Integrated proteomics reveals autophagy landscape and an autophagy receptor controlling PKA-RI complex homeostasis in neurons

Autophagy is a conserved, catabolic process essential for maintaining cellular homeostasis. Malfunctional autophagy contributes to neurodevelopmental and neurodegenerative diseases. However, the exact role and targets of autophagy in human neurons remain elusive. Here we report a systematic investigation of neuronal autophagy targets through integrated proteomics. Deep proteomic profiling of multiple autophagy-deficient lines of human induced neurons, mouse brains, and brain LC3-interactome reveals roles of neuronal autophagy in targeting proteins of multiple cellular organelles/pathways, including endoplasmic reticulum (ER), mitochondria, endosome, Golgi apparatus, synaptic vesicle (SV) for degradation. By combining phosphoproteomics and functional analysis in human and mouse neurons, we uncovered a function of neuronal autophagy in controlling cAMP-PKA and c-FOS-mediated neuronal activity through selective degradation of the protein kinase A - cAMP-binding regulatory (R)-subunit I (PKA-RI) complex. Lack of AKAP11 causes accumulation of the PKA-RI complex in the soma and neurites, demonstrating a constant clearance of PKA-RI complex through AKAP11-mediated degradation in neurons. Our study thus reveals the landscape of autophagy degradation in human neurons and identifies a physiological function of autophagy in controlling homeostasis of PKA-RI complex and specific PKA activity in neurons.


Integrated proteomics reveals autophagy landscape and an autophagy receptor controlling PKA-RI complex homeostasis in neurons
Xiaoting Zhou 1,2,3 , You-Kyung Lee 1,2 , Xianting Li 1,2 , Henry Kim 1,2 , Carlos Sanchez-Priego 2,4 , Xian Han 5,6,7 , Haiyan Tan 5,6,8 , Suiping Zhou 8 , Yingxue Fu 5,6,8 , Kerry Purtell 1,2 , Qian Wang 1,2 , Gay R. Holstein 1,2 , Beisha Tang 9,10 , Junmin Peng 5,6,8 , Nan Yang 2,4 & Zhenyu Yue 1,2,11 Autophagy is a conserved, catabolic process essential for maintaining cellular homeostasis.Malfunctional autophagy contributes to neurodevelopmental and neurodegenerative diseases.However, the exact role and targets of autophagy in human neurons remain elusive.Here we report a systematic investigation of neuronal autophagy targets through integrated proteomics.Deep proteomic profiling of multiple autophagy-deficient lines of human induced neurons, mouse brains, and brain LC3-interactome reveals roles of neuronal autophagy in targeting proteins of multiple cellular organelles/ pathways, including endoplasmic reticulum (ER), mitochondria, endosome, Golgi apparatus, synaptic vesicle (SV) for degradation.By combining phosphoproteomics and functional analysis in human and mouse neurons, we uncovered a function of neuronal autophagy in controlling cAMP-PKA and c-FOS-mediated neuronal activity through selective degradation of the protein kinase A -cAMP-binding regulatory (R)-subunit I (PKA-RI) complex.Lack of AKAP11 causes accumulation of the PKA-RI complex in the soma and neurites, demonstrating a constant clearance of PKA-RI complex through AKAP11mediated degradation in neurons.Our study thus reveals the landscape of autophagy degradation in human neurons and identifies a physiological function of autophagy in controlling homeostasis of PKA-RI complex and specific PKA activity in neurons.
Macroautophagy (herein referred to as autophagy) is a lysosomedependent degradation pathway.Mammalian cells require autophagy to maintain cellular homeostasis and respond to various stresses and injuries through digestion and recycling.During autophagy, cytoplasmic components are engulfed in a double-membraned vesicle known as the autophagosome, which is shuttled to the lysosome for degradation 1 .While autophagy degrades intracellular components in the non-selective manner in response to nutrient starvation, autophagy can also selectively degrade cytoplasmic proteins and organelles under basal conditions.More specifically, Atg8family proteins such as LC3A/B, which are conjugated on the growing autophagosome membrane, selectively recognize autophagy cargos 2 .The identification of specific interactions between Atg8family proteins and autophagy cargos has enabled the utilization of affinity purification and proximity labeling to profile selective autophagy substrates in cancer cell lines [3][4][5] .
In contrast to most mammalian cell types, neurons are postmitotic, long-living cells of the central nervous system (CNS).To maintain longevity and physiological function, neuronal autophagy removes and recycles misfolded proteins, protein aggregates, and injured organelles that would otherwise be detrimental to the cell.In fact, when essential autophagy-related genes (ATGs) Atg5 and Atg7 are conditionally deleted in mouse brains, cerebral and cerebellar neurons degenerate, causing motor dysfunctions in mice 6,7 .Using mice with disrupted Atg7 expression in specific neuron types, we previously demonstrated a compartmental function of autophagy in neurons by maintaining axonal homeostasis and preventing axonopathy 8,9 .Indeed, autophagy biogenesis can occur in the axon and participates in axonal transport to remove dysfunctional organelles and protein aggregates [10][11][12] .Furthermore, autophagy was shown to regulate synaptic activity [13][14][15] .The significance of autophagy in the CNS is further underscored in humans, where a recent study has identified biallelic, recessive variants in human ATG7 that results in impaired autophagic flux and complex neurodevelopmental disorders 16 .Furthermore, dysfunctional autophagy may contribute to major neurodegenerative diseases [17][18][19] .However, the physiology of autophagy in human neurons and detailed mechanisms for how its disruption leads to various neurological disorders remains poorly understood.
To understand the autophagy targets, multiple studies performed systematic analysis to dissect autophagy cargo in non-neuronal cells 4,[20][21][22][23][24][25][26] .However, the investigation of neuronal autophagy cargo, particularly in humans, remains limited.Profiling the differentially accumulated proteins from Atg5-deficient primary mouse neurons has revealed the function of autophagy in regulating axonal endoplasmic reticulum (ER) and luminal Ca 2+ stores 27 .Examination of isolated autophagic vesicles (AVs) from mouse brains using differential centrifugation showed the presence of mitochondrial fragments and synaptic proteins in neuronal autophagosomes 10 .
Herein we present a systematic investigation of human neuronal autophagy cargo by using human pluripotent stem cell (PSC)-induced neurons (iNeurons) through integrated proteomics and functional analysis.To enrich autophagy cargos, we suspended autophagy in iNeurons using CRISPR-interference technology to knockdown ATG7 or ATG14, both of which are essential for autophagy functions.By combining the study of the neurons from multiple lines of mouse autophagy deficient brains, our data reveals the landscape of autophagy degradation in neurons and highlights selective sets of proteins from multiple organelles as potential cargos of neuronal autophagy.Moreover, we identified a function of autophagy in controlling PKA-RI complex homeostasis through autophagy receptor AKAP11 and regulating neuronal activity.Our study thus provides a global view of autophagy degradation in human neurons and an insight into the mechanisms of neurological disorders linked to autophagy deficiency.

