mo inhibited both endogenous and Wnt3a-stimulated expression of luciferase (Fig. 2C). The requirement for PRR was specific for Wnt signaling. Nodal, fibroblast growth factor (FGF), and bone morphogenetic protein (BMP) signaling in Xenopus (fig. S3) and transforming growth factor–β (TGFβ), BMP, tumor necrosis factor–α (TNFα), FGF, or phorbol myristyl acetate (PMA) signaling in HEK293T cells (fig. S4) were not affected by depletion of PRR.

Marker gene analysis in embryos depleted of PRR showed down-regulation of the forebrain marker otx2 and the mid-hindbrain boundary marker engrailed2 (en2), whereas the hindbrain marker Krox20 was unaffected (Fig. 2D). Such embryos showed no changes in cell proliferation or apoptosis in the central nervous system (CNS) (fig. S5, A and B). They also showed no change in the mesodermal markers Xbra and chordin (fig. S5C). Taken together, these results indicate that loss of anterior marker gene expression represents a defect of antero-posterior neural patterning, wherein Wnt signaling is prominently involved (20). Injection of PRR Mo in the animal region, which gives rise to the neuroectoderm, reduced expression of en2 and of the direct Wnt target gene Axin2, but not expression of otx2. Conversely, injection of PRR Mo in the vegetal region, thus targeting mesendodermal precursors, reduced expression of otx2 but not that of en2 (Fig. 2E). These results suggest an indirect PRR requirement for otx2 expression and forebrain development through its effect to promote formation of dorsal mesoderm and a cell-autonomous requirement of PRR for en2 expression and development of the mid-hindbrain. We conclude that PRR is required for Wnt signaling and anterior CNS patterning in Xenopus embryos.

Because en2 is a well-characterized, direct Wnt target gene (21, 22), we focused on the requirement of PRR for en2 expression. Xenopus animal caps were neuralized by injection with the BMP inhibitor chordin, which increased transcription of the gene encoding the pan-neural marker NCAM as well as the gene en2 (Fig. 2F). PRR Mo specifically abolished expression of en2 and Axin2 but left NCAM unaffected. Expression of en2 and Axin2 was rescued by coinjection of human PRR mRNA as well as by β-catenin, but not by Wnt3a DNA. These results (i) corroborate that PRR is essential for en2 expression because of its requirement in Wnt signaling and (ii) indicate that PRR functions downstream of Wnts and upstream of β-catenin, consistent with the cell culture data, which placed PRR action at the level of Wnt receptors.

We therefore tested whether PRR bound to frizzled 8 (Fz8) or LRP6. In coimmunoprecipitation (CoIP) experiments with transfected cells, PRR bound to both Fz8 and LRP6 but not to the control transmembrane protein FLRT3 (Fig. 3A). Deletion of the cytoplasmic domain (ΔC), which mediates renin signaling, had no effect on Wnt receptor binding (Fig. 3A) or the ability of the protein to rescue Wnt signaling in PRR siRNA-treated cells (Fig. 3C) or Mo-treated embryos (Fig. 2B). Although transmembrane domain–deleted protein (ΔTMC) also bound to the receptors, albeit weakly (Fig. 3A), it failed to rescue Wnt signaling (Figs. 3C and 2B), suggesting that PRR transmembrane localization is essential. In contrast, the extracellular domain (ECD) was necessary for binding LRP6 or Fz8 and for Wnt signaling (Figs. 3A and 2B). These results corroborate the specificity of physical PRR-Wnt receptor interactions, which require the PRR ECD.

We confirmed these interactions in binding assays using soluble, recombinant proteins (fig. S6), as well as by CoIP of endogenous LRP6 and PRR (Fig. 3D). We also tested whether PRR binds Wnt but found no significant interaction. We conclude that both Wnt receptor binding and Wnt function require the PRR ECD but not the intracellular domain.

