HTT (huntingtin) and RAB7 co-migrate retrogradely on a signaling LAMP1-containing late endosome during axonal injury

ABSTRACT

HTT (huntingtin) is a 350-kDa protein of unknown function. While HTT moves bidirectionally within axons and HTT loss/reduction causes axonal transport defects, the identity of cargo-containing vesicles that HTT helps move remain elusive. Previously, we found an axonal retrogradely moving HTT-Rab7 vesicle complex; however, its biological relevance is unclear. Using Drosophila genetics, in vivo microscopy, membrane isolation and pharmacological inhibition, here we identified that adaptors Hip1 and Rilpl aid the retrograde motility of LAMP1-containing HTT-Rab7 late endosomes, not autophagosomes. Reduction of Syx17 and chloroquine- or bafilomycin A₁-mediated pharmacological inhibition, but not reduction of Atg5, disrupted the in vivo motility of these vesicles. Further, because HTT-Rab7 vesicles colocalized with long-distance signaling components (BMP signaling: tkv-wit, injury: wnd) and move in a retrograde direction after Drosophila nerve crush, we propose that these vesicles likely traffic damage signals following axonal injury. Together, our findings support a previously unknown role for HTT in the retrograde movement of a Rab7-LAMP1-containing signaling late endosome.

Abbreviations

Atg5: Autophagy-related 5; Atg8a: Autophagy-related 8a; AL: autolysosome; AP: autophagosome; BAF1: bafilomycin A₁; BDNF: brain derived neurotrophic factor; BMP: bone morphogenetic protein; Cyt-c-p: Cytochrome c proximal; CQ: chloroquine; DCTN1: dynactin 1; Dhc: dynein heavy chain; EE: early endosome; DYNC1I1: dynein cytoplasmic 1 intermediate chain 1; HD: Huntington disease; HIP1/Hip1: huntingtin interacting protein 1; HTT/htt: huntingtin; iNeuron: iPSC-derived human neurons; IP: immunoprecipitation; Khc: kinesin heavy chain; KIF5C: kinesin family member 5C; LAMP1/Lamp1: lysosomal associated membrane protein 1; LE: late endosome; MAP1LC3/LC3: microtubule associated protein 1 light chain 3; MAP3K12/DLK: mitogen-activated protein kinase kinase kinase 12; MAPK8/JNK/bsk: mitogen-activated protein kinase 8/basket; MAPK8IP3/JIP3: mitogen-activated protein kinase 8 interacting protein 3; NGF: nerve growth factor; NMJ: neuromuscular junction; NTRK1/TRKA: neurotrophic receptor tyrosine kinase 1; NRTK2/TRKB: neurotrophic receptor tyrosine kinase 2; nuf: nuclear fallout; PG: phagophore; PtdIns3P: phosphatidylinositol-3-phosphate; puc: puckered; ref(2)P: refractory to sigma P; Rilpl: Rab interacting lysosomal protein like; Rip11: Rab11 interacting protein; RTN1: reticulon 1; syd: sunday driver; SYP: synaptophysin; SYT1/Syt1: synaptotagmin 1; STX17/Syx17: syntaxin 17; tkv: thickveins; VF: vesicle fraction; wit: wishful thinking; wnd: wallenda

KEYWORDS:

• Axonal injury
• axonal transport
• Drosophila
• endolysosomes
• huntingtin
• signaling endosome

Introduction

The Huntington disease (HD) protein, HTT (huntingtin), is ubiquitously expressed; however, its neuronal function remains unknown despite being essential for development [1]. Although more than 350 binding partners for HTT have been identified with functions across a wide-range of cellular processes including gene expression, metabolism, protein turnover, endocytosis, and trafficking [2–6], the molecular mechanisms by which HTT and its binding partners function remain elusive. Previously, we showed that HTT moves bi-directionally within axons by interacting with kinesin 1 and dynein [7], and htt can mediate the motility of several Rab-GTPase-containing vesicles, suggesting that htt can serve as a scaffolding protein for vesicular trafficking. Reduction of htt disrupts the bi-directional movement of Rab3, Rab4, Rab11, Rab19, the retrograde motility of Rab7 and enhanced the anterograde motility of Rab2 [8–10]. Therefore, while it is likely that htt can differentially facilitate the axonal motility of Rab-GTPase-containing vesicles, the functional significance of these distinct HTT-Rab containing axonal complexes remain ambiguous.

Upon endocytic internalization into an early endosome (EE), proteins and lipids are sorted for recycling or degradation. Cargo destined to be degraded remain in an EE, which then mature to a late endosome (LE), marked by RAB7/Rab7 [11,12]. In neurons, LEs are thought to be retrogradely moved from distal regions of the axon to the cell body for either lysosome-mediated degradation [13,14] and/or long-distance signaling [15,16]. Neurons are suggested to maintain a spatial gradient of LE/lysosome sub-populations, marked by Rab7 and/or LAMP1/Lamp1, with highly degradative lysosomes being enriched near the cell body while pre-degradative LEs move from distal regions of the axons and dendrite to the cell body for degradation [14,17,18]. In parallel, ATG5/Atg5 (autophagy related 5)-dependent phagophore (PG) expansion [19,20] forms a nascent autophagosome (AP) in distal axons, marked by Atg8/MAP1LC3/LC3 (Autophagy-related 8/microtubule associated protein 1 light chain 3) [21,22]. An AP can form an amphisome through STX17/Syx17 (syntaxin 17)-dependent fusion [23,24] with an LE, which then are proposed to retrogradely move to the cell body for degradation upon fusing with a lysosome to form an autolysosome (AL [14,25]). HTT has been proposed to be present in macroautophagy/autophagy during cargo loading of an elongating PG [26] and during the transport of an AP [27]. However, since Atg8/LC3 is present in all autophagic compartments from the PG to the AL [28], this ambiguity creates a challenge in isolating HTT/htt during autophagy.

Following neuronal injury, retrograde movement is critical for damage signaling [29] and axonal regeneration [30]. In fact, an anterograde-to-retrograde conversion has been observed following rat sciatic nerve crush [31]. In response to axonal injury, retrogradely moving damage signals such as MAP3K12/DLK/wnd (mitogen-activated protein kinase kinase kinase 12/wallenda) and/or MAPK8/JNK (mitogen-activated protein kinase 8)-containing vesicles promote upregulation of pro-regenerative genes essential for axon regeneration [29,32–34]. Interestingly, Rab7/RAB7 associates with syd/MAPK8IP3/JIP3 (sunday driver/mitogen-activated protein kinase 8 interacting protein 3)-containing endosomes that traffic MAPK/JNK [29,35] and with retrogradely moving MAP3K12/DLK-containing vesicles [36]. Although HTT was recently shown to be upregulated during mouse corticospinal motor neuron regeneration following spinal cord injury [37], and has been proposed to be an injury-response signaling hub in axotomized retinal ganglion cells [38], how HTT functions during axonal injury remains unknown.

Because Rab7 is present with LEs and during autophagy, here we test the hypothesis that the putative retrogradely moving HTT-Rab7 vesicle complex is a LE and/or is part of autophagy. We found that the retrograde motility of the HTT-Rab7 complex is aided by HTT or Rab7-associated proteins Hip1 (Huntingtin-interacting protein 1) and Rilpl (Rab interacting lysosomal protein like), and that this complex is distinct from the bidirectionally-moving HTT-RAB4/Rab4 synaptic vesicle complex we previously isolated [10]. HTT-Rab7 vesicles are likely LAMP1/Lamp1-containing LEs, and not APs, since the motility of HTT-Rab7 vesicles were impaired by loss/reduction of LE-AP fusion mediated by mutant Syx17 as well as pharmacological inhibition (chloroquine or bafilomycin A₁) of lysosome acidification and lysosome-AP fusion. Further, the retrogradely moving HTT-Rab7 vesicles colocalize with long-distance signaling components (the BMP [bone morphogenetic protein]- receptors tkv [thickveins] and wit [wishful thinking]), axonal injury signals (wnd/MAP3K12/DLK), and move retrogradely even after Drosophila nerve crush, in contrast to perturbed motility of synaptic vesicles. Taken together, our observations propose a previously unknown role for HTT in the retrograde movement of a Rab7-LAMP1-containing signaling LE, with a functional relevance during axonal injury

Results

The retrogradely moving HTT-Rab7 vesicle complex is aided by adaptor proteins Rilpl and Hip1

We previously found that adaptor proteins Rip11 (Rab11-interacting protein) and Hip1, but not milton, nuf (nuclear fallout), or nmo (nemo), likely aid the bi-directional motility of the HTT-Rab4 vesicle complex [10]. To identify adaptor proteins that aid the retrograde motility of the HTT-Rab7 vesicle complex, we performed a candidate screen similar to what we have done previously [10]. Using the pGAL4-62B-SG26-1 driver, which expresses in a small population of motor neurons [39,40], we generated Drosophila larvae simultaneously expressing Rab7-GFP and HTT15Q-mRFP (nonpathogenic form of human HTT) within larval axons and observed co-migrating retrogradely moving vesicles (Figure 1a). Colocalized trajectories (yellow) from merged kymographs revealed that 39% of vesicles containing HTT and Rab7 moved retrogradely while only 14% vesicles moved anterogradely (Figure 1b). Reduction of htt (htt-KO/+, Df(98E2);CG999075 [41], or htt-RNAi [7,9]) with Rab7-GFP or reduction of Rab7 (Rab7^EY10675/+ or Rab7^d1/+ [null allele [42]]) with HTT-mRFP disrupted Rab7 or HTT motility; significantly decreasing the retrograde populations of Rab7-GFP or HTT-mRFP vesicles (Fig. S1AB). Decreased retrograde motility corresponded to increased distribution of Rab7-GFP at neuromuscular junctions (NMJs) and decreased localization at cell bodies (Fig. S1 CD). Likewise, expression of a dominant negative form of Rab7 (Rab7^T22N) also disrupted HTT motility (Fig. S1B). Western blot analysis showed that Rab7 protein levels in homozygous Rab7^EY10675/Rab7^EY10675 flies were reduced to 28% of WT levels (Fig. S2A), similar to what was previously observed for the Rab7^d1 null allele [42]. Further, reductions in either kinesin 1 (Khc⁸/+) or dynein (Dhc^6-10/+) decreased both the anterograde and retrograde motility of either HTT-mRFP or Rab7-GFP-containing vesicles (Fig. S1EF).

