{"paper_id":"ffb30036-1e37-46c0-ae77-e7ab16f59156","body_text":"1 \nNonuple atg8 mutant provides genetic evidence for \nfunctional specialization of ATG8 isoforms in Arabidopsis thaliana \nAlessia Del Chiaro1,2, *, Nenad Grujic1, *, Jierui Zhao1,2, Ranjith Kumar Papareddy1, Peng Gao1, Juncai \nMa3, Christian Lofke1, Anuradha Bhattacharya 1, Ramona Gruetzner 4, Pierre Bourguet1, Frédéric \nBerger1, Byung-Ho Kang3, Sylvestre Marillonnet4, Yasin Dagdas1  \nAffiliations \n1Gregor Mendel Institute, Austrian Academy of Sciences, Vienna BioCenter, Vienna, Austria. \n2Vienna BioCenter PhD Program, Doctoral School of the University at Vienna and Medical University of Vienna, Vienna, \nAustria \n3School of Life Sciences, Centre for Cell & Developmental Biology and State Key Laboratory of Agrobiotechnology, The 10 \nChinese University of Hong Kong, Shatin, New Territories, Hong Kong, China \n4Leibniz-Institut für Pflanzenbiochemie, Department of Cell and Metabolic Biology, Halle, Germany. \n* These authors contributed equally \nCorrespondence: Yasin Dagdas (yasin.dagdas@gmi.oeaw.ac.at) \n \nAbstract \nAutophagy sustains cellular health by recycling damaged or excess components through \nautophagosomes. It is mediated by conserved ATG proteins, which coordinate autophagosome \nbiogenesis and selective cargo degradation. Among these, the ubiquitin-like ATG8 protein plays a \ncentral role by linking cargo to the growing autophagosomes through interacti ng with selective 20 \nautophagy receptors. Unlike most ATG proteins, the ATG8 gene family is significantly expanded in \nvascular plants, but its functional specialization remains poorly understood.  Using \ntranscriptional and translational reporters in Arabidopsis thaliana , we revealed that ATG8 \nisoforms are differentially expressed across tissues and form distinct autophagosomes within the \nsame cell. To explore ATG8 specialization, we generated the nonuple Δatg8 mutant lacking all \nnine ATG8 isoforms. The mutant displayed hypersens itivity to carbon and nitrogen starvation, \ncoupled with defects in bulk and selective autophagy as shown by biochemical and \nultrastructural analyses. Complementation experiments demonstrated that ATG8A could rescue \nboth carbon and nitrogen starvation phenotypes, whereas ATG8H could only complement carbon \nstarvation. Proximity labeling proteomics further identified isoform -specific interactors under 30 \nnitrogen starvation, underscoring their functional divergence. These findings provide genetic \nevidence for functional specialization of ATG8 isoforms in plants and lay the foundation for \ninvestigating their roles in diverse cell types and stress conditions. \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 2 \nIntroduction \nAutophagy is an evolutionarily conserved cellular  quality control  mechanism essential for \nmaintaining homeostasis and adapting to environmental stresses (1-3). It functions by selectively \ndegrading and recycling damaged, redundant, or harmful cellular components, ensuring cellular \nintegrity and energy balance (4-6). Although autophagy occurs constitutively, it is highly inducible 40 \nunder stress conditions such as nutrient deprivation, hypoxia, or infection  (7-10). During these \nchallenges, autophagy sustains survival by facilitating the degradation of intracellular material, \nwhich is sequestered into specialized double -membrane compartments called \nautophagosomes (11, 12). These structures subsequently fuse with lytic organelles—the vacuole \nin plants and yeast or the lysosome in animals —where their contents are broken down and \nrecycled (3). \nContrary to initial views of autophagy as a non -selective bulk degradation process, it is now \nrecognized as a highly selective pathway  (13-16). This selectivity is mediated by specific \ninteractions between cargo receptors, known as selective autophagy receptors (SARs), and \nautophagy-related proteins such as ATG8 (17, 18). These interactions enable the precise targeting 50 \nof a wide range of substrates, from protein aggregates to damaged organelles, thus tailoring \nautophagic responses to specific cellular needs (19). \nAutophagosome biogenesis progresses through three tightly regulated stages: initiation, \nexpansion, and maturation. This process is orchestrated by the autophagy-related (ATG) protein \nfamily, comprising approximately 40 conserved members (20, 21). Central to autophagy is ATG8, \na ubiquitin-like protein crucial for autophagosome formation, cargo recruitment, and membrane \ntrafficking (22, 23 ). Once processed and lipidated, ATG8 associates with the autophagosome \nmembrane, acting as a scaffold for the assembly of other core autophagy machinery and cargo \nreceptors (24, 25). \nInterestingly, unlike most ATG proteins, which exist as single or a few isoforms, the ATG8 gene 60 \nfamily has undergone significant expansion in vascular plants (26, 27 ). While yeast and \neukaryotes from early branching groups encode a single ATG8, vascular plants possess multiple \nisoforms, with Arabidopsis thaliana containing nine distinct ATG8 genes (AtATG8a-i) forming two \nmajor clades. Clade I contains ATG8A -G, whereas Clade II contains ATG8H and ATG8I  (28). \nDespite moderate sequence divergence and differential expression patterns  (29, 30 )the \nbiological significance of ATG8 isoform expansion and its implications for selective autophagy \nremain poorly understood. \nHere, we present genetic evidence for the functional specialization of ATG8 isoforms in A. \nthaliana by generating a nonuple ATG8 mutant lacking all nine ATG8 genes. Our study reveals that \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 3 \nATG8 isoforms not only exhibit distinct tissue-specific expression and subcellular localization but 70 \nalso differ in their ability to mediate autophagic responses under specific stress conditions. \nThese findings highlight the complex regulatory landscape of autophagy in plants and provide a \nfoundation for unraveling the mechanisms underlying ATG8 specialization. \n \nResults and Discussion \nArabidopsis thaliana  ATG8 isoforms exhibit tissue -specific expression patterns and form \ndistinct autophagosomes within root cells \nTo explore ATG8 specialization, we first checked the expression patterns of all the nine \nArabidopsis ATG8 isoforms. We generated ATG8 promoter -GFP-GUS (pATG8X::GFP -GUS) \nexpressing lines and performed β-glucuronidase (GUS) staining (Fig. 1A). Interestingly,  AtATG8s 80 \nexhibit distinct expression patterns and we only observed a partial overlap between different \nisoforms, indicating a certain degree of tissue specificity. ATG8E, ATG8F and ATG8G, show a more \nwidespread expression pattern over different tissues and organs, whereas other isoforms, \nincluding ATG8A, ATG8B and ATG8I, appear to be restricted to the root (Fig. 1A). Furthermore, \nsome isoforms exhibit peculiar tissue - or cell-type specificities, such as ATG8D being strongly \ninduced in the apex of the cotyledon or AT G8C appearing specifically expressed in guard cells \n(Fig. 1A). These observations hint at a potential tissue - or cell -type-specific function of the \ndifferent ATG8 isoforms.  \nThen, we decided to test the subcellular compartmentalization of ATG8 isoforms. We co -\nexpressed mCherry-ATG8E translational fusion constructs with GFP -tagged versions of ATG8A, 90 \nATG8D and ATG8I and assessed their co -localization upon bulk autophagy inducing chemical \nTorin1 (31) and ER-stress inducer Tunicamycin (32) treatments (Fig. 1B and C). Irrespective of the \ntreatment, ATG8E colocalized almost entirely with ATG8A. Conversely, ATG8D and ATG8I exhibit \na lower degree of colocalization with ATG8E during Torin 1 treatment and even more \npronouncedly upon Tunicamycin t reatment. ATG8I, representative of the clade II of ATG8 \nisoforms, has the weakest level of colocalization with ATG8E. These data indicate there are \ndistinct pools of autophagosomes that are labelled with different ATG8 isoforms.  In sum, the \nexpression sit e polymorphism as well as varying levels of colocalization upon different \ntreatments, support the hypothesis that ATG8 isoforms might fulfill different functions or respond \nto different stimuli.  100 \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 4 \natg8 nonuple mutant (Δatg8) is deficient in autophagy \nThe partial overlap in expression and colocalization patterns of ATG8 isoforms prompted us to \ngenerate an ATG8-free Arabidopsis thaliana line that we could use as a tool to investigate ATG8 \nspecialization. Using multiplex CRISPR mutagenesis, we combined 9 guide RNAs ( one for each \nATG8 isoform) in a construct containing an intronized Cas9 (33) and transformed it to generate a \nnonuple knock-out of atg8. We first confirmed the mutations using whole genome sequencing \n(Fig. 2A). To functionally verify Δatg8 as an autophagy deficient mutant, we performed the typical \nnutrient starvation assays  (34). We transferred 9 -days old Arabidopsis seedlings to carbon or \nnitrogen deprived ½ MS liquid medium for 4 or 6 days, respectively. Similar to the autophagy -110 \ndeficient mutants atg2 and atg5 (35, 36), Δatg8 exhibits reduced growth and discoloration of the \ncotyledons (Fig. 2B and C). To biochemically validate Δatg8 mutant, we performed autophagic \nflux assays under carbon and nitrogen starvation conditions, by measuring the endogenous \nlevels of the stereotypical autophagy substrate NBR1 (Fig. 2D and E) (37). NBR1 accumulated at \na comparable level in both Δatg8 and atg5 under all conditions and was insensitive to \nconcanamycin A treatment that blocks vacuolar degradation (38), denoting that both mutants are \ndefective in autophagic degradation (Fig. 2D and E). Collectively, these results suggest that atg8 \nnonuple mutant is deficient in autophagy. \nΔatg8 mutant is defective in selective autophagy \nNitrogen and carbon starvation are considered to trigger bulk autophagy (39, 40). To test if Δatg8 120 \nis also defective in selective autophagy, we tested its ability to perform mitophagy and pexophagy. \nFor mitophagy, we measured the levels of the outer mitochondrial membrane voltage dependent \nanion channel I (VDAC) and the matrix protein isocitrate dehydro genase (IDH) upon 2,4 -\ndinitrophenol (DNP) treatment. DNP is an uncoupler that leads to mitochondrial depolarization \nand triggers mitophagy (41, 42) (Fig. 3A). Although both VDAC and IDH levels were decreased in \nthe DNP treated wild type Col-0 plants, Δatg8 behaved similar to atg5 and showed no change in \nVDAC or IDH levels (41). Likewise, when we assessed peroxisome degradation using the catalase \nantibody, Δatg8 mutant behaved similar to atg5 mutant and was unable to perform pexophagy \n(43) (Fig. 3B). Altogether, these results suggest Δatg8 mutant is defective in mitophagy and \npexophagy. 130 \nNext, we performed transmission electron microscopy (TEM) analysis of mitophagy in \nArabidopsis root cells. Although we could detect mitophagosomes in wild type Col -0 plants, we \ndid not observe any mitophagosomes in Δatg8 mutant (Fig. 3C). However, we could still detect \ndouble-membraned structures that resemble autophagosomes in the Δatg8 mutant. These \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 5 \nvesicles appear to non-specifically engulf various types of cellular components. Although further \nstudies are necessary to understand the nature of these compartments, a plausible source could \nbe provacuoles, compartments that appear during vacuole biogene sis (44). Indeed, provacuole \nformation has been reported to be independent of autophagy as autophagy -deficient mutants \ncan still form provacuoles (44). In summary, our data suggest Δatg8 mutant is unable to carry out \nautophagic recycling.  140 \nComplementation of Δatg8 with ATG8A or ATG8H reveals functional specialization \nAfter confirming the autophagy-deficient phenotype of the Δatg8 mutant, to assess the functional \nspecialization of ATG8 isoforms, we complemented it with GFP tagged ATG8A and ATG8H, \nrepresenting both clades. First, we analyzed the complemented lines with confocal microscopy. \nBoth GFP-ATG8A and GFP-ATG8H formed cytoplasmic bright puncta, which further accumulated \nin the vacuole upon Concanamycin A treatment and underwent autophagic flux (Fig. S2A-B).  \nTo assess to what extent they were able to recover the autophagic function and evaluate potential \nisoform-specific responses to different stressors, we subjected the two complementation lines \nto starvation assays. Upon carbon starvation, both complementati on mutants exhibit a similar \nphenotype to Col -0 (Fig. 4A). This suggests that both ATG8A and ATG8H could mediate the 150 \nautophagic recycling of cellular material that ensure survival during carbon deprivation. In \ncontrast, during nitrogen starvation only GFP -ATG8A expression recovered the sensitivity to \nnitrogen starvation. GFP-ATG8H expressing lines were similar to the Δatg8 mutant (Fig. 4B). These \nresults provide functional genetic evidence for ATG8H specialization in Arabidopsis thaliana. \nTo support these findings, we performed autophagic flux assays under carbon and nitrogen \nstarvation. Under carbon starvation conditions, both Δatg8/+GFP-ATG8A and Δatg8/+GFP-ATG8H \nhad similar NBR1 flux, in contrast to the Δatg8 mutant (Fig. 4C). This is consistent with the \nphenotyping results and indicates that both ATG8A and ATG8H are able to trigger bulk autophagy \nin response to carbon deprivation. However, during nitrogen starvation NBR1 flux in Δatg8/+GFP-\nATG8H was similar to the Δatg8 mutant (Fig. 4D), further corroborating the hypothesis that ATG8H, 160 \nunlike ATG8A, is not able to fully operate autophagy in response to nitrogen deprivation.  \nATG8A and ATG8H have distinct proxitomes during nitrogen starvation \nFollowing the observation that ATG8A and ATG8H do not respond equally to N starvation stress, \nwe reasoned that the two isoforms may be interacting with different proteins that are involved in \nautophagy signaling or cargo recognition. Indeed, our results co uld be explained by the inability \nof ATG8H to engage with the nitrogen starvation response signaling, or its failure to associate with \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 6 \nthe selective autophagy receptors recognizing the cargoes that need to be degraded to cope with \nnitrogen deprivation. To test these hypotheses, we complemented the Δatg8 mutant with the \nbiotinylating enzyme TurboID fused ATG8A and ATG8H and determined the ATG8 proxitomes \nduring nitrogen starvation (Fig. 4E). We used TurboID alone as negative control. 47 proteins 170 \nexhibited specific association with both or only one of the two ATG8 isoforms  (Supplementary \nTable 3) . Among these, numerous well -known interactors were found, including several ATG \nproteins. Interestingly, whereas some ATG proteins showed similar levels of association with \nATG8A and ATG8H, such as ATG3, ATG7, ATG14B and ATG18F , other ATG proteins appear to \ninteract prevalently or exclusively with just one isoform  (Supplementary Table 3). In the case of \nATG1, ATG1A seems to interact with ATG8A uniquely, whi le ATG1B can associate with both \nisoforms but still exhibits a stronger association with ATG8A in the conditions tested. ATG1 kinase \ninitiates autophagosome biogenesis and constitutes one of the major targets for autophagy \nregulation (45, 46). In light of this, our results may indicate that during N starvation ATG1A and \nATG1B recruit ATG8A preferentially to promote nutrient replenishment. It is also interesting to 180 \nobserve that less than one third of the total ATG8 interactors are shared between ATG8A and \nATG8H (Fig. 4F), whilst 22 proteins specifically interact with ATG8A, and 8 proteins interact with \nonly ATG8H. As a proof on concept, our proximity labeling analysis suggested that the adaptor \nprotein CFS1 specifically interacts with ATG8A but not with ATG8H, consistent with our previous \nresults (47). While other interactors need to be further validated, these results suggest single -\nisoform TurboID lines provide an effective tool to study ATG8 specialization in a wide range of \nstress conditions. \nSince the core autophagy machinery is shared across various selective autophagy pathways, the \nquestion of how cells achieve subcellular compartmentalization of these concurrent \nmechanisms remains unresolved. A plausible explanation is ATG8 isoform specialization, 190 \nwhereby distinct ATG8 variants interact with specific adaptors, receptors, or ATG proteins to \ndirect and compartmentalize autophagic processes. Previous biochemical and proteomic \nanalyses in potato supported this hypothesis, revealing isoform -specific inte ractomes (48). \nHowever, the co-occurrence of multiple ATG8 isoforms within individual autophagosomes (Fig. \n1) suggests that some unique interactors might have been overlooked. \nIn this study, we provide genetic evidence for ATG8 specialization in plants using an Arabidopsis \nΔatg8 nonuple mutant complemented with individual ATG8 isoforms. Unlike mitophagy observed \nin HeLa cells lacking multiple ATG8 genes, the Arabidopsis Δatg8 mutant failed to perform \nmitophagy and pexophagy, as evidenced by the absence of mitophagosomes and the \naccumulation of autophagy substrates (Fig. 3). While double -membraned vesicles were 200 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 7 \nobserved in Δatg8 cells, these are likely provacuolar compartments, which are independent of \nautophagy. \nInterestingly, complementation experiments revealed functional divergence among ATG8 \nisoforms. While both ATG8A and ATG8H restored carbon starvation sensitivity, only ATG8A was \nable to complement nitrogen starvation sensitivity (Fig. 4). Consistent with th eir specialization, \nproximity labeling proteomics demonstrated distinct interactomes for ATG8A and ATG8H. Further \nstudies are necessary to link the nitrogen sensitivity phenotype to the differentially interacting \nproteins. Nevertheless, our findings establ ish the functional specialization of ATG8 isoforms in \nplants, providing a framework for understanding how cells fine -tune autophagic processes in \nresponse to diverse and overlapping signals. 210 \n \nMaterial and Methods \nPlant material and cloning \nAll Arabidopsis thaliana lines used originate from the Columbia (Col -0) ecotype. The Δatg8 \nmutant was generated with two rounds of CRISPR editing. First, the CRISPR/Cas9 construct \nincluding gRNAs for all nine AtATG8 genes and an intronized Cas9 was assembled according to a \nprotocol previously described (33). The construct was assembled in binary vector pAGM62636 \n(33) that contains a p15 origin of replication for low copy number replication in E. coli  and an \nAgrobacterium rhizogenes  A4 ori for single copy replication in Agrobacterium, resulting in \nplasmid pAGM70811. This vector backbone was chosen to minimize the risk or recombination 220 \nbetween the 9 guides RNA cassettes present on the same plasmid. The gRNAs sequences are the \nfollowing: \nATG8A (AT4G21980): AGCTTACGGGAATTCTGTCA \nATG8B (AT4G04620): GAACTCAATACAGGTGATTG \nATG8C (AT1G62040): CAGTTAGATCAGCTGGAACA \nATG8D (AT2G05630): AAGAGGATGTTCATGCTTG \nATG8E (AT2G45170): CTGTTAGGTCTGATGGCACA \nATG8F (AT4G16520): TTCAGAGAAGAGAAGGGCAG \nATG8G (AT3G60640): GAGGAGACAGTACCGGTGGG \nATG8H (AT3G06420): AAACGCAGATCTGCCAGACA 230 \nATG8I (AT3G15580): GATGAAAGGCTCGCGGAGTCG \nPrimary transformants were genotyped by amplicon sequencing. The data was analysed using a \ncustom-built pipeline to retrieve indel frequencies in the ATG8 genes, using bwa and samtools \nmpileup (49, 50 ). Code and documentation are available at https://github.com/pierre-\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 8 \nbourguet/CRISPR_genotyping. Following sequencing, we could only confirm the successful \nmutation of eight of the nine isoforms. Therefore, a second CRISPR/Cas9 was employed to target \nthe last gene (ATG8D – AT2G05630). The construct was assembled according to the protocol \npreviously de scribed (51), on pHEE401E plasmid, with the following gRNAs sequences: \nGAACAACAGAGACTCGACCA, GTGATGTCCCGGATATTGAT. The nonuple mutant Δatg8 contains \nboth T-DNA CRISPR/Cas9 cassettes and is resistant to BASTA and Hygromycin. The mutations of 240 \nnine ATG8 isoforms were confirmed via single cell sequencing.  \nAll the plasmids, except pAGM7081, were assembled through the GreenGate cloning procedure \n(52) and were constructed as follows: pGGZ003_ATG8X::GFP -GUS (X represents the 9 ATG8 \nisoforms, from A to I), pGGSUN_RPS5::mCherry-TurboID; pGGSUN_RPS5::mCherry-TurboID-\nATG8A; pGGSUN_RPS5::mCherry -TurboID-ATG8H; pGGSUN_HTR5::GFP -ATG8A; \npGGSUN_HTR5::GFP-ATG8H. Apart from pATG8X::GFP -GUS expressing plants, which are \nhygromycin resistant, t ransformants were selected via seed coat fluorescence. The coding \nsequences of ATG8A and ATG8H carry silent mutations to avoid CRISPR/Cas9 targeting. The point \nmutations from the start codon are the following: ATG8A, 81bp T > A, 84bp C > T, 87bp A > G, 93bp \nC > G; for ATG8H, 123bp C > T, 126bp A > C, 129bp T > C, 135bp A > G, 138bp C > T. 250 \nDNA sequencing and analysis \nHigh-quality DNA was extracted using the cetyltrimethylammonium bromide (CTAB) method and \nused to construct Illumina-compatible libraries with the Nextera XT DNA Library Preparation Kit, \nfollowing the manufacturer’s instructions. Sequencing was performed o n an Illumina NextSeq \ninstrument in paired-end 150 bp mode. \nRaw FASTQ files obtained from sequencing were quality -checked and adapter -trimmed using \nTrimGalore ( https://github.com/FelixKrueger/TrimGalore) with default settings. The trimmed \nFASTQ files were aligned to the TAIR10 genome using Bowtie2 (51) with the parameters -D 15 -\nR 2 -N 0 -L 22 -i S,1,1.15. The resulting aligned BAM files were sorted and indexed using \nSAMtools (52) and manually inspected for insertions or deletions using the Integrative Genomics 260 \nViewer (IGV). Deletions and insertions were manually inspected, and the corresponding changes \nin cDNA and protein sequences were catalogued (Supplementary Table 1). \nPlant growth and treatments \nFor standard plant growth, Arabidopsis seeds were gas sterilized with sodium hypochlorite + HCl \n(10:1 v/v), sown on water-saturated soil and grown in 16h light/8h dark photoperiod with 165 μmol \nm^-2 s^ -1 light intensity. For in vitro growth, Arabidopsis seeds were surface sterilized in 70% \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 9 \nethanol for 10 minutes twice, then rinsed in absolute ethanol and dried on sterile paper. Seeds \nwere sown in ½ MS liquid medium (Murashige and Skoog salt + Gamborg B5 vitamin mixture \n[Duchefa] supplemented with 0.5 g/liter MES and 1% sucrose, pH 5.7), vernalized at 4 °C in the \ndark for 2 days , and then grown under LEDs with 85 μM/m²/s with a 14 h light/10 h dark 270 \nphotoperiod.  \nFor drug treatments, all drugs were dissolved in DMSO and added to the desired concentration: \n3 µM Torin 1 (Santa Cruz Biotechnology – CAS 1222998-36-8), 10 µg/mL Tunicamycin (Santa Cruz \nBiotechnology – CAS 11089 -65-9), 1-2 µM Concanamycin A (Santa Cruz Biotechnology – CAS \n80890-47-7), 50 µM DNP (Sigma -Aldrich, D198501-1KG). An equal amount of pure DMSO was \nadded to control samples.  \nFor confocal microscopy, Arabidopsis seeds were sterilized by 70% ethanol + 0.05% Tween 20 \n(Sigma-Aldrich) for 5 min and were subsequently sterilized by 100% ethanol for 10 min. Sterilized \nseeds were stored in sterile water at 4°C for 1 d for vernalization. Vernalized seeds were s pread \non 1/2 MS media plates (+1% plant agar [Duchefa]) and vertically grown at 21°C at 60% humidity 280 \nunder LEDs with 50 mM/m 2s a and a 16 h light/8 h dark photoperiod for 5 d ays. 5 -days old \nseedlings were incubated in ½ MS media containing either DMSO for 2 h, 3 μM Torin 1 for 2 h,  10 \nμg/mL Tunicamycin for 4 h or 2 μM Concanamycin for 2.5 h before imaging.  \nCarbon and Nitrogen starvation assays \nA. thaliana seeds (~30 per sample, 3 replicates per condition) were sterilized with ethanol, sown \nin ½ MS liquid medium, vernalized at 4 °C in the dark for 2 days and grown at 21 °C under LEDs \nwith 85 μM/m2/s and a 14 h light/10 h dark photoperiod.  The starvation treatments were \nperformed on 9-days old seedlings, by replacing ½ MS liquid medium with the same medium (as \ncontrol), ½ MS liquid medium without sucrose for C starvation or ½ MS liquid medium without \nnitrogen (Murashige & Skoog without Nitr ogen, Caisson Laboratories – MSP21) for N starvation. 290 \nPrior to medium replacement, the seedlings were washed twice with 1 mL of the new medium to \nensure proper removal of the previous medium. For C starvation, the seedlings were kept in the \ndark. Pictures of the samples were taken after 4 days of C starvation and 6 days of N starvation. \nConfocal microscopy \nFor confocal microscopy, Arabidopsis seedlings were placed on a microscope slide with water \nand covered with a coverslip. The epidermal cells of root transition and elongation zone were \nused for image acquisition. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 10 \nConfocal images were acquired via an upright point laser scanning confocal microscope ZEISS \nLSM800 Axio Imager.Z2 (Carl Zeiss) equipped with high -sensitive GaAsP detectors (Gallium \nArsenide), a LD C-Apochromat 40X objective lens (numerical aperture 1.1, water immersion), and 300 \nZEN software (blue edition 3.8, Carl Zeiss). GFP signals were excited at 488 nm and detected \nbetween 488 and 545 nm. mCherry signals were excited at 561 nm and detected between 570 \nand 617 nm.  \nImage processing and statistics \nConfocal images were processed and quantified by Fiji (version 1.52, Fiji).  