Proteomic profiling of enriched cellular pathways in autophagydeficient human induced neurons
Previous studies demonstrated that autophagy is constitutively active in neurons; basal autophagy constantly degrades cellular cargo to maintain neuronal homeostasis and protect neurons 8,28 .To systematically investigate neuronal autophagy targets under basal conditions, we suspended autophagic flux in neurons, performed integrated proteomic profiling of accumulated proteins, and identified autophagy cargo, receptors, and targeted pathways (Fig. 1a).To ensure the rigor of the analysis and isolate changes specific to autophagy inhibition, we elected to disrupt two essential autophagy genes, ATG7 and ATG14, separately in human iNeurons as well as Atg7 and Atg14 in mouse brains in a neuron-specific manner.ATG7 is an E1-like enzyme, which is required for the lipidation of LC3 and Atg5-Atg12-Atg16 conjugation 29 .
ATG14 is an essential subunit of the Beclin1-PI3K complex, which facilitates the tethering and formation of phagophore 30,31 .Blocking either ATG7 or ATG14 leads to disruption of autophagosome formation and the consequent autophagy cargo accumulation in the cells.We posit that examination of the results from the inactivation of the two functionally distinct autophagy genes facilitates the identification of the proteins related to autophagy degradation.We also performed transcriptomic analysis by RNA sequencing (RNA-seq) to discern the change of protein levels not caused by transcriptional alteration.Furthermore, we investigated the interactome of autophagy protein LC3 in the brains by including the green fluorescence protein-tagged LC3 (GFP-LC3) transgenic mice 32 in our study (Fig. 1a).
Autophagy in human neurons is poorly characterized.We sought to investigate autophagy in human glutamatergic iNeurons, representing the most common neuron type in the CNS.We used our previously established approach to generate iNeurons by transiently expressing the transcription factor neurogenin 2 (Ngn2) 33 .Fusion of a catalytically inactive dCas9 to the KRAB repressor domain enables robust knockdown (KD) of endogenous genes in iNeurons 34 .To repress ATG7 or ATG14 expression, we transduced human PSCs with a lentiviral construct expressing dCas9-KRAB and sgRNA targeting ATG7 or ATG14.Reduction of ATG7 or ATG14 protein levels was robust in PSCs and iNeurons (Fig. 1b, Supplementary Fig. 1a, b).At 6 weeks postdifferentiation (PD), the iNeurons had obtained stable expression of dendrite and synapses markers (Fig. 1c), and robust electrophysiological properties and formed functional synapses 33 .
We next profiled the proteome of ATG7 and ATG14 KD iNeurons (6-week PD) by the multiplexed tandem mass tag (TMT)-based, twodimensional liquid chromatography (LC/LC) and tandem mass spectrometry (MS/MS) analysis 35,36 .Isobaric labeling methods like TMT have become an important tool for in-depth proteomics profiling and are increasingly used for the multiplexed analysis of samples, offering high throughput and consistency 37,38 .A known limitation of isobaric tags is the co-isolation of target peptide ions with co-eluting peptides during MS2-based quantification, which can introduce considerable noise and potentially skew quantitative ratios.Fortunately, this ratio suppression also reduces experimental variations, minimally affecting the statistical analysis of significant proteome alterations after z-transformation 39,40 .The ratio suppression can be alleviated by approaches such as intensive peptide fractionation, employing a narrower MS2 isolation window, and computational correction 37 .Moreover, the MS3 approach can virtually eliminate these inaccuracies, but it necessitates additional duty cycles and specific MS configurations, along with low-resolution MS2 data for identification, which negatively impact peptide/protein identification rates 38,41 .To balance the pros and cons of the TMT method, we have optimized the method by employing extensive high-resolution LC/LC-MS/MS, which significantly improves proteome coverage and reduces ion suppression effects during quantification [42][43][44][45][46] .We identified approximate 8000 proteins across all the human iNeuron samples (Supplementary Data 1b), demonstrating the high coverage of the proteome.Principalcomponent analysis (PCA) revealed reproducible replicate data, with the variance driven mainly by the difference in ATG gene expression.Similarly, clustering samples by protein expression similarity showed two main groups representing the mutant and control iNeurons (Supplementary Fig. 1c, d).We identified 1460 and 1172 differentially expressed proteins (DEPs) in ATG7 KD and ATG14 KD iNeurons, respectively (p value < 0.05, |log 2 FC|>2-fold standard deviation (SD), SD = 0.18, Supplementary Data 1b).Our quantitative analyses reveal 1113 and 637 upregulated DEPs of proteins in ATG7 KD and ATG14 KD iNeurons, respectively, as shown in volcano plot analysis (Supplementary Data 1b, Fig. 1d, e).Gene ontology (GO) enrichment analysis of the upregulated DEPs highlighted cellular pathways associated with endoplasmic reticulum (ER), Golgi apparatus, synaptic vesicle (SV), mitochondria, endosome, and lysosome in both ATG7 KD and ATG14 KD iNeurons (Fig. 1d, e).By plotting the fold increase of the upregulated DEPs, we obtained the mean values and inferred the abundance of the proteins for different organelles or pathways (Fig. 1f, g).Golgi, SV and ER DEPs noticeably displayed the greatest abundance in both ATG7 and ATG14 KD iNeurons relative to DEPs of mitochondria and endosome.Indeed, nearly one-third of the total upregulated DEPs are ER-related proteins in both ATG7 and ATG14 KD iNeurons (Fig. 1h), while 9.2% and 15.3% of upregulated DEPs are SV proteins in ATG7 KD and ATG14 KD iNeurons, respectively (Fig. 1i).Among the upregulated DEPs in ATG7 KD and ATG14 KD iNeurons, at least thirteen were known autophagy receptors as previously reported [47][48][49] , including five known ER-phagy receptors (Fig. 1j).The identification of the previously described autophagy receptors validated the feasibility and robustness of our approach.To ask whether the increased levels of the proteins are due to transcriptional upregulation of their encoding genes, we performed RNA-seq analysis of ATG7 KD iNeurons.Strikingly, ATG7 was found as the only differentially expressed gene (DEG) (FDR < 0.05) (Supplementary Fig. 1h).Furthermore, we found that 339 upregulated DEPs are shared between ATG7 KD and ATG14 KD iNeurons (Fig. 1k), suggesting that they are unlikely ATG7 or ATG14 -specific but related to autophagy degradation.Consistently, these shared proteins were enriched in ER, Golgi apparatus, mitochondria, and SV-related pathways (Fig. 1l, m).Noticeably ER proteins are top ranked as shown in both GO analysis (FDR) and the percentage of ER proteins in the upregulated DEPs (Fig. 1l, m).Interestingly, we found the enrichment of "cAMP-dependent protein kinase complex" among the shared DEPs (Fig. 1l).
Furthermore, we identified 347 and 535 down-regulated proteins in ATG7 KD and ATG14 KD iNeurons, respectively (p < 0.05, Log 2 FC <−2SD).106 down-regulated DEPs are shared by the two mutant cells (Supplementary Fig. 1f).We found significant enrichment of the down-regulated pathways related to protein binding, synapse, and neuron development (Supplementary Fig. 1g).

Impaired clearance of selective ER and SV proteins in autophagy-deficient human iNeurons
Our quantitative proteomics suggest that many ER proteins are potential autophagy cargo in human neurons, as indicated by highranking GO terms and abundance analysis in human iNeurons (Fig. 1d, e and Supplementary Data 1b).Multiple ER membrane proteins, including RTN3, ATL1, SEC62, and TEX264, function as receptors responsible for the selective degradation of the ER by autophagy (ERphagy) 4, [50][51][52] .Immunoblot analysis confirmed that these ER-phagy receptors, along with ER resident proteins REEP5 and Calnexin (CNX), are increased or have a trend of increase in ATG7 KD and ATG14 KD iNeurons (Fig. 2a, c and supplementary Fig. 1e).A previous proteomic study classified ER proteins into four groups, such as ER membrane, ER lumen, ER-Golgi intermediate compartment (Ergic/cisGolgi), and ER curvature 53 .An overall increase of all categories was identified in both mutant iNeurons (Fig. 2e, f).Immunofluorescence (IF) analysis showed a remarkable accumulation of CNX-labeled inclusions (arrows) particularly in the neurites of ATG7 KD iNeurons (Fig. 2g, h), suggesting an expansion of ER network in the processes.Using an ER-phagy reporter, which expresses an N-terminal ER signal sequence followed by tandem monomeric RFP and GFP sequences and the ER retention sequence KDEL 4 , we found that the signal of the RFP was significantly reduced in ATG7 KD iNeurons upon nutrient-withdrawal, implying an impairment of ER-phagy (Supplementary Fig. 1i-k).
Collectively, our findings indicate that specific sets of ER and SV proteins require autophagy to maintain their homeostasis or degradation in human neurons under basal conditions.ER-phagy is impaired in human ATG7 KD iNeurons.

