G Protein Signaling from Activated Rat Frizzled-1 to the β-Catenin-Lef-Tcf Pathway

See allHide authors and affiliations

Science  01 Jun 2001:
Vol. 292, Issue 5522, pp. 1718-1722
DOI: 10.1126/science.1060100


The frizzled receptors, which mediate development and display seven hydrophobic, membrane-spanning segments, are cell membrane–localized. We constructed a chimeric receptor with the ligand-binding and transmembrane segments from the β2-adrenergic receptor (β2AR) and the cytoplasmic domains from rat Frizzled-1 (Rfz1). Stimulation of mouse F9 clones expressing the chimera (β2AR-Rfz1) with the β-adrenergic agonist isoproterenol stimulated stabilization of β-catenin, activation of a β-catenin–sensitive promoter, and formation of primitive endoderm. The response was blocked by inactivation of pertussis toxin–sensitive, heterotrimeric guanine nucleotide–binding proteins (G proteins) and by depletion of Gαq and Gαo. Thus, G proteins are elements of Wnt/Frizzled-1 signaling to the β-catenin–lymphoid-enhancer factor (LEF)-T cell factor (Tcf) pathway.

Wnts constitute a family of vertebrate genes encoding ligands essential to signaling in early development, signaling that includes control of cell proliferation, cell fate, and embryonic patterning (1). These secreted glycoproteins act via members of the frizzled gene family (2–4). Signaling downstream of some Frizzled homologs in response to Wnt-1 or Wnt-8 leads to activation of the phosphoprotein Dishevelled (Dsh/Dvl), which then represses the function of glycogen synthase kinase-3β (GSK-3) activity (5,6). In the absence of Wnt, GSK-3 phosphorylates β-catenin, reducing its stability and abundance. Wnt signaling represses GSK-3 activity, thereby increasing the stability and intracellular accumulation of β-catenin, which then accumulates in the nucleus where it binds to members of the Lef-Tcf classes of architectural high-mobility group box transcription factors to activate genes involved in early development. Analysis of this pathway has been hindered by the lack of easy experimental methods for tightly controlling the receptor activation state or inhibition of the receptor.

To allow the rapid activation and inhibition of Frizzled coupled to the β-catenin pathway, we created a chimeric receptor consisting of the extracellular and transmembrane segments of the hamster β2AR and the cytoplasmic domains of the Rfz1 (Fig. 1A). The sequence of the Rfz1 cytoplasmic domains diverge from those of the β2AR (7, 8). This chimeric receptor has the potential to be activated by soluble drugs of well-known pharmacology. Mouse F9 teratocarcinoma cells were stably transfected with an expression vector harboring the β2AR-Rfz1 chimera (9). Clones expressing mRNA encoding the Rfz-1 chimeric receptor in large amounts were identified by reverse transcription–polymerase chain reaction (RT-PCR) and propagated (Fig. 1B). Expression of the chimeric receptor was quantified readily using labeled iodocyanopindolol (ICYP), a high-affinity β-adrenergic antagonist ligand that binds specifically to the transmembrane domain of the β2AR. ICYP-binding studies of Chinese hamster ovary (CHO) clones (which do not express endogenous β2AR) stably transfected with pβ2AR-Rfz1 vector demonstrate a K dof ∼80 pM and a maximal binding capacity (B max) of 2 to 4 pmol ICYP binding per milligram of protein (10, 11). Immunoblots of cell membranes from F9 clones expressing the β2AR-Rfz1 chimera stained with antibodies to an extracellular epitope of the β2AR identified the endogenous 65-kD β2AR and a 55-kD molecular species with the predicted size of the chimera (Fig. 1C). ICYP binding to the β2AR-Rfz1 chimera displays a rightward shift of the affinity of the chimera for the β-adrenergic agonist isoproterenol (ISO) in the presence of a GTP analog (GTP-γ-S,Fig. 1D). All heptihelical receptors known to operate via heterotrimeric G proteins display this characteristic GTP-dependent shift in agonist affinity (12).

