A subset of endosome-associated Rabs are expressed in olfactory projection neurons (PNs) and are involved in dendrite targeting

(A) Diagram of Rab GTPases and the endosomal compartments they predominantly localize to. (B) Schematic of the adult Drosophila brain, highlighting the olfactory system. ORNs, olfactory receptor neurons. PNs, projection neurons. Glomerulus on the far left depicts one-to-one matching between PNs and ORNs. (C) Schematic representing PN development at three distinct timepoints. APF, after puparium formation. (D) tSNE plots of developing PN single-cell RNA-seq (scRNA-seq) depicting stages profiled: 24h APF, 48h APF, and adult (leftmost plot) and expression of endosome-associated Rab GTPases. Expression is in log2(CPM +1), where CPM stands for transcript counts per million reads. scRNA-seq data are from Xie et al., 2021. (E) Schematic of ventromedial (VM) screening line used in Rab GTPase dominant negative screen. GTP-binding defective dominant negative Rabs were expressed in all PNs using VT033006-GAL4, and targeting of VM5d/v was monitored with GMR86C10-LexA>LexAop-mtdTomato. (F–I) Representative images of indicated genotypes depicting phenotypes observed in dominant negative screen. Red dotted lines outline the VM5d/v glomeruli, and red arrows denote ectopic targeting. Scale bar, 20 μm. (J) Percent of antennal lobes with mistargeting in the Rab dominant negative screen; Fisher’s exact test was used to assess statistical significance. In this and all subsequent figures: D, dorsal; L, lateral. NCad (in blue) is used to label neuropil/glomeruli. * p < 0.05; ** p < 0.01; *** p < 0.001; n.s., not significant. n, the number of the antennal lobes quantified. Images of the antennal lobe were taken of 0–5 day-old adults unless otherwise noted.

Rab5 regulates multiple features of PN dendrite development

(A) Volume rendering of a subset of glomeruli in the adult antennal lobe depicting those targeted by larval-born PNs from the anterodorsal lineage, hereafter adPNs, in grey, and other glomeruli in blue. (B) Representative images of dendrites from adPN neuroblast clones of indicated genotypes. Red dotted lines encircle adPN glomeruli. (C) Quantification of the number of cell bodies in adPN neuroblast clones in controls (n=10) and Rab5−/− mutants (n=9). Medians (IQR); Mann-Whitney U test was used to assess statistical significance. (D, E) Quantification of percent of antennal lobes with each category of dendrite innervation to adPN glomeruli (D) and glomeruli targeted by PNs other than larval-born adPNs (E). Fisher’s exact test was used to assess statistical significance. (F–I) Representative images of DL1-PN single-cell clone dendrites of indicated genotypes at 48h APF (F, G) and in the adult stage (H, I). Red dotted lines outline ectopically innervated glomeruli and green dotted lines outline DL1 glomerulus, and red arrows denote DL1-PN axon. (J, K) Quantification of percent of antennal lobes with altered dendrite innervation compared to controls in 48h APF (J) and adult (K) PNs. Fisher’s exact test was used to assess statistical significance. Scale bar, 20 µm (B, F, H).

Rab5 is required for axon guidance and morphogenesis

(A) Schematic depicting DL1-PN dendrites innervating the DL1 glomerulus and axon projecting to the mushroom body and lateral horn. (B) Representative images of a control DL1 PN depicting its axon in the mushroom body and lateral horn (top left) and of the axon guidance defects observed in Rab5 mutants where the axon does not project to higher order brain centers (middle) and quantification of the number of antennal lobes where this phenotype was observed (right). Red arrows denote DL1-PN axons terminating in ectopic regions. Fisher’s exact test was used to assess statistical significance. (C, D) Representative image of control (C) or Rab5−/− mutant (D) DL1-PN axons at the mushroom body at 48h APF. Dotted lines denote the border of the mushroom body. (E, F) Quantification of number of terminal boutons (E) and average terminal bouton width (F) in each DL1-PN mushroom body axon for controls (n=9) and Rab5−/− mutants (n=8) at 48h APF. Data are presented as medians (IQR); statistical significance was assessed using the Mann-Whitney U test. (G, H) Representative image of control (G) or Rab5−/− mutant (H) DL1-PN axons at the mushroom body at the adult stage. Dotted lines denote the border of the mushroom body. (I, J) Quantification of number of terminal boutons (I) and average terminal bouton width (J) in each DL1-PN mushroom body axon for controls (n=9) and Rab5−/−mutants (n=6) at the adult stage. Data are presented as medians (IQR). Mann-Scale bar, 20 µm (B); 10 µm (C, G).