The mCherry-ATG8E \ncolocalization ratio was calculated as the ratio of the number of mCherry -ATG8E puncta that \ncolocalized with GFP -ATG8 isoforms to the total number of mCherry -ATG8E puncta. Statistics \ntests were performed via GraphPad Prism 8.1.1. \nProtein extraction and western blotting  \n20-40 A. thaliana seeds per sample were surface sterilized with ethanol, sown in ½ MS liquid 310 \nmedium, vernalized at 4 °C in the dark for 2 days and grown for 7 days at 21 °C under LEDs with \n85μM/m2/s with a 14 h light/10 h dark photoperiod. For starvation treatments, performed \novernight, the ½ MS liquid medium was replaced with the same medium (as control), ½ MS liquid \nmedium without sucrose for C starvation or ½ MS liquid medium without nitrogen (Murashige & \nSkoog without Nitrogen, Caisson Laboratories – MSP21) for N starvation. For C starvation, the \nsamples were kept in the dark. When required, 1 µM Concanamycin A (Santa Cruz Biotechnology \n– CAS 80890 -47-7) was added to the new medium. The seedlings were harvested in safe lock \nEppendorf tubes containing 2 mm Ø glass bead, flash frozen in liquid nitrogen and pulverized \nusing a Silamat S7 (Ivoclar vi vident). Total proteins were extracted in 2X Laemmli buffer by \nagitating the samples in the Silamat S7 for 20 s. The samples were boiled at 70 °C and 1000 rpm 320 \nshaking for 10 min, then centrifuged at max speed with a benchtop centrifuge. Total proteins were \nquantified with the amido black method. 10 µl of supernatant was added to 190 µl of deionized \nwater, vortexed and then mixed with 1 ml of Amido Black Buffer (10% acetic acid, 90% methanol, \n0.05% [w/v] Amido Black (Napthol Blue Black, Sigma N3393)) by inverting the tubes. After 10 min \ncentrifugation at max speed, pellets were washed with 1 ml of Wash Buffer (10% acetic acid, 90% \nethanol), mixed by inversion, and centrifuged for another 10 minutes at max speed. Pellets were \nresuspended in 0.2N NaOH and OD630 nm was measured, with NaOH solution as blank, to quantify \nprotein concentration with the OD = a[C] + b determined curve. 15 µg of total protein extracts \nwere separated on SDS -PAGE gels and blotted onto PVDF Immobilon -P membrane (Millipore). \nNBR1 was detected using the anti -NBR1 antibody (Rabbit polyclonal; Agrisera – AS14 2805) 330 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 11 \ndiluted 1:10000. Catalase proteins were detected with Anti -Cat antibody (Rabbit polyclonal; \nAgrisera – AS09 501), diluted 1:1000. GFP was detected with the anti -GFP antibody (Mouse \nmonoclonal; Roche – 11814460001), diluted 1:5000. Rabbit polyclonal antibod y was detected \nwith a goat anti -rabbit IgG HRP-linked antibody (Invitrogen, 65 -6120) diluted 1:5000. Hybridized \nmembranes were reacted with SuperSignal ™ West Pico PLUS Chemiluminescent Substrate \n(Thermo Fisher Scientific) and imaged using an iBright CL1500 Imaging System (Invitrogen).  \nProtein bands were quantified with ImageJ according to the protocol previously described (53) \nand normalized on the loading control. The values reported on the figures correspond to mean \nvalues for 3 biological replicates. \nFor mitophagy assays, 5 -days old Arabidopsis seedlings were treated with 50 µM DNP (Sigma -340 \nAldrich, D198501-1KG), or an equal amount of DMSO, for 2 -3 hours in the dark, then moved to \nliquid ½ MS medium for 1h recovery under light. Protein extraction and immunoblot analysis was \nperformed as previously reported (41). \nGUS staining \n10-days old Arabidopsis seedlings expressing pATG8X:: GFP-GUS (X represents the 9 ATG8 \nisoforms, from A to I) were first immersed in acetone for 20 minutes and then washed with the \nGUS buffer (50 mM NaPO 4, 2 mM K -ferrocyanide, 2mM K -ferricyanide, 0.2% Triton X -100). The \nwashed samples were subsequently incubated in GUS staining buffer [GUS buffer + 2 mM X-Gluc \n(Thermo Scientific)] under 37℃ until a blue coloration was visible. The stained samples were then \nwashed and discolored with 100% ethanol and were ready for photographing. 350 \nAffinity purification of biotinylated proteins and nanoLC-MS/MS Analysis \nA. thaliana seeds were surface sterilized with ethanol, stratified for 2 days at 4 °C in the dark and \nthen grown in ½ MS (Duchefa)/0.5% MES/1% sucrose under LEDs with 85  μM/m2/s and a 14 h \nlight/10 h dark photoperiod. 7-days old seedlings were washed and treated with N deficient ½ MS \nmedium (or control ½ MS medium) overnight and the following morning 50 μM biotin was added \nto the medium. After 1 hour of biotin incubation, the seedlings were quickly rinsed in ice cold \nwater, dried and frozen in liquid nitrogen. Around 1 gram of plant tissue was used for each sample \nand the affinity purification of biotinylated proteins was performed as previously described (54). \nFor MS Analysis, the nano HPLC system (UltiMate 3000 RSLC nano system) was coupled to an \nOrbitrap Exploris 480 mass spectrometer, equipped with a Nanospray Flex ion source (all parts 360 \nThermo Fisher Scientific). Peptides were loaded onto a trap column (PepMap Acclaim C18, 5 mm \n× 300 μm ID, 5 μm particles, 100 Å pore size, Thermo Fisher Scientific) at a flow rate of 25 μl/min \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 12 \nusing 0.1% TFA as mobile phase. After loading, the trap column was switched in line with the \nanalytical column (PepMap Acclaim C18, 500 mm × 75 μm ID, 2 μm, 100 Å,Thermo Fisher \nScientific). Peptides were eluted using a flow rate of 230 nl/min, starting wit h the mobile phases \n98% A (0.1% formic acid in water) and 2% B (80% acetonitrile, 0.1% formic acid) and linearly \nincreasing to 35% B over the next 120min. This was followed by a steep gradient to 95% B in 1 \nmin, stayed there for 6 min and ramped down in 2 min to the starting conditions of 98% A and 2% \nB for equilibration at 30°C. The Orbitrap Exploris 480 mass spectrometer was operated in data -\ndependent mode, performing a full scan (m/z range 350-1200, resolution 60,000, normalized AGC 370 \ntarget 300%) at 3 different compensation voltages (CV -45V , -60V and -75V), followed by MS/MS \nscans of the most abundant ions for a cycle time of 0.9 seconds for each. MS/MS spectra were \nacquired using an isolation width of 1.2 m/z, normalized AGC target 200%, HCD collision energy \nof 30 %, maximum injection time mode set to custom and resolution of 30,000. Precursor ions \nselected for fragmentation (include charge state 2 -6) were excluded for 45 s. The monoisotopic \nprecursor selection (MIPS) mode was set to peptide and the exclude isotopes feature was \nenabled. \nMS Data processing \nFor peptide identification, the RAW -files were