Figure 1

Design and expression of a β2AR-Rfz1 chimera in F9 stem cells. (A) Schematic of the chimera. (B), RT-PCR of β2AR-Rfz1 mRNA expression in F9 clones. The RNA of F9 clones harboring either the empty expression vector (EV) or the vector expressing the β2AR-Rfz1 chimera were reverse transcribed and amplified. Two primers, CCGGCCTTACCTCCTTCTTGCCC and CCGTTCTGCGACTTGAGCACCTCC, were employed in the PCR amplification. The molecular markers (Mk) indicate the relative size in base pairs (bp) of the amplified products. (C) Immunoblotting of immunoprecipitates of F9 clones stably transfected with either empty vector (EV) or vector expressing the β2AR-Rfz1 chimera. Blots were stained with an antibody to an exofacial domain of the β2AR, making visible both the native β2AR of F9 cells (65 kD) as well as the β2AR-Rfz1 chimeric receptor species (55 kD). (D) Dose-response curve for inhibition of ICYP binding by the β-adrenergic agonist isoproterenol, performed in the absence (−) and presence (+) of 10 μM GTPγS. The results are mean values ±SEM from six determinations.

Before testing the signaling activity of β2AR-Rfz1, we needed an assay for signaling by Rfz1. F9 cells stably transfected to express wild-type Rfz1 form primitive endoderm (PE) when treated with conditioned medium containing Xwnt-8, as monitored by positive staining for the PE markers cytokeratin endo-A (i.e., antigen for the TROMA monoclonal antibody) and tissue plasminogen activator (13,14). Clones expressing the β2AR-Rfz1 chimera were therefore treated with the β-adrenergic agonist ISO or the β-adrenergic antagonist propranolol, or both. Isoproterenol stimulated formation of PE in stably transfected F9 stem cells and propranolol blocked this response (Fig. 2A). Taken together these data demonstrate the ability of the Rfz1 chimera to be stably expressed, to bind β-adrenergic ligands with normal pharmacological properties, to display a GTP-dependent agonist shift of affinity, and to signal to PE formation.

Figure 2

ISO-stimulated formation of primitive endoderm and activation of the β-catenin–Lef-Tcf pathway in cells expressing the β2AR-Rfz1 chimera. (A) F9 clones treated with β-agonist ISO (10 μM) or β-antagonist PROP (10 μM). Clones were treated with ligand for 4 days and then fixed and stained with TROMA-1 antibody to reveal PE formation. Indirect immunofluorescence (IIF) and phase-contrast (PC) images are displayed. (B) Xwnt-8–induced expression of pTOPFLASH. Conditioned medium was collected from clones stably transfected with Xwnt-5a, Xwnt-8, or the empty vector (EV) and used to supplement 1:9 (ratio of conditioned medium of target Rfz-expressing clones to that of the Wnt-expressing clones) the medium of the F9 cells stably expressing either Rfz1, Rfz2, or the empty vector, in tandem with transient cotransfection with pTOPFLASH. (C) ISO-induced expression of pTOPFLASH. Clones stably expressing either β2AR-Rfz1 or β2AR-Rfz2 chimera were transiently transfected with either pTOPFLASH or pFOPFLASH for 24 hours and then treated with ISO, PROP, or both, for 6 hours, and then activation of the luciferase reporter genes was measured. The data shown in (B) and (C) are the mean values ± SEM of four separate experiments and are reported in relative light units (RLU; 10,000 RLU equals expression of 1 pg of luciferase). * denotes P< 0.05 and ** P < 0.01 for the difference from unstimulated values (SigmaStat software, SPSS Science).

Because PE formation in culture measured above requires 3 to 4 days, we also tested whether activation of the wild-type Rfz1 in F9 cells would more rapidly induce the Lef- and Tcf-sensitive luciferase reporter of β-catenin signaling, pTOPFLASH (15, 16). Conditioned medium containing Xwnt-8, but not Xwnt-5a or empty vector EV, stimulated TOPFLASH expression (Fig. 2B). Wnt-5a signals via Rfz2 in a manner that is largely independent of β-catenin (5). Xwnt-5a stimulation of clones expressing rat Frizzled-2 (Rfz2) did not induce pTOPFLASH. We also measured the effect of ISO on the activation of the β-catenin–Lef-Tcf pathway in clones expressing the β2AR-Rfz1 (Fig. 2C). Transcription of pTOPFLASH reporter was increased in response to ISO in clones expressing the chimera, but not clones harboring either the mutated pFOPFLASH, a control luciferase reporter vector lacking the Lef-Tcf–sensitive elements, or the empty vector. The β-adrenergic antagonist propranolol (PROP) did not activate the transcription of pTOPFLASH, but did antagonize the stimulation by ISO. ISO failed to stimulate the pTOPFLASH reported in clones expressing the β2AR-Rfz2 version of the chimera, further demonstrating the specificity of the Rfz1 chimera activation of the Lef-Tcf pathway.

