RAPGEF5 Regulates Nuclear Translocation of β-Catenin

Abstract

Canonical Wnt signaling coordinates many critical aspects of embryonic development, while dysregulated Wnt signaling contributes to common diseases, including congenital malformations and cancer. The nuclear localization of β-catenin is the defining step in pathway activation. However, despite intensive investigation, the mechanisms regulating β-catenin nuclear transport remain undefined. In a patient with congenital heart disease and heterotaxy, a disorder of left-right patterning, we previously identified the guanine nucleotide exchange factor, RAPGEF5. Here, we demonstrate that RAPGEF5 regulates left-right patterning via Wnt signaling. In particular, RAPGEF5 regulates the nuclear translocation of β-catenin independently of both β-catenin cytoplasmic stabilization and the importin β1/Ran-mediated transport system. We propose a model whereby RAPGEF5 activates the nuclear GTPases, Rap1a/b, to facilitate the nuclear transport of β-catenin, defining a parallel nuclear transport pathway to Ran. Our results suggest new targets for modulating Wnt signaling in disease states.

Keywords: GTPase; RAPGEF5; Ran independent; Rap; Wnt signaling; Xenopus; congenital heart disease; heterotaxy; nuclear transport; β-catenin.

Figure 1. Rapgef5 depletion disrupts left-right development

(A) Percentage of Rapgef5 depleted embryos with abnormal cardiac looping (A or L loops). (B) Pitx2c is expressed in the left lateral mesoderm of stage 28 control Xenopus embryos (red arrow, lateral view dorsal to the top), but is abnormally, typically bilaterally, expressed following MO or CRISPR mediated depletion of Rapgef5. Co-injection of human RAPGEF5 mRNA can rescue pitx2c in Rapgef5 morphants. (C) coco expression in the LRO of control and Rapgef5 depleted embryos at stages 16 and 19. Ventral view with anterior to the top. Graphs depict the percentage of embryos displaying abnormal coco expression. Note the reduced expression in rapgef5 morphants at both stages. (D) xnr1 and gdf3 expression is reduced in the LRO of rapgef5 morphants at stage 16, prior to the onset of cilia driven flow. Ventral views with anterior to the top. A single asterisk indicates statistical significance of P<0.05, while double and triple asterisks indicate P<0.01 and P<0.005, respectively. Also see Figure S1 and Figure S2.

Figure 2. Depletion of Rapgef5 impairs canonical Wnt signaling

(A) Depletion of Rapgef5 using MOs or CRISPR impairs foxj1 and xnr3 expression in the dorsal blastopore lip of stage 10 embryos. (B) Simplified schematic of the canonical Wnt signaling pathway. Left side: in the absence of Wnt ligand, a destruction complex containing Axin and Gsk3 phosphorylates β-catenin, which marks it for cytoplasmic destruction. Right side: Once the pathway is activated by Wnt ligand binding to Frizzled and Lrp receptors, phosphorylation of β-catenin by GSK3 is inhibited allowing β-catenin to accumulate in the cytoplasm and translocate into the nucleus to initiate transcription of Wnt target genes. (C, D) Levels of total and active β-catenin protein are essentially unchanged in Rapgef5 depleted embryos at stage 9 as assayed by western blot. However, both forms of β-catenin are reduced at stages 10 and 12. Note that levels of active β-catenin are much more severely affected. Conversely, overexpression of human RAPGEF5 mRNA results in a mild increase in total β-catenin levels and a more pronounced increase in active β-catenin levels. Triple asterisks indicates P<0.005. Also see Figure S3.