Rab5 restrains axonal morphogenesis in the lateral horn

(A, B) Representative image of a control (A) or Rab5−/− mutant (B) DL1-PN axon at the lateral horn at the adult stage. Red arrows denote enlarged boutons. (C–E) Quantification of number of secondary branches (C), number of boutons (D), and average bouton width (E) in each DL1-PN axon at the lateral horn for controls (n=10) and Rab5−/− mutants (n=7) at the adult stage. Data are presented as median (IQR); statistical significance was assessed using the Mann-Whitney U test. (F, G) Representative image of a control (F) or Rab5−/− mutant (G) DL1-PN axon at the lateral horn at 48h APF. Red arrows denote enlarged boutons. (H–J) Quantification of number of secondary branches (H), number of boutons (I), and average bouton width (J) in each DL1-PN axon at the lateral horn for controls (n=9) and Rab5−/− mutants (n=7) at 48h APF. Data are presented as median (IQR); statistical significance was assessed using the Mann-Whitney U test. Scale bar, 10 µm (A, F).

Developmental dynamics of early endosomes across neuronal compartments

(A) Airyscan super resolution images depicting the localization of FYVE-mCherry expressed in a DL1-PN single-cell clone (left), single-channel FYVE image (middle), and 3D volume rendering (right) at 24–30h APF. (B) Representative images depicting the localization of the early endosome marker FYVE-mCherry expressed in a DL1-PN single-cell clone (left), single-channel FYVE image (middle), and 3D volume rendering (right) at the adult stage. (C) Airyscan super resolution images depicting the mushroom body localization of the early endosome marker FYVE-mCherry expressed in a DL1-PN single-cell clone (left), single-channel FYVE image (middle), and 3D volume rendering (right) at 24–30h APF. (D) Representative images depicting the mushroom body localization of FYVE-mCherry expressed in a DL1-PN single-cell clone (left), single-channel FYVE image (middle), and 3D volume rendering (right) at the adult stage. (E) Airyscan super resolution images depicting the lateral horn localization of FYVE-mCherry expressed in a DL1-PN single-cell clone (left), single-channel FYVE image (middle), and 3D volume rendering (right) at 24–30h APF. (F) Representative images depicting the lateral horn localization of FYVE-mCherry expressed in a DL1-PN single-cell clone (left), single-channel FYVE image (middle), and 3D volume rendering (right) at the adult stage. Note that puncta outside of lateral horn axons are in other nearby cell types. Scale bar, 5 µm (A–F).

Rab7 plays a minor role in PN development

(A, B) Representative images of dendrite targeting of DL1-PN single-cell clones in control (A) and Rab7−/− mutant (B) antennal lobes. Red numbers in the right corner of images denote DL1-PN dendrite mistargeting phenotypic penetrance. (C) Airyscan super-resolution image depicting localization of mCherry-Rab7 in developing DL1-PN dendrites (left), single-channel image of Rab7 (middle), and 3D volume rendering of Rab7 distribution (right). (D) Representative image of localization of mCherry-Rab7 in a Rab7−/− mutant background in adult DL1-PN single-cell dendrites (left), single-channel image of Rab7 (middle), and 3D volume rendering of Rab7 distribution (right). (E, F) Representative images of mushroom body projections of control (E) and Rab7−/− mutant (F) DL1-PN single-cell clones. Red numbers in the right corner of images denote DL1-PN axon morphogenesis phenotypic penetrance. (G) Airyscan super-resolution image depicting localization of mCherry-Rab7 in a developing DL1-PN axon at the mushroom body (left), single-channel image of Rab7 (middle), and 3D volumes rendering of Rab7 distribution (right). (H) Representative image of localization of mCherry-Rab7 in a Rab7−/− mutant background in an adult DL1-PN axon at (I) Airyscan super-resolution image depicting localization of mCherry-Rab7 in a developing DL1-PN axon at the lateral horn (left), single-channel image of Rab7 (middle), and 3D volume rendering of Rab7 distribution (right). (J) Representative image of localization of mCherry-Rab7 in a Rab7−/− mutant background in an adult DL1-PN axon at the lateral horn (left), single-channel image of Rab7 (middle), and 3D volume rendering of Rab7 distribution (right). Scale bar, 20 µm (B); 10 µm (F); 5 µm (C, D, G–J).