Proteomic profiling of enriched cellular pathways in autophagydeficient neurons from mutant mouse brains
To validate our findings from human iNeurons, we sought to profile the targets of neuronal autophagy in vivo by using mouse brains.We selectively deleted Atg7 or Atg14 in CNS neurons using the combination of synapsin1 promoter-driven Cre recombinase mice (SynCre) and the Atg7 f/f or Atg14 f/f mice to generate conditional knock-out (cKO) mice.Immunoblot analysis of whole brain lysates from Atg7 f/f -SynCre (Atg7 cKO) and Atg14 f/f -SynCre (Atg14 cKO) mice confirmed the reduction of ATG7 and ATG14 protein levels, respectively, and an increase of autophagy substrate p62 and ubiquitinated protein levels (Supplementary Fig. 3a-d).Both autophagy-deficient mice exhibited a progressive reduction in body weight (Fig. 3a, b) and viability (Fig. 3c, d) over time, while Atg14 cKO mice are more vulnerable than Atg7 cKO mice.Atg7 cKO and Atg14 cKO brains (6-8-week-old) were then subjected to the TMT-LC/LC-MS/MS analysis.We profiled 9638 and 6252 proteins from Atg7 cKO and Atg14 cKO mice, respectively (Supplementary Data 1c and Supplementary Data 1d).Most of the DEPs showed an increase in abundance (as opposed to a decrease) in the mutant brains (Fig. 3e, f, Supplementary Fig. 3e, f), revealing a profound impact of autophagy deficiency on protein degradation.Indeed, 459 and 343 proteins were increased in Atg7 cKO and Atg14 cKO mouse brains compared to wild-type, respectively (p < 0.05, | Log 2 FC | >2-SD, SD = 0.1 and 0.16 for Atg7 cKO and Atg14 cKO, respectively) (Fig. 3e, f, Supplementary Data 1c and Supplementary Data 1d).Like the findings in human iNeurons (Fig. 1), the fraction of proteins whose abundance increased in the mutant mouse brains are enriched for GO terms, such as ER, mitochondria, Golgi apparatus, endosome, SV, autophagosomes, and cAMP-dependent kinase (PKA) (Fig. 3e, f, Supplementary Fig. 3g, h).Indeed, the upregulated DEPs of ER proteins account for 37% and 31% of the total upregulated DEPs from Atg7 cKO and Atg14 cKO brains, respectively (Fig. 3i).We found that 102 upregulated DEPs are shared between Atg7 cKO and Atg14 cKO brains (Fig. 3g), while ER components are the most significant in GO term analysis (Fig. 3h).We detected increased abundance of SV proteins as well as cPKA pathways in both Atg7 cKO and Atg14 cKO mouse brains (Fig. 3e, f, i, j).Known ER-phagy receptors, such as TEX264, SEC62, and FAM134B, were found among upregulated DEPs in Atg7 cKO or Atg14 cKO mouse brains (Fig. 3k).Furthermore, transcriptomic analysis using RNA-seq of Atg7 cKO brains did not reveal an overlap between the upregulated DEGs and DEPs in Atg7 cKO brains (Supplementary Fig. 3e, i).For example, upregulated DEGs are activated glial marker genes, such as Gfap, C1qc, Clec7a, and Tyrobp (Supplementary Fig. 3i).Thus, the increase of many proteins identified in Atg7 cKO brains is unlikely due to the increase in transcriptional upregulation.
Consistent with human iNeuron results (Figs. 1 and 2), immunoblot analysis demonstrated an increase or a strong tendency of increase in the abundance of ER proteins and ER-phagy receptors, such as RTN3, ATL1, SEC62, CNX and REEP5, in Atg7 cKO and Atg14 cKO brains; and similar observations were found for SV proteins, including SV2A, SV2C, SYNGR1, SYNGR3, and SYP1 (Fig. 3l, m, Supplementary Fig. 3j, k) but to a lesser degree.To determine if the ER protein accumulation is associated with ER structure alteration in mutant neurons, we performed electron microscopy (EM) analysis of Atg7 cKO mouse brains and found a remarkable expansion or stacking of tubular ER in the axons, corroborating with the increased levels of a specific set of ER proteins, such as tubular ER proteins RTN3 and REEP5, in mutant neurons (Fig. 3n).
Taken together, the results from mouse brains validated the observations in human iNeurons that basal autophagy maintains the homeostasis of proteins of multiple pathways and/or organelles in neurons, including ER, Golgi, mitochondria, SV, endosomes, lysosomes, and cAMP-PKA kinase complex.

Investigation of LC3-interactome in autophagy-deficient neurons
LC3 is the best-characterized mammalian Atg8 homolog and it contains the LC3-interation region (LIR) that mediates cargo recruitment for degradation 54 .The binding of LC3 to the cargo does not necessarily depend on the lipidation of LC3 47 .GFP-LC3 transgenic mice were used to investigate LC3-binding proteins in mouse brains 32 .Under basal conditions, however, GFP-LC3 is rarely detected in puncta or association with autophagosomes in the CNS neurons 32,55 , possibly due to rapid cleavage of GFP-LC3 or turnover of autophagosomes 56 .To facilitate the identification of LC3-interacting cargos using anti-GFP affinity isolation in neurons, we inhibited neuronal autophagy in GFP-LC3 transgenic mice by crossing GFP-LC3 mice with the Atg7 f/f -Syn-Cre mice (GFP-LC3; Atg7 cKO) and then performed GFP-pull-down from GFP-LC3; Atg7 cKO brains (Fig. 4a).Immunoblot and silver staining confirmed the enrichment of proteins with GFP affinity purification, particularly in the GFP-LC3; Atg7 cKO mice (Supplementary Fig. 4a).Mass spectrometry analysis was performed to identify GFP-LC3 binding proteins remaining on the beads.The hierarchical clustering analysis confirmed similarity across biological replicates (Supplementary Fig. 4b).Next, we filtered out non-specific binding proteins by comparing GFP-LC3 and non-transgenic mice and focused on interactions We then performed pairwise comparisons between proteins with increased GFP-LC3 interaction in Atg7 cKO brains and upregulated DEPs in ATG7 KD iNeurons, ATG14 KD iNeurons, Atg7 cKO mouse brain or Atg14 cKO mouse brains.The results revealed a considerable overlap of the increased protein levels (Fig. 4c, d, e, f).GO enrichment analyses revealed that the molecular functions and cellular component of the overlapping proteins primarily comprised ER, Golgi trafficking, ER stress response, SV cycle, response to unfolded proteins, the autophagosome, and cAMP-dependent kinase (PKA) (Fig. 4c, d, e, f).The reproducible enrichment of the PKA pathways as shown in the above multiple datasets is particularly intriguing, as we previously reported that autophagy selectively degrades PKA-RI subunits and regulates PKA activity through the AKAP11 receptor in tumor cells 57 .Unsurprisingly, autophagy receptors (p62/SQSTM1, NBR1), ER-phagy receptors (TEX264, RTN3, SEC62, CCPG1), and mitophagy receptors (FUNDC1, NIPSNAP1/2) were detected in GFP-LC3 interactome validating our detection system (Supplementary Fig. 4d).By integrating GO enrichment pathways results from the above five datasets, we observed that the ER pathway is the most significant across most of the datasets, supporting the notion that ER is a prominent target of autophagy in neurons (Fig. 4g).The ER membrane is the most enriched among all categories of ER related proteins, which are shared by the five datasets (Fig. 4h, j).Furthermore, we identified 45 proteins, which are shared by the four autophagy-deficient neuron datasets, and 15 shared by all five proteomic datasets (with LC3-interactome) (Fig. 4i).These shared proteins include known autophagy cargo, such as SQSTM1, SEC62, NBR1, TEX264, PRKAR1A, and PRKAR1B.By performing protein-protein interaction (PPI) network analysis with selective upregulated DEPs shared by ATG7 KD and ATG14 KD iNeurons (Fig. 1k), we found specific and extensive interconnectivity among functional modules representing autophagy, ER, Golgi apparatus, SV, PKA pathway and mitochondrion (Fig. 4k, Supplementary Fig. 4e).Many components of the above modules were validated for their dependence on autophagy for clearance based on this study, including AKAP11, PRKAR1A, REEP5, ATL1, RTN3, SV2, SYNGR3, SYP (highlighted in red) (Fig. 4k).
We next validated the interaction of LC3 with ER, SV, and PKA pathway proteins using the GFP-LC3 transgenic brains and iNeurons stably expressing GFP-LC3 (Supplementary Fig. 4c).Co-IP with GFP antibody confirmed interactions of LC3 with SV proteins (SV2A and SYNGR3) and AKAP11 in the mouse brain (Fig. 4l).Moreover, we treated GFP-LC3 expressing iNeurons with chloroquine to block autophagy flux and performed IF analyses.Chloroquine apparently enhanced the number of GFP-LC3 puncta (indicating autophagosomes), which partially co-localize with the endogenous SV2A, and SYNGR3 (Fig. 4m).These observations suggest that many LC3-binding proteins identified above are recruited to autophagosomes via LC3 for autophagy degradation in iNeurons.

