ACSL3 Is a Novel GABARAPL2 Interactor That Links Ufmylation And

bioRxiv preprint doi: https://doi.org/10.1101/2020.01.01.892521; this version posted January 1, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 ACSL3 is a novel GABARAPL2 interactor that links ufmylation and

2 lipid droplet biogenesis

4 Franziska Eck1, Manuel Kaulich2, and Christian Behrends1

6 1Munich Cluster for Systems Neurology (SyNergy), Medical Faculty, Ludwig-Maximilians-

7 University München, Feodor-Lynen Strasse 17, 81377 Munich, Germany

8 2Institute of Biochemistry II, Goethe University School of Medicine, Theodor-Stern-Kai 7, 60590

9 Frankfurt am Main, Germany

11 Correspondence to: [email protected]

22 Abstract

23 While studies of ATG genes in knockout models led to an explosion of knowledge about the

24 functions of autophagy components, the exact roles of LC3/GABARAP proteins are still poorly

25 understood. A major drawback for their understanding is that the available interactome data

26 was largely acquired using overexpression systems. To overcome these limitations, we

27 employed CRISPR/Cas9-based genome-editing to generate a panel of cells in which human

28 ATG8 genes were tagged at their natural chromosomal locations with an N-terminal affinity

29 epitope. This cellular resource was exemplarily employed to map endogenous GABARAPL2

30 protein complexes in response to autophagic modulation using interaction proteomics. This

31 approach identified the ER transmembrane protein and lipid droplet biogenesis factor ACSL3

32 as a stabilizing GABARAPL2-binding partner. Through this interaction, the GABARAPL2-

33 interacting protein and UFM1-activating enzyme UBA5 becomes anchored at the ER

34 membrane. Functional analysis unveiled ACSL3 and lipid droplet formation as novel regulators

35 of the enigmatic UFM1 conjugation pathway.

45 Introduction

46 From yeast to humans ATG8s are highly conserved proteins. While there is only a single Atg8

47 in yeast, the human ATG8 (hATG8) family is subdivided into the orthologs microtuble-

48 associated protein 1A/1B light chain 3 (MAP1LC3) including LC3A, LC3B, and LC3C as well

49 as -aminobutyric acid receptor-associated protein (GABARAP) including GABARAP,

50 GABARAPL1 and GABARAPL2 (1). All six hATG8 proteins share the same, ubiquitin-like fold

51 although they do not exhibit any sequence homologies with ubiquitin. However, within and

52 between the ATG8 subfamilies, the amino acid sequences show high similarities (2). A major

53 feature of LC3 and GABARAP proteins is their covalent conjugation to the phospholipid

54 phosphatidylethanolamine (PE). This process is initiated by the cysteine proteases ATG4A-D

55 that cleave all hATG8 family members to expose a C-terminal glycine residue and is followed

56 by the activation of LC3s and GABARAPs through the E1-like activating enzyme ATG7. PE-

57 conjugation of hATG8 proteins is subsequently accomplished in a concerted action of the E2-

58 like conjugating enzyme ATG3 and the E3-like ligase scaffold complex ATG12-ATG5-

59 ATG16L1. PE-hATG8 conjugates are reversible through cleavage by ATG4A-D (3).

60 The best understood function of hATG8s is in macroautophagy (hereafter referred to as

61 autophagy) which is a highly conserved degradation pathway that eliminates defective und

62 unneeded cytosolic material and is rapidly upregulated by environmental stresses such as

63 nutrient deprivation. In the past years, it was shown that autophagy is capable of selectively

64 recognizing and engulfing divers cargo such as aggregated proteins (aggrephagy), pathogens

65 (xenophagy) or mitochondria (mitophagy) with the help of specific receptor proteins (4).

66 Initiation of autophagy leads to the formation of phagophores (also called isolation

67 membranes) from preexisting membrane compartments, such as the ER. Elongation and

68 closure of isolation membranes leads to engulfment of cargo inside double membrane vesicles

69 termed autophagosomes. Fusion of autophagosomes with lysosomes forms autolysosomes in

70 which captured cargo is degraded in bulk by lysosomal hydrolases (5). During this process,

71 GABARAPs and LC3s are associated with the outer and inner membrane of phagophores and

72 regulate membrane expansion (6), cargo receptor recruitment (7), closure of phagophores (8)

73 and the fusion of autophagosomes with lysosomes (9).

74 Besides autophagy, GABARAPs and LC3s are implicated in a number of other cellular

75 pathways. For example, GABARAP was found as interactor of the GABA receptor and involved

76 in its intracellular transport to the plasma membrane (10, 11), while GABARAPL2 was identified

77 as modulator of Golgi reassembly and intra-Golgi trafficking (12, 13). GABARAPs were also

78 found as essential scaffolds for the ubiquitin ligase CUL3KBTBD6/KBTBD7 (14). Among others, LC3s

79 have regulatory functions in RhoA dependent actin cytoskeleton reorganization (15) as well as

80 in the regulation of ER exit sites (ERES) and COPII-dependent ER-to-Golgi transport (16). This

81 high functional diversity of GABARAPs and LC3s implies that these proteins are more than

82 autophagy pathway components and that there are possible other unique functions of

83 individual hATG8 proteins to be unraveled.

84 So far, interactome and functional analyses of LC3s and GABARAPs were mostly done in cells

85 overexpressing one of the six hATG8 family members (17, 18). This raises the concern that

86 an overexpressed hATG8 protein might take over functions or interactions of one of the other

87 family members due to their high sequential and structural similarity. A lack of isoform specific

88 antibodies further complicates the analysis of distinct functions of hATG8s. To facilitate the

89 study of endogenous GABARAPs and LC3s, it is important to generate alternative resources

90 and tools such as the multiple hATG8 knockout cell lines (9) or the hATG8 family member-

91 specific peptide sensors (19). To circumvent the hATG8 antibody problem, we used

92 CRISPR/Cas9 technology to seamlessly tag hATG8 genes at their natural chromosomal

93 locations. The generated cell lines express N-terminally hemagglutinin (HA)-tagged hATG8

94 family members at endogenous levels and are a powerful tool to study the functions of

95 individual GABARAPs and LC3s. All created cell lines were tested for their correct sequence

96 and functionality. As a proof of concept, we performed interaction proteomics with the

97 GABARAPL2endoHA cell line and characterized the interaction with the novel binding partner

98 ACSL3.

99 Results

100 Establishment of cells carrying endogenously HA tagged LC3s and GABARAPs

101 Complementary to our previously reported LC3CendoHA HeLa cell line (20) we sought to employ

102 CRISPR-mediated gene-editing to generate a panel of cells in which the remaining five hATG8

103 family members are seamlessly epitope tagged at their natural chromosomal locations. To this

104 end, we directed Cas9 to cleave DNA at the vicinity of the start codon of LC3 and GABARAP

105 genes in order to stimulate microhomology-mediated integration of a sequence encoding for a

106 single HA-tag using a double-stranded DNA donor molecule containing short homology arms

107 (21). Briefly, we designed PCR homology templates in which the blasticidine resistant gene, a

108 P2A sequence and the open reading frame of the HA-tag were flanked by homology arms to

109 the 5’UTRs and first exons of the LC3/GABARAP genes (Supplementary Figure S1A). In

110 parallel, we designed single guide RNAs (sgRNAs) for all hATG8 genes except LC3C and

111 cloned them into pX330 (Addgene 42230), a SpCas9 expressing vector (Supplementary

112 Figure S1A). We then transfected HeLa cells with corresponding pairs of homology template

113 and sgRNA for each LC3/GABARAP gene. After selection with blasticidine, single cell clones

114 were SANGER sequenced to confirm seamless and locus-specific genomic insertion of the

115 HA-tag. While we obtained correct clones for GABARAP, GABARAPL1, GABARAPL2 and

116 LC3B (Supplementary Figure S1B), cells that received the homology template and gRNA for

117 LC3A did not survive the antibiotic selection. We assume that this is due to the lack of LC3A

118 in HeLa cells as it is reported that LC3A expression is suppressed in many tumor cell lines

119 (22). Immunoblot analysis of the sequence-validated clones and the parental cells revealed

120 the presence of the HA-tag in the generated cell lines that corresponded to the size of the

121 tagged LC3/GABARAP protein (Figure 1A, Supplementary Figure S2A-C). Gene specific

122 CRISPR/Cas9-editing was further confirmed by RNAi-mediated depletion of endogenous

123 LC3B or GABARAP proteins in the corresponding HA-tagged hATG8 cell lines (Figure 1B,

124 Supplementary Figure S2D-F). Consistently, confocal microscopy of GABARAPL2endoHA cells

125 showed a substantially decreased HA immunolabeling upon knockdown of GABARAPL2

126 (Figure 1C). Next, we examined the integrity of the tagged LC3/GABARAP proteins by

127 monitoring their conjugation to PE in response to treatment with small molecule inhibitors which

128 either increase lipidation (Torin1), block autophagosomal degradation (Bafilomycin A1

129 (BafA1)) or prevent ATG8-PE conjugate formation (ATG7 inhibitor). As expected,

130 GABARAPL2endoHA, GABARAPendoHA and LC3BendoHA cell lines showed treatment-specific

131 lipidation levels of the respective tagged hATG8 protein (Figure 1D; Supplementary Figure

132 S2G-I). We also detected lipidated GABARAPL1, though in a manner that was independent

133 from induction or blockage of autophagy (Figure 4C). However, autophagy induction robustly

134 decreased HA-GABARAPL1 protein levels in GABARAPL1endoHA cells while blockage of

135 autophagosomal degradation led to the opposite phenotype (Supplementary Figure S2H).

136 Next, we analyzed the subcellular distribution of one of the HA-tagged hATG8 proteins (i.e.

137 GABARAPL2) in basal and autophagy-modulating conditions using confocal microscopy. In

138 GABARAPL2endoHA cells, HA-GABARAPL2 was indeed found to colocalize with the

139 autophagosomal and -lysosomal markers p62, LC3B and LAMP1 and this colocalization

140 increased upon combination treatment with Torin1 and BafA1 (Figure 1E, Supplementary

141 Figure S2J,K). Together, we successfully engineered cell lines to carry epitope tagged hATG8

142 family members which retain their functionality.