Rab11 has cell-type-specific roles in PN dendrite targeting and innervation

(A–C) Representative images of dendrite targeting of adPN neuroblast clones in control (A), Rab11−/− mutant (B), and Rab11 rescue (C) antennal lobes. (D, E) Quantification of percent of antennal lobes with each category of dendrite innervation to adPN glomeruli (D) and glomeruli targeted by PNs other than larval-born adPNs (E). Fisher-Freeman-Halton test was used to assess statistical significance. See Table S3 for complete statistical analysis. (F–H) Representative images of control (F), Rab11−/− mutant (G), and Rab11 rescue (H) DL1-PN single-cell clone dendrites. (I) Percent of the DL1 glomerulus filled by DL1-PN dendrites of controls (n=9), Rab11−/− mutants (n=7), and Rab11 rescues (n=4). Data are presented as medians (IQR); statistical significance was assessed using a Kruskal-Wallis test followed by Dunn’s test post-hoc. (J, K) Representative images of control (J) and Rab11−/− mutant (K) DL1-PN single-cell clone dendrites at 42–48h APF. (L) Percent of the DL1 glomerulus filled by DL1-PN dendrites of controls (n=5) and Rab11−/− mutants (n=7) at 42–48h APF. Data are presented as medians (IQR); statistical significance was assessed using the Mann-Whitney U test. Scale bar, 20 µm (A, F); 15 µm (J).

Rab11 promotes mushroom body axon maturation

(A–C) Representative images of adult control (A), Rab11−/− mutant (B), and Rab11 rescue (C) DL1-PN axons in the mushroom body. (D–G) Quantification of the number of secondary branches (D), number of terminal boutons (E), average terminal bouton width (F), and percent of terminal boutons with ectopic processes (G) in each DL1-PN axon for controls (n=9), Rab11−/− mutants (n=10), and Rab11 rescues (n=4). Data are presented as medians (IQR) and Kruskal-Wallis test followed by Dunn’s test post-hoc was used to assess statistical significance. (H–K) Representative images of developing control (H, J) and Rab11—/— mutant (I, K) DL1-PN axons in the mushroom body at 24-30h and 42-48h APF. (L) Quantification of the number of secondary branches in each DL1-PN axon at the mushroom body for 24-30h APF controls (n=4) and Rab11—/— mutants (n=6) and 42-48h APF controls (n=9) and Rab11—/— mutants (n=10). Data are presented as medians (IQR); statistical significance was assessed using the Mann-Whitney U test. (M) Quantification of the number of terminal boutons in each DL1-PN axon at the mushroom body for 42-48h APF controls (n=9) and Rab11—/— mutants (n=10). Data are presented as medians (IQR); statistical significance was assessed using the Mann-Whitney U test. Scale bar, 10 µm (B, I, K). Dotted lines denote the border of the mushroom body.

