Dendritic arbours are labelled by mCD8-GFP (anterior is to the left and dorsal is up). Scale bars: 30 μm (a–d) and 10 μm (a′,a″,b′,b″,h).
Error bars represent Standard Deviation (S.D.) in this and all subsequent figures.
(a–d) Clones of class IV ddaC neurons: (a) wt, (b) dlic2¹¹⁵⁷, (c) lis1^G10.14 and (d) dlic2¹¹⁵⁷; dlic2-eGFP (dlic2+ rescue). The size and pattern of the dlic2¹¹⁵⁷ and lis1^G10.14 arbours are grossly abnormal (b,c). The number of terminal branches in dlic2¹¹⁵⁷ is substantially reduced (compare a′ to b′) whereas the number of proximal branches is greatly increased (compare a″ to b″). Expressing dlic2-eGFP specifically in dlic2¹¹⁵⁷ clones fully rescues the dendrite and axon phenotypes (d).
(e) Sholl analysis of dendritic arbours of wt and dlic2¹¹⁵⁷ ddaC clones. Concentric circles with 5 μm increments were drawn around the soma and the number of dendritic branches that intersected each circle was tallied. In control ddaC neurons the number of branches increases progressively from proximal to distal and the maximal number of dendrite branches are found between 225 and 250 μm from the cell body. The dlic2¹¹⁵⁷ neurons exhibited a dramatic proximal shift in dendrite distribution such that the majority of dendrites are located within 100 μm of the soma. Blue diamonds: wt, red squares: dlic2¹¹⁵⁷.
(f) Dendrite length of wt ddaC (16,902 ± 3,292 μm, n=4) and dlic2¹¹⁵⁷ ddaC (3,671 ± 681 μm, n=4, ***p<0.001, Student’s unpaired t- test). The dendrite length of dlic2¹¹⁵⁷ ddaC is greatly reduced.
(g) The number of dendritic branch points is decreased in dlic2¹¹⁵⁷ ddaC (140 ± 29, n=4, **p<0.01, Student’s unpaired t- test) compared to wt (499 ± 124, n=4).
(h) Axons of individual wt (top) and dlic2¹¹⁵⁷ (bottom) ddaC clones coursing through the intersegmental nerve (ISN). The wt axon is a single process whereas the dlic2¹¹⁵⁷ axon has separated into multiple neurites, as revealed in cross sections of the ISN (rightmost panels) at the points indicated by the dashed line. Dotted lines in the rightmost panels delineate the boundary of the nerve. Green: mCD8; magenta: HRP, which labels all nerve processes.

Green: RFP (a–d) or Ppk (e,f), magenta: mCD8. ppk-Gal4 UAS-mCD8-GFP (a–f) was used to visualize dendrites and axons of ddaC neurons expressing the endosomal marker UAS-Rab4-RFP (a,b) or UAS-Spin-RFP (c,d). Arrows indicate dendrites, arrowheads point to axons. Scale bar: 30 μm.
(a,b) Rab4-RFP is present in dendrites and axons in (a) wt; however, its dendritic localization is reduced in (b) dic¹²²⁹/dic^ts.
(c,d) Spinster-RFP localizes to dendrites and axons in (c) wt, but its dendritic distribution is reduced in (d) dic¹²²⁹/dic^ts.
(e,f) Ppk is found specifically in dendrites in (e) wt, but is present in both axons and dendrites of (f) dlic2¹¹⁵⁷ ddaC clones.
(g) Table summarizing the axonal and dendritic localization of Golgi outposts, endosomal markers and Ppk in wt (dynein +) and dynein loss-of-function (dynein-) neurons.

Kymographs (a,c) and movie frames (b,d) showing the trajectory of EB1-GFP comets in the axons of ddaC neurons. UAS-EB1-GFP was expressed in class IV neurons by 4-77-Gal4 (a–e) and in class I neurons by 2-21-Gal4 (e). Time (in seconds) is as indicated. The bar to the right of the kymogaph indicates the portion of the movie from which the frames were taken. The soma is to the left.
(a,c) Kymographs showing that EB1-GFP always moves away from the soma in (a) control axons, but moves both towards and away from the soma in the axons of (c) dic¹²²⁹/dic^ts neurons. A portion of the dic¹²²⁹/dic^ts movie was out of focus (109–225 sec) and this section of the kymograph was removed for clarity.
(b,d) Single frames from movies showing the movement of individual EB1-GFP comets. Arrowheads and diamonds indicate comets moving anterogradely (yellow) or retrogradely (magenta). Scale bar: 5 μm. In (b) control axons all EB1-GFP comets move anterogradely whereas in (d) dic¹²²⁹/dic^ts axons EB1-GFP moves both anterogradely and retrogradely. EB1-GFP comets moved in both directions in dic¹²²⁹/dic^ts axons with or without ectopic neurites, and within individual neurites EB1-GFP comets moved retrogradely (42%) and anterogradely (58%).
(e) Bar graph illustrating the percentage of EB1-GFP comets that move anterogradely (yellow) or retrogradely (magenta) in the axons of class IV and class I control and dic¹²²⁹/dic^ts neurons in 3^rd instar larvae.