143

144 Mapping the endogenous GABARAPL2 interactome

145 Next, we selected GABARAPL2endoHA cells for a proof-of-principle immunoprecipitation (IP)

146 followed by mass spectrometric (MS) analysis to identify new binding partner candidates of a

147 hATG8 family member at endogenous levels. To distinguish between candidates that bind

148 preferentially to PE-conjugated versus unconjugated GABARAPL2 we treated stable isotope

149 labeling with amino acids in cell culture (SILAC)-labeled GABARAPL2endoHA cells with Torin1

150 and BafA1 (light) or ATG7 inhibitor (heavy). Equal amounts of heavy and light SILAC cells

151 were mixed, lysed and subjected to HA-IP. Immune complexes were eluted and size separated

152 by gel electrophoresis followed by in-gel tryptic digest, peptide extraction and desalting prior 6

153 to analysis by liquid chromatography tandem MS. SILAC labeled parental HeLa cells

154 differentially treated with Torin1/BafA1 or ATG7 inhibitor served as negative controls. In

155 duplicate experiments, we identified a total of 168 proteins whose abundances in GABARAPL2

156 immunoprecipitates were altered by at least 2.8-fold (log2 SILAC ratio ≥1.5 or ≤-1.5) in

157 response to modulation of the hATG8 conjugation status (Figure 1F). Among these regulated

158 proteins were well-characterized hATG8 binding proteins such as ATG7, CCPG1 and

159 SQSTM1 (also known as p62) as well as several candidate interaction proteins previously

160 found in large-scale screening efforts such as the mitochondrial outer membrane protein

161 VDAC1, the nucleoprotein AHNAK2, the translation initiation factor EIF4G1 and the small

162 GTPase IRGQ (23, 24) (Figure 1F). In addition, a number of known hATG8 interactors

163 including UBA5, HADHA, HADHB, RB1CC1, TRIM21 and IPO5 was found to bind

164 GABARAPL2 independent of its lipidation status since these proteins did not display

165 substantial changes in their SILAC ratios.

166

167 ACSL3 is a novel binding partner of GABARAPL2

168 Since functional annotation analysis using DAVID revealed ‘fatty acid metabolism’ as a term

169 previously not associated with LC3/GABARAP-interacting proteins (Supplementary Figure

170 S3A), we focused on the proteins found in this category. In particular, the long-chain-fatty-acid-

171 CoA ligase 3 (ACLS3) attracted our attention as it was the only ER-localized transmembrane

172 protein among these candidates. To validate ACSL3 as novel GABARAPL2 interacting protein,

173 we transiently transfected parental and GABARAPL2endoHA cells with C-terminally myc-tagged

174 ACSL3 followed by HA- or myc-IP and immunoblotting. Transfection with N-terminally tagged

175 ATG7 or p62 served as positive controls. Indeed, ACSL3-myc as well as myc-ATG7 and -p62

176 associated with endogenous GABARAPL2 (Figure 1G-I, Supplementary Figure S3B). Thus,

177 these results indicate that our hATG8endoHA cells are valuable tools to examine the LC3 and

178 GABARAP interactome at endogenous levels and to identify novel binding partners such as

179 ACSL3. 7

180 ACSL3 recruits GABARAPL2 to the ER membrane

181 ACSL3 is one of five acyl-CoA synthetases and catalysis the conjugation of CoA to long chain

182 fatty acids to form acyl-CoA (25). Besides ACSL3 was found to regulate the formation, the size

183 and the copy number of lipid droplets (26, 27). Consistent with its cellular role, ACSL3 is

184 located with its N-terminal transmembrane helix region inserted midway into the lipid bilayer of

185 the ER membrane or integrated into the monolayer of lipid droplets (LD) while its C-terminal

186 part encompassing the AMP-binding domain is facing to the cytoplasm (28-30). To further

187 validate the GABARAPL2-ACSL3 interaction, we sought to examine the subcellular

188 localization of both proteins by confocal microscopy. However, as there were no suitable

189 antibodies for immunofluorescence staining of endogenous ACSL3, we gene-edited

190 GABARAPL2endoHA cells to express ACSL3 tagged at its C-terminus with NeonGreen

191 (Supplementary Figure S1A,C). Immunoblot analysis of these newly established

192 GABARAPL2endoHA/ACSL3endoNeonGreen cells in comparison with GABARAPL2endoHA and parental

193 Hela cells transfected with TOMM20-NeonGreen confirmed the correct size of the ACSL3-

194 NeonGreen fusion (Figure 2A). Furthermore, colocalization of ACSL3-NeonGreen with the ER-

195 membrane localized chaperone Calnexin demonstrated that the NeonGreen tag did not alter

196 the presence of ACSL3 at the ER (Figure 2B). As ACSL3 is essential for LD formation, we

197 tested whether the ACSL3-NeonGreen chimera is fully functional. Thereto,

198 GABARAPL2endoHA/ACSL3endoNeonGreen cells were treated with oleic acid to induce LD formation

199 or EtOH as control prior to fixation and immunolabeling of phospholipids and neutral lipids.

200 Confocal microscopy showed a clear colocalization of ACSL3 with phospholipids and neutral

201 lipids in control cells while ACSL3 redistributed in the phospholipid monolayer of LDs when

202 cells were treated with oleic acid for 24 hrs (Figure 2C). Next, we microscopically analyzed

203 fixed and HA-immunolabeled GABARAPL2endoHA/ACSL3endoNeonGreen cells that were grown in

204 the absence and presence of Torin1 and BafA1 or ATG7 inhibitor. Consistent with our IPs,

205 these experiments revealed a partial co-localization of endogenous GABARAPL2 and ACSL3

206 (Figure 2D). Together, these results show that NeonGreen tagged ACSL3 is correctly localized

207 at the ER membrane, integrates into the monolayer of LDs upon free fatty acid treatment and

208 associates with GABARAPL2 at the ER.

209

210 GABARAPL2 is stabilized by ACSL3

211 Since GABARAPL2 is involved in autophagic cargo engulfment (31), we tested whether

212 ACSL3 is an autophagy substrate or serves as selective autophagy receptor. However,

213 stimulation of GABARAPL2endoHA cells with Torin1, BafA1, a combination of both or with ATG7

214 inhibitor showed that ACSL3 protein levels did not change upon autophagy induction or

215 blockage (Supplementary Figure S3D). Likewise, depletion of GABARAPL2 had no effects on

216 ACSL3 abundance (Supplementary Figure S3E). Thus, these results indicate that ACSL3 is

217 neither a substrate nor a receptor of autophagy under these conditions. Next, we examined

218 the effects of ACSL3 knockdown on GABARAPL2. Treatment of GABARAPL2endoHA cells with

219 two different ACSL3 siRNAs showed a significant decrease of GABARAPL2 protein levels

220 (Figure 3A). To rule out that this phenotype is due to a global perturbation of the ER, we probed

221 for the integrity of this organelle in cells depleted of ACSL3 using immunolabeling with Calnexin

222 and the ER exit site marker SEC13. However, neither the meshwork appearance nor the exist

223 sites of the ER showed any overt alterations (Supplementary Figure S3F,G). Given the high

224 structural and functional similarity between LC3 and GABARAP family members we addressed

225 whether ACSL3 depletion likewise impacts on the protein abundance of the other hATG8 family

226 members. Unexpectedly, ACSL3 knockdown experiments in GABARAPendoHA,

227 GABARAPL1endoHA and LC3BendoHA cells did not show any significant reduction in the respective

228 HA-tagged hATG8 proteins (Figure 3B-D). In contrast, we found that LC3B protein levels

229 significantly increased upon ACSL3 depletion (Figure 3D), suggesting that reduced

230 GABARAPL2 levels might be compensated by increased expression of LC3B. Intriguingly, we

231 observed that GABARAPL2 protein levels are restored in RNAi-treated GABARAPL2endoHA

232 cells treated with BafA1 but not with the proteasome inhibitor Bortezomib (Btz) (Figure 3E).

233 Thus, these results indicate that ACSL3 is not degraded by autophagy but rather serves as a

234 specific stabilizing factor of GABARAPL2 at the ER.

235

236 ACSL3 anchors UBA5 to the ER membrane

237 To better understand the biological significance of the GABARAPL2-ACSL3 interaction, we

238 turned our attention to known GABARAPL2 binding proteins and in particular to ubiquitin-like

239 modifier activating enzyme 5 (UBA5) (32) which was recently shown to be recruited to the ER

240 membrane in a GABARAPL2-dependent manner (33). Using HA- and myc-IPs, we confirmed

241 the GABARAPL2-UBA5 interaction in GABARAPL2endoHA cells transfected with myc-UBA5

242 (Figure 4A, Supplementary Figure S3C). Since ACSL3 binds GABARAPL2 at the ER

243 membrane, we investigated whether ACSL3 also associates with UBA5. To this end, we

244 generated HeLa cells stably overexpressing C-terminally HA-tagged ACSL3 and transiently

245 transfected these and parental HeLa cells with myc-UBA5. Following differential treatment with

246 oleic acid, cells were lysed and subjected to IP with anti-myc agarose. Intriguingly, we found

247 that UBA5 associates with ACSL3 independent of its activity during LD formation (Figure 4B).

248 Next, we examined whether ACSL3, GABARAPL2 and UBA5 form a ternary complex.

249 Therefore, GABARAPL2endoHA/ACSL3endoNeonGreen cells were treated with oleic acid or EtOH as

250 control, followed by anti-UBA5 and anti-HA immunolabeling. Consistent with our binding

251 assays, confocal microscopy showed colocalization of all three proteins irrespective of the

252 treatment condition (Figure 4C). Overall, these results suggest that ACSL3, GABARAPL2 and

253 UBA5 form a complex at the ER membrane independent of ACSL3’s activity in response to

254 induction of LD formation.

255

256 ACSL3 regulates ufmylation pathway components

257 Since we found that ACSL3 stabilizes GABARAPL2, we investigated whether ACSL3 depletion

258 has similar effects on UBA5 protein abundance. For this purpose, GABARAPL2endoHA cells

259 were transfected with siACSL3 or sictrl and grown in the absence or presence of BafA1 or Btz.

260 Indeed, we observed that protein levels of UBA5 decreased upon ACSL3 depletion and were

261 restored by blockage of autophagy (Figure 5A). While depletion of GABARAPL2 had no effects

262 on UBA5 protein levels (Supplementary Figure S3E). This supports the notion that UBA5 and

263 GABARAPL2 form a functional unit which is regulated by ACSL3. UBA5 is part of the

264 conjugation system, termed ufmylation, that covalently attaches the ubiquitin-like protein

265 ubiquitin fold modifier 1 (UFM1) to target proteins through an E1-E2-E3 multienzyme cascade.