Rab11 is required for the maturation of lateral horn axons

(A–C) Representative images of adult control (A), Rab11−/− mutant (B), and Rab11 rescue (C) DL1-PN axons at the lateral horn. (D, E) Quantification of the percent of axons missing a main branch (D) or percent with an overextended dorsal branch (E). Statistical significance was assessed using the Fisher-Freeman-Halton test, followed by pairwise Fisher’s exact tests post-hoc.(F, G) Quantification of number of endpoints (F) or average process length (excluding main branches) (G) of each DL1-PN axon in the lateral horn of controls (n=11), Rab11−/−mutants (n=8), and Rab11 rescues (n=4). Data are presented as medians (IQR). Kruskal-Wallis test followed by Dunn’s test post-hoc was used to assess statistical significance. Some measures did not reach statistical significance due to a low sample number inherent with MARCM analysis. (H–K) Representative images of developing control and Rab11—/— mutant DL1-PN axons in the lateral horn at 24–30h and 42–48h APF. (L) Quantification of the percentage of lateral horn axons that have missing branches in each genotype at indicated developmental timepoints. Fisher’s exact test was used to assess statistical significance Scale bar, 10 µm. Dotted lines denote the border of the loss of lateral branches. Arrow heads indicate overextension of the dorsal branch.

Developmental dynamics of Rab11 distribution in PNs

(A) Airyscan super resolution images depicting the localization of mCherry-Rab11 in a Rab11−/− mutant background in a DL1-PN single-cell clone dendrites (left), single-channel Rab11 image (middle), and 3D volume rendering (right) at 24–30h APF. Zoom panels are single optical sections of areas indicated by white boxes. (B) Representative images depicting mCherry-Rab11 localization in a Rab11−/− mutant background in a DL1-PN single-cell clone dendrites (left), single-channel Rab11 image (middle), and 3D volume rendering (right) at the adult stage. Zoom panels are single optical sections of areas indicated by white boxes. (C, D) Airyscan super resolution images depicting mCherry-Rab11 localization at the mushroom body (C) and lateral horn (D) in a Rab11−/− mutant background in a DL1-PN single-cell clone (left), single-channel Rab11 image (middle), and 3D volume rendering (right) at 24–30h APF. Zoom panels are single optical sections of areas indicated by white boxes. (E, F) Representative images depicting mCherry-Rab11 localization at the mushroom body (E) and lateral horn (F) in a Rab11−/− mutant background in a DL1-PN single cell clone (left), single-channel Rab11 image (middle), and 3D volume rendering (right) in the adult stage. Zoom panels are single optical sections of areas indicated by white boxes. Scale bar, 5 µm (A–F).

Endosome-associated Rab GTPases regulate distinct aspects of neuronal morphogenesis and circuit assembly.

Summary model of the roles of endosome-associated Rab GTPases in PN development.

Expression of endosome-associated Rab GTPases in olfactory projection neurons

(Expression of Rab GTPases in developing PNs at 24h after puparium formation (APF) and 48h APF from single-cell RNA-seq (scRNA-seq) data. Expression is in log2(CPM +1), where CPM stands for transcript counts per million reads. scRNA-seq data are from Xie et al., 2021.

Expression of endosome-associated Rab GTPases in olfactory receptor neurons

Expression of Rab GTPases in developing ORNs at 24h after puparium formation (APF) and 48h APF from single-cell RNA-seq (scRNA-seq) data. Expression is in log2(CPM +1), where CPM stands for transcript counts per million reads. scRNA-seq data are from McLaughlin et al., 2021.

Extended analysis of endosome-associated Rab GTPases

(A–D) Representative images of indicated genotypes depicting phenotypes observed in dominant negative screen. Red dotted lines outline the VM5d/v glomeruli, and red arrows denote ectopic targeting. Scale bar, 20 μm. (F) Percent of antennal lobes with mistargeting in the Rab dominant negative screen. Fisher’s exact test compared to control.

Schematic of MARCM-based mosaic analysis

(A) Schematic depicting heatshock flp induction of single-cell or neuroblast MARCM clones. MARCM can be used to generate GFP-labeled single-cell or neuroblast clones in PNs. All clones were induced by heat shock applied to newly hatched larvae (0–24h after larval hatching), so our analyses are primarily restricted to the adPNs and DL1-PN single-cell clones. (B) Schematic of MARCM analysis. A mutant (Rab5, Rab7, Rab11, or Rab21) allele is placed on a chromosome arm in trans to the chromosome arm with a GAL80 transgene. Heterozygous cells express GAL80, which represses GAL4 activity and thus inhibits GFP expression in these cells. Following, FLP-mediated mitotic recombination and X-segregation (bottom row) one of the daughter cells becomes homozygous for the mutant allele and loses the GAL80 transgene. Thus, homozygous mutant cells will be labeled with membrane-bound GFP (and can also express any UAS-based rescue transgene). (C) Quantification of adPN mistargeting in additional non-adPN glomeruli. Fisher’s exact test.

