A beta-catenin-dependent Wnt pathway mediates anteroposterior axon guidance in C. elegans motor neurons

Abstract

Background: Wnts are secreted glycoproteins that regulate diverse aspects of development, including cell proliferation, cell fate specification and differentiation. More recently, Wnts have been shown to direct axon guidance in vertebrates, flies and worms. However, little is known about the intracellular signaling pathways downstream of Wnts in axon guidance.

Methodology/principal findings: Here we show that the posterior C. elegans Wnt protein LIN-44 repels the axons of the adjacent D-type motor neurons by activating its receptor LIN-17/Frizzled on the neurons. Moreover, mutations in mig-5/Disheveled, gsk-3, pry-1/Axin, bar-1/beta-catenin and pop-1/TCF, also cause disrupted D-type axon pathfinding. Reduced BAR-1/beta-catenin activity in D-type axons leads to undergrowth of axons, while stabilization of BAR-1/beta-catenin in a lin-23/SCF(beta-TrCP) mutant results in an overextension phenotype.

Conclusions/significance: Together, our data provide evidence that Wnt-mediated axon guidance can be transduced through a beta-catenin-dependent pathway.

Figure 1. Posterior D-type axons are misguided in lin-44 mutants.

(A,B) The posterior-most D-type axons, DD6, VD12 and VD13, visualized with DsRed in wyIs75 animals, terminate at stereotyped positions along the dorsal nerve cord. A bundle of axons terminates at the anteroposterior position corresponding to where these cell bodies are located (arrowhead), and is extended by a process (arrow) that reaches the anterior border of the anal depressor muscle (ADM) labeled with GFP in (A). (C,D) DD6 axon is labeled with a Pflp-13::GFP (ynIs37) and terminates anterior to VD13 axon. (E,F) D-type axons overextend (asterisk) past the ADM into the tail in lin-44 mutant animals. The arrow and arrowhead point termination positions of the thick bundle and thin process in wild-type animals, respectively. (G) Overextension index for mutants in the three Wnts expressed in the tail region, compared to those of wild-type animals. lin-44, but not egl-20, cwn-1, cwn-2 and mom-2 single mutants, show a dramatic axon guidance defect. The severity of the phenotype is enhanced in the lin-44; egl-20 double mutant. Scale bar, 10 µm. (H) Schematic diagram of D-type axons (red) and ADM (green) in the tail region, as seen in wyIs75 animals. Other cells labeled in wyIs75 animals include the intestinal muscles (on each side of the intestine, not depicted), PDA and DVB neurons. a.u.: arbitrary units.

Figure 2. LIN-17 Fz receptor mediates the response to Wnts in D-type axons.

(A,B) lin-17 mutant animals show a phenotype similar to that of lin-44 mutants, with axons terminating posterior (asterisk) to their normal positions (arrow and arrowhead). (C) Extension index for mutants in each known Fz orthologue (lin-17, mom-5, mig-1, cfz-2), in the Ryk/Derailed orthologue lin-18, and in the tyrosine kinase Ror2 homologue cam-1. Mild overextension phenotypes were observed in less than 30% and 15% of the lin-18 and cam-1 mutant animals, respectively. No increase in the severity of the extension phenotype of lin-17 mutants was observed in lin-17; lin-44 double mutants. (D) Comparison of the extension index of lin-17 mutants and lin-17 mutants carrying a Punc-47::lin-17 transgene. p<0.001. (E) Fluorescent distribution pattern of a LIN-17::YFP fusion protein is enriched at the posterior tip of D-type axons, here shown for an L1 animal. (F) Schematic diagram corresponding to the confocal image shown in (D). Scale bars, 10 µm (A,B) and 5 µm (E). a.u.: arbitrary units.

Figure 3. Wnts act instructively to repel posterior axons.