Figure 1. The retrogradely moving HTT-Rab7 vesicle is aided by adaptor proteins Hip1 and Rilpl. (A) Schematic diagram of larval nervous system showing the brain and segmental nerves. Red box = imaged area (90 µm). Arrows depict location of the cell bodies (retrograde) and synapses (anterograde). Representative images and kymographs from simultaneous dual-color movies from larvae expressing both HTT-mRFP (red) and Rab7-GFP (green). Co-migrating tracks that contain HTT and Rab7 are seen (yellow arrow). Note that green or red only tracks are also seen (arrowheads). Bar: 5 µm. (B) Quantification of the directional analysis of trajectories containing HTT and Rab7 (yellow). n = 10. (C-D) Representative images from movies and kymographs from larvae expressing either Rab7-GFP or HTT-mRFP alone, with a Hip1 reduction, or with a Rilpl reduction. Bar: 10 µm. (E-F) Directional analysis of Rab7-GFP or HTT-mRFP trajectories compared to Rab7-GFP or HTT-mRFP trajectories with a Hip1 or Rilpl reduction. n = 10. (G) Schematic diagram of homogenate fractionation into the perinuclear supernatant (PNS), vesicle fraction (VF), soluble fraction (SF), and heavy membrane pellet (P1) by ultra-centrifugation and sucrose gradient separation. Representative western blot of an immunoprecipitation of RAB7-containing mouse VF, probed with HTT, DYNC1I1, KIF5C, DCTN1, HIP1, LAMP1, SYT1, and SYP. HTT, DYNC1I1, KIF5C, DCTN1, HIP1, and LAMP1 show presence in the RAB7 IP, while SYT1 and SYP do not. (H) Representative western blot of an immunoprecipitation of HTT-containing mouse VF, probed with RAB7, which shows presence in the HTT IP. (I) Representative western blot of an immunoprecipitation of HTT-containing human iNeuron VF, probed with HTT, DYNC1I1, KIF5C, HIP1, RAB7, LAMP1, and SYT1, which all show presence in the HTT IP. No bands are seen in the negative no antibody control (−crtl). n = 3. Statistical analysis was conducted using the two-sample two-sided Student’s t-test. Data represented as mean±sem. ns=p > 0.01, *p < 0.01, **p < 0.001, ***p < 0.0001, ****p < 0.00001. Also see Fig. S1 and S2.

To screen for adaptor proteins that aid the motility of the HTT-Rab7 vesicle complex, we generated larvae expressing Rab7-GFP or HTT-mRFP in the context of reductions in Rab-associated proteins nemo (nmo^P1/+) [43,44], Rip11 (Rip11^KG02485/+) [45], or Rilpl (Rilpl^EY06476/+) [46,47], Hip1 (Hip1^MB04365/+) [48–50] or milton (milt^k04704/+). Although HTT moves bi-directionally within larval axons, Drosophila lacks HAP1, but milton shares a HAP1-like N-domain [51–53]. Reduction of Rilpl or Hip1 disrupted Rab7-GFP/HTT-mRFP motility causing Rab7 or HTT blockages, while reductions in milton, nemo, or Rip11 did not (Figure 1c,d, Fig. S1GH). Reduction of Hip1 significantly decreased both the anterograde and retrograde movement of Rab7-HTT-containing vesicles (Figure 1e), while reduction of Rilpl only decreased the retrograde movement of Rab7 or HTT-containing vesicles (Figure 1f). Biochemical analysis coupled with immunoprecipitation (IP) of mouse brains or iPSC-derived human neurons (iNeurons) showed that RAB7, HTT, HIP1, and motor proteins kinesin 1 and dynein were present on RAB7-containing or HTT-containing membranes (Figure 1G-i). Homogenized brains/neuron lysates were first fractionated and the PNS (post-nuclear supernatant), VF (vesicle fraction), SF (soluble fraction), and P1 (heavy-membrane pellet) were isolated using sucrose-gradient centrifugation (Fig. S2B), and then HTT or RAB7 was immunoprecipitated from the VF. Both full length and fragmented HTT were seen in the RAB7-VF IP (Figure 1G). Likewise, RAB7 was seen in the HTT-VF IPs (Figure 1h,i). In addition, HIP1, KIF5C and DYNC1I1 were also present in both the HTT-VF IPs (Figure 1i) and the RAB7-VF IP (Figure 1g). Interestingly, SYT1 (synaptotagmin 1) and SYP (synaptophysin) were not observed with the RAB7-VF IP, but LAMP1 was present (Figure 1g), indicating that the HTT-RAB7 vesicle is likely distinct from the putative HTT-RAB4 synaptic vesicle complex we previously isolated [10].

The retrogradely moving HTT-Rab7 vesicle complex is likely a LAMP1/Lamp1-containing late endosome

Because RAB7/Rab7 is seen with retrogradely moving LEs [47], lysosomes [46], and APs [25], and HTT/htt is implicated during autophagy [26,27], we next tested the hypothesis that the HTT-Rab7 vesicle complex is either a LE, lysosome and/or an AP (Fig. S3A). Larvae simultaneously expressing either Rab7-GFP/HTT-mRFP, mCherry-Atg8a/HTT-eGFP, or LAMP1-GFP/HTT-mRFP showed retrogradely co-migrating Rab7-HTT, Atg8a-HTT or LAMP1-HTT containing vesicles (Figure 2a). Colocalized trajectories (yellow) revealed that 25% of vesicles contained HTT and Rab7, 17% of vesicles contained HTT and Atg8a, and 22% of vesicles contained HTT and LAMP1 (Figure 2b). Furthermore, reduction of htt (htt-RNAi) or Rab7 (Rab7^EY10675/+) with either mCherry-Atg8a or LAMP1-GFP disrupted Atg8a or LAMP1 motility (Figure 2c,d). While reduction of htt significantly decreased the retrograde movement of Atg8a-containing vesicles (Figure 2e), reduction of Rab7 decreased both the anterograde and retrograde motility of Atg8a-containing vesicles (Figure 2f). Reduction of htt or Rab7 decreased both the anterograde and retrograde motility of LAMP1-containing vesicles (Figure 2e). Therefore, HTT and Rab7 are likely present together during autophagy and with lysosomes or LEs.

Figure 2. The HTT-Rab7 vesicle complex is likely a LAMP1/Lamp1-containing late endosome and/or autophagosome. (A) Arrows depict location of the cell bodies (retrograde) and synapses (anterograde). Representative images and kymographs from simultaneous dual-color movies from larvae co-expressing either HTT-mRFP and Rab7-GFP, HTT-eGFP and mCherry-Atg8a, HTT-mRFP and LAMP1-GFP. Co-migrating tracks that contain both red and green trajectories are seen (yellow arrow). X axis=distance (µm), Y axis=time (s). Bar: 5 µm. (B) Quantification of the average co-migratory trajectories reveal the percent co-migration between HTT-Rab7, HTT-LAMP1, and HTT-Atg8a. n = 6. (C-D) Representative images from movies and kymographs from larvae expressing mCherry-Atg8a or LAMP1-GFP alone, in the context of htt (htt-RNAi) or Rab7 (Rab7^EY10675/+) reduction. Bar: 10 µm. (E-F) Directional analysis of mCherry-Atg8a or LAMP1-GFP trajectories alone compared to mCherry-Atg8a or LAMP1-GFP trajectories with a htt or Rab7 reduction. n = 10. Statistical significance was determined using the two-sample two-sided Student’s t-test. Data represented as mean±sem. ns=p > 0.01, *p < 0.01, **p < 0.001, ***p < 0.0001. ****p < 0.00001. Also see Fig. S3.

LEs can fuse with lysosomes [54], forming endolysosomes, which can converge onto the autophagic pathway upon AP-lysosome fusion to form ALs (Fig. S3A [23]). Since HTT and Rab7 co-migrated with Atg8a and LAMP1, the retrogradely moving HTT-Rab7 vesicle complex could be a LE, an AP, a lysosome, or all three. To isolate these specific compartments, we focused on Atg5 and Syx17. Atg5 is a key protein involved in the assembly of an AP from a PG [19,55], and is a marker for early autophagic compartments (Fig. S3A), while Syx17 is a SNARE protein identified in Drosophila that aids LE-AP fusion to form amphisomes [23,24] and is a candidate marker for middle autophagic compartments (Fig. S3A). APs fuse with lysosomes to become ALs (Fig. S3A [23,56]). APs can also fuse with LEs to form amphisomes [57,58], which can then fuse with lysosomes (Fig. S3A [23]). Interestingly, Syx17-dependent LE-AP fusion is thought to be required for the retrograde motility of APs within cultured dorsal root ganglion neurons [25]. We found that Atg5, Syx17, and LAMP1 mark discrete compartments in larval axons that do not show colocalization with each other (Fig. S3BE). To further define the HTT-Rab7 retrogradely co-migrating vesicle, we generated larvae expressing either HTT-mRFP or Rab7-GFP in the context of Atg5 reductions (Atg5^d04577/+ or Atg5^5cc5/+) or Syx17 reductions (Syx17^f01971/+ or Syx17^LL/+). Similar to what was seen previously in Atg5^5cc5 and Syx17^LL null alleles [24,59], western blot analysis showed that Atg5 protein levels in Atg5^d04577/+ flies were reduced to 64% and Syx17 protein levels were reduced to 81%= 34KD, 62%= 40KD in Syx17^f01971/+ flies compared to WT levels (Fig. S2A). Further, Atg5^d04577 mutant flies also showed an ataxic phenotype (Fig. S2C), similar to what was observed in null Atg5^5cc5 mutant flies [59], while Syx17^f01971 was previously verified by PCR [60,61]. Intriguingly, ref(2)P protein levels were significantly increased in Atg5^d04577/+, Syx17^f01971/+ and Rab7^EY10675/Rab7^EY10675 flies (Fig. S2D), similar to what was previously observed in Atg5^5cc5, Syx17^LL, and Rab7^d1 null mutants [24,42,59], indicating the disruption of degradation in these mutant lines. Reduction of Atg5 significantly decreased the bi-directional motility of Rab7-GFP-containing vesicles (Figure 3a,c), however the motility of HTT-mRFP containing vesicles was unaffected (Figure 3b,c). Immunolocalization analysis showed that while Rab7 and Atg5 colocalized (Fig. S3CE), HTT and Atg5 did not (Fig. S3DE). Reduction of Syx17 disrupted the retrograde motility of both HTT-mRFP or Rab7-GFP, with increased populations of stationary/stalled vesicles (Figure 3a–d), decreased populations of retrograde vesicles (Figure 3a–d), and an increased frequency of HTT-Rab7 colocalization (Figure 3e,f), but not HTT-Lamp1 colocalization (Fig. S4C). Note that both HTT and Rab7 colocalize with Syx17 (Fig. S3CDE). Reduction of htt increased the colocalization of Rab7-Syx17 (Fig. S4A) and reduction of Rab7 increased HTT-Syx17 colocalization (Fig. S4B), while the colocalization of Rab7-Lamp1 (Fig. S4A) and HTT-Lamp1 (Fig. S4B) were decreased. Further, STX17 (syntaxin 17) co-immunoprecipitated with HTT-containing vesicles isolated from mouse brain tissue, but not with ATG5 in mice (Figure 3g). STX17 also co-immunoprecipitated with RAB7-containing vesicles (Figure 3h). Therefore, while the retrogradely moving HTT-RAB7/Rab7 vesicle is likely a LE, disruption of LE-AP fusion through Syx17 reduction/loss likely disrupts the fusion of the putative HTT-RAB7/Rab7 LE with APs.

