RESEARCH ARTICLE

Septins are involved at the early stages of macroautophagyin S. cerevisiaeGaurav Barve1, Shreyas Sridhar1, Amol Aher1, Mayurbhai H. Sahani1, Sarika Chinchwadkar1, Sunaina Singh1,Lakshmeesha K. N.1, Michael A. McMurray2 and Ravi Manjithaya1,*

ABSTRACTAutophagy is a conserved cellular degradation pathway whereindouble-membrane vesicles called autophagosomes capture long-livedproteins, and damaged or superfluous organelles, and deliver them tothe lysosome for degradation. Septins are conserved GTP-bindingproteins involved in many cellular processes, including phagocytosisand the autophagy of intracellular bacteria, but no role in generalautophagy was known. In budding yeast, septins polymerize into ring-shaped arrays of filaments required for cytokinesis. In an unbiasedgenetic screen and in subsequent targeted analysis, we foundautophagy defects in septin mutants. Upon autophagy induction,pre-assembled septin complexes relocalized to the pre-autophagosomal structure (PAS) where they formed non-canonicalseptin rings at PAS. Septins also colocalized with autophagosomes,where they physically interacted with the autophagy proteins Atg8 andAtg9. When autophagosome degradation was blocked in septin-mutant cells, fewer autophagic structures accumulated, and anautophagy mutant defective in early stages of autophagosomebiogenesis (atg1Δ), displayed decreased septin localization to thePAS. Our findings support a role for septins in the early stages ofbudding yeast autophagy, during autophagosome formation.

This article has an associated First Person interview with the firstauthor of the paper.

KEY WORDS: Autophagy, Noncanonical ring, Septin,Autophagosome biogenesis, Pre-autophagosomal structure, PAS,Atg9 trafficking

INTRODUCTIONMacroautophagy (herein autophagy) is an evolutionarily conservedintracellular waste disposal and recycling process that is critical fornormal cellular and organismal homeostasis. Autophagy involvesthe formation of double-membrane vesicles called autophagosomesthat engulf intracellular material destined for degradation.Autophagosomes eventually fuse with vacuoles or lysosomes,resulting in cargo degradation and recycling of cellular building

blocks, such as amino acids, back to the cytoplasm. The biogenesisof autophagosomes remains incompletely understood.

In budding yeast cells, the site of autophagosome formationis known as the pre-autophagosomal structure (PAS) and isperivacuolarly located. Recent work has shown that the PASis tethered to endoplasmic reticulum (ER) exit sites where multipleautophagy proteins colocalize in a hierarchical sequence (Graefet al., 2013; Suzuki et al., 2007). The membrane source for thedeveloping autophagosome is contributed by the trafficking of Atg9along with its transport complex (Atg1–Atg11–Atg13–Atg23–Atg27–Atg2–Atg18–TRAPIII) to help build the initial cup-shapedstructure, the phagophore (Legakis et al., 2007; Reggiori et al., 2004;Tucker et al., 2003). Additional recruitment of the Atg5–Atg12–Atg16 complex as well as Atg8 allows the completion of theautophagosome (Feng et al., 2014).

Septin proteins bind guanine nucleotides and co-assemble inhetero-oligomers capable of polymerizing into cytoskeletal filaments(Mostowy and Cossart, 2012). Septin filaments associate directlywith membranes in a curvature-dependent manner (Bridges et al.,2016) and regulate membrane dynamics, including vesicle fusionevents (Mostowy and Cossart, 2012). In immune cells, septins alsolocalize transiently to the phagocytic cup and are functionallyinvolved in phagocytosis (Huang et al., 2008). Septins have beenimplicated in autophagy in mammalian cells infected by intracellularbacteria, where they form cage-like structures around the bacterialcells that colocalize with the autophagosome marker autophagosomemarker MAP1LC3A, the homolog of yeast Atg8. It is believed thatthese structures entrap bacteria, restricting their motility and targetingthem for autophagy-mediated degradation (Mostowy et al., 2009,2010). During Shigella infection, assembly of septin cages and theautophagosome in the host mammalian cells are interdependent(Mostowy et al., 2010, 2011; Sirianni et al., 2016). Despite thesefindings, it remains unclear to what extent septins contribute toautophagy outside the context of bacterial infection (Torraca andMostowy, 2016).

In S. cerevisiae cells undergoing mitotic proliferation, five septinproteins – Cdc3, Cdc10, Cdc11, Cdc12 and Shs1 – comprise anarray of filaments that is directly associated with the plasmamembrane at the mother–bud neck, and controls cell polarity, budmorphogenesis and cytokinesis (Glomb and Gronemeyer, 2016; Ohand Bi, 2011). Upon nitrogen starvation, diploid yeast cells undergomeiosis and sporulation, during which a cup-shaped double-membrane structure, the prospore membrane (PSM), engulfshaploid nuclei and other organelles to form stress-resistant spores(Neiman, 2005, 2011). Yeast septins are required for proper PSMbiogenesis (Heasley and McMurray, 2016), but there was no knownrole for septins in yeast autophagy. Here, we describe autophagydefects in septin-mutant strains and physical interactions betweenseptins and established autophagy factors that support a functionalrole for septins in yeast autophagy.Received 29 July 2017; Accepted 10 January 2018

1Molecular Biology and Genetics Unit, Jawaharlal Nehru Centre for AdvancedScientific Research, Jakkur, Bangalore 560064, India. 2University of Colorado,Anschutz Medical Campus, Department of Cell and Developmental Biology,Aurora, CO 80045, USA.

*Author for correspondence (ravim@jncasr.ac.in)

M.H.S., 0000-0001-8534-2197; M.A.M., 0000-0002-4615-4334; R.M., 0000-0002-0923-5485

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1

© 2018. Published by The Company of Biologists Ltd | Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience

Fig. 1. See next page for legend.