266 The E1-like enzyme UBA5 activates UFM1 by forming a thioester bond between its active site

267 and the exposed C-terminal glycine of UFM1 (32). The UFM1-conjugating enzyme 1 (UFC1)

268 then transfers UFM1 from UBA5 to the UFM1-protein ligase 1 (UFL1) which mediates the

269 attachment to target proteins (32, 34). The ER-membrane bound protein DDRGK1 anchors

270 UFL1 to the ER membrane (35) and is besides RPL26 (36) and ASC1 (37) one of the few

271 known ufmylation targets (34). While the consequences of ufmylation remains poorly

272 understood at the mechanistic level, the UFM1 conjugation pathway has been linked to the ER

273 stress response (38, 39), erythrocyte differentiation (40, 41), cellular homeostasis (42) and

274 breast cancer progression (37). Since the stability of UBA5 and its ER-recruiting factor

275 GABARAPL2 was controlled by ACSL3, we probed whether it also regulates the abundance

276 of the other proteins in the ufmylation cascade. Knockdown experiments revealed that the

277 protein levels of UFM1, UFL1 and DDRGK1 were likewise decreased upon ACSL3 depletion

278 while the abundance of UFC1 was unaffected. Notably, UFC1 is the only ufmylation component

279 that is not localized at the ER membrane. The observation that the protein levels of UFM1,

280 UFL1 and DDRGK1 were not restored by blockage of autophagy or blockage of the

281 proteasome (Figure 5B) indicates that these ufmylation factors are regulated at the

282 transcriptional level. Together, our findings suggest that ACSL3 not only anchors UBA5 but

283 might act as novel regulator of the ufmylation cascade.

284

285 LDs regulate UFM1 conjugation

286 These findings raise the question whether LD biogenesis and ufmylation are functionally

287 coupled. To test this hypothesis, we monitored the ufmylation pathway in response to induction

288 and completion of LD formation in GABARAPL2endoHA cells grown in the absence and presence

289 of oleic acid for 30 min and 24 hrs, respectively. While UBA5 levels drastically increased after

290 30 min and 24 hrs oleic acid treatment, there was no effect on UFC1 (Figure 5C, left panel). In

291 contrast, the protein levels of DDRGK1 and UFL1 firstly decreased after 30 min incubation with

292 oleic acid but after 24 hrs when LDs were formed DDRGK1 and UFL1 levels were partially

293 restored (Figure 5C, left panel). Consistent with a restoration of the UFM1 conjugation pathway

294 upon LD completion, we detected less unconjugated UFM1 (~9kDa) and more conjugated

295 UFM1 (~35kDa) in response to 24 hrs oleic acid treatment (Figure 5C, right panel). Together,

296 these results indicate that the ufmylation cascade is differentially regulated during induction

297 and completion of LD and that the ACSL3-GABARAPL2-UBA5 axis plays an important part in

298 this regulation.

299

300 Discussion

301 In this study, we identified the ER-resident transmembrane protein ACSL3 as novel binding

302 partner of GABARAPL2 and UBA5 using a CRISPR/Cas9 generated GABARAPL2endoHA cell

303 line. Furthermore, we provide evidences for the regulation of ufmylation through ACSL3 and

304 LD biogenesis.

305 In our interactome screen with endogenously tagged GABARAPL2 we found ACSL3, which

306 we confirmed as GABARAPL2 interactor by immunoprecipitation and confocal microscopy.

307 Typically, interaction between GABARAPs or LC3s and their binding partners involves an

308 ATG8 family-interacting motif (AIM; also known as LC3-interacting region (LIR)) in the hATG8

309 interactors and the LIR-docking site (LDS) in LC3 or GABARAP proteins (43-45). Amino acid

310 sequence analysis of ACSL3 revealed four potential LIRs (LIR1: 65-71, LIR2: 135-140, LIR3:

311 589-594, LIR4: 643-648). While LIR2 is localized within the AMP-binding domain of ACSL3

312 and therefore is unlikely accessible, one (or all) of the other three ACSL3 LIR candidates may

313 in principle mediate the binding to GABARAPL2. However, in addition to the LIR/LDS pairing

314 Marshall and colleagues recently reported an alternative hATG8 interaction modus in which

315 binding partners employ a ubiquitin-interacting motif (UIM) to bind to an UIM-docking site

316 (UDS) in LC3 and GABARAP proteins (46, 47). By sequence inspection we found one potential

317 UIM in ACSL3 (663-670). Given that the results from our colocalization studies and binding

318 assays points to a possible complex formation of GABARAPL2, ACSL3 and UBA5 and that

319 the LDS of GABARAPL2 is likely to be occupied by the atypical LIR of UBA5 (33, 48), it is

320 highly plausible that GABARAPL2 and ACSL3 interact in a UIM/UDS-dependent manner.

321 GABARAP proteins were shown to mediate ER recruitment of UBA5 to bring it in close

322 proximity to the membrane bound UFM1 E3 enzyme complex composed of UFL1, DDRGK1

323 and CDK5R3, thereby facilitating ufmylation (33). However, since GABARAPs are not known

324 to be conjugated to PE at the ER, the molecular basis of this recruitment process was not

325 clear. Here, we provided evidence that ACSL3 function to anchor UBA5 at the ER membrane.

326 Given that UBA5 employs an atypical LIR to bind both GABARAPL2 and UFM1 and that the

327 latter is able to outcompete GABARAPL2 binding of UBA5 in vitro (48), it is tempting to

328 speculate that GABARAPL2 interacts with UBA5 until UFM1 conjugation is triggered. In this

329 scenario, GABARAPL2 is only a recruiting factor that hands UBA5 over to ACSL3 (Figure 5D).

330 However, the binding mode of ACSL3 and UBA5 remains to be explored.

331 While targets of ufmylation are still largely unknown, two of the three known UFM1-modified

332 proteins are linked to the ER. Firstly, UFM1 conjugation of DDRGK1 is essential for the

333 stabilization of the serine/threonine-protein kinase/endoribonuclease IRE1 (inositol-requiring

334 enzyme 1) (37, 49). Secondly, it was shown that RPL26 (60S ribosomal protein L26) is

335 exclusively ufmylated and de-ufmylated at the ER membrane (36). Overall, emerging evidence

336 points to a role of the UFM1 conjugation system as regulator of ER homeostasis, ER stress

337 response and ER remodeling. Disruption of protein folding and accumulation of unfolded

338 proteins in the ER are hallmarks of ER stress which leads to the induction of the unfolded

339 protein response (UPR) via one of these three key factors: IRE1, PKR-like ER protein kinase

340 (PERK) or activating transcription factor 6 (ATF6). Protein degradation, reduction of protein

341 synthesis and enlargement of the ER capacity are part of the UPR (50). In different cell lines

342 and animal models, it was reported that ufmylation is upregulated via IRE1 or PERK upon ER

343 stress, while depletion of ufmylation components induce the UPR (38, 39, 42, 51, 52). Upon

344 re-established ER homeostasis, ufmylation coordinates the elimination of extended ER

345 membranes through ER-phagy (53, 54).

346 In our present study, we identified LD formation stimulated by oleic acid treatment as novel

347 regulator of ufmylation. LD biogenesis starts with lens formation, an accumulation of neutral

348 lipids between the ER membrane leaflets until LDs eventually bud from the ER. The

349 hydrophobic neutral lipid core of a LD is surrounded by a phospholipid monolayer with the

350 origin of the outer ER membrane leaflet (55). ACSL3 was identified as LD associated protein

351 and essential for LD biogenesis, expansion and maturation (27, 56). During initiation of LD

352 biogenesis ACSL3 is translocated and concentrated to pre-LDs to drive LD expansion by

353 mediating acyl-CoA synthesis. However, cells with enzymatically inactive ACSL3 are still able

354 to form LDs, suggesting additional functions of ACSL3 in LD biogenesis (27, 57). Induction of

355 LD formation induced by oleic acid resulted in an immediate (after 30 min) reduction of UFL1

356 and DDRGK1 protein levels and thus shut down of UFM1 conjugation (Fig 6A). Interestingly,

357 depletion of ACSL3 led to a similar phenotype with regard to these two ufmylation components.

358 Together, these results suggest that ACSL3 regulates DDRGK1 and UFL1 protein levels and

359 therefore ufmylation. The observation that inhibition of proteasomal or lysosomal degradation

360 only partly rescued this phenotype suggests that the ufmylation machinery is probably

361 downregulated at the transcriptional level. To what extend this involves one of the three UPR

362 factors IRE1, PERK or ATF6 remains to be examined. Considering that ER-phagy is blocked

363 by inhibition of the interaction between DDRGK1 and UFL1 (53), we hypothesize that LD

364 biogenesis inhibits the remodeling of ER membranes by ER-phagy. Intriguingly, completion of

365 LD formation (~24 hrs) almost restored DDRGK1 and UFL1 protein levels, which in turn led to

366 increased ufmylation (Figure 5D). Whether these UFM1 targets are linked to remodeling of ER

367 membranes by re-established ER-phagy remains to be tested.

368 Collectively, these findings underline the potential of our CRISPR/Cas9 gene-edited cell lines

369 to uncover novel cellular pathways involving hATG8 family members without the need of

370 overexpression systems, thereby complementing the recently generated LC3 and GABARAP

371 knockout cell lines (9). Together with the LC3CendoHA cell line that we previously reported (20)

372 this cellular resource circumvents the drawback of unspecific LC3 and GABARAP antibodies

373 and hence will greatly facilitate the functional dissection of individual hATG8 proteins.