Figure 3. Disrupting autophagosome-late endosome fusion with mutant Syx17 impairs the motility of the HTT-Rab7 vesicle complex. (A-B) Arrows depict location of the cell bodies (retrograde) and synapses (anterograde). Representative images and kymographs from larvae expressing (A) Rab7-GFP or (B) HTT-mRFP alone, with a reduction of Atg5 (Atg5^d04577/+ or Atg5^5cc5/+), or with a reduction of Syx17 (Syx17^f01971/+ or Syx17^LL/+). X axis=distance (µm), Y axis=time (s). Bar: 10 µm. (C-D) Directional analysis of (C) Rab7-GFP or (D) HTT-mRFP trajectories alone compared to Rab7-GFP or HTT-mRFP trajectories with Atg5 or Syx17 reductions. n = 10. (E) Representative images from larvae expressing HTT-mRFP alone or with a reduction of Syx17 (Syx17^f01971/+) that have been immunostained for Rab7. Note the presence of colocalized puncta (yellow) in the merged overlay panels with larger accumulates with Syx17 reduction. Bar: 10 µm. (F) Quantification of colocalization (mander’s correlation) between HTT and Rab7 in larval axons in the context of Syx17 reduction compared to WT. n = 6. (G-H) Representative western blot of an immunoprecipitation of HTT-containing mouse VF (G) or an immunoprecipitation of RAB7-containing mouse VF (H), probed for HTT, RAB7, LAMP1, STX17, ATG5, and SYT1. Note that while RAB7, LAMP1, STX17, and SYT1 are present in the HTT IP, ATG5 is not. While HTT, LAMP1, and STX17 are present in the RAB7 IP, SYT1 and ATG5 are not. No bands are seen in the negative no antibody control (−crtl). Statistical significance was determined using the two-sample two-sided Student’s t-test. Data represented as mean±sem. ns=p > 0.01, *p < 0.01, **p < 0.001, ***p < 0.0001, ****p < 0.00001. Also see Fig. S2, S3, and S4.

To further test the proposal that HTT-RAB7/Rab7-containing LEs fuse with APs, we next pharmacologically inhibited lysosome acidification and AP-lysosome fusion using chloroquine (CQ) or bafilomycin A₁ (BAF1) [28,62–64]. Previous studies show that CQ or BAF1 treatment cause LAMP1 accumulations [65–71]. Indeed, we also found accumulations of LAMP1 within Drosophila larval neurons expressing LAMP1-GFP fed on food laced with CQ [72] or BAF1 [22,70] for 24 h prior to in vivo imaging compared to buffer (Fig. S5A). Accumulations of Rab7 were also seen within CQ or BAF1 treated larvae expressing Rab7-GFP (Fig. S5A). However, CQ or BAF1 treatment had no effect on Syt1-containing synaptic vesicles in larvae expressing Syt1-eGFP (Fig. S5A). Further, larvae simultaneously expressing HTT-mRFP and Rab7-GFP treated with either CQ or BAF1 showed impaired motility of HTT-Rab7-containing vesicles (yellow trajectories, Figure 4a). CQ or BAF1 treatment significantly increased the co-migration of HTT-mRFP and Rab7-GFP (Figure 4b), but significantly decreased the retrograde motility of HTT-Rab7 vesicles (Figure 4c) compared to buffer-treated larvae. While impaired motility of HTT-Rab7 vesicles corresponded to increased levels of colocalization between Rab7-GFP and HTT-mRFP at cell bodies (Figure 4d,e S5B), decreased colocalization was observed at NMJs (Fig. S5C), suggesting that perhaps CQ also disrupts endocytosis [73] and the packaging of Rab7 and HTT into LEs/lysosomes at NMJs. Furthermore, we found that HTT-Rab7 vesicles colocalized with Lamp1 in larval axons, which significantly increased upon treatment with CQ (Figure 4f,g). Taken together, these observations support the proposal that the retrogradely moving HTT-RAB7/Rab7 vesicle complex is likely a LAMP1/Lamp1-containing LE, that can fuse with an amphisome or an AP in a STX17/Syx17-dependent manner.

Figure 4. CQ- and BAF1-mediated inhibition of lysosome acidification and lysosome-autophagosome fusion impairs the motility of HTT-Rab7 vesicles. (A) Schematic diagram of larval nervous system showing the brain and segmental nerves. Red box=imaged area (90 µm). Arrows depict location of the cell bodies (retrograde) and synapses (anterograde). Representative images and kymographs from simultaneous dual-color movies from larvae expressing both HTT-mRFP (red) and Rab7-GFP (green) that have been fed either buffer-, CQ-, or BAF1-laced food. Co-migrating tracks that contain HTT and Rab7 are seen (yellow arrow). X axis=distance (µm), Y axis=time (s). Bar: 5 µm. (B) Quantification of co-migratory (yellow) trajectories (%) between HTT and Rab7 in buffer-, CQ-, and BAF1-treated larvae. (C) Quantification of directional analysis of yellow (HTT and Rab7) trajectories (%) in buffer-, CQ-, and BAF1-treated larvae. (D) Schematic diagram of larval nervous system showing the brain and segmental nerves. Red box=imaged area. Representative images from simultaneous dual-color images of neuron cell bodies from larvae co-expressing HTT-mRFP (red) and Rab7-GFP (green) that have been fed either buffer-, CQ-, or BAF1-laced food. Colocalized puncta that contain HTT and Rab7 are seen (yellow arrows). Bar: 2 µm. (E) Quantification of percent colocalization (mander’s) between HTT and Rab7%) in nerves of buffer-, CQ-, and BAF1-treated larvae. n = 10. (F) Representative images of segmental nerves from larvae expressing HTT-mRFP (red) that have been simultaneously immunostained with Lamp1 (green) and Rab7 (blue) following treatment with either buffer or CQ. Note that the merged panel of buffer-treated larvae reveals heterogenous overlays containing either HTT-Rab7 (purple arrows), Rab7-Lamp1 (teal arrows), or HTT-Rab7-Lamp1 (white arrows), while the merged panel of CQ-treated larvae show more overlays containing HTT-Rab7-Lamp1 (white arrows) Bar: 10 µm. (G) Quantification of colocalization (mander’s) between HTT-Rab7, HTT-Lamp1, and Rab7-Lamp1%) in nerves of buffer-, CQ-, or BAF1-treated larvae. Data represented as overlaid histogram-dot plots (left) as well as proportional Venn diagrams (right). n = 6. Statistical significance was determined using the two-sample two-sided Student’s t-test. Data represented as mean±sem. ns=p > 0.01, *p < 0.01, **p < 0.001, ***p < 0.0001, ****p < 0.00001. Also see Fig. S5. .

Evidence for a retrogradely moving HTT-Rab7 signaling late endosome during axonal injury

Rab7 has been implicated in facilitating the retrograde axonal movement of long-distance signaling components [12,16,36,74]. To examine whether the HTT-Rab7-containing LE co-migrates with long-distance signaling components, we evaluated the retrograde BMP-signaling receptors tkv and wit [75–77]. We also examined phosphatidylinositol-3-phosphate (PtdIns3P) phospholipids (FYVE), which are predominantly found within EEs, LEs, or lysosomes and perform key functions during signaling and membrane trafficking [78–82]. Drosophila larval axons expressing the type-I receptor tkv-eGFP showed strong colocalization (80%) with the type-II receptor wit, which are known to form a complex upon ligand binding (Fig. S2E). Simultaneous expression of HTT-mRFP and tkv-eGFP showed co-migrating HTT-tkv vesicles (Figure 5a), with 17% of vesicles containing both HTT and tkv (Figure 5b). Further, immunolocalization showed that 20–24% of HTT colocalized with tkv while 42–50% of Rab7 colocalized with tkv vesicles (Figure 5g,i). Reduction of htt (htt-RNAi) or reduction of Rab7 (Rab7^EY10675/+) impaired the motility of tkv-eGFP (Figure 5c), while reduction of Rab7 significantly decreased only the retrograde motility of tkv-eGFP (Figure 5e), and reduction of htt significantly decreased both the anterograde and retrograde motility of tkv-eGFP (Figure 5f). Further in Drosophila larval axons simultaneously expressing HTT-mRFP and GFP-2xFYVE, which probes for PtdIns3Ps, 24% of vesicles showed co-migration (yellow trajectories, Figure 5a). Immunolocalization showed that 17–24% of HTT-2xFYVE or 2xFYVE-HTT vesicles were colocalized, while 43–58% of Rab7-2xFYVE or 2xFYVE-Rab7 vesicles were colocalized (Figure 5h,j, S2B). Moreover, reduction of htt or Rab7 with GFP-2xFYVE significantly impaired 2xFYVE motility (Figure 5d), decreasing both the anterograde and retrograde movement of GFP-2xFYVE (Figure 5e,f). Intriguingly, treatment of larval axons with CQ increased the colocalization of HTT-Rab7 vesicles with tkv or 2xFYVE, suggesting that inhibition of lysosome acidification and/or AP-lysosome fusion (Figure 5g-j) perturbs the motility of the HTT-Rab7-tkv-wit-2xFYVE-containing vesicle complex. Taken together, these observations support the proposal that the retrogradely moving HTT-Rab7-containing LE is likely a long-distance signaling endosome.

Figure 5. The HTT-Rab7 vesicle complex co-transports with long-distance signaling components. (A) Arrows depict location of the cell bodies (retrograde) and synapses (anterograde). Representative images and kymographs from simultaneous dual-color movies from larvae co-expressing HTT-mRFP with either tkv-eGFP or GFP-2xfyve (PtdIns3ps). Co-migrating tracks that contain both red and green trajectories are seen (yellow arrow). X axis=distance (µm), Y axis=time (s). Bar: 5 µm. (B) Quantification of the average co-migratory trajectories reveal the percent co-migration (%) between HTT-tkv and HTT-2xfyve. n = 6. (C-D) Representative images from movies and kymographs from larvae expressing tkv-eGFP or GFP-2xfyve alone, in the context of htt reduction (htt-RNAi), or a genetic reduction of Rab7 (Rab7^EY10675/+). Bar: 10 µm. (E-F) Directional analysis of tkv-eGFP or GFP-2xfyve trajectories alone compared to tkv-eGFP or GFP-2xfyve trajectories with a htt or Rab7 reduction. n = 10. (G-H) Representative images of segmental nerves from larvae co-expressing HTT-mRFP (red) and either (G) tkv-eGFP (green) or (H) GFP-2xfyve (green) that have been immunostained with Rab7 (blue) following treatment with buffer or CQ. Note that the merged panel of buffer-treated larvae reveals heterogenous overlays containing either HTT-Rab7 (purple arrows), Rab7-tkv or Rab7-2xfyve (teal arrows), HTT-tkv or HTT-2xfyve (yellow arrows) or HTT-Rab7-tkv and HTT-Rab7-2xfyve (white arrows), while the merged panel of CQ-treated larvae show more overlays containing HTT-Rab7-tkv and HTT-Rab7-2xfyve (white arrows) Bar: 10 µm. (I-J) Quantification of colocalization (mander’s, %) between (I) HTT-Rab7, HTT-tkv, and tkv-Rab7 or (J) HTT-Rab7, HTT-2xfyve, and 2xfyve-Rab7 in nerves of buffer-, CQ-, and BAF1-treated larvae. Data represented as overlaid histogram-dot plots (left) as well as proportional Venn diagrams (right). n = 6. Statistical significance was determined using the two-sample two-sided Student’s t-test. Data represented as mean±sem. ns=p > 0.01, *p < 0.01, **p < 0.001, ***p < 0.0001, ****p < 0.00001. Also see Fig. S2.

