Targeted proteomics for monitoring autophagy

Abstract

The invention relates to a peptide mix comprising peptides for use in monitoring autophagy, the peptides being derived from proteins involved in autophagy.

Description

[0001] Targeted proteomics for monitoring autophagy

[0002] Field of the invention

[0003] The invention relates to peptide mixes for use in mass spectrometry.

[0004] Technical background

[0005] Eukaryotic cells maintain homeostasis and remove damaged or superfluous cellular components through a conserved degradation pathway called macroautophagy (hereafter: autophagy). This highly conserved catabolic pathway occurs at basal levels but is enhanced by a variety of stress signals, such as nutrient starvation or organelle damage. The process starts with the de-novo formation of a double-membrane organelle called autophagosome from membranes being recruited from various sources, the endoplasmic reticulum (ER) likely being principle donor. Autophagosome biogenesis and turnover is a directional process and can be divided into five principle stages.

[0006] In phase 1 “autophagosome initiation” kinase complexes are activated that start the process. In phase 2 “membrane nucleation” initial membrane sources are recruited to the phagophore assembly site, the principal site of autophagosome biogenesis.

[0007] In phase 3 “membrane expansion” and phase 4 “pore closure”, the growing cup-shaped membrane engulfs parts of the cytoplasm and closes up to form a double membrane vesicle. The autophagosome can fuse with other vesicles from endocytic pathways forming an amphisome and finishes in phase 5 by “fusing with the lysosome / vacuole”. Lysosomal fusion exposes autophagosomal cargo including the inner membrane to acidic hydrolases and enables the degradation of its content. The generated building blocks are recycled and transported back to the cytosol by lysosomal permeases to generate energy or fuel anabolic pathways.

[0008] While basal autophagy is often considered a “bulk process”, i.e., a non-selective process recycling cellular components in a random manner, autophagy can also be selective. Well studied cases of selective autophagy include the targeted removal of damaged organelles such as the mitochondrion or ER (termed mitophagy and reticulophagy / ER-phagy, respectively) as well as of protein aggregates (aggrephagy). This particular selectivity is enabled by a set of proteins called selective autophagy receptors (SARs), bridging the autophagy cargo to proteins of the ATG8 family, which are lipidated proteins coating the autophagosomal membrane and functioning as docking sites.

[0009] Numerous methods and protocols have been described for the study of protein turnover by autophagy, several of them relying on immunodetection or fluorescent microscopy of single proteins, which are often ectopically expressed as tagged variants facilitating downstream analyses. While these methods are well established, they often require a significant protein amount or protein tagging, making them unsuitable for some sample types, e.g., primary cells being difficult to transfect, and difficult to adapt for large-scale screening approaches. Additionally, some of these methods infer the activity of the whole pathway based on the quantification of single markers. Whereas this might be relevant for the analysis of basal, i.e., non-selective, autophagy flux, such approaches likely fail to capture the complexity of stress-induced autophagy in which different cargoes might be degraded at different rates and through different pathways.

[0010] To address the complexity of autophagy regulation and activity in an unbiased manner, approaches monitoring the activity of multiple genes by RNA-seq have been developed that are amendable for higher sample throughput and might also be used in prognostic / diagnostic settings. However, to infer autophagy activity based on changes in gene transcription biological samples have to be treated for several hours, commonly 2-8 hours, which is in contrast to the rapid cellular response to stress conditions. Autophagy signaling is activated within minutes, initial autophagosome biogenesis does not require de novo protein synthesis, and in mammalian cells autophagosomes and autophagydependent protein degradation can be detected as early as 30 min post stimulus.

[0011] Objective problem to be solved

[0012] To address this gap in analytical approaches, an assay is required that is able to detect changes in protein abundance, which is amenable to high throughput and which monitors abundance changes of multiple endogenous proteins reflecting the complexity and multitude of autophagy subtypes.

