Genetic screen in Drosophila muscle identifies autophagy-mediated T-tubule remodeling and a Rab2 role in autophagy

  1. Naonobu Fujita  Is a corresponding author
  2. Wilson Huang
  3. Tzu-han Lin
  4. Jean-Francois Groulx
  5. Steve Jean
  6. Jen Nguyen
  7. Yoshihiko Kuchitsu
  8. Ikuko Koyama-Honda
  9. Noboru Mizushima
  10. Mitsunori Fukuda
  11. Amy A Kiger  Is a corresponding author
  1. University of California, San Diego, United States
  2. Graduate School of Life Sciences, Tohoku University, Japan
  3. The University of Tokyo, Japan
9 figures and 1 additional file

Figures

Figure 1 with 1 supplement
Detection of T-tubule membrane organization and remodeling in intact Internal Oblique Muscle (IOM) of live Drosophila.

(A) mCD8:GFP showed a mesh-like pattern in pharate/pre-adult dorsal abdominal IOMs at 4d APF by live imaging, with both transversal and longitudinal membrane elements as indicated. (B) Schematic of IOM and z-section regions imaged in panel C. (CD) Colocalization between mCD8:GFP (green) and Dlg1 (pink) at T-tubules in 4d APF IOMs quantified as Pearson’s correlation between Dlg1 and GFP or mCD8:GFP; ± SEM of pooled data for 10 images from three experiments. (E) Time line of fly development from third instar larva to adult at 25°C; days after puparium formation (d APF). (F) Time course microscopy of mCD8:GFP in dorsal muscles imaged through the cuticle of live wildtype animals from third instar larva (3IL) to 4d APF, showing membrane remodeling in abdomens (top), central sections of individual IOMs (middle) and magnified view of boxed regions (bottom). See Figure 1—figure supplement 1 for related data.

https://doi.org/10.7554/eLife.23367.002
Figure 1—source data 1

Relates to Figure 1D.

Pearson correlation indicating colocalization between GFP or mCD8:GFP with Dlg1 at T-tubules in IOMs at 4d APF (.xlsx file).

https://doi.org/10.7554/eLife.23367.003
Figure 1—figure supplement 1
mCD8:GFP partially colocalizes with Zormin, a Z-line marker protein.

IOM central z-sections at 4d APF. (A) Colocalization between cytosolic GFP and anti-Dlg1. (B) Colocalization between mCD8:GFP and Zormin. Magnified regions from peripheral area as indicated by box.

https://doi.org/10.7554/eLife.23367.004
T-tubules disassemble and reassemble with IOM remodeling during metamorphosis.

(A) Time course microscopy of Dlg1:GFP (T-tubule), Rtnl1:GFP (sarcoplasmic reticulum) or GFP:actin (myofibril) in wildtype animals at the indicated time points. (B) Schematic of an IOM TEM transverse section, as shown in 2CG. (CG) TEM images of IOM transverse sections in wildtype animals. Organized myofibrils and T-tubules were observed in both 3IL and 4d APF stages (C and G). At 1d APF, myofibrils were partially lost with mostly disorganized membranes (D). At 2d APF, myofibrils were completely absent with obvious appearance of autophagosomes and electron-dense lysosomal compartments (E). At 3d APF, myofibrils were reassembled but not well organized with a lack of obvious T-tubules (F).

https://doi.org/10.7554/eLife.23367.005
Figure 3 with 1 supplement
A unique T-tubule remodeling phenotype with knockdown of a set of known and unknown gene functions in autophagy.

All IOMs imaged at 4d APF. (A) Muscle-targeted RNAi screen of IOM remodeling. In primary screen of 300 selected muscle-targeted RNAi lines (see text), 77 lines exhibited eclosion or adult mobility defects; these lines were used in a secondary screen for mCD8:GFP organization by confocal imaging. Three abnormal phenotype categories were identified for 37 lines. The shared ‘small vesicle’ phenotype was found for 10 RNAi lines for five genes presented here. (B) Rab2, Rab7, Stx17, SNAP29 or Vamp7/8 RNAi resulted in IOMs filled with small, mCD8:GFP-marked vesicles. Top row, brightly-marked dorsal IOMs in whole abdomen. Bottom row, magnified image of mCD8:GFP in single IOM. (C) Schematic of IOM and regions imaged in panel DE. (D) T-tubule (Dlg1, green) and myofibril (F-actin, pink) organization in IOMs from control and Rab2, Rab7, Stx17, SNAP29, or Vamp7/8 RNAi conditions. (E) RNAi of HOPS components, Vps39, Vps18, or Vps11, exhibited shared phenotypes of (top) many mCD8:GFP-marked small vesicles and (bottom) lack of T-tubules (Dlg1, green) and myofibrils (F-actin, pink). See Figure 3—figure supplement 1 for related data.

https://doi.org/10.7554/eLife.23367.006
Figure 3—source data 1

Relates to Figure 3—figure supplement 1B.