Autophagy degrades AKAP11 and RI proteins and regulates PKA and neuronal activity in human iNeurons
The above proteomics profiling and LC3-interactome analysis suggest that autophagy may selectively degrade the components of the cAMP-PKA pathway in human and mouse neurons.The PKA holoenzyme is a tetramer consisting of two regulatory subunits (R-subunits, RI and RII) and two catalytic subunits (PKAc).Each R subunit binds and inhibits a catalytic subunit at the resting state 58 .cAMP binding to the R-subunits releases and activates PKAc 59,60 .
Given the important functions of cAMP-PKA in neurons 61,62 , we next explored the role of autophagy in the regulation of PKAdependent neuronal activity.Immunoblotting revealed that the protein levels of AKAP11, RI, and PKA Catalytic α subunit (PKA-Cα), but not RII, were significantly increased in the two types of autophagydeficient iNeurons (Fig. 5a).Blockage of autophagic flux by chloroquine leads to the accumulation of RI proteins colocalized with autophagosome markers GFP-LC3 and p62 (Fig. 5b-d).To investigate whether autophagy regulates PKA-dependent neuronal activity, we analyzed forskolin-dependent phosphorylation of CREB (p-CREB) responses and expression of the immediate early response gene c-FOS in control and ATG7 KD iNeurons.Immunoblot analysis revealed a reduction of forskolin-induced p-CREB in ATG7 KD iNeurons (Supplementary Fig. 5a).Intriguingly, we found attenuated forskolindependent c-FOS induction (the neuronal activity early response gene), suggesting a decrease of neuronal activity in ATG7 KD iNeurons (Fig. 5e and Supplementary Fig. 5b).IF analyses showed the reduced staining intensity of p-CREB and c-FOS in the nuclei of ATG7 KD iNeurons (Fig. 5f, g).Studies of the primary cortical mouse neurons derived from Atg7 cKO mice showed the similar results as in human iNeurons (Supplementary Fig. 5c).
PKA phosphorylates diverse substrates to regulate critical functions in neurons 63 .We posited that autophagy deficiency could lead to reduced phosphorylation of many PKA substrates (Fig. 5h) due to the aberrant retention of the RI that sequesters the catalytic subunit.We then performed a phospho-proteomics analysis of ATG7 KD and ATG14 KD iNeurons (Supplementary Fig. 5d, e).We identified 458 downregulated phosphorylated protein sites in ATG7 KD and 217 in ATG14 KD iNeurons, respectively.There were 86 commonly phosphorylated sites between ATG7 and ATG14 KD iNeurons, of which many proteins containing the phosphorylated sites were predicted as PKA substrates (Fig. 5i, Supplementary Data 1f).For example, phosphorylation of PKA substrates, including ADD1, ADD2, and NGFR [64][65][66] , were reduced in ATG7 KD and ATG14 KD iNeurons.The proteins with reduced phosphorylation in the autophagy-deficient iNeurons showed strong enrichment in GO terms associated with neuronal functions such as neuronal cellular compartments, neurogenesis, and neuron development (Fig. 5j).By screening PKA consensus motif -containing proteins (R/K-X-X-pS/T) 67 and those with reduced phosphorylation in autophagy-deficient iNeurons, we identified 16 putative PKA substrates in the 86 shared down-regulated phospho-proteins (Fig. 5k).Future studies will be needed to illustrate further their function in mediating neuronal activity.a, c) and SV (b, d) proteins in ATG7 KD and ATG14 KD iNeurons, respectively.Relative protein levels were normalized to the loading control β-actin.N = 6 (for RTN3, ATL1, SEC62, CNX, REEP5, SV2A, SYNGR1, SYNGR3, and SYP1 for ATG7 KD iNeurons), and n = 3 (for TEX264 and SV2C) biologically independent replicates for the proteins.The controls' means were set to 1. Two-sided unpaired t-test.Ns, no significance.e, f Violin plots of ER sub-compartment proteins enriched in ER curvature, ER-Golgi intermediate compartment (Ergic/cisGolgi), ER lumen, ER membrane in ATG7 KD iNeurons.Each dot represents one protein.Solid black bars indicate the median Log 2 FC, and the dashed bars specify the 25th and 75th interquartile range.g Immunofluorescence images of Calnexin (CNX, magenta) and MAP2 (green) staining in ATG7 KD and control iNeurons.Insets, magnified images of the boxed region.h Quantification of CNX+ large inclusions per μm 2 of total neuronal dendrite in ATG7 KD iN and control iN.Two-sided unpaired t-test; Scale bar, 10 μm.(n = 15).i Bar graph of top ranked GO terms generated by using SynGO from ATG7 KD vs. control iN (p value < 0.05, Log 2 FC > 2 SD, SD = 0.18).j Bar graph of top ranked GO terms generated by using SynGO from ATG14 KD vs. control iN.k Immunofluorescence images of ATG7 KD human iNeurons stained with anti-SV2A, anti-synaptophysin1 (SYP1), and anti-synaptogyrin3 (SYNGR3), and anti-MAP2 antibodies.The highlighted areas were enlarged in inset.Scale bar, 10 μm.(3 biological replicates, n = 10 per batch).l Quantification of synaptic "puncta" density and size labeled with SV2A, SYP1, and SYNGR3 from k. Unpaired t-test; Scale bar, 10 μm.(3 biological replicates, n = 10 per batch).All data are shown as mean ± SEM.

Loss of neuronal autophagy causes an aberrant increase of RI subunits and impairment of c-FOS activity in mouse brains
We next asked if autophagy-deficiency in neurons affects PKA subunit degradation and PKA activity in vivo.Immunoblots demonstrated that AKAP11, RΙα, and PKA-Cα subunit levels are significantly increased in Atg7 cKO and Atg14 cKO mouse brains (Fig. 6a, b).To explore the brain region specificity of the increased AKAP11 and RIα expression, we performed IF staining for AKAP11 and RIα in Atg7 cKO and control brains.The overall fluorescence intensities of AKAP11 and RΙα staining were markedly increased across many brain regions of the mutant mice.For example, we found the accumulation of AKAP11 and RΙαpositive cytosolic aggregate-like structures in the hippocampus (CA3 and DG), cortex (layer IV), amygdaloid, and thalamic nucleus (Fig. 6c-h).Additionally, we found AKAP11 colocalized with p62 in the cytosolic aggregates in neurons (Fig. 6e, f).IF staining of the mutant mouse brains revealed a reduced number of neurons labeled with c-FOS, indicating a decrease in neuronal activity at the basal level (Fig. 6i,  j).Together, our work demonstrates that basal autophagy plays a critical role in maintaining the homeostasis of PKA-RI complex in neurons; loss of autophagy disrupts PKA signaling and impairs c-FOS mediated neuronal activity in vivo.

AKAP11 mediates the degradation of PKA-RI protein complex in iNeuron
Finally, we asked whether AKAP11 plays a role in mediating the degradation of PKA RI (α and β) and PKA-Cα in neurons.We knocked down AKAP11 gene expression in iNeurons using CRISPR interference technology and found a significant increase of PKA RIα, RIβ, and Cα subunit levels in two independent lines of iNeurons through immunoblotting.In contrast, the protein level of PKA RII subunit remains unaltered (Fig. 7a, b).The IF staining intensity for both RIα and RIβ proteins was markedly enhanced in the soma and the neuritic processes of AKAP11 KD iNeurons, compared to the control iNeurons (Fig. 7c, d).Thus, the above data demonstrates a critical role for AKAP11 in mediating selective degradation of PKA-RI complex through autophagy in human neurons.Such a remarkable increase of the RIα and RIβ levels may have a profound impact of PKA-mediated signaling and neuronal activity.