Work shows that Rab7 colocalizes with MAP3K12/DLK, a long-distance signaling component which is activated after axon injury to support axon regeneration and neuron survival [33,36]. Moreover, HTT is implicated in the reprogramming of the regenerative transcriptome following axonal injury in mouse corticospinal neurons [37]. To test the hypothesis that the HTT-Rab7-containing LE moves retrogradely and can be activated by axonal injury, we first evaluated wallenda (wnd) a conserved MAPKKK, homologous to MAP3K12/DLK which functions as an upstream mediator of injury [30]. Larvae expressing either Rab7-GFP or HTT-mRFP in the context of endogenous wnd tagged with mCherry or GFP (wnd-mCherry or wnd-GFP [83,84]) showed wnd retrogradely co-migrating with HTT or Rab7-containing vesicles (Fig. S6E). Intriguingly, similar to what was observed with tkv (Figure 5a,g), treatment with CQ disrupted HTT-wnd or Rab7-wnd vesicle motility, suggesting that wnd is also present with the HTT-Rab7-tkv-wit-containing vesicle.

To directly test the hypothesis that the HTT-Rab7-containing LE moves retrogradely after axonal injury we crushed Drosophila larval nerves expressing GFP using the puc^GAL4 driver ([30,85,86], Fig. S6B). In these larvae, puc (puckered) regulatory sequences control the expression of GAL4, which will in turn express GFP after injury, and increased GFP expression will be seen in injured neuron cell bodies compared to non-injured neurons. Indeed, significant increases in GFP were observed in motor neuron cell bodies 16–24 h after nerve crush compared to 0–12 h (Figure 6a-c, S6B). Next, larvae expressing HTT-mRFP or Rab7-GFP were imaged at three axonal segments (Figure 6d, S6A, far-proximal to crush site-red box, proximal to the crush site-blue box, distal to crush site-green box) 16 h after nerve crush. As a control, larvae expressing Syt1-eGFP were imaged 16 h after nerve crush. Interestingly, while Syt1-eGFP vesicles were stalled/immobile in all three axonal segments imaged (far-proximal to crush site (red box), proximal to crush site (blue box), and distal to crush site (green box)), HTT-mRFP or Rab7-GFP vesicles showed motility at the far-proximal axonal segment to the crush site (red) and at the proximal crush site (blue box), but not at the distal axonal segment to the crush site (green box, Figure 6d). Quantification of the anterograde:retrograde flux ratio revealed a significant retrogradely biased motility for both HTT-mRFP or Rab7-GFP vesicles (Figure 6f). Note that an increase in Rab7 and HTT localization is seen in cell bodies after nerve crush compared to uninjured controls (Fig. S6D). Moreover, Drosophila larval axons simultaneously expressing HTT-mRFP or Rab7-GFP imaged 16 h after nerve crush also showed an increase in the retrogradely co-migrating HTT-Rab7 vesicle population proximal to the crush site (red box, Figure 6g,h), indicating that axonal injury does not perturb the retrograde motility of the HTT-Rab7 LE but rather stimulates its motility.

Figure 6. Axonal injury encourages enhanced retrograde movement of HTT-vesicles: evidence for a HTT-Rab7 signaling late endosome. (A) Schematic diagram of larval nervous system showing the brain. Red box=imaged area. Representative images from simultaneous dual-color images of neuron cell bodies from larvae expressing UAS-eGFP under the puc^GAL4E69 driver that have been immunostained with elav at 0, 12, 16, 18, or 24 h after being subjected to larval nerve crush in the presence of fly food laced with buffer or CQ. The presence of eGFP in motor neuron cell bodies corresponds to increased expression of puckered (puc). Bar: 2 µm. (B) Quantification of the intensity ratio of GFP:elav normalized to the uninjured-buffer condition at motor neuron cell bodies at t₂₄ of uninjured and injured/crushed larvae in buffer- or CQ-laced fly food. n = 5. (C) Representative line plot of the GFP:elav intensity ratio at motor neuron cell bodies of larvae subjected to nerve crush in either buffer- or CQ-laced fly food at 0, 12, 16, 18, or 24 h. n = 5. (D) Schematic diagram of larval nervous system showing the brain and segmental nerves. Red box=imaged area (90 µm). Arrows depict location of the cell bodies (retrograde) and synapses (anterograde). Representative images and kymographs of movies from larvae expressing Syt1-eGFP, Rab7-GFP, or HTT-mRFP that have been either left uninjured or injured/crushed for 16 h prior to in vivo imaging. Injured larvae were imaged at the far-proximal to crush site (red), proximal to crush site (blue), and distal to crush site (green). X axis=distance (µm), Y axis=time (s). Bar: 5 µm. (E) Representative images and kymographs from movies of larvae expressing HTT-mRFP (red) or Rab7-GFP (green) that have been injured for 16 h in CQ-laced fly food prior to in vivo imaging at the far-proximal to crush site (red) or the proximal to crush site (blue). (F) Quantification of the avg. anterograde:retrograde flux ratio per larvae (AU) of HTT-mRFP or Rab7-GFP trajectories. A retrograde bias is depicted as less than zero (left on x-axis), while an anterograde bias is depicted as greater than zero (right on x-axis). n = 5. (G) Representative images and kymographs from simultaneous dual-color movies from larvae expressing HTT-mRFP (red) and Rab7-GFP (green) that have been either left uninjured, injured for 16 h in buffer-laced fly food, or injured for 16 h in CQ-laced fly food prior to in vivo imaging at the far-proximal to crush site (red). Co-migrating tracks that contain HTT and Rab7 are seen (yellow arrow). X-axis=distance (µm), Y-axis=time (s). Bar: 5 µm. (H) Quantification of directional analysis (%) of HTT-Rab7 (yellow) trajectories of uninjured larvae compared to buffer-treated injured larvae and/or CQ-treated injured larvae. n = 5. (I) Representative western blot of an immunoprecipitation of HTT-containing human iNeuron VF, probed with HTT, RAB7, BDNF, and MAPK8/JNK, which all show presence in the HTT VF-IP. No bands are seen in the negative no antibody control (−crtl). n = 3. Statistical significance was determined using the two-sample two-sided Student’s t-test. Data represented as mean±sem. ns=p > 0.01, *p < 0.01, **p < 0.001, ***p < 0.0001. Also see Fig. S5 and S6.

We next postulated that if the retrogradely moving HTT-Rab7 LE carries signaling components to the cell body after axonal injury then CQ treatment should disrupt this motility. Indeed, CQ treatment perturbed the retrograde motility of HTT-mRFP or Rab7-GFP containing vesicles (Figure 6e, S5A), with significant decreases in puc^GAL4-driven GFP expression in motor neuron cell bodies 16–24 h after nerve-crush (Figure 6a-c, Fig. S6C). Quantification analysis of the anterograde:retrograde flux ratio revealed a significant anterogradely biased motility for both HTT-mRFP or Rab7-GFP vesicles with CQ treatment after nerve crush (Figure 6e,f). Note that Rab7-GFP and HTT-mRFP localization to cell bodies was also decreased with CQ treatment after nerve crush (Fig. S6D). Further, larvae co-expressing HTT-mRFP and Rab7-GFP with CQ treatment showed decreased co-migrating retrograde populations with increased stalling and reversing populations (Figure 6g,h) after nerve crush. Additionally, biochemical analysis of HTT VF-IPs from human iNeurons revealed, in addition to RAB7, the presence of long-distance signaling components BDNF (brain-derived neurotrophic factor), which promotes axonal regrowth after sciatic nerve crush [87], and MAPK8/JNK (Figure 6i), which is activated [88,89] and transported retrogradely following injury [29,32] to promote regeneration. Therefore, taken together, the retrogradely moving HTT-RAB7/Rab7-LAMP1-containing LE we isolated is likely a signaling endosome containing wnd/MAP3K12/DLK, BDNF, MAPK8/JNK, and/or the BMP-receptors tkv and wit which are carried to the cell body after axonal injury (Fig. S6F).

Discussion

Despite more than 350 binding partners for HTT identified across a wide-range of cellular processes, including trafficking, the molecular mechanisms by which HTT and its binding partners function remain elusive. Here we provide evidence for a retrogradely moving HTT-Rab7 vesicular complex using Drosophila genetics, in vivo imaging in Drosophila larval axons coupled with a custom particle tracking analysis and pharmacological inhibitors. Specifically, we found that the retrogradely moving HTT-RAB7/Rab7 vesicle is likely a LAMP1/Lamp1-containing LE that can traffic long-distance signaling components such as the BMP-receptors tkv and wit, neurotrophic factor BDNF and axonal damage response components wnd/MAP3K12/DLK and MAPK8/JNK following axonal injury. Taken together, our observations unravel a previously unknown role for HTT in the retrograde movement of a RAB7-LAMP1-containing signaling LE (Fig. S6F), which has functional relevance during axonal injury.