Quantification of disorganized IOM mutant phenotypes observed from Rab2 RNAi co-expressed with lacZ negative control or with YFP:Rab2 wildtype rescue at 4d APF (.xlsx file).

https://doi.org/10.7554/eLife.23367.007
Figure 3—figure supplement 1
A unique T-tubule remodeling phenotype identifies gene functions in autophagosome fusion.

(AB) Rab2 RNAi defects in 4d APF IOMs were rescued by co-overexpression of wildtype YFP:Rab2, but not LacZ control (A). Mean percentages ± SD of disorganized IOMs from total in 10 animals (B). (C) The effect of HOPS RNAi on IOM morphology. Whole dorsal abdomens imaged with muscle-targeted mCD8:GFP expression at 4d APF. Control and Vps39, Vps18 or Vps11 RNAi that resulted in swollen and misshapen IOMs.

https://doi.org/10.7554/eLife.23367.008
Autophagosomes accumulate in IOMs with Rab2, Rab7 or Stx17 knockdown.

All IOMs imaged at 4d APF. (A) Autophagic flux assay using tandem-tagged mCherry:GFP:Atg8 (mCherry (C), pink; GFP (G), green; colocalization, white). Peripheral IOM z-sections with magnified regions from indicated boxed areas shown below. In control IOMs, mCherry-positive only puncta were primarily detected, indicative of Atg8 flux to autolysosomes. In Rab2, Rab7 or Stx17 RNAi IOMs, dual-positive Atg8 puncta were primarily detected, indicating block in autophagic flux. (B) Pearson correlation between GFP and mCherry of mCherry:GFP:Atg8 from pooled data for 10 images from three experiments, ± SD. (C) GFP:Stx17 distribution in IOMs from control or with Rab2, Rab7 or SNAP29 RNAi, which show increased GFP:Stx17 localization at puncta and small rings. (D) Colocalization of GFP:Atg8 and mCherry:Stx17 (Pearson correlation, 0.53) in Rab7 RNAi IOMs. (EF) TEM images of IOM transverse-sections. Control IOMs show expected myofibrils and T-tubule membranes, while Rab2, Rab7 or Stx17 RNAi IOMs were filled mostly with autophagosomes. (F) Quantification of the mean number of autophagosomes per IOM area, ± SD.

https://doi.org/10.7554/eLife.23367.009
Figure 4—source data 1

Relates Figure 4B.

Pearson correlation indicating colocalization between GFP and mCherry from mCherry:GFP:Atg8 expressed in IOMs at 4d APF of control and RNAi conditions shown (.xlsx file).

https://doi.org/10.7554/eLife.23367.010
Figure 4—source data 2

Relates to both Figure 4F and Figure 5C.

Quantification of the number of autophagosomes (Figure 4F) or mitochondria (Figure 5C) manually counted per calculated IOM area at 4d APF for the control and RNAi conditions shown (.xlsx file).

https://doi.org/10.7554/eLife.23367.011
Autophagy is required for IOM T-tubule remodeling and mitochondrial clearance.

(A) mCD8:GFP in 4d APF dorsal abdominal muscles (top) and IOM section (bottom) for control and Atg1, Atg3 or Atg18 RNAi conditions. (B) T-tubule (Dlg1, green) and myofibril (F-actin, pink) organization in IOMs of control and Atg1, Atg3 or Atg18 RNAi that show fragmented and disorganized T-tubules. (C) TEM images of IOM transverse-sections show disorganized contractile system, lack of T-tubules and many mitochondria in Atg1 or Atg18 RNAi conditions at 4d APF. Quantification of the mean number of mitochondria per area, ± SD. (E) Mitochondria (YFP:Mito, green) and myofibril (F-actin, pink) organization in control and Atg1, Atg18 or Rab2 RNAi IOMs at 4d APF.

https://doi.org/10.7554/eLife.23367.012
Figure 6 with 2 supplements
Autophagy induction is coincident with and required for proper T-tubule membrane disassembly.