Discussion
Our integrated proteomic profiling and multidisciplinary experimental validation have delineated the landscape of autophagy degradation in human and mouse neurons.We show that basal autophagy targets various cellular pathways and organelles, such as ER, mitochondria, Golgi, SV proteins, endosome, and cAMP-PKA pathway, for degradation in neurons.We find a function of neuronal autophagy in maintaining the homeostasis of PKA-RI complex through the autophagy receptor AKAP11 and regulating cAMP/PKA activity.Thus, our study offers a valuable resource for understanding the physiological function of autophagy in human neurons.
Previous investigations of neuronal autophagy cargo have provided useful information for the understanding of autophagy functions in neuron 10,25,27,68 .Our integrated proteomic analysis with multiple mutant iNeuron lines and mice expands the investigations by providing insight on the neuronal autophagy functions.One key finding is the remarkable abundance of proteins from ER subcompartments 53 resulted from autophagy block, suggesting that ER is a most prominent target of neuronal autophagy.The significance of ER-phagy in neurons has been proposed in previous studies by using Atg5 cKO mice 27 and human ATG12-deficient, NGN2-induced neuron 68 .Supporting the robustness of our approach for autophagy cargo identification, our results uncovered all six previously reported ER-phagy receptors, including TEX264, FAM134B, RTN3L, SEC62, CCPG1, and ATL3, suggesting that these receptors are themselves the substrates of autophagy.Future study of the many ER protein cargoes identified from our study will assist in understanding the incredible complexity and selectivity of ER-phagy particularly in neurons 69 .
Our results demonstrate that autophagy plays a role in maintaining the homeostasis of many SV trafficking proteins in neurons.Our data suggests that a set of SV-related proteins, including synap-togyrin1 (SYNGR1), synaptogyrin3 (SYNGR3), synaptophysin1 (SYP1),

Control
Atg7 f/f -SynCre SV2A, and SV2C are potential autophagy targets in neurons.Although it was proposed that autophagy may recycle SV at presynaptic terminals 15 , one study showed that disruption of autophagy has little impact on the number of SV.Our EM study of ATG7 KD iNeuron also suggests no significant change of SV number (Supplementary Fig. 3L), arguing that autophagy may not directly turn over SV per se at the synapse 27 .Nonetheless, our observation may provide insight into the mechanism for how autophagy regulates presynaptic activity 14 .Another significant finding in our study is the identification of the specific PKA-RI complex as a target of autophagy in neurons under basal condition.Our results demonstrate that neuronal autophagy constantly degrades RI regulatory (α/β) and Cα catalytic subunits of PKA, therefore controlling PKA homeostatic activity.A recent study reported that autophagy regulates cAMP/PKA signaling pathway in neuron specific Atg5 KO mouse excitatory neurons 70 , corroborating our observations.Our study further showed a function of AKAP11 as an autophagy receptor in selective degradation of PKA RI regulatory (α/β) and Cα catalytic subunits in neurons, a function likely conserved in different cell types as we previously showed 57 .We observe discrete AKAP11 and RΙα puncta, which colocalize with p62, in neurons at multiple brain regions, such as the CA3 and DG regions of the hippocampus, layer IV of the cortex, amygdaloid, and thalamic nucleus from Atg7 cKO mice, suggesting that p62 may act as a co-receptor for AKAP11-mediated degradation, and together, they regulate PKA activity in the brain.Previous studies have shown that PKA regulates neuronal signaling pathways which is essential for learning and memory in mice and flies 61,62 .Our findings thus provide additional insight into the mechanism whereby neuronal autophagy control synaptic activity underlying learning and memory 71 .
Our study shed light on the molecular mechanism underlying multiple neurological disorders.A recent study reported 12 individuals from 5 independent families that carried biallelic, recessive variants in human ATG7 associated with severe reduction or complete loss of ATG7 expression.These individuals suffered complex neurodevelopmental disorders 16 .Our systematic study, particularly with human iNeurons carrying ATG7 deletion, provides a comprehensive view for what cellular pathways and functions could be disrupted in the patients' neurons carrying ATG7 mutations.In Alzheimer's disease (AD), dystrophic neurites and amyloid plaques are the important hallmarks of neuropathologies in the affected brains 72 .Morphological and molecular analysis of AD brains and mouse models showed a sequential and distinct enhancement of staining of ATG9, RTN3, REEP5, RAB7, LC3, and LAMP1 proteins during dystrophic neurite and amyloid plaque development 73 .The increased ATG9, RTN3 and REEP5 protein levels in AD neurites surrounds the early-stage amyloid plaques, suggesting that a block of ER-phagy occurs early and may contribute to the formation of dystrophic neurites and amyloid plaques in affected AD brains.Furthermore, rare AKAP11 truncating variants were recently identified as significant risk factors for both bipolar disorder and schizophrenia from exome sequencing 74 .Our observation raises a possibility that the pathogenic mechanism of the psychiatric disease involves disruption of PKA-RI complex homeostasis and cAMP-PKA signaling in neurons.
We also recognized the limitations of our study.For example, the long-term inhibition of autophagy function through deletion of the essential autophagy genes may cause compensatory responses with specific changes in certain accumulated proteins resulting from disturbing cellular homeostatic mechanisms, such as ER stress, posttranscriptional modifications, and protein synthesis, etc 75 .Our approaches, however, is unable to exclude those proteins with increased levels caused potentially by the above conditions.Furthermore, our study has yet to examine neuronal interactome of additional ATG8 homologs (e.g., GABARAP, GABARAPL1/2), despite our analysis of the LC3-interactome.Therefore, our studies await In summary, our comprehensive approach of deep proteomic profiling provides a global view of autophagy degradation in human and mouse neurons, offers a valuable resource of neuronal autophagy research, and shed light on the molecular mechanisms of several neurological disorders.

Animals
All animal experiments were approved by the Icahn School of Medicine at Mount Sinai Institutional Animal Care and Use Committee (IACUC-2015-0046) and were conducted in compliance with the relevant ethical regulations.Mice were maintained in social cages on a 12 h light/dark cycle with free access to food and water; animals were monitored daily for food and water intake.

Cell culture
All human ESCs and PSCs were maintained on Geltrex-coated plates in feeder-free Stemflex medium and a 5% CO 2 environment at 37 °C.Cells were passaged using 0.5 mM EDTA/PBS in Stemflex supplemented with 2 mM Thiazovivin or 50 nM Chroman1.Thiazovivin/Chroman1 was removed from the medium on the following day.Research performed on samples of human origin was conducted according to protocols approved by the Institutional Review Boards of Icahn School of Medicine at Mount Sinai.H1 (WA01) ES cells (from a male parental cell line) were obtained from WiCell Research Resources (#Ae001-A); The inducible i3N iPS cell line (from the WTC11 iPSC parental cell line, male) was kindly provided by Dr. Martin Kampmann's lab 34 .Mono-ALLELIC mEGFP-Tagged MAP1LC3B WTC iPSC Line were purchased from Coriell Institute (#AICS-0030-022) which is derived from a male parental iPS cell line, human embryonic kidney (HEK) 293 T/17 cells were purchased from ATCC (CRL-11268).
Mouse glial cultures were generated from cortical hemispheres at postnatal day 3 (P3).Briefly, the cortices from 3 pups were incubated in 5 mL of 20 Units/mL Papain, 0.5 mM EDTA, and 1 mM CaCl 2 in HBSS for 15 minutes.After incubation, the tissues were manually dissociated by forceful trituration.Cells were filtered using a 70 μm cell strainer.The resulting cells were grown in MEF medium at 37 °C with 5% CO 2 .

Cell line generation
H1-Lenti-ATG7/14 KD cell line.Human H1 stem cells were dissociated using 0.5 mM EDTA/PBS and plated in Stemflex supplemented with 50 nM Chroman1.Lentiviruses (Addgene #71236) expressing ATG7 sgRNA #3, ATG14 sgRNA #3, or an empty vector were used to transduce H1 cells during passaging.The medium was changed to Stemflex supplemented with puromycin (1 μg/ml) for 3 consecutive days.Puromycin resistant cells were dissociated using Accutase for 15 min and seed at a density of 2000 cells per 10 cm dish.Individual clones were harvested and expanded.Protein expression was confirmed by immunoblotting assay.ATG7 KD clone #1 and #6, ATG14 KD clones #3 and #5, and empty vector clones #2 and # 6 were confirmed by western blotting and used for subsequent experiments.