Prorenin is not expressed in early Xenopus embryos, and addition of renin had no effect in Wnt luciferase assays (fig. S7, A and B). Further, the PRR cytoplasmic domain, which mediates renin signaling, is not required for its role in Wnt signaling. Moreover, there are PRR homologs in Drosophila and Hydra that have no renin. Thus, PRR may function in Wnt signaling in a renin-independent manner. Indeed, a PRR fragment, identified as adenosine triphosphatase (ATPase), H+-transporting, lysosomal accessory protein 2 (ATP6AP2), interacts with the vacuolar H+-ATPase (V-ATPase) (23). We confirmed that PRR binds (directly or indirectly) to the V-ATPase subunits ATP6V0C and ATP6V0D1 but not to control transmembrane proteins (Fig. 3, E and F). We mapped the interaction domain of PRR with ATP6V0C and show that the transmembrane and the ECD of PRR are required for binding ATP6V0C (fig. S8). The results indicate that PRR is associated with the V-ATPase, although the subunit directly contacting PRR remains to be determined.

The V-ATPase is a multiprotein complex localized in intracellular organelles and at the plasma membrane. It is involved in diverse processes such as phagocytosis, virus entry, metastasis, and embryonic left-right patterning. Its main mechanism is to pump protons and acidify vesicles, thereby promoting vesicular traffic, notably endocytosis (24, 25). Disruption of pH homeostasis in V-ATPase mutants leads to lethality in various organisms (26).

LRP6 signal transduction involves receptor aggregation in signalosomes and phosphorylation, for example at Thr1479, a process which requires Dvl (5). LRP6 phosphorylation is accompanied by receptor internalization in caveolin-containing vesicles, and endocytosis is essential for Wnt/β-catenin signaling (27, 28). This raised the possibility that PRR and V-ATPase may influence LRP6 endocytosis, phosphorylation, and β-catenin activation.

Loss of function and pharmacological inhibition of V-ATPase in vitro and in vivo showed that this enzyme is required for Wnt/β-catenin signaling. In reporter assays, treatment of HEK293T cells with siRNAs targeting two subunits of V-ATPase (ATP6V1C2 and ATP6V0C) inhibited Wnt signaling (Fig. 4A). Likewise, two pharmacologic V-ATPase inhibitors, apicularen and bafilomycin (29), inhibited Wnt signaling (Fig. 4B). Neither V-ATPase siRNAs nor the pharmacologic inhibitors affected Wnt signaling stimulated by Dvl, constitutively active LRP6 (LRP6ΔE1-4), or β-catenin (Fig. 4A and fig. S9), consistent with a specific requirement for Wnt receptor activation. Furthermore, in Xenopus embryos, injection of mRNA encoding YCHE78, a well-characterized dominant-negative V-ATPase subunit E (30), elicited a very similar phenotype to that observed by treatment with PRR Mo. It inhibited expression of otx2 and en2 but not that of Krox20 (Fig. 4C). In Wnt reporter assays in Xenopus embryos, YCHE78 inhibited both endogenous as well as Wnt3a-stimulated reporter activity, and low YCHE78 doses synergized with PRR Mo in Wnt inhibition, indicative of functional interaction (fig. S10). Lastly, in chordin-neuralized animal caps, YCHE78 inhibited expression of en2 and Axin2, and this was rescued specifically by β-catenin but not by Wnt3a (Fig. 4D).

Fig. 4

Requirement of V-ATPase and acidification for Wnt/β-catenin signaling. (A and B) Wnt luciferase reporter assays in HEK293T cells stimulated with Wnt3a-conditioned medium or by transfection with Wnt1 and the indicated constructs in the presence of the indicated siRNAs or the V-ATPase inhibitors apicularen A and bafilomycin A1. Error bars indicate SDs; N = 2 and N = 3, respectively. (C) Tadpole-stage Xenopus embryos that had been injected in all animal blastomeres at the four-cell stage as indicated. (Bottom) In situ hybridization of otx2, en2, and Krox20 at neurula stage. Injection of YCHE78 mRNA unilaterally reduced expression of otx2 (67%, n = 60) and en2 (70%, n = 27) but not significantly Krox2 (11%, n = 27), as indicated by arrowheads. Co-injected β-galactosidase mRNA was used as lineage tracer (red). (D) qPCR analysis of the indicated mRNAs was performed as described in Fig. 2F. (E) Western blot of endogenous LRP6 and PRR from NTERA2 cells treated with the indicated siRNAs and stimulated for 1 hour with Wnt3a-conditioned medium or control medium. tot. LRP6, total LRP6. (F) Live-cell confocal microscopy of ratiometric LRP6 in acidic compartments. HEK293T cells were transfected with Ra-LRP6 or Ra-LRP6ΔE1-4 and membrane-anchored RFP (red) and treated for 1 hour with control or Wnt3a-conditioned medium, respectively, in the absence or presence of apicularen (Api). Images were acquired by excitation at 405 and 488 nm and subtracting frames (f488 – f405) to monitor reporter proteins in acidic compartments (green).