The biological relevance of the putative retrogradely moving HTT-RAB7-LAMP1-containing late endosome

Rab7 is an established LE marker [11]. Several studies also show that Rab7 has a substantial role in the maturation of APs and LEs, by aiding in their retrograde motility and by participating in fusion with lysosomes [11,90–92]. Further, HTT facilitates the retrograde axonal movement of Atg8a/LC3-containing autophagic vesicles (Figure 2 [27]). However, Atg8a/LC3 is used as a marker for autophagy and this is problematic because Atg8a/LC3 is present in multiple autophagy compartments from PGs to ALs [28] as well as during LE-AP fusion in distal axons [25]. Therefore, knowledge of the exact location of where HTT enters the autophagy pathway is ambiguous. To overcome these issues, we probed HTT and Rab7 with markers specific for PGs, APs, amphisomes and lysosomes together with genetic mutants and pharmacological inhibitors to identify the cellular compartment where HTT and Rab7 merge. While HTT co-migrates with retrogradely moving Rab7-LAMP1 vesicles, reduction/loss of LE-AP fusion with mutant Syx17 or by pharmacological inhibition using CQ/BAF1 impaired the retrograde movement of axonal HTT-Rab7-LAMP1-containing vesicles. In contrast, reduction/loss of Atg5-dependent PG assembly to APs had no effect (Figure 3). Despite HTT being implicated as a molecular scaffold during induction and cargo loading of elongating PGs [93] through associations with ULK1 [26], the lack of colocalization between HTT and Atg5 (Fig. S3) within axons and the failure to disrupt HTT motility with loss/reduction of Atg5 (Figure 3) suggests that HTT likely has no role at the PG within axons. Therefore, our observations suggest that the only location that HTT and Rab7 can enter the autophagy pathway is when HTT-Rab7 containing LEs fuse with Atg8a/LC3-containing APs (Fig. S3A). It is also possible that the HTT-Rab7 LE also contains LAMP1/Lamp1 since recent work shows that LAMP1 may not only mark degradative lysosomes [14]. Therefore, it is clear that Atg8a/LC3 and LAMP1/Lamp1 cannot be used to conclusively label APs and lysosomes.

The presence of HTT with Rab7 in the endocytic pathway is further supported by our finding that loss/reduction of Hip1, a membrane bound component of clathrin-mediated endocytosis and endosomal trafficking [50,94], disrupted the motility of both HTT and Rab7 (Figure 1, S1). Upon endocytic internalization, cargo marked to be degraded remains at EEs, which mature to LEs prior to long distance retrograde movement within axons to the cell body for lysosome-mediated degradation. In fact, pharmacological inhibition of lysosome acidification and lysosome-AP-LE fusion fusion impaired the retrograde co-migration of the putative HTT-Rab7-LAMP1 vesicle complex (Figure 4). Similar to work that HTT facilitates kinesin 1 [95] and dynein-mediated axonal motility [96], and that HTT moves bidirectionally by interacting with kinesin 1 and dynein [7], one potential model for HTT is to function as a molecular scaffold to coordinate dynein-mediated retrograde movement of RAB7/Rab7-LAMP1/Lamp1-containing LEs. Furthermore, while the Rab7-interacting protein FYCO1 (FYVE And coiled-coil domain autophagy adaptor 1) could mediate kinesin 1-mediated movement of LEs [97], we observed that loss/reduction of Rilpl disrupted the retrograde movement of axonal HTT-Rab7 vesicles (Figure 1, S1). Indeed, RAB7-interacting/accessory proteins RILP [46,47], ORP1 L [98,99], and PLEKHM1 [100,101] were shown to function during dynein-mediated movement of LEs. Therefore, we propose that a retrogradely moving axonal HTT-RAB7/Rab7-LAMP1 LE exists in vivo and is likely involved in autophagy through LE-AP fusion (Fig. S3A).

A novel role for HTT during axonal injury

RAB7 is implicated in the retrograde trafficking of neurotrophin signaling receptors NGF (nerve growth factor)-activated TRKA [102,103], and BDNF-activated TRKB [12] within a LE, while a RAB7 mutant causing Charcot-Marie-Tooth disease showed dysregulated TRKA-NGF retrograde signaling resulting in axonal degeneration [15,104]. Work has shown that a majority of the retrograde TRKA signaling endosomes are ultra-structurally and molecularly defined multivesicular bodies [16]. HTT is also implicated in the retrograde trafficking of TRKB-BDNF signals [105,106], with defective TRKB signaling seen in HD [107,108]. In fact, we also found BDNF with HTT-associated vesicles isolated from WT human iNeurons (Figure 6i), which was decreased in HD iNeurons (data not shown). Additionally, the BMP-receptors tkv and wit co-migrated with HTT and Rab7 (Figure 5, S2), with pathogenic HTT disrupting tkv-containing vesicle motility [109] causing synaptic growth defects [76]. Therefore, the putative retrogradely moving HTT-RAB7-LAMP1 containing endosome we isolated here likely traffics long distance signaling components such as NGF and BDNF (Fig. S6F).

Further, a role for Rab7 in mediating the retrograde transport of injury signals has also been proposed. In response to axonal injury, transcriptional reprogramming is thought to occur via several protective signaling pathways, such as the MAPK8/JNK pathway [110,111] and DLK [30,33,34,112,113]. DLK was shown to move retrogradely with Rab7-containing LEs following palmitoylation of DLK [36]. We also found the axonal damage response signal wnd co-migrating with HTT and Rab7 (Fig. S6E). Additionally, work has suggested that retrograde signaling following axonal injury is mediated by vesicle transport complexes, such as the syd/JIP3-containing endosome that carries phosphorylated- MAPK8/JNK to neuronal cell bodies [29,35] and mass spectrometry of isolated syd/JIP3-associated vesicles showed the presence of Rab7/RAB7 [35]. Macrophage-derived NADPH oxidase 2 complexes were shown to incorporate into RAB7-positive endosomes after injury in dorsal root ganglion neurons, influencing PtdIns3K signaling to stimulate neurite outgrowth and axon regeneration after sciatic nerve lesion [114]. Moreover, the RAB7 mutant associated with Charcot-Marie-Tooth showed reduced protein synthesis in the axon with compromised axon integrity after initial axon outgrowth [115]. A non-canonical BMP-signaling pathway has been suggested to mediate retrograde injury signaling from spinal cord injury sites [116]. Additionally, HTT was recently proposed to influence the reprogramming transcriptome after spinal cord injury with deletion of HTT attenuating regeneration of mouse corticospinal axons [37]. Further, loss/reduction of HTT displayed an altered proteomic profiling network in axotomized retinal ganglion cells, suggesting a role for HTT as an injury-response signaling hub [38]. Taken together, the HTT-RAB7/Rab7-containing signaling endosome we identified likely has an important role during axonal injury (Fig. S6F). Indeed, we found that the retrograde motility of the HTT-RAB7/Rab7-LAMP1 LE was not perturbed but stimulated in Drosophila larval nerve crush experiments, which activates retrograde injury signaling reported by puc, a MAPK8/JNK-phosphatase, compared to uninjured larvae (Figure 6, S6). Therefore, our observations identify a previously unknown role for HTT during axonal injury: to facilitate the movement of retrograde signals, such as BMP-receptors (tkv and wit), MAPK8/JNK, BDNF and the canonical damage signal wnd/DLK within a RAB7/Rab7-LAMP1/Lamp1 signaling LE to the cell body for transcriptional reprogramming of neurons. Future work will investigate a more direct mechanism by which HTT helps promote axonal regeneration following injury.

Materials and methods

Drosophila genetics

Neuronal drivers Appl-GAL4 (pan neuronal) and pGAL4-62B-SG26-1 (8 motor neurons) were used for neuronal expression of transgenic lines [39,40,117]. For dual-color imaging, males from UAS-Rab7-GFP, UASp-YFP.rab7^T22N, UAS-Syt1-eGFP, UAS-LAMP1-GFP, UAS-GFP-myc-2xFYVE, UAS-tkv-eGFP, MiMIC-wnd-mCherry, MiMIC-wnd-GFP, or UAS-mCherry-Atg8a (BDSC) were crossed to Appl-GAL4;T(2:3),CyO,TM6B,Tb/Pin^88K virgin females. The chromosome carrying T(2:3),CyO,TM6B,Tb is referred to as B3 and carries the dominant markers, Hu, Tb and CyO. The larval Tb (Tubby) marker is used to select larvae of interest. Appl-GAL4/Y;UAS-/Cyo,TM6B males were then crossed with UAS-HTT15Q-mRFP (gift from Troy Littleton [118]), UAS-HTTex1-25Q-eGFP (gift from Norbert Perrimon [119]), or UAS-Rab7-GFP (BDSC) virgins. Additionally, UAS-Rab7-GFP males were crossed to pGAL4-62B SG26–1;T(2:3),CyO,TM6B,Tb/Pin^88K virgin females and progeny that were pGAL4-62B-SG26-1/Y,CyO,TM6B,Tb/UAS-Rab7-GFP males were then crossed to UAS-HTT15Q-mRFP virgin females. For genetic interaction analysis, males from milt^k04704/CyO [52,53], Rilpl^EY06476 [120], Hip1^MB04365 [121], Rip11^KG02485/FM7c [122], nmo^P1/TM6 [44,123], Khc⁸/cyO (null allele [124]), Dhc^6-10/TM6B (null allele [125]), Atg5^d04577 (BDSC), Atg5^5cc5/FM7 (null allele [59], BDSC), Syx17^f01971/TM6B (BDSC [60,61]), Syx17^LL/TM6 (null allele [24], a gift from Gábor Juhász) Rab7^EY10675 (BDSC [42]), Rab7^d1/TM6 (null allele [42], a gift from Gábor Juhász), Df(98E2);CG9990 (htt-KO/+, [41], gift from Sheng Zhang), or UAS-htt-rnai [7] were crossed to Appl-GAL4;T(2:3),CyO,TM6B,Tb/Pin88K virgin females. Males were then crossed to UAS-HTT15Q-mRFP or UAS-Rab7-GFP. Note that Appl-GAL4 (× chromosome) and UAS-HTT15Q-mRFP (3^rd chromosome) are on the same chromosomes as Atg5 (× chromosome), Syx17, or Rab7 (both on the 3^rd chromosome) mutants. Western blot analysis showed that rab7 was reduced to 28% of WT levels in homozygous Rab7^EY10675/Rab7^EY10675 and Atg5 was reduced to 64% of WT levels in heterozygous Atg5^d04577/+ (Fig. S2A). Additionally, significant increases in ref(2)P protein levels were seen in homozygous Rab7^EY10675/Rab7^EY10675, heterozygous Atg5^d04577/+, heterozygous Syx17^f01971/+, and the HTT lines htt-rnai and htt-KO/+ (Fig. S2D), similar to what was previously seen [24,26,42,59]. Rab7^EY10675 showed an enhanced ectopic midline crossing phenotype [126]. Syx17^f01971 showed an Atg8-rnai phenotype of small perinuclear dots [24]. Similar to Atg5^5cc5 [59], Atg5^d04577 flies showed an ataxic phenotype (Fig. S2C). Progeny that were Appl-GAL4/Y;Cyo,TM6B males were then crossed to UAS-Rab7-GFP, UAS-LAMP1-GFP, UAS-GFP-myc-2xFYVE, UAS-tkv-eGFP, or UAS-mCherry-Atg8a, or UAS-HTT15Q-mRFP virgin females. Conversely, Rip11^KG02485/FM7c or UAS-htt-RNAi virgins were crossed to Appl-GAL4/Y;UAS-/Cyo,TM6b males. X/Y; +/+; puc^GAL4E69/TM6B (BDSC) males were crossed with UAS-eGFP (BDSC) virgin females and female progeny were selected for larval nerve crush confirmation assays. UAS-Rab7-GFP, UAS-Syt1-eGFP, UAS-LAMP1-GFP, UAS-GFP-myc-2xFYVE, UAS-HTT15Q -mRFP, or UAS-eGFP males were crossed to Appl-GAL4 or pGAL4-62B-SG26-1 virgin females. In all cases non-tubby female 3^rd instar larvae were dissected for in vivo imaging or immunohistochemistry. Sibling tubby larvae were evaluated as controls. Reciprocal crossings were also done to confirm observations. A comprehensive list of Drosophila lines used for this study is organized in Table 1.

