Lineage-related neuronal types undergo intial axon growth and regrowth in shared spatiotemporal conditions.

(A) A model of MB development highlighting that the regrowth of γ-axons temporally and spatially overlaps with the initial growth of α/β-axons. (B) A Venn diagram depicting the theoretical and predicted comparison between α/β initial axon growth and γ-axon regrowth. (C) Confocal Z-projections of R16B01-Gal4 driving expression of membranal GFP (mCD8::GFP), which specifically labels developing α/β-KCs during the pupal stage. The images were captured at five time points between 18h until 36h APF, when the axons reach the end of the lobes. At 18h APF, four groups of cell bodies can be observed, corresponding to newly born cells from the four neuroblasts, each sending out short axons that progressively extend in each subsequent image. The axon tips are marked with arrows. (D) Confocal Z-projections of γ-KCs labeled with R71G10-Gal4 driving mCD8::GFP, as identified by (Alyagor et al., 2018), between 18h to 30h APF. At 21h APF, the axons are completely pruned and begin to re-extend for a second time. By 24h APF, the regrowing axons are visible (arrow). (E) PCA of α/β and γ-KC expression profiles during development and in adult. Color code depicts neuronal type (green for γ and magenta for α/β) and developmental stage (from light to dark as development progresses).

Comparative transcriptomic analysis of α/β-axon initial growth and γ-axon regrowth.

(A) Heatmap showing k-means clustering of 3345 differentially expressed genes during α/β-KC development. Light blue and magenta represent low and high relative expression, respectively. (B) Heatmaps showing k-means re-clustering of 842 genes from cluster 1 or 320 genes from cluster 2 based on their γ-KC transcriptional profiling. Blue and coral represent low and high relative expression, respectively. For all the heatmaps, each horizontal line describes the relative expression of a single gene. Gene expression was determined in five timepoints after pupa formation (18, 21, 24, 27 and 30 hours APF) and in adult. The number of genes within selected clusters are specified to the right. A flow chart detailing the criteria used to select candidate genes for the loss-of-function (LOF) screen.

RNAi-based screen identifies 12 genes that are essential for γ-and/or α/β-axon morphology.

(A) Schematic of the WT α/β-and γ-axonal lobes, illustrating labeling patterns that match the corresponding confocal images. γ-axons are labeled in green using mtdT:3xHA under the control of R71G10-QF2. FasII is visualized with antibody staining shown in magenta, and also presented separately in greyscale. (B) Confocal Z-projection of a control MB in which γ-KCs are labeled by mtdT:3xHA (green) driven by R71G10-QF2, and OK107-Gal4 drives the expression of Luciferase. Magenta (left panel) or greyscale (right panel) is anti-FasII staining, which strongly labels α/β-axons and weakly labels γ-axons. (C,G,P) Schematics of α/β-and/or γ-axon lobe defects. (D-F,H-O,Q) Confocal Z-projections of adult MBs in which γ-KCs are labeled by mtdT:3xHA (green) driven by R71G10-QF2, with OK107-Gal4 driving expression of RNAi targeting beag (beag; D); chameau (chm; E); Heat shock gene 67Ba (Hsp67Ba; F); Chaperonin containing TCP1 subunit 5 (CCT5; H); Chaperonin containing TCP1 subunit 7 (CCT7; I); embargoed (emb; J); Dipeptidase B (Dip-B; K); Brahma associated protein 60kD (Bap60; L); Phosphomevalonate kinase (Pmvk; M); twisted bristles roughened eye (twr; N); Cadherin-N (cadN; O); miranda (mira; Q). (R) Pie charts showing the phenotypic distribution of the 300 screened genes (left), and of the axon growth phenotypes of the 12 ‘positive hits’ (right). The putative WT γ-and α/β-lobes are outlined in white and blue, respectively. Asterisks mark the missing part of α/β or γ lobes. Frame colors in (C-Q) represent the phenotypic group as depicted in (R). Numbers represent the fraction of MBs displaying the presented phenotype. Scale bar is 10 μm.