(A) Time course microscopy of autophagy-related markers in live wildtype animals over 1 day intervals during metamorphosis; GFP:Atg8 (autophagosomes), Lamp1:GFP (endolysosomes/autolysosomes), GFP:Rab7 (late endosomes/amphisomes). (B) Quantification of mean GFP:Atg8 puncta number per tissue area ± SD from at least 10 randomly selected IOMs. (CD) Time course microscopy of GFP:Atg8 (autophagosomes) and GFP:Tubby-Cter [PI(4,5)P2, T-tubules] in live wildtype animals at the indicated hours APF (C). Mean percentages IOM area covered by T-tubules (blue, ± SD) and mean number of autophagosomes (red, ± SD) quantified from at least 10 randomly selected IOMs. (E) Time course microscopy during metamorphosis of mCD8:GFP in IOMs of control, Rab2 RNAi or Atg1 RNAi (as in Figure 1F, bottom). Both Rab2 and Atg1 RNAi show normal T-tubules in 3IL muscle and initial defects in membrane organization by 1d APF that persist as a block in remodeling through 4d APF. (F) mCD8:GFP membrane in central regions of IOMs at 20h APF upon T-tubule disassembly in control, Rab2 RNAi or Atg1 RNAi, with mCD8:GFP membrane reorganization into bright patches (arrows), big membrane patches/stacks (red arrows) or membrane vesicles (arrowheads). (G) GFP:Atg8 at 1d APF in control and Rab2 RNAi IOMs. See Figure 6—figure supplements 1 and 2 for related data.

https://doi.org/10.7554/eLife.23367.013
Figure 6—source data 1

Relates to Figure 6B.

Quantification of the number of GFP:Atg8 puncta manually counted per calculated IOM area over the indicated timecourse from 3IL through 4d APF in wildtype myofibers (.xlsx file).

https://doi.org/10.7554/eLife.23367.014
Figure 6—source data 2

Relates to Figure 6D.

Quantification of the number of GFP:Atg8 puncta manually counted per calculated IOM area over the indicated timecourse from 3IL through 24h APF in wildtype myofibers (.xlsx file).

https://doi.org/10.7554/eLife.23367.015
Figure 6—source data 3

Relates to Figure 6D.

Quantification of the percent IOM area covered by minimal threshold for GFP:TubbyC-positive T-tubules over the indicated timecourse from 3IL through 24h APF in wildtype myofibers (.xlsx file).

https://doi.org/10.7554/eLife.23367.016
Figure 6—figure supplement 1
Characterization of autophagy levels during IOM remodeling.

(A) Endogenous anti-Atg8 puncta in 1d APF IOM.

https://doi.org/10.7554/eLife.23367.017
Figure 6—figure supplement 2
Characterization of autophagy requirement for IOM membrane remodeling.

(AB) Time course microscopy of muscle-expressed mCD8:GFP at 1d intervals from third instar larval stage through metamorphosis. (A) Dorsal view of whole abdomens in control or with muscle-targeted Rab7, Rab2, Stx17 or Atg18 RNAi. (B) mCD8:GFP-positive membrane network in IOM central sections from control or Stx17 or Atg18 RNAi animals.

https://doi.org/10.7554/eLife.23367.018
Figure 7 with 2 supplements
Rab2 has a conserved function required for autophagic clearance in MEFs.

(A) Rab2A/B double knockout (DKO) MEFs generated by CRISPR-Cas9. Parental, Rab2A-KO, Rab2B-KO, or Rab2A/B-DKO MEFs were cultured in regular growth medium and lysates were analyzed by immunoblots with indicated antibodies. Asterisks denote non-specific bands. (BC) LC3 puncta in Rab2A/B_DKO MEFs. Parental or Rab2A/B_DKO MEFs were cultured in regular medium and analyzed by immunofluorescence microscopy (B). Mean number of LC3 puncta from 30 cells ± SEM of pooled data from three experiments (C). (D) LC3-II flux assay in parental or Rab2A/B_DKO. Parental or Rab2A/B_DKO MEFs were cultured in regular medium (Fed) or EBSS (Stv) for 2h with or without 100 nM Bafilomycin A1 and analyzed by immunoblotting. (EG) TEM analysis of parental (E) or Rab2A/B_DKO MEFs (F) cultured in regular medium. Autolysosomes, red arrowheads; Autophagosomes, green arrowhead. Quantification of the number of autophagosomes or autolysosomes from 20 randomly selected areas (G). (HI) Mosaic analysis of parental (GFP-) and Rab2A/B_DKO MEFs (GFP+) with Lysotracker Red (H) or Magic Red (I). See Figure 7—figure supplements 1 and 2 for related data.