Generation of neurons from human PSCs
Glutamatergic neurons were generated by overexpression of the transcription factor Ngn2 as previously described 33 .Briefly, human PSCs were dissociated using 0.5 mM EDTA/PBS and plated at a density of 88,000 cells/cm 2 in N2 medium supplemented with 50 nM Chro-man1.At the same time, cells were mixed with FUW-TetO-Ngn2-P2Apuromycin and FUW-rtTA lentiviruses by adding them directly to the medium.24 h later, the medium was replaced by N2 medium (1×N2 supplement, 1×NEAA in DMEM-F12 medium) containing Doxycycline (2 mg/mL) to induce transgene expression.Transduced cells were enriched with puromycin (1 mg/mL) for 2 days.4-5 days post Doxycycline, neurons were dissociated and plated together with mouse glial cells (100,000 cells/cm 2 ) on Geltrex-coated plates.Two weeks after transgene induction, Doxycycline was removed and the neuronal culture was maintained in Neurobasal A medium supplemented with 1× B27, 1× Glutamax, and 1% fetal bovine serum.Mature neurons were used for various experiments on day 42 (after 5 weeks of co-culture with mouse glia).

RNA extraction and mRNA sequencing
To determine gene expression level, ATG7 KD and Control iNeurons (Day 42, 1 well of a 6-well plate around 2.5 × 10 5 cells total) were washed with PBS and lysed in 1 mL of Trizol for RNA extraction.Following lysis, DNA contamination was removed using TURBO DNA-free kit treatment.3 biological replicates were obtained for both ATG7 KD and Control iNeurons.Three pairs of mouse whole brain tissue from Atg7 f/f and Atg7 f/f -SynCre mice (6~8 weeks old, both male and female were included) were dissected and snapped frozen using liquid nitrogen.Mice whole brains were ground using the Mortar & Pestle Set (Thermofisher, Cat# FB961N and 50-195-4054054).

Western blot analysis of mouse brain lysate
Whole brains were collected from mice described above at 6-8 weeks of age and homogenized with a lysis buffer containing 50 mM Tris HCl, pH 7.5, 150 mM NaCl, 1% Triton X-100, and proteinase/phosphatase inhibitor.Brain lysates were subjected to immunoblotting assay following the same protocol for cultured cells.

Forskolin treatment of human iNeurons
Human i3N PSC cell lines were transduced with Lenti-ATG7-hygromycon-BFP viruses to knock-down ATG7 and Lenti-NTC-hygromycin-BFP viruses as control.Cells were selected using hygromycin B (Life Technologies, #10843555001, 1:250) every two generations.PSC cells were plated in the 6-well plate in Stemflext medium supplemented with 50 nM Chroman1.24 h later, the medium was replaced by N2 medium containing Doxycycline (2 mg/mL) to induce neuron differentiation.For immunofluorescence staining, 4-5 days post Doxycycline, neurons were dissociated and plated together with mouse glial cells (neuron density at 100,000 cells/cm 2 ) on Geltrex-coated plates.Two weeks after transgene induction, Doxycycline was removed and the neuronal culture was maintained in Neurobasal A medium supplemented with 1× B27, 1× Glutamax, and 1% fetal bovine serum.Mature neurons on day 42 (after 5 weeks of co-culture with mouse glia) were treated with 4 μm Forskolin (Sigma-Aldrich, # F3917) for 5 min, 1 h and 2 h.Neurons were fixed and subjected to immunofluorescence staining as described above.For western blot, PSC cells were plated in 6-well plates at density of 150 k/cm 2 in Stemflext medium supplemented with 50 nM Chro-man1.24 h later, the medium was replaced by N2 medium containing Doxycycline (2 mg/mL) to induce neuron differentiation and was changed to fresh N2 plus Doxycycline every 2 days.Medium was changed to Neurobasal medium supplemented with 1× B27, 1× Glutamax, and 1% fetal bovine serum until 14 days post Doxycycline treatment.Neurons on day 14 were treated with 4 μm Forskolin for 5 min, 1 h and 2 h.Cells were harvested for western blot as described above.

Screening strategy for ideal candidate gene from autophagydeficient human iNeurons MS data
A multi-step screening process was used to determine the most ideal candidate gene for our study.The first step in establishing a list of candidate genes was to screen for genes in the ATG7 and ATG14 KD human iNeurons proteomics data that had a Log 2 FC greater than the standard deviation (Log 2 FC > SD) and a Log 2 FC greater than 0 (Log 2 FC > 0).These genes were then subjected to GO:TERM analysis by a web-based tool to determine which genes were related to the ER.SynGo was performed in the webtool (https://www.syngoportal.org/).

Electron microscopy (EM)
Three pairs of Atg7 f/f and Atg7 f/f -PCP2-Cre mice (6~8 weeks old) were sacrificed, and samples were processed following the procedures.
Briefly, 3 pairs of Atg7 f/f and Atg7 f/f -PCP2-Cre mice were perfused and fixed with 2% paraformaldehyde and 2% glutaraldehyde in 0.1 M PB, pH 7.4.40 mm vibrotome-cut utra-thin (∼70 nm) sections were collected on uncoated 200-mesh grids.Grids were viewed with a TecnaiSpiritBT Transmission Electron Microscope (FEI) at 80 KV and pictures were taken with Gatan 895 ULTRASCAN Digital Camera.
For morphology EM on cells, 3 pairs of independently prepared ATG7 KD and control human iNeuron cells (42 days post differentiation) grown on Permanox slides (Electron Microscopy Sciences (EMS), Hatfield, PA) were taken from incubation, and directly placed in 2% glutaraldehyde and 2% paraformaldehyde /0.1 M sodium cacodylate buffer for a minimum of 2 h at 4 °C.Cells were washed, fixed in 1% aqueous osmium tetroxide at RT for 1 h, washed, and transferred to 2% aqueous uranyl acetate at RT for 1 h.Slides were washed with distilled water, dehydrated in an ascending aqueous ethanol series, and then embedded in Epon resin.Inverted BEEM capsules (#3) were placed directly over regions of interest, filled with fresh resin, and transferred to a vacuum oven for heat polymerization at 60 °C for 12-24 h.To separate the cells from the slides, a hot plate was heated to 60 °C, and the slides were placed directly on a pre-heated hot plate for exactly 3 min and 30 s.The capsules were removed from the hot plate and carefully dislodged free from the slide using a plier.Ultrathin (85 nm) sections were collected onto 300 mesh copper grids (EMS) using a Leica UC7 ultramicrotome (Leica Biosystems Inc., Buffalo Grove, IL), contrast stained with uranyl acetate and lead citrate, and imaged on a Hitachi 7700 transmission electron microscope (Hitachi High Technologies America, Inc., Dallas, TX) equipped with an AMT 2 K × 2 K digital camera (Advanced Microscopy Techniques, Corp., Woburn, MA).

Sample preparation and quantitative proteomics
Autophagy-deficient human iNeurons and mouse brains.Three batches of H1-Lenti-ATG7/14 KD derived Ngn2 iNeurons were included for the quantitative proteomics and phosphoproteomics analyses (day 42 post differentiation, two subclones for each genotype.Three samples of ATG7 KD clone # 1 and clone #6, one sample of ATG14 KD clone #3 and two samples of ATG14 KD clone#5, three samples of Vector clone #2 and four samples of clone #6).The cells were washed with PBS once and manually dissociated by gentle trituration in PBS.Cells were then centrifuged at 500 g for 3 min.The supernatant was removed, and cell pellet was snap-frozen by liquid nitrogen and stored at −80 °C before being subject to proteomics analyses.For mouse brain samples, 6-8-week-old autophagy-deficient mice (3 Atg7 f/f -SynCre and 4 Atg7 f/f mice; 3 Atg14 f/f -SynCre and 4 Atg14 f/f mice) were sacrificed and whole brains were dissected, and snap frozen in liquid nitrogen and stored in −80 °C before proteomics analyses.
Protein extraction, quantification of iNeurons and mouse brain samples.iNeuron cell pellets and mouse brain samples were lysed, and the protein concentrations were quantified.Briefly, the frozen samples were homogenized in the lysis buffer (50 mM HEPES, pH 8.5, 8 M urea, and 0.5% sodium deoxycholate) with 1X PhosSTOP phosphatase inhibitor cocktail (Sigma-Aldrich).Protein concentration was measured by the BCA assay (Thermo Fisher, # 23225) and then confirmed by Coomassie-stained short SDS gels.
Protein digestion and TMT labeling for iNeurons and mouse brain samples.The analysis was performed with a previously optimized protocol 45 ~0.1 mg of iNeurons protein samples and ~0.1 mg of mouse brain protein samples (in lysis buffer with 8 M urea) were first digested by Lys-C (Wako, 1:100 w/w) at 21 °C for 2 h and then diluted by 4-fold to reduce urea to 2 M by 50 mM HEPES followed by the addition of trypsin (Promega, 1:50 w/w) to continue the digestion at 21 °C overnight.The digestion was terminated by the addition of 1% trifluoroacetic acid.After centrifugation, the supernatant was desalted and then dried by Speedvac.Each sample was resuspended in 50 mM HEPES (pH 8.5) for TMT labeling, and then mixed equally followed by desalting.The iNeuron samples were labeled by TMT 16.The Atg7 cKO and Atg14 cKO mouse brain samples were labeled by two sets of TMT 10 respectively, but only 7 TMT channels was discussed here.All TMT channel information can be found in Supplementary Data 1a.