A hallmark of the Rfz1 pathway is the stabilization of β-catenin in response to Wnt. The time-course for β-catenin stabilization (Fig. 3A) and pTOPFLASH activation (Fig. 3B) was studied in clones expressing β2AR-Rfz1 (17). Stabilization of β-catenin was observed in clones expressing the chimera, within 1 to 2 hours of stimulation with ISO. The stabilization of β-catenin peaked at 3 hours after treatment with ISO, thereafter declining to basal levels within the next 3 hours. ISO stimulated increased activation of the Lef-Tcf–sensitive luciferase reporter gene within the first several hours. Unlike the β-catenin stabilization, the activation of the Lef-Tcf response continued to increase for up to 6 hours, 2 to 3 hours after the β-catenin levels had returned to basal levels.

Figure 3

Activation of β2AR-Rfz1 chimera in F9 stem cells provokes stabilization of β-catenin and the activation of the Lef-Tcf–sensitive luciferase reporter gene. Activation of β2AR-Rfz1 chimera in F9 stem cells after clones stably expressing the β2AR-Rfz1 chimera were treated with ISO at time = 0. Over the next 6 hours, the intracellular concentration of β-catenin and the activation of the Lef-Tcf–sensitive luciferase reporter gene pTOPFLASH were measured. (A) The stability of β-catenin was assessed by SDS-PAGE of the intracellular complement of β-catenin. The proteins were transferred to blots, stained with antibodies to β-catenin, and the amount of stain was quantified. (B) Transcriptional activation of pTOPFLASH. The data shown are the mean values ± SEM of five separate experiments and are reported in relative light units (RLU).

One of the fundamental questions in understanding how signaling pathways regulate developmental processes is whether or not the activation of β-catenin–Lef-Tcf pathway by Frizzledhomologs involves heterotrimeric G proteins. It has been reported that Frizzled signaling through the Wnt/Ca++ pathway, that does not involve β-catenin, requires G proteins (18, 19), but no data have shown a comparable role for G proteins in the Wnt–β-catenin pathway. The GTP-induced shift in agonist affinity of the β2AR-Rfz1 chimera (Fig. 1D) suggests that the chimera is a G protein–linked receptor. Pertussis toxin (PT) which ADP-ribosylates and inactivates several subfamilies of heterotrimeric G protein α-subunits, including Gi, Go, and Gt, provided an additional test. Treating β2AR-Rfz1 chimera-expressing clones with PT for 4 hours reduced the activation of pTOPFLASH by ISO (Fig. 4A). Further elucidation of specific G proteins involved in the Rfz1–β-catenin–Lef-Tcf pathway was performed via suppression of various G protein subunits with antisense oligodeoxynucleotides [ODN (20, 21)]. Suppression of either Gαo or Gαq resulted in a suppression of signaling from β2AR-Rfz1 to the activation of pTOPFLASH. These data reveal an obligate role for specific G proteins in signaling to the β-catenin–Lef-Tcf pathway. The overall pathway is sensitive to PT treatment and suppression of either Gαo (a PT substrate) or Gαq precludes the activation of the pathway, as highlighted in a 24-hour time-course for ISO-stimulated activation of pTOPFLASH transcription in clones expressing the β2AR-Rfz1 chimera (Fig. 4C). Thus, the β2AR-Rfz1 chimeric receptor displays an agonist-specific, GTP-dependent shift in receptor affinity (Fig. 1D) and a response sensitive both to PT and to suppression of Gαo and Gαq (Fig. 4A), properties characteristic of members of the superfamily of G protein–linked receptors. Results from the expression of constitutively active, GTPase forms of Gαo (Q205L) and of Gαq (Q209L) show that expression of Q209L Gαq and, to a lesser extent, expression of Q205L Gαo constitutively activate the Lef-Tcf pathway (Fig. 4D). Expression of constitutively active versions of Gα11 and Gαs, in constrast, had no effect on pTOPFLASH activity.