loaded into Proteome Discoverer (version \n2.5.0.400, Thermo Scientific). All MS/MS spectra were searched using MSAmanda v2.0.0.19924 380 \n(Dorfer V . et al., J. Proteome Res. 2014 Aug 1;13(8):3679- 84). The peptide mass tolerance was set \nto ±10 ppm and fragment mass tolerance to±10 ppm, the maximum number of missed cleavages \nwas set to 2, using tryptic enzymatic specificity without proline re striction. The RAW -files were \nsearched against the Arabidopsis database (32,785 sequences; 14,482,855 residues), \nsupplemented with common contaminants and sequences of tagged proteins of interest. The \nfollowing search parameters were used: Oxidation on met hionine, phosphorylation on serine, \nthreonine, and tyrosine, deamidation on asparagine and glutamine, iodoacetamide derivative on \ncysteine, beta-methylthiolation on cysteine, biotinylation on lysine, ubiquitinylation residue on \nlysine, ubiquitination on lysine, pyro-glu from q on peptide N -terminal glutamine, acetylation on \nprotein N -Terminus were set as variable modifications. The result was filte red to 1 % FDR on 390 \nprotein level using the Percolator algorithm (Käll L. et al., Nat. Methods. 2007 Nov; 4(11):923 -5) \nintegrated in Proteome Discoverer. The localization of the post -translational modification sites \nwithin the peptides was performed with the tool ptmRS, based on the tool phosphoRS (Taus T. et \nal., J. Proteome Res. 2011, 10, 5354-62). Additionally, an Amanda score cut-off of at least 150 was \napplied. Protein areas have been computed in IMP-apQuant (Doblmann J. et. al, J Proteome Res \n2019, 18(1):535 -41) by summi ng up unique and razor peptides. Resulting protein areas were \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 13 \nnormalized using iBAQ (Schwanhäusser B. et al., Nature 2011, 473(7347):337−42) and sum \nnormalization was applied for normalization between samples. Match-between-runs (MBR) was \napplied for peptides with high confident peak area that were identified by MS/M S spectra in at \nleast one run. Proteins were filtered to be identified by a minimum of 2 PSMs in at least 1 sample 400 \nand quantified proteins were filtered to contain at least 3 quantified peptide groups. Statistical \nsignificance of differentially expressed proteins was determined using limma (Smyth, G. K. - \nLinear models and empirical Bayes methods for assessing differen tial expression in microarray \nexperiments. Statistical Applications in Genetics and Molecular Biology, 2004, Volume 3, Article \n3.) (Supplementary Table 2).  \nTo identify potential interactors, log2FC was calculated comparing average PSMs in TID -ATG8A \nsamples (n = 3) or TID-ATG8H samples (n = 3) against the TID control (n = 3). Proteins with average \nPSMs > 3, log2FC > 1 and p -value > 0,05 were selected as potential interactors. The list of \npotential interactors is provided in Supplementary Table 3. \nTEM 410 \nThe TEM assay was performed following the previously established method (41, 55). Briefly, 5 -\ndays old Arabidopsis seedlings were germinated on ½ MS plates and dissected under microscopy \nbefore freezing. For high -pressure freezing, the root tips were collected and immediately frozen \nwith a high-pressure freezer (EM ICE, Leica). For freeze substitution, the root tips were substituted \nwith 2% osmium tetroxide in anhydrous acetone and maintained at −80 °C for 24 hours using an \nAFS2 temperature -controlling system (Leica). Subsequently, the samples were subjected to \nthree washes with precool ed acetone and slowly warmed up to room temperature over a 60 -h \nperiod before being embedded in EPON resin. After resin polymerization, samples were mounted \nand trimmed. For the ultrastructure studies, 100 nm thin sections were prepared using an \nultramicrotome (EM UC7, Lecia) and examined with a Hitachi H -7650 TEM (Hitachi -High 420 \nTechnologies) operated at 80 kV . \n \nData availability \nAll the source data used to generate the main and supplementary figures are deposited to Zenodo \n(10.5281/zenodo.14277422). Genomic sequencing data generated in this study are deposited at \nthe National Center for Biotechnology Information Gene Expression Omnibus (NCBI GEO, \nhttps://www.ncbi.nlm.nih.gov/geo/) under accession number GSE283481. The mass \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 14 \nspectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the \nPRIDE partner repository. \n 430 \nAcknowledgements \nWe thank Vienna Biocenter Core Facilities (VBCF), particularly Proteomics, BioOptics and Plant \nSciences. We thank the CLIP cluster (https://clip.science) for access for image analysis. We \nacknowledge funding from Austrian Academy of Sciences, Austrian Science Fund (FWF , P32355, \nP34944, SFB F79, DOC 111), Vienna Science and Technology Fund (WWTF , LS17‐047, LS21‐009), \nand European Research Council Grant (Project number: 101043370). Peng Gao is supported by \nthe Vienna International Postdoctoral Program (VIP2) and Marie Curie Fellowship (Project \nnumber: 847548).  \n \nThe authors declare no competing financial interest.  440 \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 15 \nReferences \n1. D. C. Bassham et al., Autophagy in development and stress responses of plants. \nAutophagy 2, 2-11 (2006). \n2. T. Su et al., Autophagy: An Intracellular Degradation Pathway Regulating Plant Survival \nand Stress Response. Front Plant Sci 11, 164 (2020). \n3. Y . Liu, D. C. Bassham, Autophagy: pathways for self-eating in plant cells. Annu Rev Plant \nBiol 63, 215-237 (2012). \n4. J. Zhou et al., NBR1-mediated selective autophagy targets insoluble ubiquitinated 450 \nprotein aggregates in plant stress responses. PLoS Genet 9, e1003196 (2013). \n5. M. Izumi, K. Yoshimoto, H. Batoko, Editorial: Organelle Autophagy in Plant Development. \nFront Plant Sci 11, 502 (2020). \n6. R. S. Marshall, R. D. Vierstra, Autophagy: The Master of Bulk and Selective Recycling. \nAnnu Rev Plant Biol 69, 173-208 (2018). \n7. E. A. Merkulova, A. Guiboileau, L. Naya, C. Masclaux-Daubresse, K. Yoshimoto, \nAssessment and optimization of autophagy monitoring methods in Arabidopsis roots \nindicate direct fusion of autophagosomes with vacuoles. Plant Cell Physiol 55, 715-726 \n(2014). \n8. L. Chen et al., Autophagy contributes to regulation of the hypoxia response during 460 \nsubmergence in Arabidopsis thaliana. Autophagy 11, 2233-2246 (2015). \n9. A. Ismayil, M. Yang, Y . Liu, Role of autophagy during plant-virus interactions. Semin Cell \nDev Biol 101, 36-40 (2020). \n10. P . Wang, Y . Mugume, D. C. Bassham, New advances in autophagy in plants: Regulation, \nselectivity and function. Semin Cell Dev Biol 80, 113-122 (2018). \n11. R. Le Bars, J. Marion, R. Le Borgne, B. Satiat-Jeunemaitre, M. W. Bianchi, ATG5 defines a \nphagophore domain connected to the endoplasmic reticulum during autophagosome \nformation in plants. Nat Commun 5, 4121 (2014). \n12. W. G. van Doorn, A. Papini, Ultrastructure of autophagy in plant cells: a review. \nAutophagy 9, 1922-1936 (2013). 