DFz2 inhibits Wave axon extension towards the posterior end

(A, A’) Schematic diagram of Wave neuron morphology and function. Wave neurons are segmentally repeated in the VNC, receive inputs from tactile sensory neurons, and drive distinct behaviors depending on the segment. The anterior Wave (a-Wave) comprises anteriorly polarized axon/dendrite and drive backward locomotion, whereas posterior Wave (p-Wave) has axon/dendrite extension towards the posterior end and drive forward locomotion. (B-E) Lateral view of single-cell images of Wave neurons revealed by mosaic analyses in 3rd instar larvae. The green channel shows Wave neurons, and the blue channel shows anti-HRP staining that visualizes the neuropil. SEZ: subesophageal zone, T1-3: thoracic VNC, A1-8/9: abdominal VNC. (B, C) Control A2 (B) and A6 (C) Wave neuron. Scale bar = 50 µm. White arrows indicate axon processes. (D, E) A2 (D) and A6 (E) Wave neuron in which DFz2 is knocked-down using the TRiP RNAi line. Note the elongation of the axons towards the posterior end (dotted boxes). Orange arrows indicate the putative presynaptic varicosities in the ectopic axons. (F-J) Tiling profile of the axons of A2 to A6 Wave neurons in control and DFz2 knocked-down animals. n indicates the number of neurons. Tiling percentage indicates the fraction of samples in which the Wave axon innervates the corresponding neuromere. (K, L) Quantification of axon extension towards the posterior end in A2 (K) and A6 (L) Wave neuron measured in MCFO images, following KD and overexpression of DFz2, respectively. **: p < 0.01, ***: p < 0.001, Welch’s t-test with Bonferroni correction. (M) Summary of DFz2 KD and overexpression phenotypes.

DFz4 specifically promotes axon extension towards the posterior end in A6-Wave

(A, B) DFz4 KD A2 (A) and A6 (B) Wave neuron using the TRiP RNAi line. Note the shortening of A6 axon towards the posterior end (dotted box in B, see Figures 1B and C for control). Scale bar = 50 µm. (C-G) Tiling profile of A2 to A6 Wave axons in control and DFz4 knocked-down animals. n indicates the number of neurons. Tiling percentage indicate the fraction of samples (single Wave neurons) whose Wave axon innervated the corresponding neuromere. The same analyses as in Figures 1F-J but for DFz4 knocked-down animals. The control data in Figures 1F-J are shown as references. (H, I) Quantification of axon extension towards the posterior end in A2 (H) and A6 Wave (I) measured in MCFO images, following KD and overexpression of DFz4, respectively. *: p < 0.05, Welch’s t-test with Bonferroni correction. The control data shown in Figures 1K and L are reused for this analysis. (J) Summary of DFz4 KD and overexpression phenotypes.

DWnt4 regulates A-P extension of Wave axons

(A-C) DWnt4 promotes posterior axon extension of p-Wave. (A, B) Examples of axon morphologies towards the posterior end (white arrows) in control (A) and DWnt4C1/EMS23 mutant (B) animals. The posterior end of the axon, derived from p-Wave, is shortened in DWnt4 mutant. Dotted boxes indicate the abnormal shortening of the axon. (C) Quantification of A6-Wave axon extension to segment A6 or A7. n = 6 (control) and 3 (DWnt4C1/EMS23) neurons, respectively: p = 0.0119, Fisher’s exact test. (D-G) DWnt4 regulates axon extension of a-Wave. (D) A2-Wave in DWnt4C1/EMS23 is visualized using heat-shock FlpOut. Note the elongation of its axon towards the posterior end, as well as the shortening of its axon towards the anterior end (dotted boxes). Scale bar = 50 µm. See Figure 1B for control. (E) Tiling profile of A2-Wave axons in control and DWnt4C1/EMS23mutant animals. n indicates the number of neurons. Tiling percentage indicates the fraction of samples (single Wave neurons) whose Wave axon innervated the corresponding neuromere. The control data in Figure 1F are shown as references. (F, G) Quantification of the relative length (see Method details) of the anterior and posterior fraction of A2-Wave axons between control (visualized by MCFO) and mutant (visualized by heat-shock FlpOut) animals. (F) Quantification of the anterior fraction of A2-Wave axons. (G) Quantification of the posterior fraction of A2-Wave axons. n indicates the number of neurons. *: p < 0.05, **: p < 0.01, Welch’s t-test. The control data are replotted from Figure 1K for different comparisons. (H) Summary of DWnt4 phenotypes.

Complementary graded expression of DWnt4 and DFz2 along the A-P axis

(A, B) DWnt4-GFP cells in the embryonic CNS. (A) Distribution of DWnt4-GFP cells. Nuclear-localized GFP is expressed under the regulation Wnt4MI03717-Trojan-GAL4. anti-Elav counterstaining was performed to label all neurons and normalize the GFP signal. GFP expression is stronger in posterior than anterior VNC. Scale bar = 50 µm. (B) Quantification of normalized GFP signals with respect to Elav signals in each neuromere (where 1 denotes the maximum). n = 5 animals. Not having a common alphabet indicated above the plots between groups indicates statistical significance (α=0.05, Tukey’s HSD test).