Figure 4

Activation of the β-catenin–Lef-Tcf–sensitive pathway: effects of PT and suppression of G protein subunits. Sensitivity of the ISO-stimulated pTOPFLASH activation to pretreatment with PT and ODNs antisense to G protein subunits. Clones expressing the β2AR-Rfz1 chimera were treated either for 4 hours with PT (10 ng/ml) or for 48 to 72 hours with antisense ODNs (A) and sense/missense (24) ODNs to suppress specific G protein subunits. Clones were then treated with ISO for an additional 6 hours, and the pTOPFLASH activation was measured. The data are the mean values of triplicate determinations of a representative experiment and are reported in relative light units (RLU). * denotesP < 0.05 for the difference from the cells not exposed to ISO (SigmaStat software, SPSS Science). (B) Immunoblots of F9 clones treated without (–) and with (+) ODNs antisense to the indicated G protein subunit and stained with antibodies against the same subunit (29). (C) Time-course for the activation of Lef-Tcf–sensitive luciferase reporter gene in F9 clones expressing the β2AR-Rfz1 chimera and treated with 10 μM ISO for 24 hours. The F9 clones were treated either with or without pertussis toxin (PT, 10 ng/ml, 4 hours prior) or with ODNs antisense to Gαo or to Gαq (48 hours prior). The results shown are mean values from three or four experiments. Treatment of F9 cells with antisense ODNs provokes a suppression of the expression of the targeted G protein subunits (13,19). (D) Effects of expression of constitutively activated mutant forms of the alpha subunits of Go (Q205L), Gq (Q209L), G11 (Q209L), and Gs (Q227L) in F9 clones on the activity of the Lef-Tcf–sensitive luciferase reporter gene. (E) Pertussis toxin inhibition of Wnt-dependent gene transcription in Xenopus embryos. Xenopus embryos were injected with synthetic RNAs, then blastula animal caps were isolated and analyzed for expression of the known Wnt target genesXnr-3 and Siamois by RT-PCR. Lane 1, animal caps of uninjected embryos as controls; lane 2, animal caps of embryos injected with PT A protomer; lane 3, animal caps of embryos injected with Xwnt-8 RNA to induce target genes, as a positive control; lane 4, animal caps of embryos injected with both RNAs for Xwnt-8 and the PT A protomer. The “–RT” is a control establishing the lack of genomic DNA when the reverse transcriptase was omitted. Histone H4, which is expressed throughout development, is a control establishing that the treatment with PT A promoter does not simply stop transcription. These experiments are representative of three separate experiments.

Our results above implicate PT-sensitive G proteins as playing a direct role in the activation of β-catenin target genes. However, Wuet al. (22) have recently reported that PT does not block the ability of Xwnt-8 to induce an ectopic axis inXenopus embryos, a process that relies on the expression of β-catenin target genes. Given the apparent discrepancy, we directly tested whether the ability of Xwnt-8 to induce expression of two known direct targets of β-catenin, Siamois and Xnr-3(23), could be blocked by pertussis toxin (Fig. 4E). We confirmed that blastula animal cap explants do not express either gene (Fig. 4E), and that PT does not provoke expression of either gene by itself. Expression of Xwnt-8 in explants caused increased transcription of both Siamois and Xnr-3. Injection of RNA encoding the A protomer of pertussis toxin before the injection of Xwnt-8 RNA, however, attenuated the activation of both of the target genes. Similarly, expression of EF1-alpha was equivalent in all four lanes (24). These data demonstrate that the PT treatment specifically inhibits transcription of β-catenin target genes. The data, measuring activation of endogenous direct targets of β-catenin signaling, confirm our conclusions in F9 cells that pertussis toxin attenuates activation of β-catenin target genes.

The classical Wnt–β-catenin signaling paradigm is based on the ability of frizzled signals to enhance the cellular accumulation of β-catenin in the nucleus and to activate the Lef-Tcf transcriptional complex (1, 5, 6). What has not been evident is whether there is a requirement for specific G protein subunits early in the response, before stabilization of β-catenin. Our studies with a chimeric receptor show that it binds agonist and provokes formation of PE. Activation of the chimera provokes stabilization of β-catenin and rapid transcription of a β-catenin–Lef-Tcf reporter gene. These effects are sensitive to PT as well as depletion of Gαο and Gαq. The ability of the cytoplasmic domains of the Rfz1 to enable the chimera to signal to the β-catenin–Lef-Tcf pathway and to enable an agonist-specific GTP-dependent shift of chimera receptor affinity argue that the Frizzled gene product is a G protein–linked receptor. Further evidence in support of our proposal is the recently reported ability of RGS proteins to inhibit Xwnt-8 signaling inXenopus (22). Our observations suggest that the activation of the classical β-catenin–Lef-Tcf pathway in mammalian cells is G protein–linked and that Rfz1 is a G protein–linked receptor.

  • * To whom correspondence should be addressed. E-mail: craig{at}


View Abstract

Navigate This Article