Table 1. Key resource table.

Resource	Source	Identifier
Antibodies and Dyes
Mouse anti-Rab7	Developmental Studies Hybridoma Bank	Rab7 RRID: AB_2722471
Rabbit anti-Lamp1	Abcam	ab30687 RRID: AB_775973
Mouse anti-RAB7	Santa Cruz Biotechnology	sc -13,156 RRID: AB_627385
Rat anti-Syx17 (Syntaxin 17)	Laboratory of Gábor Juhász	[24]
Rabbit anti-ref(2)P	Laboratory of Gábor Juhász	[127]
Rabbit anti-ATG5 (N-terminal)	Sigma-Aldrich	A0856 RRID: AB_1078238
Rabbit anti-SYT1	Phosphosolutions	1975-STG RRID: AB_2492251
Mouse anti-SYP (SY38)	ThermoFisher Scientific	MAB5258 RRID: AB_2313839
Mouse anti-DYNC1I1 (74.1)	Abcam	ab23905 RRID: AB_2096669
Mouse anti-HIP1 (4B10)	Novus biological	NB300-203 RRID: AB_10000880
Mouse anti-KIF5C	Laboratory of Lawrence Goldstein	[128]
Mouse anti-elav	Developmental Studies Hybridoma Bank	Elav-9F8A9 RRID: AB_528217
Mouse anti-wit (23C7)	Developmental Studies Hybridoma Bank	23C7 anti-wit RRID: AB_528513
Mouse anti-BDNF #1	Developmental Studies Hybridoma Bank	BDNF-#1 RRID: AB_2617198
Mouse anti-BDNF #9	Developmental Studies Hybridoma Bank	BDNF-#9 RRID: AB_2617199
Rabbit anti-MAPK8/JNK	Cell Signaling Technology	9252 RRID: AB_2250373
Rabbit anti-phospho-MAPK8/JNK	Cell Signaling Technology	9251 RRID: AB_331659
Rabbit anti-ACTA1/Actin	ThermoFisher Scientific	MA5 -32,479 RRID: AB_2809756
Mouse anti- TUBA4A (tubulin, alpha 4A)	Abcam	ab7291 RRID: AB_2241126
Mouse anti-TUBB3 (tubulin, beta 3 class III)	BioLegend	801,213 RRID: AB_2728521
Mouse anti-HTT (1HU-4C8)	EMD Millipore	MAB2166 RRID: AB_2123255
Rabbit anti-HTT (EP867Y)	Abcam	ab45169 RRID: AB_733062
Mouse anti-RTN1 (MON160)	Santa Cruz Biotechnology	sc -23,880 RRID: AB_672542
Mouse anti-Cyt-c-p (A-8)	Santa Cruz Biotechnology	sc -13,156 RRID: AB_627385
Rabbit anti-DCTN1 (dynactin 1)	Abcam	ab96004 RRID: AB_10677601
Anti-Mouse Alexa Fluor® 488	Invitrogen	A11001 RRID: AB_2534069
Anti-Mouse Alexa Fluor® 568	Invitrogen	A11004 RRID: AB_2534072
Anti-Mouse Alexa Fluor® 647	Invitrogen	A21235 RRID: AB_2535804
Anti-Rabbit Alexa Fluor® 488	Invitrogen	A11008 RRID: AB_143165
Anti-Rabbit Alexa Fluor® 568	Invitrogen	A11011 RRID: AB_143157
Anti-Rabbit Alexa Fluor® 647	Invitrogen	A21244 RRID: AB_141663
Anti-Mouse secondary antibody, HRP	Invitrogen	32,430 RRID: AB_1185566
Anti-Rabbit secondary antibody, HRP	Invitrogen	32,460 RRID: AB_1185567
Alexa Fluor® 594 Goat Anti-Horseradish Peroxidase	Jackson Immuno Research Labs	123-585-021 RRID: AB_2338966
Fluorescein (FITC) Goat Anti-Horseradish Peroxidase	Jackson Immuno Research Labs	123-095-021 RRID: AB_2314647
Biological Samples
Mouse brain tissue (CBA/CaJ background)	Laboratory of Matthew Xu-Friedman	N/A
Experimental Models: Human Cell Lines
GM23279, polyQ=25, 26y, female	NIGMS Repository (Coriell Institute for Medical Research -Camden, NJ)	GM23279*C
ND38555, polyQ=17, 48y, female	NINDS Repository (Coriell Institute for Medical Research -Camden, NJ)	ND3855 RRID: CVCL_Y822
Chemicals, Peptides, and Recombinant Proteins
Chloroquine (CQ)	Cayman Chemical	14,194 PubChem: 24,278,090
Bafilomycin-A1 (BAF1)	Cayman Chemical	11,038 PubChem: 6,436,223
Rapamycin (RM)	Cayman Chemical	13,346 PubChem: 5,284,616
Paclitaxel	Millipore Sigma	58-055-65MG PubChem: 36,314
Corning Matrigel	ThermoFisher Scientific	CB40230A RRID: N/A
Advanced DMEM/F12	Invitrogen	12,634,028 RRID: N/A
Essential 8 media	Invitrogen	A1517001 RRID: N/A
Neurobasal media	Invitrogen	21,103,049 RRID: N/A
PSC neural induction media	Invitrogen	A1647801 RRID: N/A
B27 supplement media	Invitrogen	17,504-044 RRID: N/A
Protease inhibitor cocktail	Pierce	PIA32965 RRID: N/A
Phosphatase inhibitor	Pierce	PI88667 RRID: N/A
Protein A/G Magnetic Beads	Pierce	PI88802 RRID: N/A
Vecta Shield Mounting Medium	Vector Laboratories	NC9265087 RRID: N/A
Experimental Models: D. melanogaster Organisms/Strains
P{Appl-GAL4.G1a}1, y^1 w^*	Bloomington Drosophila Stock Center	BDSC: 32,040; FlyBase: FBst0032040
Appl-Gal4; T(2,3), CyO, TM6B,Ttb^1/Pin^88k	Laboratory of Lawrence Goldstein	[117]
pGal4-62B SG26–1; T(2,3), CyO, TM6B, Tb^1/Pin^88k	Laboratory of Lawrence Goldstein	[40]
w^*; P{UAS-Rab7-GFP}3	Bloomington Drosophila Stock Center	BDSC: 42,706; FlyBase: FBst0042706
Df(98E2); CG9990/TM3 (htt-KO/+)	Laboratory of Norbert Perrimon	[41]
pUAST-HTT15Q-mRFP (UAS-HTT-mRFP)	Laboratory of Troy Littleton	[118]
y^1 w^67c23; P{EPgy2}Rab7^EY10675	Bloomington Drosophila Stock Center	BDSC: 20,630; FlyBase: FBst0020630
UAS-htt-rnai (htt = Drosophila htt)	Laboratory of Lawrence Goldstein	[7]
y^1 P{EPgy2}Rilpl^EY06476 w^67c23	Bloomington Drosophila Stock Center	BDSC: 16,726; FlyBase: FBst0016726
y^1 w^67c23; P{lacW}milt^k04704 /cyO	Bloomington Drosophila Stock Center	BDSC: 10,553; FlyBase: FBst0010553
w^1118; Mi{ET1Hip1^MB04365	Bloomington Drosophila Stock Center	BDSC: 24,809; FlyBase: FBst0024809
y^1 w^*; P{lacW}nmo^P1/TM3, Sb^1	Bloomington Drosophila Stock Center	BDSC: 27,897; FlyBase: FBst0027897
w^1118; Mi{ET1}nuf^MB09772	Bloomington Drosophila Stock Center	BDSC: 27,803; FlyBase: FBst0027803
y^1 P{SUPor-P}Rip11^KG02485 /FM7c, sn^+	Bloomington Drosophila Stock Center	BDSC: 13,742; FlyBase: FBst0013742
b^1 pr^1 Khc^8 /CyO	Bloomington Drosophila Stock Center	BDSC: 1607; FlyBase: FBst0001607
mwh^1 Dhc64C^6-10 h^1 st^1 p^p e^s /TM6B,Tb^1	Bloomington Drosophila Stock Center	BDSC: 8747; FlyBase: FBst0008747
w^*; P{UAS-Syt1.eGFP}3	Bloomington Drosophila Stock Center	BDSC: 6926; FlyBase: FBst0006926
w^*; P{UAS-GFP-LAMP}2; P{nSyb-GAL4.S}3/T(2;3)TSTL, CyO: TM6B, Tb^1	Bloomington Drosophila Stock Center	BDSC: 42,714; FlyBase: FBst0042714
w^*; P{UAS-GFP-myc-2xFYVE}2	Bloomington Drosophila Stock Center	BDSC: 42,712; FlyBase: FBst0042712
w^1118; PBac{WH}Syx17^f01971 /TM6B, Tb^1	Bloomington Drosophila Stock Center	BDSC: 18,495; FlyBase: FBst0018495
w^1118 P{XP}Atg5^d04577 CG1677^d04577 lncRNA:CR44357^d04577; l(3)*^* /TM6B, Tb^1	Bloomington Drosophila Stock Center	BDSC: 19,206; FlyBase: FBst0019206
w*; P{UAS-tkv-eGFP.D}3	Bloomington Drosophila Stock Center	BDSC: 51,653; FlyBase: FBst0051653
y^1 w^1118; P{UASp-mCherry-Atg8a}2; dr^1 /TM3, Ser^1	Bloomington Drosophila Stock Center	BDSC: 37,750; FlyBase: FBst0037750
pUAST-HTTex1-25Q-eGFP (UAS-HTT-eGFP)	Laboratory of Norbert Perrimon	[119]
P{UAS-eGFP}8, w1118	Bloomington Drosophila Stock Center	BDSC: 5428; FlyBase: FBst0005428
w*; P{GAL4E69}puc^GAL4E69/TM3, Sb^1 Ser^1	Bloomington Drosophila Stock Center	BDSC: 6762; FlyBase: FBst0006762
y1 w*; Mi{PT-mCh.2}wnd^MI00494-mCh.2/TM6B, Tb^1	Bloomington Drosophila Stock Center	BDSC: 39,657; FlyBase: FBst0039657
y1 w*; Mi{PT-GFSTF.2}wnd^MI00494-GFSTF.2/TM6B, Tb^1	Bloomington Drosophila Stock Center	BDSC: 39,656; FlyBase: FBst0039656
Atg5^5cc5/FM7i, P{ActGFP}JMR3	Bloomington Drosophila Stock Center	BDSC: 91,419 FlyBase: FBst0091419 [59]
Syx17^LL/TM6	Laboratory of Gábor Juhász	[24]
Rab7^d1/TM6	Laboratory of Gábor Juhász	[42]
y1 w*; P{UASp-YFP.Rab7^T22N}06	Bloomington Drosophila Stock Center	BDSC: 9778 FlyBase: FBst0009778
Software/Algorithms
MATLAB-based particle tracking program	Laboratory of Danuser	Yang et. al., 2005
ImageJ	Schneider et. al., 2012 https://imagej.net/	RRID: SCR_003070
Metamorph/Metavue Imaging Software	Molecular Devices, Sunnyvale, CA, USA	RRID: SCR_002368
Minitab18	https://www.minitab.com/en-us/	RRID: SCR_014483
Microsoft Excel	https://www.microsoft.com/en-gb/	RRID: SCR_016137
RStudio	http://www.rstudio.com/	RRID:SCR_000432
OriginLab/OriginPro	https://www.originlab.com/	RRID: SCR_014212
BioRender	http://biorender.com	RRID:SCR_018361

Human iPSC cultures

The following cell lines were obtained from the NIGMS Human Genetic Cell Repository at the Coriell Institute for Medical Research: iPSCs from WT (polyQ = 25, 26y, female) patients were purchased from the NIGMS Human Genetic Cell Repository at the Coriell Institute for Medical Research (GM23279*C) and iPSCs from WT (polyQ = 17, 48y, female) patients were purchased from the NINDS Human Cell Repository (ND38555).