374

375 Material and Methods

376 Cell culture and treatments

377 HeLa cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM) + GlutaMAX-I

378 (Gibco) supplemented with 10 % fetal bovine serum (FBS) and 1mM sodium pyruvate (Gibco)

379 and grown at 37° C and 5 % CO2. For SILAC mass spectrometry, cells were grown in lysine-

380 and arginine-free DMEM (Gibco) supplemented with 10 % dialyzed FBS, 2 mM glutamine

381 (Gibco), 1 mM sodium pyruvate (Gibco) and 146 mg/ml light (K0, Sigma) or heavy L-lysine

382 (K8, Cambridge Isotope Laboratories) and 84 mg/ml light (R0, Sigma) or heavy L-arginine

383 (R10, Cambridge Isotope Laboratories). SILAC labeled cells were counted after harvesting,

384 mixed 1:1 and stored at -80° C. For selection Puromycin (2 µg/ml) or Blasticidine (4 µg/ml) was

385 added to the growth medium. The following reagents were used for treatments: oleic acid (EMD

386 Millipore, 4954, 600 µm in EtOH, 30 min or 24 hrs), Bafilomycin A1 (Biomol, Cay11038-1, 200

387 nM in DMSO, 2 hrs), Torin 1 (Tocris, 4247, 250 nM in DMSO, 2 hrs), Bortezomib (LC Labs B-

388 1408, 1 µM in PBS, 8 hrs), ATG7 inhibitor (Takeda ML00792183, 1 µM, 24 hrs).

389

390 Plasmids

391 attB flanked ORFs, generated by PCR were cloned into the Gateway entry vector pDONR233.

392 ORFs from pDONR233 constructs were introduced into one of the following destination vectors

393 using recombination cloning: pHAGE-CMV-C-FLAG-HA, pEZYmyc-HIS (Addgene, #18701) or

394 pDEST-myc. Stable pHAGE-ACSL3-HA expressing cells were generate by lentiviral

395 transduction followed selection with 2 µg/ml Puromycin. pEZY and pDEST constructs were

396 used for transient expression in cells (see transfection).

397

398 Genome editing

399 The N-terminal HA-tagged hATG8 cell lines were generated with homology PCR templates

400 containing 87 bp of GABARAP/GABARAPL1/GABARAPL2/LC3B-5’UTR including the start

401 codon followed by the Blasticidine resistance gene, P2A, HA and 92bp downstream of the start

402 codon of the corresponding hATG8 gene. For the C-terminal ACSL3-NeonGreen cell line, we

403 used a homology PCR template containing 75 bp of the last exon of ACSL3, the NeonGreen

404 ORF (Allele Biotech), T2A and the Blasticidine resistance gene ending with 84 bp downstream

405 of the last exon of ACSL3. sgRNAs for hATG8s and ACSL3, designed with the online design

406 tool from the Broad Institute (https://portals.broadinstitute.org/gpp/public/analysis-tools/sgrna-

407 design) were clone into BSbI digested px330 (Addgene #42230), a SpCas9 expressing

408 plasmid (sgRNA: GABARAP: GGAGGATGAAGTTCGTGTAC, GABARAPL1:

409 TGCGGTGCATCATGAAGTTC, GABARAPL2: CCATGAAGTGGATGTTCAAG, LC3B:

410 AGATCCCTGCACCATGCCGT, ACSL3: AGAAAATAATTATTCTCTTC). HeLa cells were

411 seeded in a 6-well plate and transfected with Lipofectamin 2000 according to the manufacture’s

412 instructions with sgRNA and corresponding homology PRC template. After 48 hrs, cells were

413 selected with 4 µg/ml Blasticidine and single cell selection in 96-well plates. Correct

414 introduction of the tag was verified by PCR and sequencing.

415

416 Antibodies

417 For immunoblotting the following primary antibodies were used at a concentration of 1:1000 in

418 5 % milk-TBS-T: HA (Cell Signaling 3724S), HA (Biolegend, 901501), PCNA (Santa Cruz, sc-

419 7907), ACSL3 (Santa Cruz sc-166374), mNeonGreen (Chromotek, 32F6), UBA5 (Proteintech

420 12093-1-AP/Sigma HPA017235), UFC1 (Proteintech 15783-1-AP), DDRGK1 (Proteintech

421 21445-1-AP), UFL1 (Abcam ab226216), UFM1 (Abcam ab109305) or at a concentration of

422 1:100 in 5 % milk-TBS-T: anti-myc (Monoclonal Antibody Core Facility, Helmholtz Zentrum

423 Munich, 9E1, rat IgG1), anti-myc (Monoclonal Antibody Core Facility, Helmholtz Zentrum

424 Munich, 9E10, mouse IgG). As secondary antibodies we used horseradish peroxidase coupled

425 anti-mouse (Promega W402B), anti-rabbit (Promega, W401B) antibodies at a concentration of

426 1:10 000 and anti-rat IgG1 (Monoclonal Antibody Core Facility, Helmholtz Zentrum Munich)

427 antibody at a concentration of 1:100 in 1 % milk-TBS-T. The following primary antibodies were

428 used for immunofluorescence in 0.1 % BSA-PBS: HA (Roche, 11867423001, 1:50), p62 (BD,

429 610832, 1:500), LAMP1 (DSHB, H4A3, 1:50), LC3 (MBL, PM036, 1:500), HCS LipidTOX™

430 Red Phospholipidosis Detection Reagent (Thermo Scientific, H34351, 1:1000), HCS

431 LipidTOX™ Deep Red Neutral Lipid Stain (Thermo Scientific, H34477, 1:500), UBA5

432 (Proteintech, 12093-1-AP, 1:250), Calnexin (Stressgen, SPA-860, 1:100), SEC13 (Novus,

433 AF9055-100, 1:300). The following fluorophore conjugated secondary antibodies from Thermo

434 Fisher were use at a concentration of 1:1000 in 0.1 % BSA-PBS: anti-mouse IgG Alexa Fluor

435 488 (A-11001), anti-rabbit IgG Alexa Fluor 488 (A-11008), anti-rabbit IgG Alexa Fluor 594 (A-

436 11012) and anti-rat IgG Alexa Fluor 647 (A-21247).

437

438 Transfection

439 For siRNA knockdowns, cells were reversely transfected with Lipofectamine RNAiMax

440 (Thermo Fisher Scientific) according to the manufacturer’s guidance with 30 nM of the following

441 siRNAs from Dharmacon/Horizon Discovery and harvested 72 hrs after transfection: sicrtl

442 UGGUUUACAUGUUUUCCUA, siACSL3#1 UAACUGAACUAGCUCGAAA, siACSL3#2:

443 GCAGUAAUCAUGUACACAA, siGABARAP GGUCAGUUCUACUUCUUGA, siGABARAPL1 17

444 GAAGAAAUAUCCGGACAGG, siGABARAPL2 GCUCAGUUCAUGUGGAUCA, siLC3B

445 GUAGAAGAUGUCCGACUUA. Plasmids were transiently transfected with Lipofectamine

446 2000 (Thermo Fisher Scientific) according to the instruction of the manufacturer or with 10 mM

447 PEI (Polyethylenimine) and cells were collected after 48 hrs.

448

449 Immunoblotting

450 Cell were lysed in RIPA (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 0.5% sodium desoxycholate,

451 1 % NP-40, 0.1 % SDS, 1x EDTA-free protease inhibitor (Roche), 1x phosphatase inhibitor

452 (Roche)) for 30 min. After elimination of cell debris by centrifugation, proteins were diluted with

453 3x loading buffer (200 mM Tris-HCl [pH 6.8], 6 % SDS, 20 % Glycerol, 0.1 g/ml DTT, 0.1 mg

454 Bromophenol blue) and boiled at 95°C. Proteins were size separated by SDS-PAGE with self-

455 casted 8 %, 10 %, 12 % and 15 % gels followed by protein transfer onto nitrocellulose

456 membranes (GE Healthcare Life Sciences, 0.45 µm). For better visibility of endogenous HA-

457 hATG8s membranes were boiled for 5 min in PBS after protein transfer. Blots were blocked in

458 TBS-T (20 mM Tris, 150 mM NaCl, 0.1% Tween-20) supplemented with 5 % low fat milk (Roth)

459 for 1 hr. Primary antibodies were incubated overnight followed by several wash steps with TBS-

460 T and incubation with secondary antibodies for 1 hr at room temperature. After repeated

461 washing, immunoblots were analyzed with Western Lightning Plus ECL (Perkin Elmer).

462

463 Immunofluorescence

464 All steps were carried out at room temperature. Cells growing on glass coverslips in 12-well

465 plates were fixed with 4 % paraformaldehyde in PBS for 30 min followed by permeabilization

466 with 0,1 % Trition-X-100 in PBS or 0,1 % Saponin in PBS for 30 min and 1 hr blocking in 1%

467 BSA-PBS. First and secondary antibody incubation was done sequentially for 1 hr at room

468 temperature in 0.1 % BSA-PBS followed by mounting of the coverslips with ProlongGold

469 Antifade with Dapi (Thermo Fisher). In between each step, cells were washed several times 18

470 with PBS. Cells were imaged with a LSM 800 Carl Zeiss microscope using 63x oil-immersion

471 objective and ZEN blue edition software and analyses with ImageJ (version 1.52).

472

473 Immunoprecipitation

474 Frozen cell pellets from 4x15 cm cell culture plates for mass spectrometry or 1x15 cm cell

475 culture plate for immunoblotting were lysed in Glycerol buffer (20 mM Tris [pH 7.4], 150 mM

476 NaCl, 5 mM EDTA, 0.5 % Triton-X-100, 10 % Glycerol, 1x protease inhibitor, 1x phosphatase

477 inhibitor) for 30 min at 4° C with end-over-end rotation. Lysates were cleared from cell debris

478 by centrifugation prior to adjustment of protein concentrations between the samples and

479 overnight immunoprecipitation at 4° C with pre-equilibrated anti-HA-agarose (Sigma) or anti-

480 c-myc-agarose (Thermo fisher). Agarose beads were washed five times with Glycerol buffer

481 followed by elution of proteins with 3x loading buffer and boiling of the samples at 95° C.

482 Samples were then analyzed by SDS-PAGE (self-casted or BioRad’s 4-20 % gels) followed by

483 immunoblotting or in-gel tryptic digestion.

