Autophagosomes move towards the ncMTOC in fat cells.

A: Schematic drawing of the experimental design for screening. B: Non-fused autophagosomes accumulate in the perinuclear region upon Vps16A silencing. C: The cis-Golgi compartment remains unchanged upon the expression of vps16a RNAi. D, E: The accumulation of autophagosomes is not perinuclear in α or β tubulin; vps16a double RNAi cells. The boxed areas in the main panels, marked by cyan, are enlarged in the insets (D’ and E’). F-I: Quantification of data shown in B-E; n=10 cells. J, K: Autophagosomes position at an ectopic MTOC (yellow arrows) formed upon Shot silencing. (29 out of 41 cells exhibited an ectopic MTOC = 70.73%). L: Autophagosomes are near microtubule minus-ends, marked by Khc-nod-LacZ. The boxed area in the main panels, marked by cyan, is enlarged in L’’’ (proximity sites indicated by cyan arrowheads). M, N: Correlative ultrastructural analysis shows autophagosomes accumulating near the nucleus upon Vps16A silencing (border of control and silencing cells marked by green) (M). Shot knockdown causes aggregation of autophagosomes in ectopic foci in vps16a RNAi cells (N). An ectopic cleavage furrow (hallmark of Shot depletion (Sun et al., 2019)) is also visible (cyan arrows in N’). Note: the magnification of N is higher to better show this structure. O, P: Atg8a positive autophagosomes are clustered around the ectopic MTOC (yellow arrows) which is encircled by the signal of the minus-end marker Khc-nod-LacZ in shot RNAi (O) and vps16a; shot double RNAi cells (P). The boxed areas in the main panels, marked by cyan, are enlarged in O’, O”, P’ and P”. Nuclei are outlined in blue in J’, L’-L’’’, O’, O”, P’ and P”. The GFP signal of RNAi and Khc-nod-LacZ expressing cells is false-colored blue in composite images. Q-S: Quantification of data shown in L, O, P; n=10 cells. The boundaries of RNAi cells are highlighted in magenta in the grayscale panels.

A-C: Immunostainings for Atg8a (A), Rab7 (B), and Arl8 (C) show that pre-fusion autophagosomes (Atg8a and Rab7 positive) and pre-fusion lysosomes (Arl8 positive) accumulate in the perinuclear region in vps16a RNAi cells. D, E: Atg8a (D) or Atg1 (E) knockdown eliminates the mCherry-Atg8a signal from vps16a RNAi cells. F, G: Silencing of Shot using two additional independent RNAi lines also results in the formation of an ectopic MTOC (marked by yellow arrows) in Vps16A depleted cells, around which autophagosomes accumulate. An ectopic MTOC was present in 29 out of 38 cells (RNAi/2) and 23 out of 29 cells (RNAi/3), representing 76.32% and 79.31%, respectively. H-K: Quantification of data shown in A-E; n=10 cells. The boundaries of RNAi-expressing cells are highlighted in magenta in the grayscale panels. The outlines of nuclei are drawn in blue in F’ and G’.

A dynein-dynactin complex is required for minus-end directed autophagosome transport.

A-G: Knockdown of dynein (A-D) and dynactin subunits (E-G) results in the peripheral redistribution of autophagosomes in vps16a RNAi cells (red arrows). H, I: Kinesin silencing does not affect the perinuclear accumulation of autophagosomes in vps16a RNAi cells. J: Proposed model of the suggested dynein-dynactin complex responsible for autophagosome positioning in fat cells. DHC: dynein heavy chain; DIC: dynein intermediate chain; DLIC: dynein light intermediate chain; DLC: dynein light chain. K: Quantification of data shown in A-I; n=10 cells. The boundaries of RNAi cells are highlighted in magenta in the grayscale panels.