Our data indicate that V-ATPase activity is required for activation of the Wnt receptor. To test this, we treated mouse P19 embryonal carcinoma cells with Wnt3a and monitored phosphorylation of LRP6 by immunofluorescence microscopy (6). Phosphorylation of LRP6 was inhibited in cells treated with PRR/V-ATPase siRNA (fig. S11). Immunoblotting also showed that Wnt-stimulated phosphorylation of LRP6 in human teratocarcinoma (NTERA2) cells was inhibited by depletion of PRR, Dvl1-3, or ATP6V0C (Fig. 4E) or by treatment of NTERA2 and P19 cells with apicularen and bafilomycin (fig. S12). Similarly, Wnt3a-induced expression of Axin2 was inhibited upon treatment of neuroblastoma SHEP cells with apicularen (fig. S12).

These results demonstrate that phosphorylation of LRP6 (which correlates with LRP6 activation) requires V-ATPase activity, suggesting that the receptor may need to enter an acidic intracellular compartment to become phosphorylated. To analyze acidification directly, we fused the extracellular domain of full-length LRP6 with the green fluorescent protein (GFP) pH-sensor variant pHLuorin, which can be monitored by ratiometric imaging (31) (fig. S13, A and B). Fluorometric analysis of ratiometric-LRP6 (Ra-LRP6) in cell lysates indicated that the protein shows a higher fluorescence at the excitation wavelength of 488 nm than at 405 nm, when the pH is below pH ~ 6.5 (fig. S13, C and D). In unstimulated live cells, fluorescence of transfected Ra-LRP6 was mostly undetectable. However, within minutes of Wnt treatment, Ra-LRP6 fluorescence was observed in intracellular vesicles (movie S1). The signal plateaued after 1 hour and was inhibited by apicularen treatment (Fig. 4F and fig. S13E). We also tested a fusion protein between pHLuorin and constitutively active LRP6 (Ra-LRP6ΔE1-4). LRP6ΔE1-4 is spontaneously aggregating, is constitutively phosphorylated, and resides in signalosomes (5). Consistent with this, Ra-LRP6ΔE1-4 showed punctate fluorescence even in unstimulated cells, indicating that it enters acidic vesicles spontaneously (Fig. 4F). Ra-LRP6 may also be a useful tool to monitor acute Wnt signaling in living cells.

Our data reveal an unsuspected role of the prorenin receptor PRR in Wnt/β-catenin signaling and provide evidence that this multifunctional protein interacts with V-ATPase. We propose a mechanism wherein PRR is part of the Wnt receptor complex, acting as a specific adaptor between LRP6 and V-ATPase. Upon Wnt stimulation, this signaling complex is endocytosed, and across the vesicle membrane V-ATPase generates a proton gradient that is essential for LRP6 phosphorylation and hence β-catenin activation. Our results raise intriguing questions about the role of PRR in renin signaling and mental retardation and the immediate consequence of acidification in Wnt receptor signaling. The Na+-H+ exchanger Nhe2 is required for Fz–planar cell polarity signaling in Drosophila (32), suggesting that electrochemical regulation may have multiple roles in Wnt receptor signaling. The V-ATPase may also provide a therapeutic target to modulate Wnt signaling in a disease context.

Supporting Online Material

www.sciencemag.org/cgi/content/full/327/5964/459/DC1

Materials and Methods

Figs. S1 to S13

Table S1

References

Movie S1

  • * These authors contributed equally to this work.

References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. We thank H. Clevers, R. Grosschedl, X. He, R. Moon, J. Nathans, R. Nusse, M. Levin, G. Miesenböck, Y. M. Chan, M. A. Skinner, and M. Lorizate for reagents; Y. L. Huang for confocal microscopy; A. Glinka for recombinant Wnt3a-V5; and T. Büchling and K. Bartscherer for sharing data before publication and discussion. This work was supported by the Deutsche Forschungsgemeinschaft and the European Commission (Endotrack and Marie-Curie Program). S.P.A. is a recipient of a fellowship from Gobierno Vasco.
View Abstract

Navigate This Article