484

485 Mass spectrometry

486 SDS-PAGE gel lines were cut in 12 equal size bands, further chopped in smaller pieces and

487 placed in 96 well plates (one band per well). Gel pieces were washed with 50 mM ammonium

488 bicarbonate (ABC)/50 % EtOH buffer followed by dehydration with EtOH, reduction of proteins

489 with 10 mM DTT in 50 mM ABC at 56° C for 1 hr and alkylation of proteins with 55 mM

490 iodacetamide in 50 mM ABC at room temperature for 45 min. Prior to overnight trypsin-digest

491 (12 ng/ul trypsin in 50 mM ABC, Promega) at 37° C, gel pieces were washed and dehydrated

492 as before. Peptide were extracted from gel pieces with 30 % acetonitrile/3 % trifluoroacetic

493 acid (TFA), 70 % acetonitrile and finally 100 % acetonitrile followed by desalting on custom-

494 made C18-stage tips. Using an Easy-nLC1200 liquid chromatography (Thermo Scientific),

495 peptides were loaded onto 75 µm x 15 cm fused silica capillaries (New Objective) packed with 19

496 C18AQ resin (Reprosil- Pur 120, 1.9 µm, Dr. Maisch HPLC). Peptide mixtures were separated

497 using a gradient of 5%–33% acetonitrile in 0.1% acetic acid over 75 min and detected on an

498 Q Exactive HF mass spectrometer (Thermo Scientific). Dynamic exclusion was enabled for 30

499 s and singly charged species or species for which a charge could not be assigned were

500 rejected. MS data were processed with MaxQuant (version 1.6.0.1) and analyzed with Perseus

501 (version 1.5.8.4, http://www.coxdocs.org/doku.php?id=perseus:start). IP experiments from

502 GABARAPL2endoHA and control parental HeLa cells were performed in duplicates and

503 triplicates, respectively. Matches to common contaminants, reverse identifications and

504 identifications based only on site-specific modifications were removed prior to further analysis.

505 Log2 heavy/light ratios were calculated. A threshold based on a log2 fold change of greater

506 than 1.5-fold or less than -1.5-fold was chosen so as to focus the data analysis on a smaller

507 set of proteins with the largest alterations in abundance. Additional requirements were at least

508 two MS counts, unique peptides and razor peptides as well as absence in IPs from parental

509 HeLa control cells. For functional annotations, the platform DAVID (https://david.ncifcrf.gov/)

510 was used.

511

512 Data availability

513 The mass spectrometry proteomics data have been deposited to the

514 ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the

515 PRIDE partner repository with the dataset identifier PXD016734.

516

517 Statistical analysis

518 Quantification and statistical analysis of western blots were done with imageJ and Phyton

519 (version 3.7). Statistical significance was calculated with Student’s t test and data represent ±

520 SEM (standard error of the mean). Statistical analysis of MS data was done with Perseus.

521

522 Acknowledgement

523 We would like to thank Georg Werner and all members of the Behrends lab for reagents, advice

524 and critical discussion. This work was supported by the Deutsche Forschungsgemeinschaft

525 (German Research Foundation) within the framework of the Munich Cluster for Systems

526 Neurology (EXC2145 SyNergy), the Collaborative Research Center (CRC1177), and the

527 project grant BE 4685/2-1.

528

529 Author contribution

530 FE performed all experiments. MK provided advice for CRISPR/Cas9 tagging strategy. FE and

531 CB conceived the study and wrote the manuscript.

532

533 Conflict of interest

534 The authors declare that they have no conflict of interest.

535

536 References

537 1. Slobodkin MR, Elazar Z. The Atg8 family: multifunctional ubiquitin-like key regulators of

538 autophagy. Essays Biochem. 2013;55:51-64.

539 2. Shpilka T, Weidberg H, Pietrokovski S, Elazar Z. Atg8: an autophagy-related ubiquitin-like

540 protein family. Genome Biol. 2011;12(7):226.

541 3. Mizushima N, Yoshimori T, Ohsumi Y. The role of Atg proteins in autophagosome

542 formation. Annu Rev Cell Dev Biol. 2011;27:107-32.

543 4. Kirkin V, Rogov VV. A Diversity of Selective Autophagy Receptors Determines the

544 Specificity of the Autophagy Pathway. Mol Cell. 2019;76(2):268-85. 21

545 5. Dikic I, Elazar Z. Mechanism and medical implications of mammalian autophagy. Nat Rev

546 Mol Cell Biol. 2018;19(6):349-64.

547 6. Xie Z, Nair U, Klionsky DJ. Atg8 controls phagophore expansion during autophagosome

548 formation. Mol Biol Cell. 2008;19(8):3290-8.

549 7. Stolz A, Ernst A, Dikic I. Cargo recognition and trafficking in selective autophagy. Nat Cell

550 Biol. 2014;16(6):495-501.

551 8. Weidberg H, Shpilka T, Shvets E, Abada A, Shimron F, Elazar Z. LC3 and GATE-16 N

552 termini mediate membrane fusion processes required for autophagosome biogenesis. Dev

553 Cell. 2011;20(4):444-54.

554 9. Nguyen TN, Padman BS, Usher J, Oorschot V, Ramm G, Lazarou M. Atg8 family

555 LC3/GABARAP proteins are crucial for autophagosome-lysosome fusion but not

556 autophagosome formation during PINK1/Parkin mitophagy and starvation. J Cell Biol.

557 2016;215(6):857-74.

558 10. Leil TA, Chen ZW, Chang CS, Olsen RW. GABAA receptor-associated protein traffics

559 GABAA receptors to the plasma membrane in neurons. J Neurosci. 2004;24(50):11429-

560 38.

561 11. Wang H, Bedford FK, Brandon NJ, Moss SJ, Olsen RW. GABA(A)-receptor-associated

562 protein links GABA(A) receptors and the cytoskeleton. Nature. 1999;397(6714):69-72.

563 12. Legesse-Miller A, Sagiv Y, Porat A, Elazar Z. Isolation and characterization of a novel low

564 molecular weight protein involved in intra-Golgi traffic. J Biol Chem. 1998;273(5):3105-9.

565 13. Muller JM, Shorter J, Newman R, Deinhardt K, Sagiv Y, Elazar Z, et al. Sequential SNARE

566 disassembly and GATE-16-GOS-28 complex assembly mediated by distinct NSF activities

567 drives Golgi membrane fusion. J Cell Biol. 2002;157(7):1161-73.

568 14. Genau HM, Huber J, Baschieri F, Akutsu M, Dotsch V, Farhan H, et al. CUL3-

569 KBTBD6/KBTBD7 ubiquitin ligase cooperates with GABARAP proteins to spatially restrict

570 TIAM1-RAC1 signaling. Mol Cell. 2015;57(6):995-1010.

571 15. Baisamy L, Cavin S, Jurisch N, Diviani D. The ubiquitin-like protein LC3 regulates the Rho-

572 GEF activity of AKAP-Lbc. J Biol Chem. 2009;284(41):28232-42. 22

573 16. Stadel D, Millarte V, Tillmann KD, Huber J, Tamin-Yecheskel BC, Akutsu M, et al. TECPR2

574 Cooperates with LC3C to Regulate COPII-Dependent ER Export. Mol Cell. 2015;60(1):89-

575 104.

576 17. Behrends C, Sowa ME, Gygi SP, Harper JW. Network organization of the human

577 autophagy system. Nature. 2010;466(7302):68-76.

578 18. Popovic D, Akutsu M, Novak I, Harper JW, Behrends C, Dikic I. Rab GTPase-activating

579 proteins in autophagy: regulation of endocytic and autophagy pathways by direct binding

580 to human ATG8 modifiers. Mol Cell Biol. 2012;32(9):1733-44.

581 19. Stolz A, Putyrski M, Kutle I, Huber J, Wang C, Major V, et al. Fluorescence-based ATG8

582 sensors monitor localization and function of LC3/GABARAP proteins. EMBO J.

583 2017;36(4):549-64.

584 20. Le Guerroue F, Eck F, Jung J, Starzetz T, Mittelbronn M, Kaulich M, et al. Autophagosomal

585 Content Profiling Reveals an LC3C-Dependent Piecemeal Mitophagy Pathway. Mol Cell.

586 2017;68(4):786-96 e6.

587 21. Kaulich M, Dowdy SF. Combining CRISPR/Cas9 and rAAV Templates for Efficient Gene

588 Editing. Nucleic Acid Ther. 2015;25(6):287-96.

589 22. Bai H, Inoue J, Kawano T, Inazawa J. A transcriptional variant of the LC3A gene is involved

590 in autophagy and frequently inactivated in human cancers. Oncogene. 2012;31(40):4397-

591 408.

592 23. Ewing RM, Chu P, Elisma F, Li H, Taylor P, Climie S, et al. Large-scale mapping of human

593 protein-protein interactions by mass spectrometry. Mol Syst Biol. 2007;3:89.

594 24. Rolland T, Tasan M, Charloteaux B, Pevzner SJ, Zhong Q, Sahni N, et al. A proteome-

595 scale map of the human interactome network. Cell. 2014;159(5):1212-26.

596 25. Soupene E, Kuypers FA. Mammalian long-chain acyl-CoA synthetases. Exp Biol Med

597 (Maywood). 2008;233(5):507-21.

598 26. Fujimoto Y, Itabe H, Kinoshita T, Homma KJ, Onoduka J, Mori M, et al. Involvement of

599 ACSL in local synthesis of neutral lipids in cytoplasmic lipid droplets in human hepatocyte

600 HuH7. J Lipid Res. 2007;48(6):1280-92. 23

601 27. Kassan A, Herms A, Fernandez-Vidal A, Bosch M, Schieber NL, Reddy BJ, et al. Acyl-

602 CoA synthetase 3 promotes lipid droplet biogenesis in ER microdomains. J Cell Biol.

603 2013;203(6):985-1001.

604 28. Brasaemle DL, Dolios G, Shapiro L, Wang R. Proteomic analysis of proteins associated

605 with lipid droplets of basal and lipolytically stimulated 3T3-L1 adipocytes. J Biol Chem.

606 2004;279(45):46835-42.

607 29. Ingelmo-Torres M, Gonzalez-Moreno E, Kassan A, Hanzal-Bayer M, Tebar F, Herms A,

608 et al. Hydrophobic and basic domains target proteins to lipid droplets. Traffic.

609 2009;10(12):1785-801.

610 30. Poppelreuther M, Rudolph B, Du C, Grossmann R, Becker M, Thiele C, et al. The N-

611 terminal region of acyl-CoA synthetase 3 is essential for both the localization on lipid

612 droplets and the function in fatty acid uptake. J Lipid Res. 2012;53(5):888-900.

613 31. Schaaf MB, Keulers TG, Vooijs MA, Rouschop KM. LC3/GABARAP family proteins:

614 autophagy-(un)related functions. FASEB J. 2016;30(12):3961-78.