A-C: Silencing of dynein and dynactin subunits (Dhc64C in A, DCTN1-p150 in B, and DCTN2-p50 in C) using independent RNAi lines also leads to the peripheral redistribution of mCherry-Atg8a positive autophagosomes in vps16a RNAi cells (red arrows). D-I: Peripheral accumulation (red arrows) of Atg8a (D-F) or Rab7 (G-I) positive autophagosomes can be seen in dhc64c, vps16a (D, G), dlic, vps16a (E, H), or dctn1-p150, vps16a (F, I) double RNAi cells. J, K: Overexpression of a dominant negative form of DCTN1-p150 or wild type DCTN2-p50 (which exhibits a dominant-negative effect) in Vps16A depleted cells also results in the peripheral redistribution of mCherry-Atg8a positive autophagosomes (red arrows). L, M: Silencing of the dynein light chain Dlc90F and the dynactin subunit cpa in Vps16A KD cells results in a weaker phenotype, with mCherry-Atg8a positive autophagosomes scattered in the cytosol. N, O: Knockdown of Girdin (N), a suggested dynein activator, and the dynein regulator Lis-1 (O) results in a generally dispersed distribution of mCherry-Atg8a positive autophagosomes in vps16a RNAi cells. P, Q: The perinuclear distribution of autophagosomes in vps16a RNAi cells remains unaffected by the co-expression of khc (P) or klc (Q) RNAi. The boundaries of RNAi-expressing cells are highlighted in magenta in the grayscale panels. R-V: Quantification of data shown in A-Q; n=10 cells.

The proper dynein/kinesin ratio determines the directionality of autophagosome transport.

A, B: Overexpression of kinesin motors blocks the minus-end transport of autophagosomes, leading to their accumulation in the cell periphery in vps16a RNAi cells (red arrows). C: In dynactin-kinesin-vps16a triple RNAi cells, autophagosomes are distributed throughout the cytoplasm. D: Overexpression of the recombinant minus-end motor Khc-nod-LacZ does not affect minus-end directed autophagosome transport in vps16a RNAi cells. E: Khc-nod-LacZ expression partially rescues the dynactin KD-induced peripheral redistribution of autophagosomes in vps16a RNAi cells. F: Quantification of data shown in A-E; n=10. Data for dctn1-p150 RNAi is included as a positive control for peripheral distribution (shown in Figure 2E, K). G, H: Overexpressed Klp98A-3xHA accumulates at the periphery of marginal fat body cells, specifically on the side facing the body cavity and in contact with the hemolymph (red arrows), in both control (G) and fusion-inhibited (vps16a RNAi) cells (H). The boundaries of RNAi or kinesin overexpressing cells are highlighted in magenta in the grayscale panels. Fat body edges are outlined in white in G’ and H’.

Rab7 and Rab39 small GTPases and their interactors are responsible for bidirectional movement of autophagosomes.

A-H: Knockdown of Rab7 (A), its interactor Epg5 (B), the subunits of its guanine nucleotide exchange factor Mon1 (C) and Ccz1 (D), as well as Rab39 (E) and its interactor ema (F), inhibits the perinuclear positioning of autophagosomes in vps16a RNAi cells. In contrast, other factors such as Plekhm1 (G) and prd1 (H) do not affect autophagosome positioning. I: Proposed model of Rab small GTPases with their adaptors involved in autophagosome positioning. J: Quantification of data shown in A-H; n=10 cells. The boundaries of RNAi cells are highlighted in magenta in the grayscale panels.

A: Expression of an independent rab39 RNAi in vps16a RNAi cells results in the scattering of mCherry-Atg8a positive autophagosomes. B: Atg8a positive autophagosomes accumulate throughout the cytoplasm in rab7, vps16a double RNAi cells. C-D: Atg8a positive autophagosomes are scattered throughout the cytosol in rab39, vps16a (C) or ema, vps16a (D) double RNAi cells. E: Expression of rab7 RNAi in Vps16A KD cells eliminates the Rab7 signal, indicating effective RNA interference. F: Rab7 positive structures (autophagosomes) are scattered throughout the cytosol in rab39, vps16a double RNAi cells. The boundaries of RNAi-expressing cells are highlighted in magenta in the grayscale panels. G-I: Quantification of data shown in A-D, F; n=10 cells.

A-D: Silencing the endolysosomal small GTPases Rab2 (A), Rab5 (B), Rab14 (C), and Arl8 (D) does not alter the perinuclear positioning of mCherry-Atg8a positive autophagosomes in vps16a RNAi cells. E, F: Rab11 knockdown leads to the scattering of mCherry-Atg8a positive autophagosomes in vps16a RNAi cells (E), which appears to be due to the extreme accumulation of autophagosomes, as indicated by endogenous Atg8a immunostaining (F). The boundaries of RNAi-expressing cells are highlighted in magenta in the grayscale panels. G, H: Quantification of data shown in A-F; n=10 cells.