https://doi.org/10.7554/eLife.23367.019
Figure 7—source data 1

Relates to Figure 7C.

Quantification of the number of LC3 puncta in parental or Rab2A/B double-knockout MEFs grown in fed conditions (.xlsx file).

https://doi.org/10.7554/eLife.23367.020
Figure 7—source data 2

Relates to Figure 7G.

Quantification of the number of autophagosomes and autolysosomes per cell area in TEM images of parental or Rab2A/B double-knockout MEFs grown in fed conditions (.xlsx file).

https://doi.org/10.7554/eLife.23367.021
Figure 7—source data 3

Relates to Figure 7—figure supplement 1B and D.

Quantification of the number of Atg8 puncta and Ref(2)P puncta per area of fat body from fed larvae (Figure 7—figure supplement 1B). Pearson correlation indicating colocalization between GFP and mCherry from mCherry:GFP:Atg8 expressed in fat body from starved larvae (Figure 7—figure supplement 1D). Control and Rab2 RNAi lines as shown (.xlsx file).

https://doi.org/10.7554/eLife.23367.022
Figure 7—source data 4

Relates to Figure 7—figure supplement 1F.

Quantification of the number of autolysosomes, autolysosome size and number of autophagosomes per tissue area in TEM images of control or Rab2 RNA fat body from fed or starved conditions (.xlsx file).

https://doi.org/10.7554/eLife.23367.023
Figure 7—source data 5

Relates to Figure 7—figure supplement 2I.

Quantification of EGFR protein levels normalized to actin levels and percent of ratio at 0h over timecourse as shown for control, Rab2A/2B double knockout and myc:Rab2A revertant of double knockout (.xlsx file).

https://doi.org/10.7554/eLife.23367.024
Figure 7—figure supplement 1
Rab2 RNAi blocks autophagic flux in third instar larval fat body.

(AB) Atg8 and Ref(2)P/p62 detected by immunofluorescence in control (lacZ) or Rab2 RNAi fat body from fed condition. Rab2-1 and Rab2-2 represent two different RNAi hairpins; see Materials and methods (A). Number of Atg8 or Ref(2)P objects normalized to fat body tissue area (B). (CD) Autophagic flux detected by mCherry:GFP:Atg8 in control or Rab2 RNAi fat body in starved condition (C). Percentage of autophagosomes (colocalized mCherry and GFP, or ‘yellow’ Atg8) to total number of autophagic vacuoles (total of red plus ‘yellow’ puncta) normalized to fat body area (D). (EF) TEM of control or Rab2-depleted larval fat body from fed or starved conditions (E). (F) Quantification of autophagic structures. Number of autolysosomes (top) or autophagosomes (bottom) normalized to fat body area; autolysosome size (right).

https://doi.org/10.7554/eLife.23367.025
Figure 7—figure supplement 2
Characterization of Rab2A/B double knockout MEFs.

(AB) Rescue experiment in Rab2A/B_DKO MEFs. Parental or Rab2A/B_DKO MEFs stably expressing indicated constructs were cultured in regular medium and analyzed by immunofluorescence microscopy (A). LC3-II flux assay in parental or Rab2A/B_DKO MEFs stably expressing indicated constructs. MEFs were cultured in regular medium (Fed) or EBSS (Stv) for 2h with or without 100 nM Bafilomycin A1 and analyzed by imunoblots as indicated (B). (CD) TEM images of parental (C) or Rab2A/B_DKO MEFs (D) cultured in regular medium. (EG) Mosaic analysis of parental (GFP-) and Rab2A/B_DKO (GFP+) MEFs by LC3 immunostaining (E), Lysotracker Red (F) or Magic Red (G). (H and I) EGFR degradation assay in Rab2A/B_DKO MEFs. Parental, Rab2A/B_DKO or Rab2A/B_DKO MEFs stably expressing myc-Rab2A were serum-starved for overnight and then treated with 100 nM EGF. After EGF addition, cells were chased in cycloheximide and lysed at time points indicated. The lysates were analyzed by EGFR and β-actin immunoblots (H). (I) Ratio of normalized EGFR to β-actin integrated densities from three independent experiments; mean ± SD.