2

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience

RESULTSAutophagy defects in septin mutantsTo identify autophagy defects in viable mutant yeast strains, weintroduced into a collection of temperature-sensitive (Ts−) mutantsin a POT1-GFP strain, which expresses a marker of pexophagy(Kondo-Okamoto et al., 2012), a specialized form of autophagy inwhich peroxisomes are degraded (Oku and Sakai, 2016). Targetingof Pot1–GFP to the vacuole during starvation-induced pexophagyresults in destruction of the Pot1 part of the fusion protein andaccumulation of free GFP, which is readily detected byimmunoblotting (Fig. 1A; Fig. S1A,B). Unlike in wild-type (WT)cells, where free GFP accumulated at both 22°C and 37°C, in cellsexpressing any of several Ts− mutant alleles of the septin CDC10(G100E or P3S G44D) or CDC11 (G29E, G34D or S31F S100P)more free GFP was detected at 22°C, compared to what was seen at37°C, and the Pot1–GFP fusion remained intact at 37°C (Fig. 1A;Fig. S1A). These results were also corroborated by usingfluorescence microscopy to visualize the delivery of GFP-labeledperoxisomes to the vacuole as diffuse GFP inside the vacuolarlumen (Fig. S1B). At 37°C the number of starved septin-mutantcells showing free GFP inside the vacuole was reduced significantlywhen compared to the numbers of starved WT cells, and also whencompared to numbers of mutant cells incubated at 22°C (Fig. S1C).These data point to a requirement for septin function in pexophagy.In nutrient-replete conditions, the Ts− mutants of CDC10 and

CDC11 in which we found pexophagy defects arrest cell divisionwith failed cytokinesis (Hartwell, 1971). Interestingly, we didnot observe Pot1–GFP-processing defects in cells expressing Ts−

mutant versions of CDC3 (G365R) or CDC12 (G247E) (Fig. S1D),which were originally isolated in the same cell division screen(Hartwell, 1971) as the CDC10 and CDC11 mutants that causedpexophagy defects. To explain this discrepancy, we considered thatin cdc3(G365R) or cdc12(G247E) cells, high temperature preventsde novo assembly of septin complexes but does not destabilizeexisting structures (Dobbelaere et al., 2003; Kim et al., 1991;Weemset al., 2014). Since pexophagy, like autophagy in general, occurs instarved non-dividing cells, we hypothesized that a functionalcontribution of septins to pexophagy may not require assembly ofnew septin complexes, and instead utilizes pre-existing complexesassembled prior to the nutrient withdrawal and temperature upshift.Indeed, yeast septins are exceedingly long-lived proteins, evenduring starvation (McMurray and Thorner, 2008). In support of thismodel, Pot1–GFP processing was compromised in cdc12-td cells(Fig. S1D), which express a temperature-degron-tagged Cdc12 thatis known to cause rapid disassembly of pre-existing filamentousseptin structures associated with the bud neck in dividing cells

(Weems et al., 2014). These findings indicate that the septincomplexes involved in pexophagy are composed of the same septinsthat were previously synthesized when nutrients were available andsupported cytokinesis in budding cells.

To ask whether septins are more generally involved in autophagy,we examined the processing of GFP–Atg8, which is processed inthe vacuole during autophagy (Cheong and Klionsky, 2008).Similar to the pexophagy results obtained with Pot1–GFP, wenoticed considerable slowdown of autophagic flux in septin mutantcells, as evidenced by slower processing of GFP–Atg8 (Fig. S2). Inaddition to mutant alleles harboring substitutions in specificresidues, conditionally viable septin-mutant cells can be obtainedby deleting the CDC10 gene (Flescher et al., 1993; Frazier et al.,1998; McMurray et al., 2011). In cells lacking Cdc10, septinfilament assembly and functions essential for mitotic proliferationrequire septin hetero-hexamers formed via non-native Cdc3homodimerization (McMurray et al., 2011). Cells lacking Cdc10are temperature-sensitive for mitotic proliferation due to inefficientCdc3 homodimerization at high temperatures (McMurray et al.,2011). If the same septin complexes that function in cytokinesis arealso involved in autophagy, then wewould expect to find autophagydefects in cdc10Δ cells at 37°C. Indeed, this was the case, as assayedby both Pot1–GFP and GFP–Atg8 processing (Fig. S3A,B).Similarly, cells lacking Cdc11 require non-native Cdc12homodimerization for survival, but septin function is moreseverely compromised in cdc11Δ mutants because Shs1 occupiesthe same position as Cdc11 in a subset of hetero-octamers and, indoing so, prevents efficient Cdc12 homodimerization (McMurrayet al., 2011). We found that Pot1–GFP processing was moreseverely compromised in cdc11Δ than in cdc10Δ mutants(Fig. S3A,B), consistent with the relative magnitude of the effectson mitotic proliferation. Also consistent was the relatively minoreffect of deleting SHS1 (Fig. S3A,B), which has only mild effectson mitotic proliferation at 37°C (Finnigan et al., 2015; Mino et al.,1998). Finally, deleting SHS1 in a cdc11Δ background improvesmitotic proliferation, presumably because Shs1 no longer interfereswith Cdc12 homodimerization (McMurray et al., 2011). Pexophagyand autophagy defects in cdc11Δ shs1Δ cells were equivalent to thedefects in cdc10Δ cells (Fig. S3A,B), providing additional supportfor our conclusion that cytokinesis and autophagy share similarrequirements for septin complex assembly.

In WT cells, the induction of autophagy triggers a coalescence ofmultiple small vacuoles into a single organelle (Baba et al., 1994),which is readily observed by visualizing FM4-64 labeling of thevacuolar membrane (Fig. S3C). We noticed that in septin mutantcells, particularly the viable deletion mutants, vacuolar coalescencewas largely defective (Fig. S3C). The defect in pexophagy incdc11Δ cells was rescued by introduction of a plasmid encodingWT Cdc11, confirming that the pexophagy defect resulted from theabsence of Cdc11 (Fig. S3E).