(C, D) Expression of DWnt4 protein in the neuropil of each neuromere. (C) DWnt4 and HRP immunostaining. Scale bar = 10 μm. (D) Normalized DWnt4/HRP signals in each neuromere are calculated. n = 18 samples from 9 animals *: p < 0.05, ***: p < 0.001, paired t-test (comparison to A8).

(E, F) Expression of DFz2 protein in the neuropil of each neuromere. (E) DFz2 and HRP immunostaining. Scale bar = 10 μm. (F) Normalized DFz2/HRP signal in each neuromere are calculated. n = 16 samples from 8 animals *: p < 0.05, **: p < 0.01, paired t-test (comparison to A2).

(G) Model of segment-specific axon guidance in Wave neurons. DWnt4 serves as a graded axon guidance cue (concentrated towards the posterior end of VNC) and is recognized by DFz2/DFz4 receptors as repulsive/attractive cue, respectively. DFz2 functions both in a-Wave and p-Wave, whereas DFz4 functions selectively in p-Wave.

DFz2 KD in Wave neurons alters motor commands in vivo

(A) Schematic of optogenetics assay in vivo. (B-B’’) Comparison of the larval behavior between LED OFF and ON conditions in control animals. LED illumination decreases stride duration of forward locomotion (i.e. induces fast-forward locomotion, B), induces backward locomotion (B’), and induces rolling (B’’). n = 52 animals. ***: p < 0.001, Wilcoxon’s signed-rank test. (C-C’’) Comparison of the larval behavior between LED OFF and ON conditions in DFz2 knocked-down [TRiP] animals. LED illumination decreases stride duration of forward locomotion (i.e. induces fast-forward locomotion, C), does not induce backward locomotion (C’), and induces rolling (C’’). n = 18 animals. **: p < 0.01, ***: p < 0.001, Wilcoxon’s signed-rank test. (D, E) Comparison of the stride duration of forward locomotion between control and DFz2 knocked-down animals in LED OFF (D) and ON (E) conditions. The stride duration showed no significant difference in LED OFF period (D) but was shorter in DFz2 knocked-down animals in LED ON period (E). Mean±SD stride durations: 0.85±0.32 (Ctrl, OFF), 0.71±0.22 (Ctrl, ON), 0.71±0.21 (DFz2 KD, OFF), 0.50±0.12 (DFz2 KD, ON). ***: p < 0.001, Mann-Whitney’s U-test.

Modulation of tactile-evoked behavior by DFz2 KD in Wave neurons

(A) Left panel: Scheme of gentle-touch assay using von Frey filament. Right panel: Stereotypic behavioral responses induced by the gentle head touch, categorized according to the Kernan score (Kernan et al., 1994). (B-C”) Alteration in behavior responses upon DFz2 KD using UAS-DFz2-RNAi [KK] (B-B’’) or UAS-DFz2-RNAi [GD] (C-C’’). (B, C) Distribution of the Kernan score. Driver control; R60G09-GAL4/+, effector control; UAS-DFz2-RNAi/+, and experimental group (R60G09-GAL4 > UAS-DFz2-RNAi). n = 40 trials each. (B’, C’) Fraction of behavioral responses classified as backward (scores 3 and 4) and turning (score 2). The numbers indicate the trials that induced the classified response. *: p < 0.05, **: p < 0.01, ***: p < 0.001, Fisher’s exact test with Bonferroni correction. (B’’, C’’) Quantifications of stride duration of forward locomotion following gentle touch. Note that smaller stride duration indicates faster forward locomotion. n=36 (driver control), 40 (effector control), 39 (B’’) and 40 (C’’) (experimental) trials. *: p < 0.05, ***: p < 0.001, Steel-Dwass test. (D) Summary of the current study.

R60G09-GAL4 consistently targets Wave neurons from embryonic to larval stages

(A) R60G09-GAL4 targets Wave neurons. Double labeling of MB120B-spGAL, a previously characterized Wave-specific GAL4, and newly identified R60G09-GAL4 and R77H11-LexA in 3rd instar larvae shows that these three lines commonly target Wave neurons (arrows). Scale bar = 50 μm. (B-D) R60G09-GAL4 consistently labels Wave neurons from embryonic to larval stages. (B) Expression driven by R60G09-GAL4 in 12 hr AEL (Stage 15) embryos.

At this stage, the expression is seen in Wave neurons (white arrows) and putative dMP2 neurons (black arrows), which undergo apoptosis at later stages (except in segments A6-A8, Miguel-Aliaga, 2004). Stacked image. Scale bar = 50 μm. (C) Expression in 16hAEL (Stage 16/17) embryos. White arrows: Wave neurons, black arrows: putative dMP2 neurons, gray arrows: degenerating corpses of putative dMP2 neurons. Stacked image. Scale bar = 50 μm. (D) Expression in 20hAEL (Stage 17) embryos. White arrows indicate Wave neurons. Note that dMP2 neurons (except in segments A6-A8) are no longer present and the GAL4-driven expression is now largely confined to Wave neurons at this stage. Stacked image. Scale bar = 50 μm. (E) Expression in 3rd instar larvae. In the VNC, the GAL4-driven expression is confined to Wave neurons (white arrows) and a pair of ascending neurons in T2 segment. Note that the GAL4-driven expression in dMP2 neurons (including A6-A8) is absent in larval stages. Stacked image. Scale bar = 50 μm.