https://doi.org/10.7554/eLife.23367.026
Figure 8 with 1 supplement
Rab2 localizes to completed autophagosomes.

(A) Interaction between Rab2A/B and Vps39 or Vps41. COS-7 cells were co-transfected with T7-Vps39 or -Vps41 and FLAG-tagged Rab as indicated. Two days later, cells were lysed, immunoprecipitated with anti-T7 antibody, and detected by immunoblots as indicated. (B) YFP:Rab2 distribution in IOMs at 4d APF in control, Stx17 or Rab7 RNAi conditions. The number of YFP:Rab2 puncta and rings increased when autophagosome-lysosome fusion was blocked. (CD) Colocalization between YFP:Rab2 and mCherry:Atg8 in Rab7 RNAi IOMs at 4d APF (C). Line plot profile of a yellow arrow in panel C (D). (EF) Colocalization between Rab2A and LC3 in MEFs. MEFs stably-expressing GFP:Rab2A were cultured in EBSS for 1h and immunostained for LC3 (E). Mean percent colocalization with LC3 ± SD of 10 images (F). (GH) Colocalization analysis between stably-expressed mStrawberry:Rab2A, GFP:LC3 and Atg16L1 in MEFs cultured in EBSS for 1h. Arrows: LC3-positive, Atg16L1-positive isolation membranes. Arrowheads: LC3-positive, Atg16L1-negative autophagosomes (G). Mean percent colocalization with Rab2A ± SD of 10 images (H). See Figure 8—figure supplement 1 for related data.

https://doi.org/10.7554/eLife.23367.027
Figure 8—source data 1

Relates to Figure 8F.

Quantification of percent LC3 puncta that colocalizes with GFP:Rab2A, GFP:Rab2B or GFP in MEFs (.xlsx file).

https://doi.org/10.7554/eLife.23367.028
Figure 8—source data 2

Relates to Figure 8H and Figure 8—figure supplement 1D.

Quantification of percent mStrawberry:Rab2A (Figure 8H) or mStrawberry:Rab2B (Figure 8—figure supplement 1D) puncta that colocalizes with GFP:LC3 or Atg16L1 in MEFs (.xlsx file).

https://doi.org/10.7554/eLife.23367.029
Figure 8—source data 3

Relates to Figure 8—figure supplement 1F and H.

Quantification of percent mStrawberry:Rab2A (Figure 8—figure supplement 1F) or mStrawberry:Rab2B (Figure 8—figure supplement 1H) puncta that colocalizes with GFP:LC3 or Lamp1 in MEFs (.xlsx file).

https://doi.org/10.7554/eLife.23367.030
Figure 8—source data 4

Relates to Figure 8—figure supplement 1J and L.

Quantification of percent LC3 (Figure 8—figure supplement 1J) or Lamp1 (Figure 8—figure supplement 1L) puncta that colocalizes with GFP:Rab2A or GFP:Rab7 in MEFs (.xlsx file).

https://doi.org/10.7554/eLife.23367.031
Figure 8—figure supplement 1
Rab2A and Rab2B localize to complete autophagosomes, but not isolation membranes nor lysosomes.