615 32. Komatsu M, Chiba T, Tatsumi K, Iemura S, Tanida I, Okazaki N, et al. A novel protein-

616 conjugating system for Ufm1, a ubiquitin-fold modifier. EMBO J. 2004;23(9):1977-86.

617 33. Huber J, Obata M, Gruber J, Akutsu M, Lohr F, Rogova N, et al. An atypical LIR motif

618 within UBA5 (ubiquitin like modifier activating enzyme 5) interacts with GABARAP proteins

619 and mediates membrane localization of UBA5. Autophagy. 2019:1-15.

620 34. Tatsumi K, Sou YS, Tada N, Nakamura E, Iemura S, Natsume T, et al. A novel type of E3

621 ligase for the Ufm1 conjugation system. J Biol Chem. 2010;285(8):5417-27.

622 35. Wu J, Lei G, Mei M, Tang Y, Li H. A novel C53/LZAP-interacting protein regulates stability

623 of C53/LZAP and DDRGK domain-containing Protein 1 (DDRGK1) and modulates NF-

624 kappaB signaling. J Biol Chem. 2010;285(20):15126-36.

625 36. Walczak CP, Leto DE, Zhang L, Riepe C, Muller RY, DaRosa PA, et al. Ribosomal protein

626 RPL26 is the principal target of UFMylation. Proc Natl Acad Sci U S A. 2019;116(4):1299-

627 308.

628 37. Yoo HM, Kang SH, Kim JY, Lee JE, Seong MW, Lee SW, et al. Modification of ASC1 by

629 UFM1 is crucial for ERalpha transactivation and breast cancer development. Mol Cell.

630 2014;56(2):261-74.

631 38. Lemaire K, Moura RF, Granvik M, Igoillo-Esteve M, Hohmeier HE, Hendrickx N, et al.

632 Ubiquitin fold modifier 1 (UFM1) and its target UFBP1 protect pancreatic beta cells from

633 ER stress-induced apoptosis. PLoS One. 2011;6(4):e18517.

634 39. Zhang Y, Zhang M, Wu J, Lei G, Li H. Transcriptional regulation of the Ufm1 conjugation

635 system in response to disturbance of the endoplasmic reticulum homeostasis and

636 inhibition of vesicle trafficking. PLoS One. 2012;7(11):e48587.

637 40. Cai Y, Pi W, Sivaprakasam S, Zhu X, Zhang M, Chen J, et al. UFBP1, a Key Component

638 of the Ufm1 Conjugation System, Is Essential for Ufmylation-Mediated Regulation of

639 Erythroid Development. PLoS Genet. 2015;11(11):e1005643.

640 41. Tatsumi K, Yamamoto-Mukai H, Shimizu R, Waguri S, Sou YS, Sakamoto A, et al. The

641 Ufm1-activating enzyme Uba5 is indispensable for erythroid differentiation in mice. Nat

642 Commun. 2011;2:181.

643 42. Zhang M, Zhu X, Zhang Y, Cai Y, Chen J, Sivaprakasam S, et al. RCAD/Ufl1, a Ufm1 E3

644 ligase, is essential for hematopoietic stem cell function and murine hematopoiesis. Cell

645 Death Differ. 2015;22(12):1922-34.

646 43. Noda NN, Kumeta H, Nakatogawa H, Satoo K, Adachi W, Ishii J, et al. Structural basis of

647 target recognition by Atg8/LC3 during selective autophagy. Genes Cells.

648 2008;13(12):1211-8.

649 44. Pankiv S, Clausen TH, Lamark T, Brech A, Bruun JA, Outzen H, et al. p62/SQSTM1 binds

650 directly to Atg8/LC3 to facilitate degradation of ubiquitinated protein aggregates by

651 autophagy. J Biol Chem. 2007;282(33):24131-45.

652 45. Rogov V, Dotsch V, Johansen T, Kirkin V. Interactions between autophagy receptors and

653 ubiquitin-like proteins form the molecular basis for selective autophagy. Mol Cell.

654 2014;53(2):167-78.

655 46. Marshall RS, Li F, Gemperline DC, Book AJ, Vierstra RD. Autophagic Degradation of the

656 26S Proteasome Is Mediated by the Dual ATG8/Ubiquitin Receptor RPN10 in Arabidopsis.

657 Mol Cell. 2015;58(6):1053-66.

658 47. Marshall RS, Vierstra RD. Dynamic Regulation of the 26S Proteasome: From Synthesis

659 to Degradation. Front Mol Biosci. 2019;6:40.

660 48. Habisov S, Huber J, Ichimura Y, Akutsu M, Rogova N, Loehr F, et al. Structural and

661 Functional Analysis of a Novel Interaction Motif within UFM1-activating Enzyme 5 (UBA5)

662 Required for Binding to Ubiquitin-like Proteins and Ufmylation. J Biol Chem.

663 2016;291(17):9025-41.

664 49. Liu J, Wang Y, Song L, Zeng L, Yi W, Liu T, et al. A critical role of DDRGK1 in endoplasmic

665 reticulum homoeostasis via regulation of IRE1alpha stability. Nat Commun. 2017;8:14186.

666 50. Karagoz GE, Acosta-Alvear D, Walter P. The Unfolded Protein Response: Detecting and

667 Responding to Fluctuations in the Protein-Folding Capacity of the Endoplasmic Reticulum.

668 Cold Spring Harb Perspect Biol. 2019;11(9).

669 51. Gerakis Y, Quintero M, Li H, Hetz C. The UFMylation System in Proteostasis and Beyond.

670 Trends Cell Biol. 2019;29(12):974-86.

671 52. Zhu H, Bhatt B, Sivaprakasam S, Cai Y, Liu S, Kodeboyina SK, et al. Ufbp1 promotes

672 plasma cell development and ER expansion by modulating distinct branches of UPR. Nat

673 Commun. 2019;10(1):1084.

674 53. Liang JH, Lingeman E, Luong T, Ahmed S, Nguyen T, Olzmann J, et al. A genome-wide

675 screen for ER autophagy highlights key roles of mitochondrial metabolism and ER-

676 resident UFMylation. bioRxiv. 2019.

677 54. DeJesus R, Moretti F, McAllister G, Wang Z, Bergman P, Liu S, et al. Functional CRISPR

678 screening identifies the ufmylation pathway as a regulator of SQSTM1/p62. Elife. 2016;5.

679 55. Henne WM, Reese ML, Goodman JM. The assembly of lipid droplets and their roles in

680 challenged cells. EMBO J. 2018;37(12).

681 56. Fujimoto Y, Itabe H, Sakai J, Makita M, Noda J, Mori M, et al. Identification of major

682 proteins in the lipid droplet-enriched fraction isolated from the human hepatocyte cell line

683 HuH7. Biochim Biophys Acta. 2004;1644(1):47-59.

684 57. Kimura H, Arasaki K, Ohsaki Y, Fujimoto T, Ohtomo T, Yamada J, et al. Syntaxin 17

685 promotes lipid droplet formation by regulating the distribution of acyl-CoA synthetase 3. J

686 Lipid Res. 2018;59(5):805-19.

687

688 Figure legends

689 Figure 1. Interaction analysis of endogenously tagged hATG8 proteins. A, GABARAPL2endoHA

690 and parental HeLa cell lysates were analyzed by immunoblotting using anti-HA and -PCNA

691 antibodies. The latter was used as loading control. B,C, GABARAPL2endoHA cells were

692 reversely transfected for 72 hrs with non-targeting (sicrtl) or GABARAPL2 siRNA followed by

693 lysis and immunoblot analysis (B) or fixation and immunolabeling (C) using an anti-HA

694 antibody. Scale bar: 10 µm. D,E, GABARAPL2endoHA cells were treated as indicated and

695 subjected to lysis and immunoblotting (D) or fixation and immunolabeling (E) using anti-HA

696 and -p62 antibodies. Scale bar: 10 µm. Arrowheads indicate colocalization events. F,

697 Scatterplot represents interaction proteomics of SILAC labeled GABARAPL2endoHA cells

698 differentially treated with Torin1 and BafA1 (light) or ATG7 inhibitor (heavy). Significantly

699 enriched proteins upon Torin1 and BafA1 combination treatment or ATG7 inhibition are

700 highlighted in red and blue, respectively. Proteins in grey are unchanged. G-I, Immunoblot

701 analysis of anti-HA immunoprecipitates from lysates derived from parental HeLa and

702 GABARAPL2endoHA cells transiently transfected for 48 hrs with myc-tagged ATG7 (G), p62 (H)

703 or ACSL3 (I).

704

705 Figure 2. Validation of GABARAPL2endoHA/ACSL3endoNeonGreen cells. A, GABARAPL2endoHA and

706 GABARAPL2endoHA/ACSL3endoNeonGreen cells as well as parental HeLa cells transiently

707 transfected with TOMM20-NeonGreen were lysed and analyzed by immunoblotting with

708 indicated antibodies. B, Fixed GABARAPL2endoHA/ACSL3endoNeonGreen cells were immunostained

709 with an anti-calnexin antibody. Scale bar: 10 µm. C, GABARAPL2endoHA/ACSL3endoNeonGreen cells

710 were treated with oleic acid or EtOH (control) for 24 hrs followed by fixation and

711 immunolabeling of phospholipids and neutral lipids. Scale bar: 10 µm. Two confocal planes

712 are shown for oleic acid treatment. D, GABARAPL2endoHA/ACSL3endoNeonGreen cells treated with

713 Torin1 and BafA1 or ATG7 inhibitor were fixed and immunolabeled with an anti-HA antibody.

714 Scale bar: 10 µm. Arrowheads indicate HA-NeonGreen colocalization events.

715

716 Figure 3. Stabilization of GABARAPL2 through ACSL3. A-D, GABARAPL2endoHA (A),

717 GABARAPendoHA (B), GABARAPL1endoHA (C) and LC3BendoHA (D) cells were reversely

718 transfected with two different ACSL3 siRNAs. Lysates were analyzed by immunoblotting with

719 indicated antibodies. Data represent mean ±SEM. Statistical analysis (n = 4) of the HA/PCNA

720 ratio was performed using Student’s t-test (*p<0.005, **p<0.001). E, GABARAPL2endoHA cells

721 reversely transfected with siRNAs targeting ACSL3 for 72 hrs were treated with BafA1 or Btz

722 and analyzed by immunoblotting.