Figure 3. Rapgef5 acts downstream of β-catenin cytoplasmic stabilization

(A) The ability to induce secondary axes in development can be used as a read out of Wnt signaling activity. Uninjected embryos have a single axis (dotted green line) while embryos injected with β-catenin mRNA can have a second embryonic axis (two dotted red lines) that are readily detected at st 16–19 embryos. Injection of WT, stabilized (ST), or NLS tagged β-catenin can induce secondary axes. Rapgef5 depletion significantly decreases the percentage of secondary axes induced by WT and ST β-catenin. Injection of GSK3 mRNA reduces secondary axes induced by WT β-catenin but has no effect on ST β-catenin. The ability of NLS-β-catenin to induce secondary axes is unaffected by reduction of Rapgef5 levels. (B) Rapgef5 knockdown reduces luciferase activity in embryos injected with WT or ST β-catenin mRNA in a TOPFlash assay. Data are represented as mean ± SD. (C) Wnt signaling activity is reduced in ST β-catenin MEFs (Δexon3) following siRNA depletion of Rapgef5. The schematic depicts the Catnblox(ex3) mouse allele. Exon 3 (E3), containing the GSK3 phosphorylation sites that target β-catenin for cytoplasmic degradation, is flanked by loxp sites allowing for its conditional removal and production of a stabilized (ST) β-catenin allele. WT: WT cells; FOP: ST β-catenin cells transfected with FOPFlash negative control; TOP: ST β-catenin cells transfected with TOPFlash reporter plasmid; TOP + R5 siRNA: ST β-catenin cells transfected with Rapgef5 siRNA and TOPFlash reporter plasmid; Top + C. siRNA: ST β-catenin cells transfected with control siRNA and TOPFlash reporter plasmid. A Renilla luciferase transfection control was included in each treatment to allow normalization. Data are represented as mean ± SEM. (D) Pharmacological inhibition of GSK3 by the addition of BIO between stages 9 – 11 results in increased β-catenin signaling and loss of anterior development in Xenopus embryos. Depletion of Rapgef5 can counteract this effect and rescue development of the head demonstrating that Rapgef5 regulates Wnt signaling downstream of GSK3. A single asterisk indicates P<0.05, while double and triple asterisks indicate P<0.01 and P<0.005, respectively.

Figure 4. Rapgef5 is required for the nuclear localization of β-catenin

(A,B) GFP tagged WT and ST β-catenin localize to the plasma membrane and nucleus in the dorsal blastopore lip of stage 10 control embryos but this nuclear localization is lost in rapgef5 morphants (note loss of nuclear GFP signal in merged images). (C) NLS β-catenin-GFP localizes normally into nuclei even in the absence of Rapgef5. Graphs in the center panel represent the ratio of nuclear localized GFP relative to nuclear localized NLS-Cherry control. Controls were normalized to ease comparison. The graphs on the right display ratiometric analysis of nuclear vs cytoplasmic GFP (β-catenin) and NLS-mCherry levels. Nuclear localized β-catenin is reduced in Rapgef5 morphants relative to controls for WT and stabilized β-catenin but not NLS-β-catenin. Localization of NLS-mCherry is unaltered. All data represented as mean ± SE. A double asterisk indicates P<0.01, a triple asterisk p<0.005.

Figure 5. NLS β-catenin rescues foxj1 transcription and LR patterning defects in Rapgef5 morphants

Depletion of Rapgef5 reduces the expression of the Wnt responsive gene foxj1 and alters LR patterning (pitx2c). Co-injection of 50pg of NLS-β-catenin mRNA significantly rescues (A) expression of foxj1 at stage 10 and (B) pitx2c expression in Rapgef5 depleted embryos. A single asterisk indicates P<0.05, a double asterisk P<0.01.

Figure 6. Depletion of Rapgef5 impairs nuclear translocation of endogenous β-catenin

(A) Nuclear/cytoplasmic fractionation reveals that levels of endogenous β-catenin are reduced in the nuclei of Rapgef5 morphants, while cytoplasmic levels are essentially normal. (B) BIO treatment does not rescue the nuclear localization of endogenous β-catenin. (C) Quantification of average β-catenin levels in nuclear and cytoplasmic fractions of control and morphant embryos with or without BIO treatment (error bars represent S.E.M.). All nuclear treatments (black bars) are relativized to the BIO untreated nuclear control and all cytoplasmic treatments (grey bars) are relativized to the BIO untreated cytoplasmic control. Note the reduction of nuclear localized β-catenin in Rapgef5 depleted embryos with or without BIO treatment. The BIO treated rapgef5 morphants have less nuclear localized β-catenin than BIO untreated UC controls, despite having higher levels of cytoplasmic β-catenin. The cellular compartment markers (H3 for nuclear fraction and β-actin for cytoplasmic fraction) displayed in A and B are overexposed in order to demonstrate the degree of purification. For quantitation of β-catenin (C), shorter exposures of the gel (unsaturated) were used in which these markers serve to normalize the amount of protein loaded for each compartment. A single asterisk indicates P<0.05, a double asterisk P<0.01.

Figure 7. Active Raps and β-catenin interact

(A) The GFP-^RBDRalGDS active Rap sensor reveals the presence of active Raps in the nucleus and plasma membrane of dorsal blastopore lip cells at stage 10. (B) Rap1 protein co-immunoprecipitates with β-catenin. (C) Immunoprecipitation of active Rap proteins from stage 10 Xenopus embryos using the RalGDS sensor. β-catenin immunoprecipitates with active Raps demonstrating an interaction, BC; β-catenin. (D) Injection of CA (constitutively active) Rap1b mRNA can partially rescue loss of foxj1 expression in rapgef5 morphants, while DN (dominant negative) Rap1b mRNA does not. A triple asterisk indicates P<0.0001. Also see Figure S7.