Phosphopeptide enrichment of iN samples
Phosphopeptide enrichment was carried out by TiO 2 beads as descripted previously 81 .Briefly, the pooled desalted TMT labeled peptides (~1.8 mg) were dissolved in the binding buffer (65% acetonitrile, 2% TFA, and 1 mM KH2PO4).TiO 2 beads (7.2 mg) were washed twice with the washing buffer (65% acetonitrile, 0.1% TFA), incubated with the peptide at 21 °C for 20 min.The TiO 2 beads were then washed twice again with washing buffer and loaded into a C18 StageTip (Thermo Fisher), followed by the elution of phosphopeptides by the basic pH buffer (15% NH4OH, and 40% acetonitrile).The eluates were dried respectively and dissolved in 5% formic acid for LC-MS/MS analysis.The flowthrough of phosphopeptide enrichment was desalted and dried for basic LC fractionation.

GFP-LC3 affinity purification from mouse brains, in-gel digestion, and TMT labeling
The GFP-LC3 overexpression and control mouse brains (3 pairs of 6~8 weeks old mice, both male and female mice were used) were harvested and rinsed with cold PBS.Homogenization buffer A (0.32 M sucrose, 1 mM NaHCO3, 0.25 mM CaCl2, 1 mM MgCl2, 50 mM Tris HCl, pH 7.5) supplied with halt proteinase and phosphatase inhibitors (Invitrogen, #78442) was added as 2 ml buffer A per mouse brain.The tissue was homogenized with a glass douncer 20 times before being centrifuged at 1000 × g for 10 min.The supernatant was transferred to a new 15-ml falcon tube labeled as the cytoplasmic sample.2X buffer D (300 mM NaCl, 20% NP40, 10% Sodium Deoxycholate, 2 mM EDTA, 2 mM EGTA, 100 mM Tris HCl, pH 7.5) was added and the sample was rotated for 1 h.Next, the sample was centrifuged at 20,800 × g for 20 min and the supernatant was transferred into a new falcon tube.The sample was then centrifuged for the second time using the same conditions, with the supernatant being collected.100 µl was saved as input and protein concentration was quantified.15 µl of GFP-trap beads (ChromoTek, #gmta-20) per mouse brain was added and the sample was incubated overnight.The next day, beads were washed three times (10 minutes each, 500 µl of 1X buffer D).After the final wash, the protein was eluted at 70 °C with 30 µl of 1X LDS-PAGE sample buffer (Thermo Fisher, #NP0007).3 µl of the IP sample was used for SDS-PAGE and Silver staining (Invitrogen, #LC6070).The rest of the IP samples were run on a short SDS gel followed by in-gel digestion 82 , the peptides were then labeled by TMT11 respectively, mixed equally, followed by desalting for the subsequent fractionation.Only 8 of the 11 channels' data was discussed here.All TMT channel information can be found in Supplementary Data 1a.

Extensive two-dimensional liquid chromatography-tandem mass spectrometry
For each of the 4 TMT sets (iNeuron (16 samples, genotypes and clone numbers as described above), Atg7 cKO mouse brain (3 Atg7 f/f -SynCre and 4 Atg7 f/f mice brains), Atg14 cKO mouse brain (3 Atg14 f/f -SynCre and 4 Atg14 f/f mice brains) and GFP-LC3 affinity purification samples (described above).All the human iNeurons were generated from the three independent biological replicates (n = 3).All the mice were 6~8 weeks old, both male and female mice were used.Western blots using the remaining cell lysates after the Mass-spectometry was used to confirm the genotype of the samples of all human iNeurons and mouse samples.The TMT labeled samples were fractionated by offline basic pH reverse phase LC respectively, and each of these fractions was analyzed by the acidic pH reverse phase LC-MS/MS 40 .The offline basic pH LC was performed with an XBridge C18 column (3.5 μm particle size, 4.6 mm × 25 cm, Waters), buffer A (10 mM ammonium formate, pH 8.0), buffer B (95% acetonitrile, 10 mM ammonium formate, pH 8.0), using a 2-3 h gradient of 15-35% buffer B. In the acidic pH LC-MS/MS analysis, fractions were analyzed sequentially on a column (75 μm × 15-30 cm, 1.9 μm C18 resin from Dr. Maisch GmbH, 65 °C to reduce backpressure) coupled with a Fusion or Q Exactive HF Orbitrap mass spectrometer (Thermo Fisher Scientific).Peptides were analyzed with a 1-3 h gradient (buffer A: 0.2% formic acid, 5% DMSO; buffer B: buffer A plus 65% acetonitrile).For mass spectrometer settings, positive ion mode and data-dependent acquisition were applied with one full MS scan followed by a 20 MS/MS scans.MS1 scans were collected at a resolution of 60,000, 1 × 10 6 AGC and 50 ms maximal ion time; higher energy collision-induced dissociation (HCD) was set to 32-38% normalized collision energy; ~1.0 m/z isolation window with 0.3 m/z offset was applied; MS2 spectra were acquired at a resolution of 60,000, fixed first mass of 120 m/z, 410-1600 m/z, 1 × 10 5 AGC, 100-150 ms maximal ion time, and ~15 s of dynamic exclusion.For the phosphoproteomics analysis, 1.5 m/z isolation window was used to improve sensitivity.

Protein and phosphopeptide identification and quantification by the JUMP software suite
The bioinformatics processing of protein and phosphopeptide identification and quantification were carried out with the JUMP software suites (Version 1.13.0) 83.In brief, MS/MS raw data were searched against a target-decoy database to estimate false discovery rate (FDR) 84 .We combined the downloaded Swiss-Prot, TrEMBL, and UCSC databases and removed redundancy (human: 83,955 entries; mouse, 59,423 entries; date of downloading: April 20, 2020) to create the database.Main search parameters were set at precursor and product ion mass tolerance ( ± 15 ppm), full trypticity, maximal modification sites (n = 3), maximal missed cleavage (n = 2), static mass shift including carbamidomethyl modification (+57.02146 on Cys), TMT tags (+229.16293for TMT10/11 or +304.20714 for TMT16 on Lys and Ntermini), and dynamic mass shift for oxidation (+15.99491 on Met) and Ser/Thr/Tyr phosphorylation (+79.96633).Peptide-spectrum matches (PSM) were filtered by mass accuracy, clustered by precursor ion charge, and the cutoffs of JUMP-based matching scores (J-score and ΔJn) to reduce FDR below 1% for proteins during the whole proteome analysis or 1% for phosphopeptides during the phosphoproteome analysis.Protein and phosphopeptide quantifications were performed based on the reporter ions from MS2 using our previously optimized method 40 .
The external solution used during recordings consisted of (in mM): 140 NaCl, 5 KCl, 10 HEPES, 2 CaCl 2 , 2 MgCl 2 , and 10 Glucose, with a pH of 7.4.All recordings of cell culture experiments were performed at room temperature.
To ensure data quality, any cells exhibiting a change in series resistance (Rs) exceeding 20% during the recording were excluded from the analysis.Additionally, recordings with an access resistance greater than 25 MΩ were also excluded.The electrophysiological data were collected using pClamp 10.5 software (Molecular Devices) and the MultiClamp 700B amplifier (Molecular Devices).The detection of miniature excitatory postsynaptic currents (mEPSCs) was performed using the template search algorithm in Clampfit 10.5 software (Molecular Devices).