723

724 Figure 4. UBA5 binds to ACSL3 and GABARAPL2. A, Immunoblot analysis of anti-HA

725 immunoprecipitates from lysates derived from parental HeLa and GABARAPL2endoHA cells

726 transiently transfected for 48 hrs with myc-UBA5. B, Parental HeLa and GABARAPL2endoHA

727 cells transfected with myc-UBA5 were treated with oleic acid or EtOH for 24 hrs prior to lysis,

728 anti-myc immunoprecipitation and immunoblotting. C, GABARAPL2endoHA/ACSL3endoNeonGreen

729 cells were treated with oleic acid for 30 min or 24 hrs followed by fixation and immunolabeling

730 with anti-UBA5 and -HA antibodies. Scale bar: 10 µm. Arrowheads indicate colocalization

731 events between UBA5, HA and NeonGreen. Two confocal planes are shown for the 24 hrs

732 time point.

733

734 Figure 5. Influences of ACSL3 on the ufmylation pathway. A,B, GABARAPL2endoHA cells were

735 transfected with ACSL3 siRNAs and treated with Btz or BafA1 followed by lysis and

736 immunoblot analysis using indicated antibodies. C, GABARAPL2endoHA cells were treated with

737 oleic acid for 30 min or 24 hrs or with EtOH for 24 hrs prior to lysis and immunoblotting with

738 indicated antibodies. D, Working model summarizing the impact of ACSL3 and LD biogenesis

739 on the ufmylation pathway. Upon recruitment via GABARAPL2, UBA5 is anchored at the ER

740 membrane by ACSL3. LD induction through oleic acid blocks ufmylation through degradation-

741 mediated disassembly of the UFM1 E3 enzyme complex. Completion of LD formation leads to

742 reassemble of the E3 complex and increased ufmylation. Dotted blue arrows indicate ER-

743 recruitment, black arrows indicate ufmylation cascade.

744

745 Supplementary Figure legends

746 Supplementary Figure S1. Endogenous epitope tagging of hATG8 and ACSL3 genes. A,

747 Experimental CRISPR/Cas9 workflow. B,C, Sequence data from PCR products of the tagged

748 GABARAPendoHA, GABARAPL1endoHA, GABARAPL2endoHA, LC3BendoHA cell lines (B) and the

749 GABARAPL2endoHA/ACSL3endoNeonGreen cell line (C). Sequence data from PCR products of the.

750 Introduced CRISPR sequences are indicated in bold.

751

752 Supplementary Figure S2. Validation of endogenously HA-tagged hATG8 proteins. A-C,

753 GABARAPendoHA (A), GABARAPL1endoHA (B), LC3BendoHA (C) and parental HeLa (A-C) cells

754 were lysed followed by immunoblotting and analysis with indicated antibodies. D-F,

755 GABARAPendoHA (D), GABARAPL1endoHA (E), LC3BendoHA (F) cell lines were reversely

756 transfected with indicated siRNAs prior to immunoblot analysis. G-I, GABARAPendoHA (G),

757 GABARAPL1endoHA (H), LC3BendoHA (I) were treated as indicated followed by lysis and

758 immunoblotting. J,K, GABARAPL2endoHA cells treated with indicated inhibitors were 29

759 immunolabeled with anti-LAMP1 (K) or anti-LC3 (L) antibody. Scale bar: 10 µm. Arrowheads

760 indicate colocalization events.

761

762 Supplementary Figure S3. Analysis of GABARAPL2-interacting proteins. A, Annotation

763 enrichment analysis of candidate GABARAPL2-interacting proteins with log2 SILAC H/L ratios

764 ≥1.5 or ≤-1.5. The bar graphs show significantly overrepresented UniProt keywords. B,C,

765 Immunoblot analysis of anti-myc immunoprecipitates from lysates derived from parental HeLa

766 and GABARAPL2endoHA cells transiently transfected for 48 hrs with myc-tagged ATG7 (B) or

767 UBA5 (C). Stability and knockdown of ACSL3. D, GABARAPL2endoHA cells were treated as

768 indicated and subjected to lysis and analyzed with immunoblotting and anti-ACSL3 antibody.

769 E, Reversely transfected GABARAPL2endoHA cells with non-targeting (sicrtl) or GABARAPL2

770 siRNA were lysed followed by immunoblotting and analysis with indicated antibodies. F,G,

771 GABARAPL2endoHA/ACSL3endoNeonGreen cells were transfected with indicated siRNAs prior to

772 immunolabeling with Calnexin or SEC13. Scale bar: 10 µm.

773

774

30 bioRxiv preprint doi: https://doi.org/10.1101/2020.01.01.892521; this version posted January 1, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 1

C HA HA

A endoHA B sictrl

kDa sictrl siGABARAPL2

kDa HeLa GABARAPL2 17 HA 17 HA 35 PCNA 35 PCNA GABARAPL2endoHA siGABARAPL2

GABARAPL2endoHA

E HA p62 HA p62 D DMSO

kDa DMSO Torin1 BafA1 + BafA1 Torin1 Inhibitor ATG7 17 HA

35 PCNA

GABARAPL2endoHA Torin1 / BafA1 Torin1

GABARAPL2endoHA

F 12 G myc-p62 Input HA-IP HADHA CHMP4A GABARAPL2 AHNAK2 EZR 10 RPS2 HAendo HAendo EIF4G1 ERO1L WDR1 TPT1 ATP13A1 8 ATG7 intensities VDAC1 kDa HA-beads HeLa GABARAPL2 HeLa GABARAPL2

10 EFR3A IRGQ TFG LRPPRC 63

Log myc TRIM21 6 CCPG1 p62 ACSL3 RB1CC1 IPO5 17 UBA5 HA HADHB

-8 -6 -4 -2 0 2 4 6 8 Log2 (SILAC Ratio) ATG7 Inhibitor/ Torin1 & BafA1 myc-ATG7 H Input HA-IP ACSL3-myc

I endoHA endoHA Input HA-IP HAendo HAendo

kDa HA-beads HeLa GABARAPL2 HeLa GABARAPL2

75 myc

kDa HA-beads HeLa GABARAPL2 HeLa GABARAPL2 75 myc 75 myc (l. e.)

75 myc (l. e.) 17 HA

17 HA (l. e.) 17 HA bioRxiv preprint doi: https://doi.org/10.1101/2020.01.01.892521; this version posted January 1, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 2

A endoNeonGreen B NeonGreen Calnexin NeonGreen Calnexin /ACSL3 endoHA endoHA

endoHA endoNeonGreen GABARAPL2 GABARAPL2 TOMM20-NeonGreen HeLa + GABARAPL2 /ACSL3 kDa

ACSL3 75 NeonGreen Phospholipids Neutral lipids NeonGreen Phospholipids Neutral lipids

ACSL3 (l. e.) C 75 EtOH 17 HA

35 PCNA

oleic acid 75

48 NeonGreen

oleic acid

75 GABARAPL2endoHA/ACSL3endoNeonGreen 48 NeonGreen (l. e.)

35 D PCNA NeonGreen HA NeonGreen HA

DMSO

Torin1 + BafA1

ATG7 Inhibitor

GABARAPL2endoHA/ACSL3endoNeonGreen bioRxiv preprint doi: https://doi.org/10.1101/2020.01.01.892521; this version posted January 1, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 3

A 1.2 ** *

sictrl siACSL3#1 siACSL3#2 1 kDa 75 ACSL3 0.8 0.6 17 HA 0.4 protein levels 35 PCNA 0.2 Normalized GABARAPL2 GABARAPL2HAendo 0 sictrl siACSL3#1 siACSL3#2

B n.s. 1.4 n.s. 1.2 sictrl siACSL3#1 siACSL3#2 kDa 1 75 ACSL3 0.8 HA 0.6 17

protein levels 0.4 35 PCNA 0.2 Normalized GABARAP Normalized GABARAP

HAendo 0 GABARAP sictrl siACSL3#1 siACSL3#2

C n.s. n.s. 1.2

kDa sictrl siACSL3#1 siACSL3#2 1 75 ACSL3 0.8

17 HA 0.6

protein levels 0.4 17 HA (l. e.) 0.2

35 Normalized GABARAPL1 PCNA 0 GABARAPL1HAendo sictrl siACSL3#1 siACSL3#2

* D 1.6 * 1.4 1.2

kDa sictrl siACSL3#1 siACSL3#2 1 75 ACSL3 0.8 0.6

17 protein levels HA 0.4 Normalized LC3B 0.2 35 PCNA 0 sictrl siACSL3#1 siACSL3#2 LC3BHAendo

DMSO Treatment E

kDa sictrl siACSL3#1 siACSL3#2 sictrl siACSL3#1 siACSL3#2 75 ACSL3

17 HA

35 BafA1 PCNA

75 ACSL3

17 HA Btz 35 PCNA

GABARAPL2HAendo bioRxiv preprint doi: https://doi.org/10.1101/2020.01.01.892521; this version posted January 1, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 4

A myc-UBA5 B myc-UBA5 myc-IP Input Input HA-IP EtOH + + + + 24 hrs oleic acid + + HAendo HAendo ACSL3-HA ACSL3-HA ACSL3-HA ACSL3-HA myc-beads HeLa HeLa kDa HA-beads HeLa GABARAPL2 HeLa GABARAPL2 kDa 48 myc 75 HA

48 myc 48 myc (l. e.)

myc (l. e.) 17 HA 48

C NeonGreen UBA5 HA NeonGreen UBA5 HA

EtOH

30 min oleic acid

24 hrs oleic acid

Figure 5

A DMSO Treatment

kDa sictrl siACSL3#1 siACSL3#2 sictrl siACSL3#1 siACSL3#2 75 ACSL3

48 UBA5 BafA1

PCNA 35

75 ACSL3

48 UBA5 Btz

35 PCNA

GABARAPL2HAendo

B DMSO BafA1 Btz C 30 min oleic acid 24 hrs oleic acid EtOH kDa kDa EtOH 30 min oleic acid 24 hrs oleic acid kDa sictrl siACSL3#1 sictrl siACSL3#1 sictrl siACSL3#1 75 ACSL3

75 ACSL3 35 PCNA 35 PCNA 17 HA 35 UFM1 Ubl 20 35 UFC1 E2 PCNA 25 35 PCNA 48 UBA5 E1 11 48 DDRGK1 E3 35 35 PCNA PCNA 35 PCNA 20 UFC1 E2 100 UFL1 E3 35 PCNA