Mouse brains

CBA/CaJ background mouse brains (gift from Matthew Xu-Friedman, The State University of New York at Buffalo) were dissected, halved, then stored on dry ice and used immediately or stored at −80°C for future use.

In vivo imaging and analysis of vesicle motility within whole-mount larval axons

Larvae were dissected and immediately imaged under physiological conditions as previously detailed [9]. Non-tubby, female larvae were dissected and imaged under physiological conditions in dissection buffer. Motility was visualized in the red (568 nm), green (488 nm), or both channels (568/488 nm using a dual-view beam splitter attachment) within larval segmental nerves using a Nikon TE-2000 microscope (Nikon, Melville, NY, USA). From each larva, four sets of movies at an imaging window frame size of 90 μm at 150 frames were taken from the mid-region of the larva at an exposure of 500 ms using a Cool Snap HQ cooled CCD camera (Photometrics, Tucson, AZ, USA) and the Metamorph imaging system (Molecular Devices, Sunnyvale, CA, USA). Kymographs were generated in Metamorph using the kymograph stack tool. From a total of 10 larvae a set of 40 movies were imaged for each genotype at a spatial resolution of 0.126 μm/pixel. The four movies, each lasting a 1.25 min span a total time of 5 min. Because most of the vesicles take <1 min to move they will have moved out of the 90 μm imaging window by the end of the first movie since each time frame for each movie lasts 1.25 min. Movies were analyzed using a MATLAB-based particle tracker program as previously described [129]. Briefly, vesicle trajectories were analyzed to obtain the overall distribution of cargo populations (directional analysis) and individual vesicle movement behaviors (velocities, pause frequencies/durations, run lengths). Duration-weighted segmental velocity evaluates the average velocity behavior that vesicles exhibit per time spent moving. Individual vesicles were automatically categorized as either anterograde, retrograde, reversing, or stationary. Reversing refers to a vesicle that has at least one switch event between anterograde and retrograde motility. Both anterograde or retrogradely moving mitochondria show a net movement in the respective direction without pausing or reversing.

Simultaneous dual-view in vivo imaging

A narrow single-band GFP/DsRED (488/568 nm) filter cassette was added to the Cool Snap HQ camera. Movies were taken in dual-view mode using the split view software in Metamorph at 150 frames from the mid-region of the larva at an exposure of 500 ms to simultaneously image RFP and GFP tagged vesicles. The Cool Snap HQ camera dual-view mode was aligned using Metamorph software (Split-View settings) before each imaging session. Movies were split by wavelength and each kymograph for each split movie was created, merged and analyzed for co-migration. Trajectories of vesicles with colocalized tracks were identified from kymographs using Metamorph software. For each fluorescence channel, a kymograph was generated using Metamorph as previously done [10]. Briefly, after selecting the first channel, all frames within the time-lapse image sequence of this channel were added together to produce a summed image. Then a polyline was generated on an axon. After generation of the polyline, the kymograph was created using Metamorph software. The polyline was then copied from one fluorescent channel to the other and used to create a kymograph for the other fluorescent channel. To identify vesicles with colocalized signals from both channels, the kymographs were colored in red and green, respectively, and combined into a single RGB kymograph. Non-stationary vesicles with colocalized signals were identified and counted by their yellow color in the combined kymograph. Note that to differentiate meaningful colocalization we evaluated the colocalization of the entire trajectory of a moving particle during the entire time frame of the movie. Therefore, only particles containing the same trajectory in both the red and green channels would show colocalization in yellow when merged and spurious colocalization observed in one-time frame would be avoided. The total number of colocalized full trajectories for 20 kymographs across five larvae was counted for each genotype. Pearson’s and Mander’s coefficients were obtained using Coloc2 in ImageJ. Briefly, kymographs from Drosophila larval segmental nerves expressing two fluorophores were separated into red and green channels using Metamorph and analyzed using Coloc2 in ImageJ.

Immunohistochemistry of larval axons for quantification of accumulations or colocalization

3^rd instar larvae were dissected and fixed in 8% paraformaldehyde, washed with PBT (phosphate buffered saline [PBS; 0.15 M NaCl and 10 mM Na₂HPO₄, pH 7.5] supplemented with 0.1% Tween-20 [Pierce, 28,352]) and incubated overnight with antibodies against Atg5 (1:100; Sigma-Aldrich, A0856), Syx17 (1:100; gift from Gábor Juhász, Eötvös Loránd University & Eötvös Loránd Research Network), Lamp1 (1:100; Abcam, ab30687), Syt1 (1:100; Phosphosolutions, 1975-STG), elav (1:100; DSHB [Developmental Studies Hybridoma Bank], elav-9F8A9), p-MAPK8/JNK (1:500; Cell Signaling Technology,), wit (1:100; DSHB, 23C7 anti-wit), and/or Rab7 (1:100; DHSB, Rab7). Larvae were incubated in secondary antibodies (1:100; Alexa Fluor 568 anti-mouse [Invitrogen, A11004], Alexa Fluor 488 anti-mouse [Invitrogen, A11001], Alexa Fluor 647 anti-mouse [Invitrogen, A21235], Alexa Fluor 568 anti-rabbit [Invitrogen, A11011], Alexa Fluor anti-rabbit 488 [Invitrogen, A11008], and/or Alexa Fluor 647 anti-rabbit [Invitrogen, A21244]), and mounted using Vectashield mounting medium (Vector Laboratories, NC9265087). Images of segmental nerves were collected using a Nikon Eclipse TE 2000 U microscope at 40X using the 40X objective or 90X using the 60X objective with 1.5X gain (Nikon, Melville, NY, USA) alongside the FITC (488 nm), TxRED (568 nm), both FITC/TxRED (568/488 nm using a dual-view beam splitter attachment) and/or Cy5 (647 nm) filters. Quantitative analysis on axonal blockages was carried out by collecting six confocal optical images from middle segmental larval nerves. For each genotype, at least six confocal optical images across five larvae were imaged and the number of axonal blockages was measured using NIH ImageJ software as previously done [7,117]. Quantitative analysis on colocalization of compartments in larval nerves reported with simultaneous dual-view imaging was carried out by collecting six confocal optical images from middle segmental larval nerves. For each genotype, at least six confocal optical images across five larvae were imaged. The total number of colocalized particles in larval nerves from six confocal images across five larvae was counted for each genotype. Mander’s and Pearson’s coefficients were obtained using Coloc2 in image J. Fixed imaged of larval nerves taken at 90X using the FITC, TxRED, and Cy5 filters were merged into a single RGB image to analyze colocalization noted as triple colocalization (white puncta) or double colocalization (yellow, magenta, or cyan puncta)/

Larval feeding

For chemical feeding experiments, 3^rd instar larvae expressing UAS-HTT15Q-mRFP, UAS-rab7-GFP, UAS-LAMP1-GFP, UAS-Syt1-eGFP, UAS-GFP-myc-2xFYVE, or UAS-eGFP were grown with fly flood (24 g agar, 22 g yeast, 164 g cornmeal, 12 g soy flour, and 200 mL molasses, and 5 L H₂O) containing buffer (0.01% DMSO,), 3 mg/mL chloroquine (Cayman Chemical, 14,194) dissolved in buffer/0.01% DMSO, 0.3 mg/mL bafilomycin A₁ (Cayman Chemical, 11,038) dissolved in buffer (0.01% DMSO), 1.0 µg/mL rapamycin (Cayman Chemical, 13,346) dissolved in buffer/0.01% DMSO, or 10 µm paclitaxel (Millipore Sigma, 58-055-65 MG [130]) dissolved in buffer/0.01% DMSO for 12, 16, 18, or 24 h. Larvae were dissected and imaged for in vivo motility, described above. Alternatively, larvae were dissected for immunohistochemistry, described above. For dual-view imaging, larvae simultaneously expressing UAS-HTT15Q-mRFP with either UAS-Rab7-GFP, UAS-LAMP1-GFP, UAS-tkv-eGFP, or UAS-GFP-myc-2xFYVE were subjected to either fly food feeding condition and dissected and imaged for dual-view imaging, described above. Alternatively, larvae were dissected for immunohistochemistry, described above.

Larval nerve crush

Larval nerve crush was performed as done [30]. Larvae were anesthetized with CO₂ gas for 1 min prior to pinching larval segmental nerves through the ventral cuticle for 5 s with Dumont #5 Inox forceps (Fine Science Tools, 11,252–20). Successful injury events were recorded by posterior paralysis of larvae with actively moving mouthpieces. Observational recording of successful injury was confirmed by Phase-contrast imaging and HRP-staining (HRP-FITC [Jackson Immuno Research Labs, 123-095-021] or HRP-TxRED [Jackson Immuno Research Labs, 123-585-021] at 1:50) of nerve crush site of dissected larvae as well as through reporting of increased GFP at motor neuron cell bodies in larvae expressing puc^GAL4E69; UAS-eGFP, which was validated by p-MAPK8/JNK (Cell Signaling Technology, 9251) immunostaining of neuron cell bodies (data not shown). For incubations, larvae were either immediately dissected for t₀ results, or transferred to fresh fly-food plates containing either Buffer/0.01% DMSO, 10 µm paclitaxel dissolved in Buffer/0.01% DMSO, or 3 mg/mL CQ dissolved in Buffer/0.01% DMSO. These larvae were further incubated for 12, 16, 18, or 24 h at 29ºC prior to dissection for immunohistochemistry or in vivo imaging as described above.