A-L: Overexpressing the wild-type (WT) and constitutively active (CA) YFP or GFP-tagged forms of selected endolysosomal/autophagosomal Rab GTPases (Rab7-WT: A, Rab7-CA: B, Rab2-WT: D, Rab2-CA: E, Rab39-WT: F, Rab39-CA: G, Rab11-WT: H, Rab11-CA: I, Rab5-WT: J, Rab5-CA: K, Rab14-WT: L) had no effect on the perinuclear distribution of mCherry-Atg8a positive autophagosomes in vps16a RNAi cells. The signals of both WT and CA forms of YFP-Rab7 and YFP-Rab2 strongly overlapped with mCherry-Atg8a positive autophagosomes in vps16a RNAi cells (A, B, D, E), similar to endogenous Rab7 as evidenced by Rab7 immunostaining (C). The GFP signal of RNAi-expressing cells is false-colored blue in C. In contrast, the YFP-tagged WT form of Rab39 showed weak but apparent colocalization (cyan arrowheads) (F), while Rab11 did not colocalize with mCherry-Atg8a even when its CA form was expressed (H, I). Neither form of Rab5 overlapped with mCherry-Atg8a puncta (J, K), similar to Rab14 (L), indicating that these structures are either endosomes or lysosomes (magenta arrowheads point to autophagosomes in the grayscale panels of L). The boundaries of YFP/GFP-tagged Rab or GFP expressing cells are highlighted in magenta in the grayscale panels showing mCherry-Atg8a or Rab channels. The boxed areas in the main panels, marked by cyan, are enlarged in the insets. M: Potential roles of Rab GTPases in autophagosome transport and fusion. N: Quantification of data shown in A, B, D-L; n=10 cells.

Epg5 is responsible for bidirectional movement of autophagosomes.

A-B: The distribution of Atg8a (A) or Rab7 (B) positive autophagosomes becomes dispersed upon the expression of epg5 and vps16a RNAi. C: Overexpression of YFP-tagged Rab7 does not rescue the scattered distribution of mCherry-Atg8a positive autophagosomes in the absence of Epg5 in vps16a RNAi cells, even though the colocalization of YFP-Rab7 with mCherry-Atg8a remains unaffected. Cyan arrowheads in the grayscale panels point to YFP-Rab7 and mCherry-Atg8a double positive dots. D: The localization of Arl8 positive lysosomes remains perinuclear in epg5, vps16a double RNAi cells. The boundaries of RNAi and YFP-Rab7 expressing cells are highlighted in magenta in the grayscale panels. E-H: Quantification of data shown in A-D; n=10 cells. I, J: Epg5-9xHA colocalizes with endogenous Rab7 (I) or Atg8a (J) positive structures in S2R+ cells. Cyan arrowheads within insets (marked by cyan boxes in panels I and J) point to Epg5-9xHA and Rab7 or Atg8a double positive structures, respectively. I’ and J’ show scatter plots generated from the images of cells in panels I and J, respectively, depicting the intensity correlation profiles of Epg5-9xHA with Rab7 or Atg8a. Pearson correlation coefficients (R) are indicated, with the average R (n=10 cells) also shown, indicating colocalization in both cases. K: Epg5-9xHA colocalizes with Atg8a positive, Lamp1-3xmCherry negative (pre-fusion) autophagosomes, as well as with Atg8a and Lamp1-3xmCherry double positive autolysosomes in S2R+ cells. Cyan arrowheads in insets (marked by a cyan box in panel K) point to Epg5-9xHA, Atg8a double positive, Lamp1-3xmCherry negative structures, while a yellow arrowhead marks a triple positive autolysosome. K’ and K”: Scatter plots based on the cell in panel K, show intensity correlations of Epg5-9xHA with Lamp1-3xmCherry and Atg8a, respectively. Pearson correlation coefficients indicate partial colocalizations. L, M: Coimmunoprecipitation experiments show that Epg5-9xHA binds to Rab7-FLAG (L) and endogenous Dhc64C (M) in cultured Drosophila cells. The asterisk in L marks immunoglobulin light chain. The smeared input bands of Dhc64C in panel M are due to the large size of Dhc64C, which affects its migration characteristics.