[0013] Summary of the invention

[0014] In one aspect, the invention relates to a method for monitoring autophagy selectivity, the method comprising probing a sample via mass spectrometry for the amount of at least two different selective autophagy receptor (SAR) proteins using a peptide mix as peptide standard, wherein the peptide mix comprises peptides derived from the at least two different SAR proteins.

[0015] In a second aspect, the invention relates to a mass spectrometry peptide standard comprising a peptide mix, wherein the peptide mix comprises peptides derived from at least two different SAR proteins, for use in a method according to the invention.

[0016] Brief description of the figures

[0017] Fig. 1 shows peptide responses and quantitative accuracy. A. Example calibration curve of the GABARAPL2 peptide IQLPSEK which was detected as doubly charged precursor (2+) and used to monitor GABARAPL2 protein abundance. Inlay is a zoom of the low concentration range. Black crosses indicate measured data points. B. Distribution of limits of detection and quantification of the 94 tryptic peptides used to monitor autophagy..

[0018] Fig. 2 shows the setup experiment. A. Average CV per technical replicate, B. Heatmap representing z-scored normalized peak intensities for each of the peptide that could be accurately measured. C. individual peptide values for peptides of the TAX1 BP1 protein.

[0019] Fig. 3 shows protein regulation by autophagy upon different starvation stimuli. A. Heatmap summarizing the behavior of regulated proteins reaching a p-value < 0.01 in an ANOVA test across all conditions. B. Example proteins for the main three identified clusters in the experiment. Basal conditions (DM EM) were set as 1 for reference.

[0020] Fig. 4 shows comparisons of methods and measurements of CVs. A. Euler plot of target proteins quantified in a minimum of one condition in DIA (Spectronaut normalization), DDA (MaxLFQ) and PRM. B. Average coefficients of variations at the protein level across HBSS and HBSS + BafA1 treatments using different quantification methods. Single values are represented by black dots. C.-E. Average peptides coefficients of variations distribution within biological replicates D. Detailed data from both lowest and highest CV producing peptides derived from C. ATG3 and E. STBD1, respectively, are shown as examples. F. Peptide CVs across biological replicates, representing biological variation and variation introduced by sample preparation.

[0021] Fig. 5 shows regulation of protein abundances by autophagy upon different starvation stimuli. A. Heatmap of a hierarchical cluster analysis summarizing the behavior of regulated proteins reaching a p-value < 0.01 in an ANOVA test across all conditions (2 h treatments). B.-D. Example proteins for the clusters 1, 2 and 3 from panel a, respectively. Basal conditions (DM EM) were set as 1 for reference. E. Western blot analyses of some of the targets of the proteomics assay. Shown is one representative of n=3 biological replicates. Actin was used for normalization. F. Quantification from the western blots shown in E. Asterisks represent significance level of a two-side unpaired Student t-test between conditions in B. and within a condition + / - BafA1 treatment in C., D., F. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001 , ns=not significant. Bars highlight average values and dots single experiments.

[0022] Fig. 6 shows PRM measurements of DFP treated cells. A. Immuno-fluorescence microscopy of mito-QC cells. A representative of n=5-10 biological replicates is shown for a total of 200 cells. B. Quantification of the mito-QC reporter-based experiments highlighted in (a). Black dots represent the number of mitolysosomes per cell. C. Flow cytometry analysis of mito-QC cells. Shown is one representative of n=3 biological replicates D. Heatmap summarizing the changes in protein regulation upon mitophagy induction with DFP. Represented proteins reach a p-value < 0.01 by ANOVA test. E. Detailed quantification of soluble p62-like autophagy receptors linked to mitophagy. F. Western blot analysis of indicated target proteins. Shown is one representative experiment of n=3 biological replicates. Note: dashed lines indicate cropping marks of unrelated samples. All conditions were run on the same gel and blot. G. Quantification of western blots exemplified in F. H. Detailed quantification of ubiquitin-independent autophagy receptors involved in mitophagy. I. Detailed quantification of other autophagy receptors. Only measured datapoints are represented. If not enough datapoints were measured for DMEM condition (reference point), the samples were not normalized to DMEM. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001, ns=not significant, two-sided unpaired Student's T test. Bars highlight average values and dots single experiments.