Identification of candidate Wnt receptors and ligands that regulate A-P Wave axon guidance

(A-D) Identification of candidate receptors. R60G09-GAL4 was used to drive expression of UAS-RNAi constructs for Wnt receptors and CD4-GCaMP6f in Wave neurons. Since expression of CD4-GCaMP6f visualizes all Wave axons which fasciculate together to form an axon bundle, the morphology of individual Wave axons cannot be assessed in general by this method. However, since it is known from previous single-cell analyses that the most posterior part of the Wave axon bundle contains only the axon of A6-Wave (Takagi et al., 2017), we focused on this part of the axon bundle (white arrows) to study how manipulation of Wnt receptor expression affect it morphology. As compared to the control (A), DFz2 KD [TRiP] (B) results in posterior elongation while DFz4 KD [TRiP] (C) results in shortening of this axon branch. Arrowhead: cell body of A6-Wave. Dotted boxes indicate the abnormal extension/shortening of the axon. Scale bar = 50 μm. (D) Quantification of the ending points of A6-Wave (p-Wave) axon extension in R60G09-GAL4 > UAS-RNAi [TRiP] animals. p-values were calculated by Chi-square test (p = 2.38×10-6) followed by Haberman’s adjusted residual analysis (α = 0.0019; see Method details). (E) Identification of candidate ligands. Quantification of the ending points of A5-Wave (p-Wave) axon extension in each mutant line. MB120B-spGAL4 was used for DWnt5400 mutant and R77H11-GAL4 for all the other mutant lines to express CD4-GCaMP6f. p-values were calculated by Chi-square test (p = 0.024) followed by Haberman’s adjusted residual analysis (α = 0.0064). (F, F’) Ectopic posterior extension of A2-Wave axon by DFz2 KD with an independent RNAi line. (F) Lateral view of a single Wave clone in which DFz2 was knocked-down with the KK RNAi line, visualized by mosaic analyses in 3rd instar larvae. As was observed with the TRiP RNAi line (see Figures 1B, D), the axon was ectopically extended towards the posterior (dotted box). The green channel shows the Wave neuron, and the blue channel shows anti-HRP staining that visualizes the neuropil. White arrows indicate ectopic axon processes. Scale bar = 50 μm. (F’) Quantification of the ectopic extension. Normalized posterior axon lengths were quantified in A2-Wave and categorized into normal (length < 1) and abnormally extended (length > 1) neurons. p-value was calculated by one-tailed Fisher’s exact test (p = 0.0455). (G, G’) Posterior extension of A6-Wave is shortened by DFz4 KD with two RNAi lines (TRiP and KK). (G) A6-Wave (p-Wave) axon extension in control (left) and DFz4 KD (right, KK) animals. Note that the posterior extension of A6-Wave axons can be specifically examined in R60G09-GAL4 > CD4-tdGFP animals, since only A6 Wave neurons extend axons beyond A6. Arrowhead: cell body of A6-Wave. Scale bar = 50 μm. (G’) Quantification of the ending points of A6-Wave (p-Wave) axon extension in control, DFz4 KD (TRiP), and DFz4 KD (KK) animals. p-values were calculated by Chi-square test (p= 0.0053) followed by Haberman’s adjusted residual analysis (α = 0.0085).

Effect of DFz2 KD on Wave dendrite extension

Tiling profile of A2-Wave to A6-Wave dendrites in control and DFz2 knocked-down [TRiP] animals. n indicates the number of neurons. Tiling percentage indicate the fraction of samples (single Wave neurons) whose Wave dendrite innervated the corresponding neuromere.

Expression of DWnt4 and DFz2 is suppressed in mutants

(A, B) Expression of DWnt4 protein in the neuropil of each neuromere. (A) The fluorescence of anti-DWnt4 antibody in control (yw) and mutant (DWnt4EMS23) embryos are shown (with the same contrast for each channel in each group). Scale bar = 10 μm. (B) Comparison of DWnt4 signal between control (yw) and mutant (DWnt4EMS23) embryos in each neuromere. ***: p < 0.001, Welch’s t-test. (C, D) Expression of DFz2 protein in the neuropil of each neuromere. (C) The fluorescence of anti-DFz2 antibody in control (yw) and DFz2 KD (elav-GAL4 > UAS-DFz2-RNAi [KK]) embryos are shown (with the same contrast for each channel in each group). Scale bar = 10 μm. (D) Comparison of DFz2 signal between control (yw) and DFz2 KD (elav-GAL4 > UAS-DFz2-RNAi [KK]) embryos in each neuromere. *: p < 0.05, **: p < 0.01, ***: p < 0.001, Welch’s t-test.