A-D: The size, number and distribution of Rab7 (A, B) or FYVE-GFP positive endosomes (C, D) and Lamp1 positive lysosomes (C, D) remain unaffected by epg5 RNAi expression (B, D) in nephrocytes, compared to controls (A, C). E-K: Quantification of data shown in A-D; E: n=512 (control) and 461 (epg5 RNAi) endosomes from 10 cells. F: n=10 cells. G: n=496 (control) and 461 (epg5 RNAi) endosomes from 10 cells. H: n=10 cells. I: n=15 cells. J: n=422 (control) and 397 (epg5 RNAi) lysosomes from 10 cells. K: n=10 cells. L, M: Ultrastructural analysis reveals no significant difference in the endolysosomal compartment between control (L) and epg5 RNAi (M) nephrocytes. α: late endosomes; β: lysosomes; m: mitochondria; LD: lipid droplets.

A-H: Knockdown of key regulators of autophagosome transport in a Snap29 RNAi background results in autophagosome localization patterns similar to those observed with Vps16A RNAi.

A-H: In luciferase; Snap29 double knockdown cells, non-fused autophagosomes accumulate in the perinuclear region, marked by mCherry-Atg8a (A). Shot; Snap29 double knockdown causes autophagosomes to accumulate around an ectopic MTOC (B), marked by yellow arrows in B’. Nuclear outlines are shown in blue. Dhc64C knockdown in Snap29 RNAi cells causes autophagosomes to redistribute to the cell periphery (C, red arrows in C’). Khc knockdown does not alter the perinuclear distribution of autophagosomes seen in Snap29 RNAi cells (D). Co-knockdown of Snap29 with Rab7 (E), Epg5 (F), Rab39 (G), or ema (H) results in scattered autophagosome distribution throughout the cytoplasm. In grayscale panels, the boundaries of RNAi-expressing cells are highlighted in magenta. I: Quantification of the data shown in panels A and C–H. n = 10 cells.

The positioning of pre-fusion, immature autolysosomes is very similar to autophagosomes in vps16a RNAi cells.

A-J: Lamp1 positive lysosomes accumulate around the nuclei in cells co-expressing a control (luciferase) RNAi (A). This positioning remains unaffected by the co-expression of khc (E) or arl8 RNAi (J). Depletion of Shot (B) or Dhc64C (C) in vps16a RNAi cells redistributes Lamp1 positive lysosomes from the perinuclear cytoplasm to an ectopic MTOC (yellow arrows) or to the periphery, respectively. Lamp1 positive lysosomes are scattered throughout the cytosol in vps16a RNAi cells upon the co-expression of rab7 (F), rab39 (G), and ema (H) RNAi-s. In contrast, Lamp1 positive lysosomes retain their perinuclear distribution in epg5, vps16a double RNAi cells (D). Similar to Rab7 or Rab39, the expression of rab2 RNAi in Vps16A KD cells results in the scattering of Lamp1 positive lysosomes, with a trend observed that lysosomes tend to accumulate near the periphery (red arrows) (I). The boundaries of RNAi-expressing cells are highlighted in magenta in the grayscale panels. The outlines of nuclei are drawn in blue in B’. K, L: Quantification of data shown in A, C-J; n=10 cells.

Epg5 regulates autolysosome maturation.

A, B: Epg5 knockdown results in a significant reduction in the size of 3xmCherry-Atg8a positive autolysosomes (B) compared to control RNAi (luciferase RNAi) expressing cells (A). C, D: epg5 RNAi cells lack large Rab7 (C) and Arl8 (D) positive autolysosomes, which are present in surrounding control cells (cyan arrowheads in insets point to Rab7 and Arl8 positive autolysosomes in control cells). The boundaries of RNAi cells are highlighted in magenta in the grayscale panels. E-H: Quantification of data shown in A-D; n=10 cells.

Minus-end directed transport is required for autolysosome maturation.

A-E: Autolysosome size is significantly reduced upon the loss of dynein (A-D) or dynactin (E) function. Red arrows point to autolysosomes at the cell periphery in A’-E’ and in the inset of E. F, G: Kinesin knockdowns do not significantly influence autolysosome size. H: Autolysosome size increases at ectopic foci (yellow arrows) in shot RNAi cells. The boundaries of RNAi-expressing cells are highlighted in magenta in the grayscale panels. The outlines of nuclei are drawn in blue in the inset of E’ and in H’. I-L: Quantification of data shown in A-H; n=10 cells.