Pmvk is cell autonomously required for γ-axon regrowth but not for initial axon extension of γ nor α/β.

(A-F) Confocal Z-projections of L3 (A-B) or adult (C-F) MBs with γ-KCs labeled by mCD8::GFP (green) driven by the γ-specific R71G10-Gal4 (A-D), or the α/β-specific R44E04-Gal4 (E-F). MBs in images (B-F) also express RNAi targeting Pmvk driven by R71G10-Gal4 (B,D) or R44E04-Gal4 (F). (G-L) Confocal z-projections of control (G,I,K) or PmvkE20* (H,J,L) MARCM clones in L3 (G-H) or Adult (I-L). R71G10-Gal4 (G-J) or R44E04-Gal4 (K-L) drive expression of mCD8::GFP (green). Clones were generated by heat-shocking at 24 hours after egg-laying (G-J) or at 0h APF (K-L). The full extent of the γ lobe in adult and L3 is outlined in white (as determined by the FasII staining; magenta). Arrows demarcate axon growth defects. Scale bar is 10 μm. Numbers represent the fraction of MBs displaying the presented phenotype. (M) Schematic representation of the PmvkE20* mutant allele, which includes a single-nucleotide substitution (G>T) at position 60, corresponding to amino acid 20, leading to the replacement of glutamic acid (E) with a stop codon.

The mevalonate pathway is essential for γ-axon regrowth.

(A-C) Confocal Z-projections of adult MBs from control (A) or those expressing RNAi targeting Hmgcr (B) or Mvk (C) driven by the pan-KC driver OK107-Gal4. γ-KCs are labeled by mtdT:3xHA (green) driven by R71G10-QF2. Magenta is FasII staining. The putative WT α/β- and putative γ-lobes are outlined in blue and white, respectively. Arrows indicate axon growth defects. Numbers represent the fraction of MBs displaying the presented phenotype. Scale bar is 10 μm. (D) A schematic of the mevalonate pathway, illustrating key metabolites and enzymes. Enzymes shown in red are those found to influence γ-axon development. The diagram also highlights the different downstream routes that FPP can take—either toward the synthesis of sterol isoprenoids (e.g., cholesterol) or non-sterol isoprenoids (such as ubiquinone, dolichol, and isopentenyl-tRNA). Additionally, FPP serves as a precursor for protein prenylation, which occurs in two main forms: farnesylation and geranylgeranylation. The dashed arrows emphasize that cholesterol biosynthesis is absent in Drosophila

Pmvk likely regulates axon regrowth via the TOR pathway.

(A) Scheme of developing γ-KCs in which UNF promotes regrowth via the TOR pathway (adapted from Yaniv et al., 2012). Rheb is a CAAX protein that undergoes prenylation, enabling its translocation to the plasma membrane to activate downstream effectors. (B) Quantification of axon regrowth for the genotypes displayed in (C-G). The x/y ratio indicates relative axon position along the adult γ lobe (quantification described in Figure S7). Box plots represent the median (line), interquartile range (box), and spread within 1.5×IQR (whiskers), circles indicate individual data points. Groups were compared using one-way ANOVA on logit-transformed data followed by Dunnett’s post-hoc test. A significant difference was observed only between control and PmvkE20* (***p < 0.001). Comparisons of PmvkE20* only with PmvkE20* and UAS-Rho (p = 0.696), or UAS-Rheb (p = 0.704), or UAS-S6KCA (p = 0.120) were not significant. (C-G) Confocal Z-projections of γ single-cell MARCM clones in adult MBs, either control (C, comprised of multiple single cell clones), PmvkE20* homozygous mutant (D), or PmvkE20* homozygous mutant also overexpressing Rho (E), Rheb (F) or S6KCA (G). The γ-specific driver R71G10-Gal4 drives expression of membranal GFP (mCD8::GFP; green). Magenta is FasII staining. The γ lobe is outlined by a dashed white line. Arrows indicate axon growth defects. Numbers represent the fraction of MBs displaying the presented phenotype. Scale bar is 10 μm