If septin mutant cells have defects in autophagy, then they shouldbe sensitive to starvation, survival during which requires afunctional autophagy pathway (Suzuki et al., 2011). Indeed,heterozygous diploid strains lacking one copy of CDC10 areknown to be sensitive to nutrient deprivation (Davey et al., 2012), apreviously unexplained phenotype. Haploids lacking Cdc10 arealso sensitive to rapamycin (Chan et al., 2000), which could reflect arequirement for autophagy to survive the starvation-like metabolicconditions that result from inhibition of Tor1/2, as autophagy isinduced by rapamycin even in nutrient-replete conditions (Noda andOhsumi, 1998). Finally, we note that a high-throughput geneticinteraction study previously reported negative interactions between

Fig. 1. Septins migrate from the pre-existing bud-neck ring to cytoplasmduring starvation. (A) Pexophagy was affected in cdc10P3SG44D (cdc10-5)cells as compared to WT cells at the non-permissive temperature (37°C).(B) Microscopy images of Cdc10–GFP, Cdc11–GFP and Shs1–GFP cellsunder nutrient rich, nutrient deficient and rapamycin (0.4 µg/ml) treatmentconditions. Cells from log phase (0.6 to 0.8 OD) were transferred to starvationmedium (1 OD/ml) and imaged at different time points. (C) Quantification of thenumber of cells showing rings and puncta grown in rich, starvation orrapamycin treatment medium for 24 h. For quantification, cells showing onlyring or only dots were considered. Images acquired were converted intomaximum intensity projections, deconvolved and a total of 100 cells werequantified. (D) Cdc10, Cdc11 and Shs1 all colocalize as puncta duringstarvation. Strains JTY5396 and JTY5397 were grown as in B and imaged.(E) Septin localization in presence of cycloheximide (C) and rapamycin (R).Cells were grown as described in Fig. 1B in presence of cycloheximide(50 µg/ml) and rapamycin (YPD+C+R) and in presence of rapamycin(0.4 µg/ml) alone (YPD+R). Scale bars: 5 µm.

3

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience

Fig. 2. See next page for legend.

4

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience

cdc10 mutants and mutations in the autophagy genes ATG3, ATG8and ATG9 (Costanzo et al., 2010). These findings provideindependent support for a functional requirement for septinsin autophagy.

Septinsmove from the bud neck to the PAS during starvationand associate with mature autophagosomesOur genetic findings indicate a functional role for septin complexesin autophagy and further suggested that the same septin complexesassembled prior to induction of autophagy are utilized duringautophagy, without a requirement for new septin synthesis orassembly. Based on these findings, we predicted that, duringstarvation, septins should re-localize from the bud neck to sites ofautophagosome assembly (i.e. the PAS). Consistent with thisprediction, upon starvation GFP-tagged Cdc10, Cdc11 and Shs1present at the bud neck quickly (within 5 h) transitioned to cytosolicpuncta (Fig. 1B,C; Movie 1). Addition of rapamycin to cells in richmedium had the same effect (Fig. 1B,C). mCherry-tagged Cdc10colocalized in these cytosolic puncta with Cdc11–GFP and Shs1–GFP (Fig. 1D). We could not obtain conclusive results withfluorescently tagged Cdc3 or Cdc12 due to the propensity instarvation conditions of these fusion proteins to form, and incorporateother septins into, aberrant rod-shaped structures (Fig. S1E; data notshown). Crucially, relocalization of septins from the bud neck tocytosolic puncta did not require new protein synthesis, sinceequivalent results were observed in the presence of the translationinhibitor cycloheximide (Fig. 1D). These findings demonstrate that,upon starvation, pre-existing septin proteins re-localize from the siteof cytokinesis to cytosolic puncta, consistent with a role in a cytosolicprocess in these conditions.To ask whether the cytosolic puncta to which septins relocalize

upon starvation included the PAS, we examined septin–GFPlocalization in cells also expressing an mCherry-tagged version ofAtg8, which marks the PAS prior to its processing within thevacuole (Delorme-Axford et al., 2015). Approximately 30% ofcytosolic septin–GFP puncta in starved cells were also labeled bymCherry–Atg8 (Fig. 2A,B). Whereas in WT cells autophagosomesdisappear as they fuse with the vacuole, in cells lacking Ypt7,a Rab GTPase required for autophagosome-vacuole fusion,autophagosomes persist (Ishihara et al., 2001; Kim et al., 1999).We observed a corresponding increase in the number of septin focithat were also marked by mCherry–Atg8 in cells lacking Ypt7

(Fig. 2C,D). Similar results were obtained using GFP-tagged septinsand an RFP-tagged version of Ape1/Lap4, an aminopeptidase that istethered to the PAS in its precursor form prior to proteolyticactivation in the vacuolar lumen (Delorme-Axford et al.,2015) (Fig. 2E,F). Careful imaging revealed that septin–GFPpuncta could often be resolved as rings surrounding the PAS(Fig. 2G; Fig. S4A–E). The dimensions of these rings (400–600 nmin diameter) are about half the size of septin rings at the bud neck(Okada et al., 2013) and are instead similar to the size ofautophagosomes (400–900 nm in diameter) (Suzuki and Ohsumi,2010). The autophagy protein Atg9 is also known to form ∼500-nm-wide rings around mCherry–Atg-marked autophagosomes(Yamamoto et al., 2012). Septin–GFP localization at the PAS wastransient (Movie 2), providing an explanation for our observationsthat not every PAS was associated with septins. These resultssupport a model in which septins associate with the developing PASand remain associated with mature autophagosomes.

Septins interact with autophagy proteinsTo ask whether septin localization at the PAS involves physicalinteractions between septins and known autophagy proteins, weemployed two parallel approaches. Immunoprecipitation of GFP-tagged septins using the GFP tag as an epitope resulted in the co-precipitation of untagged Atg8, and the amount of co-precipitatedAtg8 increased in the absence of Ypt7 (Fig. 3A), consistent with aprolonged interaction due to stabilization of autophagosomes.Negligible Atg8 was precipitated by the GFP antibody when GFPwas not fused to a septin (Fig. 3A). Furthermore, Cdc10 interactedin vivo with Atg9 in a bimolecular fluorescence complementation(BiFC) assay (Fig. 3B); other septins were not tested. In addition tosingle cytoplasmic puncta in starved cells, which we confirmed tobe the PAS because they colocalized with Ape1–RFP, a Cdc10–Atg9 BiFC signal was observed at the necks of budding cells(Fig. 3B,C; Movie 3). Since Atg8 and Atg9 are also involved in the‘cytoplasm to vacuole targeting’ (Cvt) pathway, a biosynthetic formof selective autophagy active even in rich medium conditions(Reggiori and Klionsky, 2013), these observations likely representotherwise transient associations between septins and the autophagymachinery during Cvt that are prolonged by the essentiallyirreversible BiFC event. The autophagy proteins are therebyartificially tethered to the septin ring at the bud neck, whichfacilitates detection of the BiFC event, but does not faithfully reporton where the interaction first took place. Taken together, thesefindings provide strong evidence that septins physically interactwith the core autophagy machinery both during Cvt and starvation-associated autophagy.