A: Silencing of Rab7 decreases mCherry-Atg8a positive autolysosome size as expected, and they remain generally dispersed. B, C: Knockdown of Rab39 (B), but not ema (C), relocates mCherry-Atg8a positive autolysosomes to the periphery (red arrows). Importantly, neither knockdown decreases the size of autolysosomes. The boundaries of RNAi-expressing cells are highlighted in magenta in the grayscale panels. D, E: Quantification of data shown in A-C; n=10 cells.

Loss of the autophagosome positioning machinery decreases autophagosome-lysosome fusion.

A-G: In starved control RNAi (luciferase) expressing cells, mCherry-Atg8a overlaps with endogenous Lamp1 (A), indicating normal autophagosome-lysosome fusion and autolysosome formation. Autolysosomes still form in shot (B) or khc RNAi (C) cells, as mCherry-Atg8a colocalizes with endogenous Lamp1 similar to controls, but these are found in ectopic foci (yellow arrows) in shot RNAi cells. The outlines of nuclei are drawn in blue in B’ and B’’. Conversely, RNAi-s targeting factors responsible for minus-end directed autophagosome transport (D-G) decrease this overlap, suggesting less effective autophagosome-lysosome fusion. The GFP signal of RNAi-expressing cells is false-colored blue in composite images. The boundaries of RNAi-expressing cells are highlighted in magenta. The boxed areas in the main panels, marked by cyan, are enlarged in the insets (M, merged image; A8, 3xmCherry-Atg8a; L1, Lamp1). Cyan arrowheads point to mCherry-Atg8a/Lamp1 double-positive structures, while magenta and green arrows indicate mCherry-Atg8a or Lamp1 single-positive dots, respectively. H: Quantification of data shown in A-G; n=10 cells.

A-F: Partial inhibition of autophagosome-lysosome fusion is observed upon knockdown of factors responsible for minus-end transport, based on 3xmCherry-Atg8a and GFP-Lamp1 colocalization experiments. In starved control RNAi (luciferase) expressing cells, mCherry-Atg8a overlaps with GFP-Lamp1 (A), indicating normal autophagosome-lysosome fusion and autolysosome formation. Autolysosomes still form in shot (B) or khc RNAi cells (C), as mCherry-Atg8a colocalizes with GFP-Lamp1, but these are found in ectopic foci (yellow arrows) in the former case. The outlines of nuclei are drawn in blue in B’, B”. Conversely, RNAi-s targeting factors responsible for minus-end directed autophagosome transport (D-F), including epg5 RNAi, decrease this overlap, suggesting less effective autophagosome-lysosome fusion. G: Quantification of data shown in A-F; n=10 cells. H: Silencing of Vps16A, used as a positive control, almost completely blocks autophagosome-lysosome fusion as indicated by the minimal overlap of the two signals. I-K: Quantification of data shown in H; n=10 cells. The boundaries of RNAi-expressing cells are highlighted in magenta in the grayscale panels. The boxed areas in the main panels, marked by cyan, are enlarged in the insets (M, merged image; A8, 3xmCherry-Atg8a; L1, GFP-Lamp1). Cyan arrowheads show overlapping dots, while magenta and green arrows point to non-colocalizing 3xmCherry-Atg8a positive and GFP-Lamp1 positive dots, respectively.

Model of the transport of autophagosomes and lysosomes in starved fat cells.

Before fusion, autophagosomes and lysosomes are transported towards the perinuclear ncMTOC by a cytosolic dynein complex in starved fat cells to ensure proper fusion and effective degradation. This process requires Rab7, Rab39, and their interactors Epg5 and ema on autophagosomes, and Rab2, Rab7, Rab39, and the Rab39 interactor ema, but not Epg5, on lysosomes. After fusion, Arl8 mediates the plus-end transport of autolysosomes, while Rab39 promotes dynein-regulated minus-end directed transport. Thus, the motility of pre-fusion and post-fusion organelles is differently regulated: pre-fusion organelles generally move towards the MTOC, while post-fusion organelles exhibit bidirectional motility. This spatial regulation ensures proper fusion rates and degradation efficiency.