470 \n13. M. Jin, X. Liu, D. J. Klionsky, SnapShot: Selective autophagy. Cell 152, 368-368.e362 \n(2013). \n14. A. Stolz, A. Ernst, I. Dikic, Cargo recognition and trafficking in selective autophagy. Nat \nCell Biol 16, 495-501 (2014). \n15. J. C. Farré, S. Subramani, Mechanistic insights into selective autophagy pathways: \nlessons from yeast. Nat Rev Mol Cell Biol 17, 537-552 (2016). \n16. T. Lamark, T. Johansen, Mechanisms of Selective Autophagy. Annu Rev Cell Dev Biol 37, \n143-169 (2021). \n17. T. Johansen, T. Lamark, Selective autophagy mediated by autophagic adapter proteins. \nAutophagy 7, 279-296 (2011). 480 \n18. V . Rogov, V . Dötsch, T. Johansen, V . Kirkin, Interactions between autophagy receptors \nand ubiquitin-like proteins form the molecular basis for selective autophagy. Mol Cell \n53, 167-178 (2014). \n19. S. Luo et al., Cargo Recognition and Function of Selective Autophagy Receptors in \nPlants. Int J Mol Sci 22,  (2021). \n20. X. Wen, D. J. Klionsky, An overview of macroautophagy in yeast. J Mol Biol 428, 1681-\n1699 (2016). \n21. H. Yamamoto, S. Zhang, N. Mizushima, Autophagy genes in biology and disease. Nat Rev \nGenet 24, 382-400 (2023). \n22. A. Kaufmann, V . Beier, H. G. Franquelim, T. Wollert, Molecular mechanism of autophagic 490 \nmembrane-scaffold assembly and disassembly. Cell 156, 469-481 (2014). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 16 \n23. A. Kaufmann, T. Wollert, Scaffolding the expansion of autophagosomes. Autophagy 10, \n1343-1345 (2014). \n24. T. Johansen, T. Lamark, Selective Autophagy: ATG8 Family Proteins, LIR Motifs and Cargo \nReceptors. J Mol Biol 432, 80-103 (2020). \n25. W. Liu et al., ATG8-Interacting Motif: Evolution and Function in Selective Autophagy of \nTargeting Biological Processes. Front Plant Sci 12, 783881 (2021). \n26. H. Popelka, D. J. Klionsky, Post-translationally-modified structures in the autophagy \nmachinery: an integrative perspective. Febs j 282, 3474-3488 (2015). \n27. E. Seo et al., Comparative analyses of ubiquitin-like ATG8 and cysteine protease ATG4 500 \nautophagy genes in the plant lineage and cross-kingdom processing of ATG8 by ATG4. \nAutophagy 12, 2054-2068 (2016). \n28. R. Kellner, J. C. De la Concepcion, A. Maqbool, S. Kamoun, Y . F . Dagdas, ATG8 \nExpansion: A Driver of Selective Autophagy Diversification? Trends Plant Sci 22, 204-214 \n(2017). \n29. D. Winter et al., An \"Electronic Fluorescent Pictograph\" browser for exploring and \nanalyzing large-scale biological data sets. PLoS One 2, e718 (2007). \n30. S. Sláviková et al., The autophagy-associated Atg8 gene family operates both under \nfavourable growth conditions and under starvation stresses in Arabidopsis plants. J Exp \nBot 56, 2839-2849 (2005). 510 \n31. C. C. Thoreen et al., Correction: An ATP-competitive mammalian target of rapamycin \ninhibitor reveals rapamycin-resistant functions of mTORC1. J Biol Chem 295, 2886 \n(2020). \n32. M. Stephani et al., A cross-kingdom conserved ER-phagy receptor maintains \nendoplasmic reticulum homeostasis during stress. Elife 9,  (2020). \n33. R. Grützner et al., High-efficiency genome editing in plants mediated by a Cas9 gene \ncontaining multiple introns. Plant Commun 2, 100135 (2021). \n34. T. Chung, A. R. Phillips, R. D. Vierstra, ATG8 lipidation and ATG8-mediated autophagy in \nArabidopsis require ATG12 expressed from the differentially controlled ATG12A AND \nATG12B loci. Plant J 62, 483-493 (2010). 520 \n35. Y . Wang, M. T. Nishimura, T. Zhao, D. Tang, ATG2, an autophagy-related protein, \nnegatively affects powdery mildew resistance and mildew-induced cell death in \nArabidopsis. Plant J 68, 74-87 (2011). \n36. A. R. Thompson, J. H. Doelling, A. Suttangkakul, R. D. Vierstra, Autophagic nutrient \nrecycling in Arabidopsis directed by the ATG8 and ATG12 conjugation pathways. Plant \nPhysiol 138, 2097-2110 (2005). \n37. S. Svenning, T. Lamark, K. Krause, T. Johansen, Plant NBR1 is a selective autophagy \nsubstrate and a functional hybrid of the mammalian autophagic adapters NBR1 and \np62/SQSTM1. Autophagy 7, 993-1010 (2011). \n38. M. Krebs et al., Arabidopsis V-ATPase activity at the tonoplast is required for efficient 530 \nnutrient storage but not for sodium accumulation. Proc Natl Acad Sci U S A 107, 3251-\n3256 (2010). \n39. J. H. Doelling, J. M. Walker, E. M. Friedman, A. R. Thompson, R. D. Vierstra, The APG8/12-\nactivating enzyme APG7 is required for proper nutrient recycling and senescence in \nArabidopsis thaliana. J Biol Chem 277, 33105-33114 (2002). \n40. Y . Xiong, A. L. Contento, D. C. Bassham, AtATG18a is required for the formation of \nautophagosomes during nutrient stress and senescence in Arabidopsis thaliana. Plant J \n42, 535-546 (2005). \n41. J. Ma et al., Friendly mediates membrane depolarization-induced mitophagy in \nArabidopsis. Curr Biol 31, 1931-1944.e1934 (2021). 540 \n42. N. D. Georgakopoulos, G. Wells, M. Campanella, The pharmacological regulation of \ncellular mitophagy. Nat Chem Biol 13, 136-146 (2017). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 17 \n43. K. Yoshimoto et al., Organ-specific quality control of plant peroxisomes is mediated by \nautophagy. J Cell Sci 127, 1161-1168 (2014). \n44. C. Viotti et al., The endoplasmic reticulum is the main membrane source for biogenesis \nof the lytic vacuole in Arabidopsis. Plant Cell 25, 3434-3449 (2013). \n45. Y . Kamada et al., Tor directly controls the Atg1 kinase complex to regulate autophagy. \nMol Cell Biol 30, 1049-1058 (2010). \n46. N. Mizushima, The role of the Atg1/ULK1 complex in autophagy regulation. Curr Opin \nCell Biol 22, 132-139 (2010). 550 \n47. J. Zhao et al., Plant autophagosomes mature into amphisomes prior to their delivery to \nthe central vacuole. J Cell Biol 221,  (2022). \n48. E. K. Zess et al., N-terminal β-strand underpins biochemical specialization of an ATG8 \nisoform. PLoS Biol 17, e3000373 (2019). \n49. H. L. Xing et al., A CRISPR/Cas9 toolkit for multiplex genome editing in plants. BMC Plant \nBiol 14, 327 (2014). \n50. A. Lampropoulos et al., GreenGate---a novel, versatile, and efficient cloning system for \nplant transgenesis. PLoS One 8, e83043 (2013). \n51. B. Langmead, S. L. Salzberg, Fast gapped-read alignment with Bowtie 2. Nat Methods 9, \n357-359 (2012). 