Septins are involved in autophagosome biogenesisTo determine at which stage of autophagosome assembly septinsfunction, we first examined septin-mutant cells to identify thestep in autophagosome formation that fails when septins aredysfunctional. We combined the cdc10(P3S G44D) mutationwith ypt7Δ, to block autophagosome degradation, and shiftedstarved cells to 37°C. We found a decrease in the number ofmCherry–Atg8 foci per cell compared to that in ypt7Δ cells withWTseptins (Fig. 3D,E), indicating that septin dysfunction perturbsautophagosome biogenesis, rather than delivery of autophagosomesto the vacuole and degradation. Next, we examined septin–GFP andmCherry–Atg8 localization in cells lacking Atg1, in which PASassembly begins but no mature autophagosomes are produced(Suzuki et al., 2007). In atg1Δ cells, colocalization between GFP-tagged septins and mCherry–Atg8 decreased significantly

Fig. 2. Septins colocalize with autophagosomes and form ring likestructures. (A) Septins colocalize with the mCherry–Atg8 (mCh-Atg8)-labeledPAS. Cells were grown as described in Fig. 1B and imaged. All images are of asingle z-section and are deconvolved. (B) Quantification of the number ofPASs that colocalize with septins. More than 300 cells were counted. Fromthese cells, only 20–30% cells showed a PAS and cells that showedcolocalization between the PAS and septin dot were quantified. Quantificationwas performed manually by using the cell counter plugin of Fiji at everyz-section. (C) Representative images showing colocalization of septins andautophagosomes (autophagosomes are highlighted by white arrowheads).Cells were grown as in Fig. 1B and were imaged. (D) Quantification of thenumber of cells showing multiple colocalizations between septins andmCherry–Atg8 puncta. For quantification, images were deconvolved andbackground subtracted, and then colocalization was checked by using thecolocalization highlighter plugin. Colocalized points were then quantified byusing the cell counter plugin of Fiji. More than 150 cells were quantified.(E) Colocalization of septins with Ape1–RFP. Cdc10-GFP, Cdc11–GFP andShs1–GFP ypt7Δ cells expressing Ape1–RFP were grown as in Fig. 1B andimaged. (F) Quantification of the number of septin puncta colocalized withApe1–RFP puncta. (G) Formation of a non-canonical ring around mCherry–Atg8 by Shs1–GFP. Cells were grown as in Fig. 1B and were imaged. Eightz-sections of the same image at 0.2 µm each are shown. Scale bars: 2 µm.

5

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience

Fig. 3. See next page for legend.

6

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience

compared to that in WT cells (Fig. 3F,G), suggesting that septinsarrive at the nascent autophagosome after the PAS has alreadybegun to assemble.The autophagy protein Atg9 accumulates at the PAS in atg1Δ cells

due to a defect in retrograde Atg9 transport to the sourcesof membrane trafficking (Sekito et al., 2009). In WT cells,Atg9–mCherry colocalized primarily with GFP-tagged septins(Fig. 4A,B). Interestingly, a single bright punctum of GFP–Atg9was also observed in cdc10(P3S G44D) cells incubated at 37°C(Fig. 4C). This punctum colocalized with the PAS marker Ape1(Fig. 4C–E), indicating that septin dysfunction prevents properAtg9 retrograde transport away from the PAS, a phenocopy of theabsence of Atg1. We further noticed that Ape1 was mislocalized inseptinmutant cells, as only a fewApe1–RFP puncta colocalized withGFP–Atg8 in cdc10(P3S G44D) cells incubated at 37°C (Fig. 4F,G).Taken together, these experiments point to a role for septin complexesin biogenesis of autophagosomes following PAS assembly.

DISCUSSIONThe original budding yeast mutants defective in autophagy wereidentified by unbiased genetic screens based on phenotypes offailure to accumulate ‘autophagic bodies’, to survive duringnitrogen starvation (Tsukada and Ohsumi, 1993) or to degradespecific cytoplasmic enzymes (Thumm et al., 1994), with anunderlying assumption that autophagy is non-essential for colonygrowth in rich medium. Another study systematically searched forautophagy defects in a collection of mutants harboringhypomorphic alleles of essential genes (Shirahama-Noda et al.,2013), but by definition these alleles provide sufficient function tosupport proliferation. We report here the identification of septinmutants in what is, to our knowledge, the first unbiased screen forautophagy (pexophagy) defects among conditionally lethalmutants. We observed severe defects under conditions non-permissive for proliferation, which explains why septin mutantswere not isolated in previous autophagy screens.Septins are functionally important for numerous processes

involving changes in membrane shape, ranging from cytokinesis(Hartwell, 1971; Kim et al., 2011) to mitochondrial fission(Pagliuso et al., 2016) to the retraction of membrane blebs(Gilden et al., 2012). Macroautophagy requires the synthesis anddirected extension of a double-bilayer ‘isolation membrane’ which,when its leading edges fuse together, becomes an autophagosome,and is destined for fusion with the vacuole or lysosome. In metazoancells, septins form cage-like assemblies around intracellular bacteriaand recruit autophagocytic machinery (Torraca and Mostowy,

2016), but these studies did not determine whether septinsparticipate in the assembly of autophagocytic membranestructures per se, and defects in general autophagy (i.e. notassociated with bacterial infection) have not been reported.