(A,B) In wild-type animals, the dorsal posterior branch of DD6 extends up to the position corresponding to that of its cell body in the VNC (arrow). (B) is a higher magnification view of the boxed region in (A). (C,D) In lin-44; egl-20 mutants, the dorsal posterior process of DD6 overextends (asterisk) as compared to its normal position (arrow). (D) is a higher magnification view of the boxed region in (C). (E,F) Expression of Wnt in the posterior tail using a Plin-44::lin-44::gfp array rescues the DD6 overextension phenotype in lin-44; egl-20 mutant animals. (F) is a higher magnification view of the boxed region in (E). (G,H) Expression of Wnt in a domain slightly anterior to the lin-44 expression domain using a Pegl-20::lin-44::gfp array rescues the DD6 overextension phenotype in lin-44; egl-20 mutant animals, and causes underextension defects in 50% of the animals. (H) is a higher magnification view of the boxed region in (G). (I–K) Ectopic expression of Wnt in the dorsal body wall muscles causes DD processes to stop prematurely (asterisks), resulting in gaps in the DNC. (J) is a higher magnification view of the boxed region in (I). (K) is a schematic corresponding to the confocal image shown in (J). (L) Quantification of the DD6 extension phenotypes (top) and the DNC gap phenotypes (bottom) observed in L1 animals expressing no Wnt, Wnt in the posterior tail, or Wnt in the dorsal body wall muscles in mutant or wild-type backgrounds. (M) Schematic showing DD morphology in L1 animals and the domain of Wnt expression expected from the rescuing and ectopic expression arrays. Scale bars, 20 µm (A,C,E,G,I), 10 µm (B,D,F,H) and 5 µm (J).

Figure 4. Mutants in the canonical Wnt pathway display D-type axon guidance defects.

(A–H) Posterior D-type axons underextend in mig-5/Dsh mutants (A,B), bar-1/β-catenin mutants (E,F) and pop-1/TCF mutants (G,H), but overextend in pry-1/Axin mutants (C,D). In all images, the arrow and arrowhead indicate the wild-type termination points, and the asterisk indicates the abnormal termination point of posterior D-type axons. (I) Extension index for mutants in components of the canonical Wnt pathway. (J) Genetic pathway controlling D-type axon guidance, inferred from mutant analysis. Note that Dsh acts negatively in this pathway, opposite to its predicted role in the canonical pathway. Scale bar, 10 µm. a.u.: arbitrary units.

Figure 5. BAR-1/β-catenin regulates axon pathfinding cell-autonomously in D-type neurons.

(A) Expression of a functional BAR-1::GFP fusion protein in D-type neurons is sufficient to partially rescue the underextension defect observed in bar-1(ga80) animals. (B) Extension index of lin-17; bar-1 double mutants, compared to either of the single mutants. The double mutant shows a phenotype intermediate between the two single mutants. ***p<0.001. (C–E) Expression of a Pflp-13::GFP reporter (ynIs37) is not affected in lin-44 (C), bar-1 (D) and pop-1 (E) mutants DD neurons. Scale bar, 10 µm. a.u.: arbitrary units.

Figure 6. The ubiquitin ligase F-box LIN-23 regulates axon guidance through β-catenin.

(A) In lin-23 mutants, D-type axons overextend past the ADM (asterisk). (B) Expression of the lin-23 cDNA specifically in D-type neurons is sufficient to completely rescue the overextension phenotype. ***p<0.001. (C) Extension index of bar-1; lin-23 double mutants, compared to either of the single mutants. The double mutant shows a phenotype very similar to bar-1 mutants, suggesting that lin-23 regulates axon guidance through bar-1. n.s. p>0.1, ***p<0.001. Scale bar, 10 µm. a.u.: arbitrary units.

Figure 7. One possible model of the signaling pathways mediating axon guidance in response to Wnt.

Because lin-17 mutants have both underextension and overextension defects, we propose that LIN-17/Fz can activate both a canonical pathway that attracts axons (or prevents axon termination) (left) and a yet uncharacterized pathway that repels axons (or triggers axon termination) (right). The existence of this additional pathway and its interaction with the canonical pathway are supported by the analysis of lin-17; bar-1 double mutants (see Figure 5). According to this model and consistent with our data, both the activation of the uncharacterized pathway and the inhibition of the canonical pathway by LIN-44/Wnt are required for the axons to stop in their correct position.