560 \n52. P . Danecek et al., Twelve years of SAMtools and BCFtools. Gigascience 10,  (2021). \n53. S. Stael, L. P . Miller, D. Fernández-Fernández Á, F . Van Breusegem, Detection of \nDamage-Activated Metacaspase Activity by Western Blot in Plants. Methods Mol Biol \n2447, 127-137 (2022). \n54. A. Mair, S. L. Xu, T. C. Branon, A. Y . Ting, D. C. Bergmann, Proximity labeling of protein \ncomplexes and cell-type-specific organellar proteomes in Arabidopsis enabled by \nTurboID. Elife 8,  (2019). \n55. B. H. Kang, Electron microscopy and high-pressure freezing of Arabidopsis. Methods \nCell Biol 96, 259-283 (2010). \n 570 \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 18 \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 19 \nFigure 1. Arabidopsis thaliana ATG8 isoforms exhibit tissue-specific expression patterns and \nform distinct autophagosomes within root cells \n(A) Representative GUS staining images showing the spatial -temporal expression patterns of 9 \nArabidopsis ATG8 isoforms.10-days old Arabidopsis seedlings expressing pATG8X::GFP -GUS (X \nrepresents the 9 ATG8 isoforms, from A to I) were stained with GUS staini ng buffer. (B) \nRepresentative confocal microscopic images showing the colocalization of mCherry-ATG8E with \ndifferent GFP -ATG8 isoforms in Arabidopsis root epidermal cells. 5 -days old Arabidopsis 580 \nseedlings co -expressing mCherry -ATG8E with GFP -ATG8E, GFP -ATG8A, GFP -ATG8D, or GFP -\nATG8I were incubated in ½ MS liquid media containing either DMSO (as mock condition) for 2 h, \n5 μM Torin1 for 2 h, or 10 μg/mL Tunicamycin for 4 h. Representative images of 10 replicates were \nshown here. Scale bars, 5 μm. Inset scale bars, 3 μm. (C) Quantification of mCherry -ATG8E \ncolocalization ratio of the Arabidopsis root epidermal cells imaged in (B). The mCherry -ATG8E \ncolocalization ratio was calculated as the ratio of th e number of mCherry -ATG8E puncta that \ncolocalized with GFP-ATG8 isoforms to the total number of mCherry-ATG8E puncta. Bars indicate \nthe mean ± SD of 10 replicates. Brown-Forsy and Welch ANOVA tests with Dunnett’s T3 multiple \ncomparisons tests were used for  statistically comparing the colocalization difference between \neach treatment group. 590 \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 20 \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 21 \nFigure 2. atg8 nonuple mutant (Δatg8) is deficient in autophagy \n(A) Schematic representation of gene models illustrating CRISPR-induced mutations in the ATG8 \ngene. Blue arrows denote insertion sites, while red arrows indicate deletion sites. The  forward \nand reverse arrows within the gene model indicate the directionality of gene  transcription, \nreflecting the strandedness of the gene. (B, C) Carbon (B) and nitrogen (C) starvation phenotypic \nassays comparing Col-0, atg5 and Δatg8 (n = 3) in carbon-rich (+C) or carbon-deficient (-C) ½ MS 600 \nliquid medium and nitrogen -rich (+N) or  nitrogen-deficient ( -N) ½ MS  liquid medium. (D, E)  \nWestern blots comparing endogenous NBR1 levels in Col-0, atg5 and Δatg8 upon carbon (D) and \nnitrogen (E) starvation, in combination with Concanamycin A (1 μM). Relative quantification of \nprotein bands is reported below the blots. \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 22 \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 23 \nFigure 3. The Δatg8 mutant is not able to perform mitophagy and pexophagy.  \n(A) Western blot s comparing endogenous IDH and VDAC levels in Col -0, atg5 and Δatg8 upon 610 \nDNP treatment. (B) Western blots comparing endogenous catalase levels in Col-0, atg5 and Δatg8 \nupon N starvation treatment. Relative quantification of protein bands is reported below the blot. \n(C) Electron micrographs of Col-0 and Δatg8 root cells treated with DNP or DMSO. Scale bars, \n500 nm.  \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 24 \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 25 \nFigure 4. Complementation of Δatg8 with ATG8A or ATG8H reveals functional specialization 620 \nof ATG8 isoforms \n (A, B) Carbon (A) and nitrogen (B) starvation phenotypic assays comparing Col -0, Δatg8 and \ncomplementation lines Δatg8 /+ATG8A, Δatg8/+ATG8H (n = 3) in carbon-deficient (-C) ½ MS liquid \nmedium and nitrogen -deficient ( -N) ½ MS liquid medium. (C, D) Western blots comparing \nendogenous NBR1 levels in Col-0, Δatg8, Δatg8 /+ATG8A and Δatg8 /+ATG8H upon carbon (C) and \nnitrogen (D) starvation, in combination with Concanamycin A (1 μM). Relative quantification of \nprotein bands is reported below the blots. (E) Schema tic representation of TurboID proximity \nlabeling analysis. (F) Venn diagram reporting common and unique interactors of ATG8A and \nATG8H under nitrogen starvation.  \n  630 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 26 \n \nSupplementary Figure 1. Replicates of western blots in Figure 2 (A, B) and Figure 4 (C, D).  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 27 \n \nSupplementary Figure 2. Arabidopsis ATG8 complementation lines have normal \nautophagosome structure and autophagic flux under control and nutrient -deficient \nconditions. (A) Representative confocal microscopic images showing the autophagosomes and \nthe autophagic bodies inside the vacuole in root epidermal cells of the complementation lines \nΔatg8 /+GFP-ATG8A and Δatg8 /+GFP-ATG8A. 5-days old Arabidopsis seedlings were incubated \nin ½ MS liquid media or nitrogen -deficient ( -N) liquid media for 3 h, or ½ MS liquid media or \nnitrogen-deficient (-N) liquid media containing 2 μM concanamycin A for 2.5 h before imaging. 640 \nRepresentative images of 3 replicates were shown here. Scale bars, 10 μm. Inset scale bars, 5 \nμm. (B) Western blot comparing autophagic flux of complementation lines Δatg8 /+GFP-ATG8A \nand Δatg8 /+GFP -ATG8A upon C starvation treatment, in combination with concanamycin A \n(1μM).  \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint \n\n   \n \n 28 \n \nSupplementary Figure 3. Replicates of western blots in Figure 3.  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted December 10, 2024. ; https://doi.org/10.1101/2024.12.10.627464doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}