Extended analysis of Rab7 loss-of-function MARCM phenotypes

(A) Quantification of number of cell bodies in adPN neuroblast clones of controls (n=5), Rab7−/− mutants (n=14), and Rab7 rescues (n=6). Data are presented as medians (IQR). Kruskal-Wallis test followed by Dunn’s test post-hoc was used to assess statistical significance. (B–D) Representative images of adPN neuroblast clones of indicated genotypes. (E, F) Quantification of percent of antennal lobes with each category of dendrite innervation to adPN glomeruli (E) and non-adPN glomeruli (F) in indicated genotypes. Fisher-Freeman-Halton test was used to assess statistical significance. See Table S2 for complete statistical analysis. (G) Representative images of DL1-PN lateral horn axons in indicated genotypes. Scale bar, 20 µm (B), 10 µm (G).

Rab7 is dispensable for DL1-PN development

(A, B) Representative images of DL1-PN dendrite targeting in indicated genotypes. Red numbers in the right corner of images denote DL1-PN dendrite mistargeting phenotypic penetrance. (C, D) Representative images of mushroom body axons in indicated genotypes. Red numbers in the right corner of images denote DL1-PN axon morphogenesis phenotypic penetrance. (E, F) Representative images of lateral axons in indicated genotypes. Red numbers in the right corner of images denote DL1-PN axon morphogenesis phenotypic penetrance. Scale bar, 20 µm (C); 10 µm (D, I)

Additional analysis of Rab11 dendrite targeting phenotypes

(A) Quantification of ectopic dendrite targeting phenotypes across all analyzed non-adPN glomeruli, excluding those present in Figure 5E. Note the non-adPN glomeruli present in Figure 5E are significantly different. Fisher-Freeman-Halton test was used to assess statistical significance. See Table S3 for complete statistical analyses. (B) Quantification of number of cell bodies in adPN neuroblast clones of controls (n=5), Rab11−/− mutants (n=8), and Rab11 rescues (n=5). Data are presented as medians (IQR). Kruskal-Wallis test followed by Dunn’s test post-hoc test was used to assess statistical significance. (C) Representative images of mistargeting observed in indicated genotypes. Red dotted outline denotes DA1 (top row) or DC3/VA1d (bottom row) glomeruli. Green arrows denote mistargeting. (D) Quantification of the percent of the indicated glomerulus filled with PN dendrites in DA1-PN controls (n=15) and Rab11−/− mutants (n=12) and VA1d/DC3-PN controls (n=14) and Rab11−/− mutants (n=5). Data are presented as medians (IQR). Mann-Whitney U test was used to assess statistical significance. (E) Quantification of the proportion of antennal lobes with mistargeting. Fisher’s exact test was used to assess statistical significance. (F) Quantification of the number of cell bodies in neuroblast clones containing labeled DA1-PNs visualized by the MZ19-GAL4 driver for controls (n=15) and Rab11−/− mutants (n=12). Data are presented as medians (IQR). Mann-Whitney U test was used to determine assess significance. (G) Quantification of the number of cell bodies in neuroblast clones containing labeled VA1d/DC3-PNs visualized by the MZ19-GAL4 driver controls (n=14) and Rab11−/− mutants (n=5). Data are presented as medians (IQR). Mann-Whitney U test was used to assess statistical significance. Scale bar, 20 µm (C).

Complete list of genotypes used in this study

Complete statistical analysis for Rab7 mutant adPN neuroblast clones

Statistical significance was assessed using the Fisher-Freeman-Halton test for overall group comparisons, followed by pairwise Fisher’s exact tests with Benjamini-Hochberg correction for multiple comparisons.

Complete statistical analysis for Rab11 mutant adPN neuroblast clones

Statistical significance was assessed using the Fisher-Freeman-Halton test for overall group comparisons, followed by pairwise Fisher’s exact tests with Benjamini-Hochberg correction for multiple comparisons.