Quantification and statistical analysis
Unless otherwise indicated, all data presented are the average of at least three biological replicates from each of at least three independent experiments.The statistical significance of differences between two groups was determined using the unpaired two-tailed Student's t test or Mann-Whitney U test based on the normality test.For multiple-means comparisons, statistical significance was determined by one-way analysis of variance followed by Dunnett's multiple comparisons test using GraphPad Prism 10 (GraphPad Software, CA, USA).All values are reported as mean ± SEM.See figure legends for details on specific statistical tests run for each experiment.Error bars were calculated using the standard deviation of all the replicates.Graphs and plots were generated using Graphpad Prism 10.0 or RStudio (2022.07.1).Figures were generated using Adobe Illustrator 2022.

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Fig. 2 |
Fig.2| Validation of the DEPs of ER and SV related proteins in autophagydeficient human iNeurons.Quantification of ER (a, c) and SV (b, d) proteins in ATG7 KD and ATG14 KD iNeurons, respectively.Relative protein levels were normalized to the loading control β-actin.N = 6 (for RTN3, ATL1, SEC62, CNX, REEP5, SV2A, SYNGR1, SYNGR3, and SYP1 for ATG7 KD iNeurons), and n = 3 (for TEX264 and SV2C) biologically independent replicates for the proteins.The controls' means were set to 1. Two-sided unpaired t-test.Ns, no significance.e, f Violin plots of ER sub-compartment proteins enriched in ER curvature, ER-Golgi intermediate compartment (Ergic/cisGolgi), ER lumen, ER membrane in ATG7 KD iNeurons.Each dot represents one protein.Solid black bars indicate the median Log 2 FC, and the dashed bars specify the 25th and 75th interquartile range.g Immunofluorescence images of Calnexin (CNX, magenta) and MAP2 (green) staining in ATG7 KD and

RTN3Fig. 3 |
Fig. 3 | Proteomic analysis of proteins from mouse brains with neuron-specific KO of Atg7 or Atg14.Graph depicting body weight (grams) of Atg7 f/f and Atg7 f/f -SynCre mice (a), Atg14 f/f and Atg14 f/f -SynCre mice (b) at different ages.Data points represent mean +/− SD.Kaplan-Meier survival curve of Atg7 f/f and Atg7 f/f -Syn-Cre (c), Atg14 f/f and Atg14 f/f -Syn-Cre (d) mice at different ages.e Volcano plot of the DEPs from proteomic data of Atg7 f/f and Atg7 f/f -SynCre mouse brains.Inset bar graph shows the enriched GO terms associated with DEPs (p < 0.05, Log 2 FC > 2 SD, SD = 0.1).f Volcano plot of the DEPs from proteomic data of Atg14 f/f and Atg14 f/f -Syn-Cre mouse brains.Inset bar graph shows the enriched GO terms among DEPs (p < 0.05, Log 2 FC > 2 SD, SD = 0.16).g Venn diagram shows the overlap between the DEPs of Atg7 cKO and Atg14 cKO mouse brains.h GO enrichment analysis of 102 DEPs shared between Atg7 cKO and Atg14 cKO mouse brains.Plot of Log 2 FC and -Lg (FDR) for ER (i) and SV (j) DEPs from Atg7 cKO and Atg14 cKO mouse brains, respectively.DEPs are indicated by blue (Atg7 cKO) and orange dots (Atg14 cKO).Inset circle plots show the percentage of ER DEPs vs. the rest of the DEPs.ER and SV proteins were defined by their GO annotations in gProfiler.k A table of the known autophagy receptors and those detected (check marks) in proteomic analysis of the enriched proteins in Atg7 cKO and Atg14 cKO mouse brains (p < 0.05, Log 2 FC > 2 SD).The list of autophagy receptors was manually collated through literature searches.Quantification of the change of the indicated ER and SV proteins from Atg7 cKO (l) and Atg14 cKO (m) mouse brains vs. Control.n = 3 biologically independent replicates.Relative protein levels were normalized to β-actin.The controls' means were set to 1. Two-sided unpaired t-test.NS, no significance.All data are shown as mean ± SEM. n Electron microscopy (EM) images of Purkinje axons at the deep cerebellar nuclei (DCN) area from Atg7 f/f and Atg7 f/f -Pcp2-Cre mice.Scale bar, 1 μm.

Fig. 4 |
Fig. 4 | Proteomic analysis of GFP-LC3-interacting proteins from neuronspecific Atg7 cKO mouse brains.a Schematic of LC3-interactome by combining genetic models and proteomics.Created with BioRender.com.b Volcano plot of proteomic results from DEPs between Atg7 f/f -SynCre; GFP-LC3 IP vs. control IP (gray dots) and DEPs between Atg7 f/f -SynCre; GFP-LC3 IP and Atg7 f/f ; GFP-LC3 IP (green dots).Threshold for enriched DEPs is p < 0.05 and Log 2 FC > 2 SD (SD = 0.21).Dashed line indicates -Lg (p-value) = 1.3.Venn diagram (left) of overlapping DEPs between ATG7 KD iNeurons (c), ATG14 KD iNeurons (d), Atg7 cKO mouse brain (e) or Atg14 cKO mouse brain (f) and Atg7 f/f -SynCre; GFP-LC3 IP.Bar graph displays the enriched GO annotations from the overlapping proteins.g Heatmap of GO annotations enriched in ATG7 KD iNeurons, ATG14 KD iNeurons, Atg7 cKO mice, Atg14 cKO mice and Atg7 f/f -SynCre; GFP-LC3 IP. h Percentage of identified ER proteins and ER sub-compartment proteins in different generic models.Unidentified: not detected in the dataset.i Venn diagram of upregulated DEPs enriched in ATG7 KD iNeurons, ATG14 KD iNeurons, Atg7 cKO mice, Atg14 cKO mice and Atg7 f/f -SynCre; GFP-LC3 IP.Numbers indicate protein numbers.15 proteins shared by all the datasets were shown in the right panel.j Heatmap shows Log 2 FC of ER compartment proteins that are identified in all 5 datasets.k Six examples of enriched PPI modules in upregulated DEPs shared by ATG7 KD and ATG14 KD iNeurons with the proteins validated in this study highlighted in red.Each dot represents a protein, and the interaction is indicated by connected lines.l Co-immunoprecipitation and immunoblot analysis through GFP-LC3 affinity pull-down from GFP-LC3 transgenic and WT mouse brains.m Immunofluorescence images (left) of GFP-LC3 knock-in human iNeurons with anti-SV proteins (SYNGR3 or SV2A).The cells were pretreated with chloroquine (CQ, 100 μM) for 24 h before fixation.Insets, magnified images of the boxed regions.Scale bar, 10 μm.Pearson's coefficient analysis of GLP-LC3 with SYNGR3 (n = 33 CQ-treated and 30 control iNeurons) or SV2A (n = 31 CQ-treated 28 control iNeurons) from 3 biologically independent replicates, two-sided unpaired ttest.All data are shown as mean ± SEM.
80, pRSV-rev (Addgene, #12253), pMDLg (Addgene, #12251), pCAG-VSVG (Addgene, #64084).To be brief, transfection mixtures A and B (one per lentivirus) were prepared as following (per each 10 cm): Tube A: 500 μl DMEM + 20 μg total DNA (e.g. 10 μg TetO-FUW-GFP or any other lentivirus plasmid, 5 μg PMDL, 2.5 μg RSV, 2.5 μg VSVg).Tube A and B were combined at the maximum speed for 10 seconds by vortex.Tube B: 500 μl DMEM + 60 μl 1X PEI, mix by vortexing.The mixture were incubated at room temperature for 20 min and vortexed one more time and added dropwise to the cells (1 ml per plate).At this point, cells were 90% confluent.6 h later, cells were washed with PBS and 6 ml of fresh prewarmed MEF media was added to each plate.30 h after the last media exchange, GFP expression was checked.The first media collection was only proceed if GFP is expressed by >than 70% of the cells.46 h post-transfection, a second collection was performed, and the plates were discarded.The I and II collections were combined.Lentiviral particles were ultracentrifuged at 75,300 × g, 4 °C for 2 h, re-suspended in DMEM, aliquoted, and stored at −80 °C.Only virus preparations with >90% infection efficiency as assessed by EGFP expression or puromycin resistance were used for experiments.