The Atg9 retrograde transport defects we found in septin mutantssuggest that septins may help Atg9 molecules to deliver membranesource for developing autophagosomes. Additionally, our findingsof dynamic septin localization to a subset of PAS structures afterinitial PAS formation, septin rings of diameters consistent with thoseof the autophagosomalmembranes, interactions between septins andthe autophagosomal membrane protein Atg9, and defects inautophagosome maturation in septin mutants are all consistentwith roles for septins in guiding isolation membrane extension. Inthis regard, the autophagy defects we observed in septin mutant cellsare reminiscent of defects in extension of theyeast PSM (HeasleyandMcMurray, 2016), another double-bilayer membrane that engulfscytoplasmic components prior to fusion of its leading edges.Notably, whereas proper septin function in PSM extension appearsto require the de novo assembly of hetero-octamers containing twosporulation-specific septin proteins (Garcia et al., 2016), ourfindings suggest that the same septin hetero-octamers assembledduring mitotic proliferation are sufficient to support autophagosomematuration. Emerging studies (Bridges et al., 2016) suggest that theability of rod-shaped septin hetero-oligomers to interact withmembranes of specific micron-scale curvatures and polymerizeinto filaments may be key to septin function in various contexts. Thedetails of interactions between septins and the established autophagymachinery, particularly membranes, and whether post-translationalmodifications to the septins drive their departure from the site ofcytokinesis are compelling subjects for future research.

MATERIALS AND METHODSYeast strains and mediaWild-type (WT) and autophagy knockout mutant yeast strains used in thisstudy are derived from BY4741, BY4742, and S288C. These strains wereobtained from EUROSCARF. Strain and primer details are listed in TablesS1 and S2, respectively. The WT Pot1–GFP strains are laboratory strainswith GFP tagged genomically to the C-terminus of Pot1, and were obtainedfrom Prof. Richard Rachubinski, University of Alberta, Canada. Septin Ts−

mutants were kindly provided by Prof. Charlie Boone, Toronto, into which aPOT1-GFP cassette was transformed to obtain strains used for pexophagyassays. Septin knockout mutants were prepared using the standardtransformation protocols (Baudin et al., 1993). GFP-ATG8 pRS316 and2xmCherry-ATG8 pRS316 plasmids were a kind gift from Prof. YoshinoriOhsumi, Tokyo Institute of Technology, Tokyo. GFP-Atg9 pRN295 was akind gift from Prof. Michael Thumm, University of Stuttgart, Germany.Vector pSUN5 was created by amplifying Atg9 promoter and the open-reading frame (ORF) from a WT strain and cloned into the pRS316vector atthe SacII and NotI sites. The tandem repeat of mCherry separated by a 45 bplinker region was cloned in two steps between the Not1, XmaI and HindIIIsites. A linker region of 13 bp was also included between the Atg9 ORF andstart codon of mCherry.

WT cells and mutants were grown in YPD medium (1% yeast extract,2% peptone and 2% dextrose) at 30°C and 22°C, respectively. Forpexophagy assays, oleate medium (0.25% yeast extract, 0.5% peptone,1% oleate, 5% Tween-40 and 5 mM phosphate buffer) was used to induceperoxisome formation, and synthetic defined medium (0.17% yeast nitrogenbase lacking amino acids and ammonium sulphate plus 2% dextrose) wasused to induce autophagy. For Ts−mutants and knockout mutants, 22°C and37°C were used as the permissive and non-permissive temperatures,respectively.

Microscopy and cytologyCells were grown in respective media, centrifuged (20817 g for 2 min).and were mounted on agarose (2% w/v) pads for microscopy. Images

Fig. 3. Septins are involved in autophagosomes biogenesis. (A) Westernblot showing the septin–Atg8 interaction in WT and in ypt7Δ strains expressingCdc10–GFP, Cdc11–GFP and Shs1–GFP. Cells were grown as described inthe Materials and Methods. IB, immunoblot; IP, immunoprecipitation.(B,C) BiFC experiments. A strain expressing Cdc10-Vc (Cdc10 C-terminustagged with C-terminus of Venus) and Atg9-Vn (Atg9 C-terminus tagged withN-terminus of Venus) with or without Ape1–RFP was grown as described inFig. 1B and imaged after 5 h of incubation in starvation medium.(D) Representative images and (E) quantification showing autophagosomenumber per cell at 37°C. All the images aremaximum intensity projections, andmore than 50 cells were quantified manually with Fiji. *P<0.05 (comparisonbetween non-Ts and Ts at 37°C); **P<0.01 (comparison between 22°C and37°C in Ts) (two-way ANOVA). (F) Representative images and(G) quantification of colocalization events between mCherry–Atg8 and thethree GFP-tagged septins in the atg1Δ strain. A total of 50 cells were quantifiedmanually at every z-plane. *P<0.05 for Cdc10–GFP, **P<0.01 for Cdc11–GFPand Shs1–GFP (two-way ANOVA). Scale bars: 2 µm (C,F); 5 µm (B,D).

7

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience

were taken in z-sections of 0.2 µm step size using a Delta Visionmicroscope (GE Healthcare) fitted with 100×1.4 NA objective and Cool-SNAP HQ2 camera. Images were acquired using FITC and TRITCfilters. Image processing and quantification were performed withSoftWorx (GE Healthcare) and Fiji (NIH) software. For colocalizationanalysis, images were de-convolved and background subtracted, and

colocalized entities were either quantified manually by using the cellcounter plugin or automatically by using the colocalization highlighterplugin in Fiji for all z-sections. Representative colocalizationsevents quantified manually were also confirmed by line profile andcolocalization measurement options in SoftWorx (GE Healthcare). Allthe supplementary movies are of a single z-plane.

Fig. 4. See next page for legend.

8

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience

Western blot analysis and quantificationWhole cell extracts were prepared via the trichloroacetic acid (12.5% TCAw/v) precipitation method followed by ice-cold acetone washes (twice).Protein extracts were then analyzed by SDS-PAGE and western blotting(mouse anti-GFP monoclonal 1:3000, cat. no. 11814460001, RocheApplied Science). Blots were visualized using anti-mouse-IgG secondaryantibody conjugated to horseradish peroxidase (HRP; Bio-Rad, 1:10,000)on a gel documentation system (G:Box chemi XT4, Syngene). Images wereanalyzed using ImageJ (NIH). Lanes were marked, followed by plotting andlabeling peaks using the analysis tool for gels. The ratios of the intensity ofthe free GFP band to total GFP (to either the Pot1–GFP plus GFP band, orGFP–Atg8 plus GFP) band was quantified and plotted as the percentage ofcleaved product.