35 PCNA 48 DDRGK1 E3 63 UFM1 Ubl 35 PCNA

100 UFL1 E3 35 35 35 PCNA 35 PCNA 25 UFM1 Ubl GABARAPL2HAendo GABARAPL2HAendo

35 PCNA

GABARAPL2HAendo

D Completion of LD formation LD component anchores LD induction blocks ufmylation increases ufmylation UBA5 at ER membrane

transcriptional inhibition? LD disassembly/ degradation

? ? cytosol LD formation

ER-lumen 30 min oleic acid 24 hrs oleic acid

ACSL3 GABARAPL2 UBA5 UFM1 UFC1 DDRGK1 UFL1 CDK5R3 Substrate bioRxiv preprint doi: https://doi.org/10.1101/2020.01.01.892521; this version posted January 1, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure S1

A genomic DNA

knocked-in DNA-sequence

px330

plasmid with gRNA and Cas9 Cas9

PCR-product gRNA

PCR-products: ATG8-genes

5‘UTR BSD P2A HA first exon

Cas9 mediated doublestrandbreak ACSL3-gene last exon NeonGreen T2A BSD 3‘UTR

GABARAPendoHA:TTCGTGGATCGCTCCGCTGAATCCGCCCGCGCGTCGCCGCCGTCGTCGCCGCCCCCCGTCCCGGCCCCCCTGGGTTCCCTCAGCCCAGCCCT B GTCCAGCCCGGTTCCGGGGAGGATGAAGCCGGCCAAGCCTTTGTCTCAAGAAGAATCCACCCTCATTGAAAGAGCAACGGCTACAATCAACAGCATCCCCATCT CTGAAGACTACAGCGTCGCCAGCGCAGCTCTCTCTAGCGACGGCCGCATCTTCACTGGTGTCAATGTATATCATTTTACTGGGGGACCTTGTGCAGAACTCGTGG TGCTGGGCACTGCTGCTGCTGCGGCAGCTGGCAACCTGACTTGTATCGTCGCGATCGGAAATGAGAACAGGGGCATCTTGAGCCCCTGCGGACGGTGCCGACA GGTGCTTCTCGATCTGCATCCTGGGATCAAAGCCATAGTGAAGGACAGTGATGGACAGCCGACGGCAGTTGGGATTCGTGAATTGCTGCCCTCTGGTTATGT

GABARAPL1endoHA:TGCACACTCGGCCCAGCGCTGTTGCCCCCGGAGCGGACGTTTCTGCAGCTATTCTGAGCACACCTTGACGTCGGCTGAGGGAGCGGGACAG GGTCAGCGGCGAAGGAGGCAGGCCCCGCGCGGGGATCTCGGAAGCGCTGCGGTGCATCATGAAGCCGGCCAAGCCTTTGTCTCAAGAAGAATCCACCCTCATT GAAAGAGCAACGGCTACAATCAACAGCATCCCCATCTCTGAAGACTACAGCGTCGCCAGCGCAGCTCTCTCTAGCGACGGCCGCATCTTCACTGGTGTCAATGT ATATCATTTTACTGGGGGACCTTGTGCAGAACTCGTGGTGCTGGGCACTGCTGCTGCTGCGGCAGCTGGCAACCTGACTTGTATCGTCGCGATCGGAAATGAGA ACAGGGGCATCTTGAGCCCCTGCGGACGGTGCCGACAGGTGCTTCTCGATCTGCATCCTGGGATCAAAGCCATAGTGAAGGACAGTGATGGACAGCCGACGGC AGTTGGGATTCGTGAATTGCTGCCC

GABARAPL2endoHA:GCCCCTTTACGTGCGGCCCCGCCCCTTGGCGTGGCGCCCTGACAAATGGCGCCGGAAGCCCCGCCCCCGGCCGGTTGCTAGGCTCCGACA GCCGGAAGTCCCGCCTGCCGTGTAGTCGCCGCCGTCGCTGCCGCTGCCGCTGCCGCCGTCGTTGTTGTTGTGCTCGGTGCGCTGAGCTCCGCGGCTCCGCGAG CCGGTTCCGTCCCCTTCCCGCCGCGGCCATGAAGCCGGCCAAGCCTTTGTCTCAAGAAGAATCCACCCTCATTGAAAGAGCAACGGCTACAATCAACAGCATCC CCATCTCTGAAGACTACAGCGTCGCCAGCGCAGCTCTCTCTAGCGACGGCCGCATCTTCACTGGTGTCAATGTATATCATTTTACTGGGGGACCTTGTGCAGAAC TCGTGGTGCTGGGCACTGCTGCTGCTGCGGCAGCTGGCAACCTGACTTGTATCGTCGCGATCGGAAATGAGAACAGGGGCATCTTGAGCCCCTGCGGACGGTG CCGACAGGTGCTTCTCGATCTGCATCCTGGGATCAAAGCCATAGTGAAGGACAGTGATGGACAGCCGACGGCAGTTGGGATTCGTGAATTGCTGCCCTCTGGTT ATGTGTGGGAGGGC

LC3BendoHA:CTGCGTGCCGCTGCTGGGTTCCGCCACGCCCGTCATGGCGGCGGCCCCGGCCGGCTCTGGCCCCGCCCCTCGGTGACGCGTCGCGAGTCACCTGA CCAGGCTGCGGGCTGAGGAGATACAAGGGAAGTGGCTATCGCCAGAGTCGGATTCGCCGCCGCAGCAGCCGCCGCCCCCGGGAGCCGCCGGGACCCTCGCGT CGTCGCCGCCGCCGCCGCCCAGATCCCTGCACCATGCCGGCCAAGCCTTTGTCTCAAGAAGAATCCACCCTCATTGAAAGAGCAACGGCTACAATCAACAGCAT CCCCATCTCTGAAGACTACAGCGTCGCCAGCGCAGCTCTCTCTAGCGACGGCCGCATCTTCACTGGTGTCAATGTATATCATTTTACTGGGGGACCTTGTGCAGA ACTCGTGGTGCTGGGCACTGCTGCTGCTGCGGCAGCTGGCAACCTGACTTGTATCGTCGCGATCGGAAATGAGAACAGGGGCATCTTGAGCCCCTGCGGACGG TGCCGACAGGTGCTTCTCGATCTGCATCCTGGGATCAAAGCCATAGTGAAGGACAGTGATGGACAGCCGACGGCAGTTGGGATTCGTGAATTGCTGCCCTCTGG TTATGTGTGGGAGGGC

ACSL3endoNeonGreen:TATTTTTTTTTAATCATCTTAGCAAGTCTGGAAAAGTTTGAAATTCCAGTAAAAATTCGTTTGAGTCATGAACCGTGGACCCCTGAAACTGGTCTGGT C GACAGATGCCTTCAAGCTGAAACGCAAAGAGCTTAAAACACATTACCAGGCGGACATTGAGCGAATGTATGGAAGAAAAGCTGGCGGCATGGTGAGCAAGGGCGA GGAGGATAACATGGCCTCTCTCCCAGCGACACATGAGTTACACATCTTTGGCTCCATCAACGGTGTGGACTTTGACATGGTGGGTCAGGGCACCGGCAATCCAA ATGATGGTTATGAGGAGTTAAACCTGAAGTCCACCAAGGGTGACCTCCAGTTCTCCCCCTGGATTCTGGTCCCTCATATCGGGTATGGCTTCCATCAGTACCTGCC CTACCCTGACGGGATGTCGCCTTTCCAGGCCGCCATGGTAGATGGCTCCGGATACCAAGTCCATCGCACAATGCAGTTTGAAGATGGTGCCT bioRxiv preprint doi: https://doi.org/10.1101/2020.01.01.892521; this version posted January 1, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure S2 endoHA

A endoHA B C endoHA HeLa LC3B HeLa GABARAPL1 HeLa GABARAP kDa kDa kDa HA HA HA 17 17 17 35 35 PCNA 35 PCNA PCNA

D E F sictrl siLC3B sictrl siGABARAP kDa kDa sictrl siGABARAPL1 kDa HA HA 17 HA 17 17 35 35 35 PCNA PCNA PCNA

GABARAPendoHA GABARAPL1endoHA LC3BendoHA G H I ATG7 Inhibitor ATG7 kDa DMSO Torin1 BafA1 + BafA1 Torin1 ATG7 Inhibitor ATG7 HA DMSO Torin1 BafA1 + BafA1 Torin1

17 DMSO Torin1 BafA1 + BafA1 Torin1 Inhibitor ATG7 kDa kDa HA HA 17 17 HA (l. e.) 17 35 35 PCNA PCNA 35 PCNA endoHA GABARAPL1endoHA LC3B

GABARAPendoHA J K HA HA LAMP1 HA LAMP1 LC3 HA LC3

DMSO

Torin1 + BafA1

GABARAPL2endoHA GABARAPL2endoHA bioRxiv preprint doi: https://doi.org/10.1101/2020.01.01.892521; this version posted January 1, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure S3

A B myc-ATG7 Uniprot Keywords Input myc-IP

Fatty acid metabolism 2.78x10-03 endoHA endoHA Protease inhibitor 2.78x10-03 Ribonucleoprotein 2.17x10-03 RNA-binding 2.01x10-03 kDa myc-beads HeLa GABARAPL2 HeLa GABARAPL2 Protein biosynthesis 2.86x10-06 75 myc Initiation factor 3.11x10-07 0 5 10 15 17 HA Number of proteins with log2 SILAC ratios ≤-1.5 or ≥1.5

C myc-UBA5 Input myc-IP D E sictrl siGABARAPL2 HAendo HAendo kDa ATG7 Inhibitor ATG7 kDa DMSO Torin1 BafA1 + BafA1 Torin1 75 ACSL3 75 ACSL3 17 HA

myc-beads HeLa GABARAPL2 HeLa GABARAPL2 17 HA kDa 35 PCNA 48 myc 48 HA UBA5 48 myc (l. e.) 17 17 HA 17 HA 35 PCNA 35 PCNA GABARAPL2endoHA GABARAPL2endoHA

Calnexin/DAPI Calnexin F G SEC13/DAPI SEC13

sictrl sictrl

siACSL3#1 siACSL3#1

endoHA endoNeonGreen GABARAPL2 /ACSL3 GABARAPL2endoHA/ACSL3endoNeonGreen