(A) Colocalization between Rab2B and LC3 in MEFs. MEFs stably expressing GFP:Rab2B were cultured in EBSS for 1h, and immunostained for LC3. (B) Immuno-electron microscopy of endogenous LC3 and GFP:Rab2A. MEFs stably expressing GFP:Rab2A were starved for 90 min and subjected to immuno-EM using mouse anti-GFP, rabbit anti-LC3 and secondary antibodies conjugated with colloidal gold particles (mouse, 12 nm; rabbit, 18 nm). (CD) Colocalization between Rab2B, LC3 and Atg16L1 in MEFs. MEFs stably expressing mStrawberry:Rab2B and GFP:LC3 were cultured in EBSS for 1h, and immunostained for Atg16L1 (B). Mean percent colocalization with Rab2B ± SD of 10 images (D). (EH) Colocalization between Rab2A (EF) or Rab2B (GH) with LC3 and Lamp1 in MEFs. MEFs stably expressing mStrawberry:Rab2A/B and GFP:LC3 were cultured in EBSS for 1h, and immunostained for Lamp1 (E, G). Mean percent colocalization with Rab2A/B ± SD of 10 images (F, H). (IL) Colocalization among Rab2A, Rab7 and LC3 (IJ) or Lamp1 (KL). MEFs stably expressing both GFP:Rab7 and mStraw:Rab2A were incubated in EBSS for 1h and then stained for LC3 or Lamp1 (I, K). Mean percent colocalization with LC3 or Lamp1 ± SD of 10 images (J, L).

https://doi.org/10.7554/eLife.23367.032
Figure 9 with 1 supplement
Hierarchal analysis of Rab2 and factors involved in autophagosome-lysosome fusion.

(A) Dynamics of EYFP:Rab2A and super-enhanced CFP (seCFP):Stx17TM in autophagy. MEFs expressing both EYFP:Rab2A and seCFP:Stx17TM were incubated in amino acid free DMEM and imaged by fluorescent microscopy. Rab2A localization preceded Stx17 by several minutes in most cases. (BC) Colocalization of YFP:Rab2 with mCherry:Atg8 in 4d APF IOMs with Rab7, Stx17 or Vps39 RNAi (B), indicating Rab2 localization to autophagosomes is independent of these functions. Mean percentage of YFP:Rab2-positive puncta colocalized with mCh:Atg8 puncta quantified from 10 randomly selected areas (C). (DE) Colocalization between mCherry:Atg8 and Dor (Vps18) in 1d APF control IOMs (D) or Rab2 RNAi IOMs (E). Line plot profiles of Atg8 and Dor intensities along white arrows in merged images. (F) Model for autophagy-mediated T-tubule remodeling and a Rab2 role in autophagosome-lysosome fusion. In myofiber remodeling, progression from T-tubule membrane disassembly requires Atg1-mediated autophagy induction and may contribute as a source of autophagosomal membranes. Mitochondria are a major autophagic cargo in this process. Rab2 localizes to completed autophagosomes and interacts with the HOPS complex to promote autophagosome-lysosome fusion, leading to cargo degradation. The autolysosomal membranes could be recycled back as a source for T-tubule membrane reassembly or other membrane structures in the cell.

https://doi.org/10.7554/eLife.23367.033
Figure 9—source data 1

Relates to Figure 9C.

Quantification of the percent YFP:Rab2 puncta colocalized with mCherry:Atg8 puncta in IOMs at 4d APF for the control and RNAi conditions shown (.xlsx file).

https://doi.org/10.7554/eLife.23367.034
Figure 9—figure supplement 1
Rab2 is recruited to mature autophagosomes and required for Dor/Vps18 HOPS localization

(A) Dynamics of EYFP:Rab2A and super-enhanced CFP (seCFP):LC3 in autophagy.

MEFs expressing both EYFP:Rab2A and seCFP:LC3 were incubated in amino acid free DMEM and imaged by a fluorescent microscopy. LC3 localization preceded Rab2 for several minutes in most cases. (B) Colocalization between GFP:Atg8 and Dor (Vps18, red) in 1d APF control or Rab2 RNAi IOMs. White boxes are areas shown in Figure 9D–E.

https://doi.org/10.7554/eLife.23367.035

Additional files

Supplementary file 1

RNAi lines screened with DMef2-GAL4 muscle-targeted expression.

https://doi.org/10.7554/eLife.23367.036

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Naonobu Fujita
  2. Wilson Huang
  3. Tzu-han Lin
  4. Jean-Francois Groulx
  5. Steve Jean
  6. Jen Nguyen
  7. Yoshihiko Kuchitsu
  8. Ikuko Koyama-Honda
  9. Noboru Mizushima
  10. Mitsunori Fukuda
  11. Amy A Kiger
(2017)
Genetic screen in Drosophila muscle identifies autophagy-mediated T-tubule remodeling and a Rab2 role in autophagy
eLife 6:e23367.
https://doi.org/10.7554/eLife.23367