Pexophagy assayCells were grown in YPD and 0.2 optical density (OD) units (measuredat 600 nm) was inoculated in fresh YPD medium. To induceperoxisome formation, cells were then incubated in oleate medium(1 OD/ml) for 14 to 16 h. Cells were then washed twice with sterilewater and SD-N medium was added (3 OD/ml) to induced pexophagy.Time points were collected for western blot analysis. For microscopy,the 6-h time point was collected and FM4-64 (1 µl/ml of 1 mg/mlstock) was added to stain the vacuoles at respective temperatures. Cellswere then imaged.

ImmunoprecipitationWT and ypt7Δ expressing Cdc10–GFP, Cdc11–GFP or Shs1–GFP togetherwith mCherry–Atg8 were grown in SD-Ura medium. After the culturesreached 0.6–0.8 OD, 400 OD cells were transferred to starvation medium(3 OD/ml) and were incubated for 5 h at 30°C. After 5 h, 100 OD cellswere lysed as per the protocol mentioned in Nagaraj et al. (2008).For immunoprecipitations, 15 µl of GFP-trap beads (Chromotek) wereadded and the manufacturer’s protocol was followed. After theimmunoprecipitation, the sample containing the beads was heated andloaded on the gel for SDS-PAGE followed by western blotting. Blots werefirst probed with either mouse anti-GFP (1:3000, cat. no. 11814460001,Roche Applied Science) or rabbit anti-Atg8 antibody (1:3000, a kindgift from Prof. Yoshinori Ohsumi) and then probed with secondary anti-mouse-IgG (Bio-Rad, 1:10,000) and anti-rabbit-IgG (Bio-Rad, 1:10,000)antibodies conjugated to HRP and were developed on a gel documentationsystem (G: Box chemi XT4, Syngene).

StatisticsAll statistical analysis was performed using GraphPad Prism. To calculatesignificance levels two-way ANOVA and Student’s t-tests were used. Themean±s.e.m. is shown in all graphs.

AcknowledgementsWe would like to thank Prof. Charles Boone, Prof. Yoshinori Ohsumi, Prof. KausikChakraborthy, Prof. Michael Snyder, Prof. Michael Thumm, Prof. RichardRachubinski, Prof. Jeremy Thorner, for generously sharing strains, plasmids andreagents. Critical reading of the manuscript and inputs from Prof. M. R. S. Rao,Aparna Hebbar, and members of the autophagy lab. We thank Veena A. andPriyadarshini Sanyal for technical help.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsConceptualization: R.M., G.B., S. Sridhar, A.A., M.A.M.; Methodology: G.B.,S. Sridhar, A.A., M.H.S., S.C., S. Singh, L.K.N.; Software: G.B.; Validation: G.B.;Formal analysis: G.B., S. Sridhar, A.A., M.H.S., S.C., S. Singh, L.K.N., M.A.M.;Investigation: R.M.; Resources: R.M., M.A.M.; Data curation: R.M., M.A.M.; Writing -original draft: R.M., G.B., S. Sridhar; Writing - review & editing: R.M., G.B., M.A.M.;Visualization: R.M., G.B., A.A., M.A.M.; Supervision: R.M.; Project administration:R.M.; Funding acquisition: R.M.

FundingThis work was supported by a Wellcome Trust DBT India Alliance IntermediateFellowship (5009159-Z-09-Z) and Jawaharlal Nehru Centre for Advanced ScientificResearch intramural funds to R.M. S. Sridhar was supported by a fellowship from theCouncil of Scientific and Industrial Research (CSIR). Deposited in PMC forimmediate release.

Supplementary informationSupplementary information available online athttp://jcs.biologists.org/lookup/doi/10.1242/jcs.209098.supplemental

ReferencesBaba, M., Takeshige, K., Baba, N. and Ohsumi, Y. (1994). Ultrastructural analysis

of the autophagic process in yeast: detection of autophagosomes and theircharacterization. J. Cell Biol. 124, 903-913.

Baudin, A., Ozier-Kalogeropoulos, O., Denouel, A., Lacroute, F. and Cullin, C.(1993). A simple and efficient method for direct gene deletion in Saccharomycescerevisiae. Nucleic Acids Res. 21, 3329-3330.

Bridges, A. A., Jentzsch, M. S., Oakes, P. W., Occhipinti, P. and Gladfelter, A. S.(2016). Micron-scale plasma membrane curvature is recognized by the septincytoskeleton. J. Cell Biol. 213, 23-32.

Chan, T. F., Carvalho, J., Riles, L. and Zheng, X. F. S. (2000). A chemicalgenomics approach toward understanding the global functions of the target ofrapamycin protein (TOR). Proc. Natl. Acad. Sci. USA 97, 13227-13232.

Cheong, H. and Klionsky, D. J. (2008). Biochemical methods to monitorautophagy-related processes in yeast. Methods Enzymol. 451, 1-26.

Costanzo, M., Baryshnikova, A., Bellay, J., Kim, Y., Spear, E. D., Sevier, C. S.,Ding, H., Koh, J. L. Y., Toufighi, K., Mostafavi, S. et al. (2010). The geneticlandscape of a cell. Science 327, 425-431.

Davey, H. M., Cross, E. J. M., Davey, C. L., Gkargkas, K., Delneri, D., Hoyle,D. C., Oliver, S. G., Kell, D. B. and Griffith, G. W. (2012). Genome-wide analysisof longevity in nutrient-deprived Saccharomyces cerevisiae reveals importance ofrecycling in maintaining cell viability. Environ. Microbiol. 14, 1249-1260.

Delorme-Axford, E., Guimaraes, R. S., Reggiori, F. and Klionsky, D. J. (2015).The yeast Saccharomyces cerevisiae: an overviewof methods to study autophagyprogression. Methods San Diego Calif 75, 3-12.

Dobbelaere, J., Gentry, M. S., Hallberg, R. L. and Barral, Y. (2003).Phosphorylation-dependent regulation of septin dynamics during the cell cycle.Dev. Cell 4, 345-357.

Feng, Y., He, D., Yao, Z. and Klionsky, D. J. (2014). The machinery ofmacroautophagy. Cell Res. 24, 24-41.

Finnigan, G. C., Takagi, J., Cho, C. and Thorner, J. (2015). ComprehensiveGenetic Analysis of Paralogous Terminal Septin Subunits Shs1 and Cdc11 inSaccharomyces cerevisiae. Genetics 200, 821-841.