Drosophila geotaxis assay

Negative geotaxis assays were performed as previously done [59,131,132]. WT and Atg5^d04577 flies were reared for 4 weeks at 25°C and 60% humidity. 20 healthy flies (10 males and 10 females) were then selected and transferred into a negative geotaxis cylinder. Flies were at rest for 5 min after transfer and then tapped to induce negative geotaxis. Fly movements were captured by images taken 3 s post-induction. Fly climbing speed and distance traveled after negative geotaxis induction were calculated using ImageJ and analyzed.

Human iPSC neuronal differentiation

Human iPSCs were grown and expanded on corning matrigel (Thermofisher Scientific, CB40230A) using E8 iPSC media (Invitrogen, A1517001) as previously done [10]. Pluripotency was analyzed using an antibody against OCT-3/4 (1:200; SCBT [Santa Cruz Biotechnology], sc-5279) and Hoechst (Thermofisher Scientific, PI62249) was used as a nuclear staining as detailed below. After 4 passages iPSCs were differentiated into NPCs (neuronal precursors) using PSC neural induction media (Invitrogen, A1647801) and published protocols (publication #MAN0008031). NPCs were identified using an antibody against Nestin (1:200; SCBT, sc -23,927) and then differentiated into mature iNeurons using neurobasal media supplemented with 1× B27 (Invitrogen, 17,504–044) and 2 mM glutamine (ThermoFisher Scientific, 25,030,081). Differentiated neurons, identified using antibodies against MAP2 (1:100, BD Biosciences, BDB556326), TUBB3 (1:100; Biolegend, 801,201) and SYP (1:100, EMD Millipore, MAB5258) exhibited an extensive neurite network after 21 days at which time they were used for biochemical experiments as previously done [10].

Preparation of protein extracts from fly brains, mouse brains, and human iNeurons

Mouse or fly brains were homogenized in homogenization buffer (10 mM HEPES, pH 7.4, 100 mM K acetate, 150 mM sucrose, 5 mM EGTA, 3 mM Mg acetate, 1 mM DTT) containing a cocktail of protease inhibitors (Pierce, PIA32965) and phosphatase inhibitors (Pierce, PI88667) with 5 mM EDTA. Neuronal extracts from Fly brains were centrifuged at 1000 g for 15 min and the supernatant (PNS) was used as detailed below for western blotting. Mouse brain extracts were then centrifuged at 1000 g for 15 min at 4°C. The PNS was then used for sucrose gradient fractionation analysis as detailed below or for western blotting. iNeurons were manually removed from 6-well or 12-well plates using ice-cold homogenization buffer (10 mM HEPES, pH 7.4, 100 mM K acetate, 150 mM sucrose, 5 mM EGTA, 3 mM Mg acetate, 1 mM DTT) containing a cocktail of protease inhibitors (Pierce, PIA32965) and 5 mM EDTA, blended for 30 s on ice using a motorized pestle, and then quickly snap-frozen in liquid nitrogen. Neuronal extracts were then centrifuged at 1000 g for 15 min and the supernatant (PNS) was used for sucrose gradient fractionation analysis as detailed below or for western blotting.

Sucrose gradient fractionations

PNS samples human iNeuron extracts or mouse brains were further fractionated into soluble fractions (SF), heavy membrane pellet (P1), and vesicle fractions (VF) by sucrose gradient ultra-centrifugations using lysis buffer (4 mM HEPES, 320 mM sucrose, pH 7.4) containing a phosphatase inhibitor (Pierce, PI88667) and protease inhibitor cocktail (Pierce, PIA32965) as previously done [10,133,134]. Briefly, 300 ul of PNS was combined with 300 ul 62% sucrose and layered onto a sucrose gradient (35% and 8% sucrose) and centrifuged at 50,000 g for 90 min. The VF (35/8 layer), SF, and P1 were removed and used in western blot analysis. 100 ul of lysis buffer was used to dissolve P1.

Co-immunoprecipitation analysis

1000 ug of total protein from the mouse VF or the iNeuron VF were incubated with a mouse monoclonal antibody to RAB7 (1:1000, SCBT, sc -13,156) or a mouse monoclonal antibody to HTT (1:500, EMD Millipore, MAB2166) overnight at 4°C. Protein homogenates were then incubated with protein A/G magnetic beads (Pierce, PI88802) with rotation for 90 min and eluted with elution buffer (0.1 M glycine, pH 2.0). Eluents were then analyzed by western blot analysis as detailed below. Alternatively, Co-IP was performed on protein homogenates using the crosslinking co-immunoprecipitation kit (Pierce, 26,147) with 0.25 mM DSS (Pierce, 21,655) for 30 min before incubating antibody-bound magnetic beads with VFs.

SDS-PAGE and western blot analysis

Fly brains, mouse brains, or human iNeuron fractions were separated by SDS-PAGE. Samples were denatured in NuPage LDS sample buffer (Invitrogen, NP008) with 4 mM β-mercaptoethanol (Sigma-Aldrich, M6250) and run on 4–12% Bis-Tris gels (Invitrogen, NP0322BOX) which were then transferred to nitrocellulose membranes (Cytiva, 10,119–996). Blots were blocked using TBST (tris buffered saline [TBS; 150 mM NaCl and 20 mM Tris-base, pH = 7.5] supplemented with 0.1% Tween-20 [Pierce, 28,352]) with 5% BSA (Lampire, 50-413-345) for 60 min at 25°C and incubated with primary antibodies: SYT1 (1:1000; Phosphosolutions, 1975-STG), ref(2)P (1:500; gift from Gábor Juhász, Eötvös Loránd University & Eötvös Loránd Research Network), RAB7 (1:1000; SCBT, sc -13,156), Rab7 (1:100; DSHB, Rab7), Syx17 (1:500; gift from Gábor Juhász, Eötvös Loránd University & Eötvös Loránd Research Network), ATG5 (1:1000; Sigma-Aldrich, A0856), Lamp1 (1:1000; Abcam, ab30687), RTN1 (1:1000; SCBT, sc -23,880), SYP (1:1000, Thermofisher Scientific, MAB5258), Cyt-c-p (1:1000; SCBT, sc -13,156), HIP1 (1:1000, Novus biological, NB300–203), DCTN1 (1:500; Abcam, ab96004), KIF5C (1:500; gift from Lawrence Goldstein, University of California, San Diego), DYNC1I1 (1:1000, Abcam, ab23905), ACTA1/Actin (1:1000; ThermoFisher Scientific, MA5– 32,479), TUBA4A (1:2000, Abcam, ab7291), MAPK8/JNK (1:1000, Cell Signaling Technology, 9252), BDNF #1 (1:250; DSHB, BDNF-#1), BDNF #9 (1:250; DSHB, BDNF-#9), HTT rabbit polyclonal (1:1000, Abcam, ab45169), and HTT mouse monoclonal (1:1000; EMD Millipore, MAB2166) for 16 h at 4°C. Blots were then incubated with appropriate secondary HRP-conjugated anti-mouse (1:1000; ThermoFisher Scientific, 32,430) or anti-rabbit (1:1000; ThermoFisher Scientific, 32,460) secondary antibodies and imaged using a ChemiDoc MP imaging system (Bio-Rad Laboratories, Hercules, CA, USA) with ECL (Pierce, PI32106) or diluted West Femto (1:5 in TBS [detailed above]; Pierce, 34,095). Images from 3 blots were quantified using ImageJ.

Statistical analysis

The statistical analysis used for each experiment is indicated in each figure legend. First power and sample size (n) calculations were performed on Minitab18 for each experimental paradigm: comparing 2 means from 2 samples, with two-sided equality to identify the sample size that corresponds to a power of 0.9 with α = 0.01.For each experiment, a stringent significance threshold of p < 0.01 (99% confidence) was used as detailed in [9]. Based on the power analysis, quantifications were performed across 5, 6, or 10 larvae. The n-value refers to the number of larvae or the number of immunoblots from 3 independent experiments. Individual data points for each analysis were averaged for each n and compared.

To select the appropriate statistical test, data distributions for each transport dynamic analyzed were first checked for normality using the nortest package of R: the Lilliefors test and Anderson – Darling test as previously detailed [9,40,129]. Statistical significance of normal distributions was calculated by one-way ANOVA/post-hoc analysis to reduce Type I error, followed by two-sample two-tailed Student’s t-tests to test to compare individual groups in Excel and Minitab18. Statistical analysis reported in figures report results from Student’s t-tests, as results from ANOVA/post-hoc and Student’s t-tests were consistent. Statistical significance of normal distributions was calculated by a two-sample two-tailed Student’s t-test and/or ANOVA while the non-normal segmental velocity distributions were compared using the non-parametric Wilcoxon – Mann–Whitney rank sum test in Excel and Minitab18. Unless otherwise specified, the data compared was found to be normally distributed. For in vivo motility quantifications, each larva (total of 4 movies, > 500 vesicles) were pooled, then the averaged before performing statistical analysis. Therefore, statistical analysis was performed on global quantifications from each larva rather than individual vesicle dynamics (n = 10 larvae, 4 movies per larvae). Overlaid dot plots were constructed for all figures using OriginLab/OriginPro to represent mean±SEM and biological models were generated in BioRender.

Data availability

All data generated or analyzed during this study are included in this published article or in supplementary information. Raw data is available from the corresponding author upon request.

Acknowledgment

We thank the members of the Gunawardena lab for constructive discussions and insight. We thank Dr. Troy Littleton for the UAS-HTT15Q-mRFP line, Dr. Norbert Perrimon for the UAS-HTTex1-25Q-eGFP line, Dr. Sheng Zhang for the htt-KO line, Dr. Matthew Xu-Friedman for mouse brains, the Coriell Cell Repository and the NINDS Repository for the WT human iPSC lines, Dr. Gábor Juhász for the Syx17^LL and Rab7^d1 fly lines as well as the Syx17 and the ref(2)P antibodies, Dr. Lawrence Goldstein for the KIF5C antibody, the Bloomington Drosophila Stock Center, and the Developmental Studies Hybridoma Bank. SG thanks Priyantha Karunaratne for constant support.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2022.2119351

Additional information

Funding

This work was supported in part by R03 NS084386, R03 NS092024 and an award from the BrightFocus Foundation to SG. TJK and JW were supported by fellowships from the Mark Diamond Research Fund. TJK was also supported by the Niciszewska Mucha Dissertation Fellowship, an ASCB Travel Award, and a Beverly Patterson and Charles W. Bishop Neuroscience Travel Award. LT and HH were supported by fellowships from the Center for Undergraduate Research and Creative Activities (CURCA) from SUNY at Buffalo. HH was also supported by a fellowship from Philip G. Miles Undergraduate Summer Research Fund; National Institute of Neurological Disorders [R03 NS084386]; National Institute of Neurological Disorders [R03 NS092024]