Green: ManII-eGFP, magenta: mCD8. Arrowheads: axon, arrows: position of Golgi outposts.
(a–h) Localization of Golgi outposts in wt and dlic2¹¹⁵⁷ ddaC clones. Boxes with dotted lines highlight axons (shown in c,d) and boxes with dashed lines highlight proximal dendrites (shown in e,f). Distal dendritic tips are shown (g,h). Scale bars: 30 μm (a,b) and 15 μm (c–h).
(a–f) In (c) wt ddaC there are virtually no Golgi outposts in the axon (box with dotted lines in a). In contrast, many are present in (d) dlic2¹¹⁵⁷ axons (box with dotted lines in b). Golgi outposts are more numerous in the proximal dendritic arbour of dlic2¹¹⁵⁷ clones (compare e and f). The signal intensity is optimized for visualizing the smaller Golgi outposts in the proximal arbour; consequently, the outposts close to the soma are over-exposed (e,f).
(g,h) Along the distal dendrite there are fewer Golgi outposts in (h) dlic2¹¹⁵⁷ compared to (g) wt.
(i,j) Frames from movies (Supplementary Information, Movies 3 and 4) following Golgi outposts in the soma and axon of (i) control and (j) dic¹²²⁹/dic^ts neurons. Time (in seconds) is as indicated. The single Golgi outpost (arrowhead) in the control axon does not move whereas in the dic¹²²⁹/dic^ts neuron there are many Golgi outposts in the axon and one Golgi outpost (arrow) moves from the soma into the axon. Scale bar: 12.5 μm.
(k) Number of Golgi outposts in proximal (control: 10.00 ± 3.39, n=5; dlic2¹¹⁵⁷: 15.45 ± 4.18, n=11, *p<0.025, Student’s unpaired t- test) and distal (control: 7.92 ± 3.11 per 100 μm, n=6; dlic2¹¹⁵⁷: 2.00 ± 2.00 per 100 μm, n=3, *p<0.025, Student’s unpaired t-test) dendritic arbours as well as axons (control: 2.33 ± 2.71 per 100 μm, n=6; dlic2¹¹⁵⁷: 34.49 ± 14.55 per 100 μm, n=5, ***p<0.001, Student’s unpaired t- test).
(l) Size of Golgi outposts in proximal (control: 3.31 ± 1.08 μm², n=5; dlic2¹¹⁵⁷: 6.00 ± 1.75 μm², n=11, **p<0.01, Student’s unpaired t- test) and distal (control: 1.12 ± 0.36 μm², n=6; dlic2¹¹⁵⁷: 0.25 ± 0.25 μm², n=3; **p<0.01, Student’s unpaired t- test) dendritic arbours as well as axons (control: 0.45 ± 0.51, n=6; dlic2¹¹⁵⁷: 16.71 ± 6.09, n=5, ***p<0.001, Student’s unpaired t- test).

Green: βgal, magenta: mCD8. UAS-nod-lacZ (a–d) and UAS-kin-lacZ (e–h) driven by ppk-Gal4, which also drives UAS-mCD8-GFP expression. Open arrowhead: proximal axon, filled arrowhead: axon shaft; arrows: proximal dendrites. Scale bar: 30 μm.
(a–d) Localization of Nod-βgal in (a) wt, (b) dic¹²²⁹/dic^ts, (c) dlic2¹¹⁵⁷ ddaC clone and (d) ddaC over-expressing dmn (dmn OE). The lower panels (a′–d′) show the Nod-βgal channel at a slightly higher magnification. In wt axons Nod-βgal enters only the very proximal axon but is not present in the axon shaft. In dic¹²²⁹/dic^ts, dlic2¹¹⁵⁷ and dmn OE neurons Nod-βgal extends into the axon shaft. The dic¹²²⁹/dic^ts axon shown in (b,b′) is not unusually wide, yet it has strong Nod-βgal signal.
(e–h) Localization of Kin-βgal in (e) wt, (f) dic¹²²⁹/dic^ts, (g) dlic2¹¹⁵⁷ ddaC clone and (h) dmn OE. Inserts show Kin-βgal channel alone. Kin-βgal is normally localized specifically to axons and this distribution is not changed by a reduction in dynein activity.