Flescher, E. G., Madden, K. and Snyder, M. (1993). Components required forcytokinesis are important for bud site selection in yeast. J. Cell Biol. 122, 373-386.

Frazier, J. A., Wong, M. L., Longtine, M. S., Pringle, J. R., Mann, M., Mitchison,T. J. and Field, C. (1998). Polymerization of purified yeast septins: evidence thatorganized filament arraysmay not be required for septin function. J. Cell Biol. 143,737-749.

Fig. 4. Septins colocalize with Atg9 and play a role in Atg9 retrogradetransport. (A) Representative images showing colocalization between septinsand Atg9. Cdc10–GFP, Cdc11–GFP and Shs1–GFP cells expressingAtg9–mCherry (Atg9-mCh) were grown as in Fig. 1B and were imaged.(B) Quantification of the number of cells showing colocalization betweenseptins and Atg9–mCherry puncta. Quantification was performed as inFig. 3G. More than 150 cells were quantified. (C) Atg9 retrograde transport isaffected in the cdc10-5 strain at 37°C. Cells were grown in starvation mediumfor 6 h and were imaged. (D) Quantification of the number of cells showing abright Atg9 punctum at 22°C (permissive temperature) and 37°C (non-permissive temperature). Quantification was performed manually by using Fijisoftware, and a total of 50 cells were quantified in each of the threeexperiments. ***P<0.001, 22°C versus 37°C inWT and cdc10-5 cells (two-wayANOVA). (E) Quantification of the number of cells showing colocalizationbetween the bright GFP–Atg9 punctum and Ape1–RFP. Quantification wasperformed manually by using Fiji software at each z-section, and a total of 30cells were quantified in each of the three experiments. ***P<0.001, atg1Δversus cdc10-5 cells (unpaired t-test.). (F) Colocalization between GFP–Atg9and mCherry–Atg8, and GFP–Atg8 and Ape1–RFP. The cdc10-5 cellsexpressing either GFP–Atg9 with mCherry–Atg8 or GFP–Atg8 with Ape1–RFP were grown in SD Ura or SD −His −Ura medium at 22°C. Logarithmicallygrowing cells were then incubated in starvation medium (1 OD/ml) for 3 h at22°C and 37°C. (G) Quantitation of the colocalization of GFP–Atg9 andmCherry–Atg8 puncta with GFP–Atg8 and Ape1–RFP puncta. 30 cells werequantified in each of the three experiments. **P<0.01 (paired t-test). Scalebars: 2 µm (A,F); 5 µm (C).

9

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience

Garcia, G., Finnigan, G. C., Heasley, L. R., Sterling, S. M., Aggarwal, A.,Pearson, C. G., Nogales, E., McMurray, M. A. and Thorner, J. (2016).Assembly, molecular organization, and membrane-binding properties ofdevelopment-specific septins. J. Cell Biol. 212, 515-529.

Gilden, J. K., Peck, S., Chen, Y.-C. M. and Krummel, M. F. (2012). The septincytoskeleton facilitates membrane retraction during motility and blebbing. J. CellBiol. 196, 103-114.

Glomb, O. and Gronemeyer, T. (2016). Septin organization and functions inbudding yeast. Front. Cell Dev. Biol. 4, 123.

Graef, M., Friedman, J. R., Graham, C., Babu, M. and Nunnari, J. (2013). ER exitsites are physical and functional core autophagosome biogenesis components.Mol. Biol. Cell 24, 2918-2931.

Hartwell, L. H. (1971). Genetic control of the cell division cycle in yeast. IV. Genescontrolling bud emergence and cytokinesis. Exp. Cell Res. 69, 265-276.

Heasley, L. R. andMcMurray, M. A. (2016). Roles of septins in prosporemembranemorphogenesis and sporewall assembly in Saccharomyces cerevisiae.Mol. Biol.Cell 27, 442-450.

Huang, Y.-W., Yan, M., Collins, R. F., Diciccio, J. E., Grinstein, S. and Trimble,W. S. (2008). Mammalian septins are required for phagosome formation.Mol. Biol.Cell 19, 1717-1726.

Ishihara, N., Hamasaki, M., Yokota, S., Suzuki, K., Kamada, Y., Kihara, A.,Yoshimori, T., Noda, T. and Ohsumi, Y. (2001). Autophagosome requiresspecific early Sec proteins for its formation and NSF/SNARE for vacuolar fusion.Mol. Biol. Cell 12, 3690-3702.

Kim, H., Haarer, B. and Pringle, J. (1991). Cellular morphogenesis in theSaccharomyces cerevisiae cell cycle: localization of the CDC3 gene product andthe timing of events at the budding site. J. Cell Biol. 112, 535-544.

Kim, J., Dalton, V. M., Eggerton, K. P., Scott, S. V. and Klionsky, D. J. (1999).Apg7p/Cvt2p is required for the cytoplasm-to-vacuole targeting, macroautophagy,and peroxisome degradation pathways. Mol. Biol. Cell 10, 1337-1351.

Kim, M. S., Froese, C. D., Estey, M. P. and Trimble, W. S. (2011). SEPT9 occupiesthe terminal positions in septin octamers and mediates polymerization-dependentfunctions in abscission. J. Cell Biol. 195, 815-826.

Kondo-Okamoto, N., Noda, N. N., Suzuki, S. W., Nakatogawa, H., Takahashi, I.,Matsunami, M., Hashimoto, A., Inagaki, F., Ohsumi, Y. and Okamoto, K.(2012). Autophagy-related protein 32 acts as autophagic degron and directlyinitiates mitophagy. J. Biol. Chem. 287, 10631-10638.

Legakis, J. E., Yen, W.-L. and Klionsky, D. J. (2007). A cycling protein complexrequired for selective autophagy. Autophagy 3, 422-432.

McMurray, M. A. and Thorner, J. (2008). Septin stability and recycling duringdynamic structural transitions in cell division and development. Curr. Biol. CB 18,1203-1208.

McMurray, M. A., Bertin, A., Garcia, G., III, Lam, L., Nogales, E. and Thorner, J.(2011). Septin filament formation is essential in budding yeast. Dev. Cell 20,540-549.

Mino, A., Tanaka, K., Kamei, T., Umikawa, M., Fujiwara, T. and Takai, Y. (1998).Shs1p: a novel member of septin that interacts with spa2p, involved in polarizedgrowth in saccharomyces cerevisiae. Biochem. Biophys. Res. Commun. 251,732-736.

Mostowy, S. and Cossart, P. (2012). Septins: the fourth component of thecytoskeleton. Nat. Rev. Mol. Cell Biol. 13, 183-194.

Mostowy, S., NamTham, T., Danckaert, A., Guadagnini, S., Boisson-Dupuis, S.,Pizarro-Cerda, J. and Cossart, P. (2009). Septins regulate bacterial entry intohost cells. PLoS ONE 4, e4196.

Mostowy, S., Bonazzi, M., Hamon, M. A., Tham, T. N., Mallet, A., Lelek, M.,Gouin, E., Demangel, C., Brosch, R., Zimmer, C. et al. (2010). Entrapment ofintracytosolic bacteria by septin cage-like structures. Cell Host Microbe 8,433-444.

Mostowy, S., Sancho-Shimizu, V., Hamon, M. A., Simeone, R., Brosch, R.,Johansen, T. and Cossart, P. (2011). p62 and NDP52 proteins targetintracytosolic Shigella and Listeria to different autophagy pathways. J. Biol.Chem. 286, 26987-26995.

Nagaraj, S., Rajendran, A., Jackson, C. E. and Longtine, M. S. (2008). Role ofnucleotide binding in septin-septin interactions and septin localization inSaccharomyces cerevisiae. Mol. Cell. Biol. 28, 5120-5137.

Neiman, A. M. (2005). Ascospore formation in the yeast Saccharomyces cerevisiae.Microbiol. Mol. Biol. Rev. MMBR 69, 565-584.

Neiman, A. M. (2011). Sporulation in the budding yeast Saccharomyces cerevisiae.Genetics 189, 737-765.

Noda, T. and Ohsumi, Y. (1998). Tor, a phosphatidylinositol kinase homologue,controls autophagy in yeast. J. Biol. Chem. 273, 3963-3966.

Oh, Y. and Bi, E. (2011). Septin structure and function in yeast and beyond. TrendsCell Biol. 21, 141-148.

Okada, S., Leda, M., Hanna, J., Savage, N. S., Bi, E. andGoryachev, A. B. (2013).Daughter cell identity emerges from the interplay of Cdc42, septins, andexocytosis. Dev. Cell 26, 148-161.

Oku, M. and Sakai, Y. (2016). Pexophagy in yeasts. Biochim. Biophys. Acta 1863,992-998.

Pagliuso, A., Tham, T. N., Stevens, J. K., Lagache, T., Persson, R., Salles, A.,Olivo-Marin, J.-C., Oddos, S., Spang, A., Cossart, P. et al. (2016). A role forseptin 2 in Drp1-mediated mitochondrial fission. EMBO Rep. 17, 858-873.

Reggiori, F. and Klionsky, D. J. (2013). Autophagic processes in yeast:mechanism, machinery and regulation. Genetics 194, 341-361.

Reggiori, F., Tucker, K. A., Stromhaug, P. E. and Klionsky, D. J. (2004). TheAtg1-Atg13 complex regulates Atg9 and Atg23 retrieval transport from the pre-autophagosomal structure. Dev. Cell 6, 79-90.

Sekito, T., Kawamata, T., Ichikawa, R., Suzuki, K. and Ohsumi, Y. (2009). Atg17recruits Atg9 to organize the pre-autophagosomal structure.Genes Cells DevotedMol. Cell. Mech. 14, 525-538.

Shirahama-Noda, K., Kira, S., Yoshimori, T. and Noda, T. (2013). TRAPPIII isresponsible for vesicular transport from early endosomes to Golgi, facilitating Atg9cycling in autophagy. J. Cell Sci. 126, 4963-4973.

Sirianni, A., Krokowski, S., Lobato-Marquez, D., Buranyi, S., Pfanzelter, J.,Galea, D., Willis, A., Culley, S., Henriques, R., Larrouy-Maumus, G. et al.(2016). Mitochondria mediate septin cage assembly to promote autophagy ofShigella. EMBO Rep. 17, 1029-1043.

Suzuki, K. and Ohsumi, Y. (2010). Current knowledge of the pre-autophagosomalstructure (PAS). FEBS Lett. 584, 1280-1286.

Suzuki, K., Kubota, Y., Sekito, T. andOhsumi, Y. (2007). Hierarchy of Atg proteinsin pre-autophagosomal structure organization. Genes Cells Devoted Mol. Cell.Mech. 12, 209-218.

Suzuki, S. W., Onodera, J. andOhsumi, Y. (2011). Starvation induced cell death inautophagy-defective yeast mutants is caused by mitochondria dysfunction. PLoSONE 6, e17412.

Thumm, M., Egner, R., Koch, B., Schlumpberger, M., Straub, M., Veenhuis, M.andWolf, D. H. (1994). Isolation of autophagocytosis mutants of Saccharomycescerevisiae. FEBS Lett. 349, 275-280.

Torraca, V. and Mostowy, S. (2016). Septins and bacterial infection. Front. CellDev. Biol. 4, 127.

Tsukada, M. and Ohsumi, Y. (1993). Isolation and characterization of autophagy-defective mutants of Saccharomyces cerevisiae. FEBS Lett. 333, 169-174.

Tucker, K. A., Reggiori, F., Dunn, W. A. and Klionsky, D. J. (2003). Atg23 isessential for the cytoplasm to vacuole targeting pathway and efficient autophagybut not pexophagy. J. Biol. Chem. 278, 48445-48452.

Weems, A. D., Johnson, C. R., Argueso, J. L. and McMurray, M. A. (2014).Higher-Order Septin Assembly Is Driven by GTP-Promoted ConformationalChanges: Evidence From Unbiased Mutational Analysis in Saccharomycescerevisiae. Genetics 196, 711-727.

Yamamoto, H., Kakuta, S., Watanabe, T. M., Kitamura, A., Sekito, T., Kondo-Kakuta, C., Ichikawa, R., Kinjo, M. and Ohsumi, Y. (2012). Atg9 vesicles are animportant membrane source during early steps of autophagosome formation.J. Cell Biol. 198, 219-233.

10

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs209098. doi:10.1242/jcs.209098

Journal

ofCe

llScience