[0023] Detailed description of the invention

[0024] The objective problem is solved by a method for monitoring autophagy selectivity, the method comprising probing a sample via mass spectrometry for the amount of at least two different selective autophagy receptor (SAR) proteins using a peptide mix as peptide standard, wherein the peptide mix comprises peptides derived from the at least two different SAR proteins. As used herein, the term “probing” refers to analyzing a sample in order to detect and / or quantify the presence of peptides and / or proteins therein.

[0025] As used herein, the term “peptide” refers to an amino acid chain having a length of between 5 and 40 amino acids. In a preferred embodiment, the peptides have length between 7 and 25 amino acids.

[0026] The peptides used in the invention are isotopically labelled heavy peptides. The skilled person knows how peptides can be isotopically labelled. In some embodiments, the peptides are labelled with13C,15N or2H. In a preferred embodiment, the peptides are labelled with13C and / or15N. In principle, any amino acid within the peptides may be labelled. In preferred embodiments, it is the arginine and lysine residues that are labelled.

[0027] Using the methods of the invention, it is possible to distinguish between different types of autophagy since the peptide mix used therein as peptide standard comprises peptides derived from at least two different SAR proteins. Since each SAR protein is specific for a different type of autophagy, detection of different SAR proteins allows to determine which types of autophagy are occurring in the sample, i.e., which structures / organelles are degraded by autophagy.

[0028] The different types of autophagy are aggrephagy, mitophagy, pexophagy, lysophagy, zymophagy, ER- / reticulo-phagy, ferritinophagy, glycophagy, xenophagy, ribophagy, midbody autophagy, golgiphagy, nucleophagy and lipophagy. Consequently, in preferred embodiments, the at least two different SAR proteins for which the sample is probed are specific for at least two different autophagy types selected from aggrephagy, mitophagy, pexophagy, lysophagy, zymophagy, ER- / reticulo-phagy, ferritinophagy, glycophagy, xenophagy, ribophagy, midbody autophagy, golgiphagy, nucleophagy and lipophagy.

[0029] In a preferred embodiment, the at least two SAR proteins are selected from the group consisting of SQSTM1, NBR1, Optineurin, CALCOCO2, TAX1BP1, Alfy, CCPG1, FAM134B, RTN3, NIX / BNIP3L, TOLLIP, TNIP1, FKBP8, BNIP3, PHB2, NCOA4, NUFIP1, STBD1, BNIP1, PHBP2, PNPLA2, BCL2L13, TEX264, SEC62, FAM134A, FAM134C, CDK5RAP3, ATL3, SPART, YIPF3, YIPF4, GOLPH3, LMNB1 and GABARAPL1 and GABARAPL2, the latter two being autophagosome membrane-bound proteins addressing bulk autophagy. In a preferred embodiment, the peptide mix comprises one or more peptides having SEQ ID NO: 1, 2, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 78, 79, 80, 81, 82, 92, 93, 115, 116, 117, 118, 119, 120, 121, 122, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 148, 149, 150, 151, 152, 153, 154, 168, 169, 174, 175, 176, 177,

[0030] 190, 191, 192, 193, 194, 206, 207, 208. In some embodiments, the peptide mix comprises at least 10, 20, 30, 40, 50, 60, 70, 80, 90 or all of these peptides.

[0031] In some embodiments, the peptide mix comprises peptides derived from at least 3, 4, 5, 6, 7, 8, 9, 10, 11 or 12 SAR proteins. In a particularly preferred embodiment, the peptide mix contains peptides derived from one SAR protein for each type of autophagy. In a particularly preferred embodiment, the peptide mix comprises one or more peptides having SEQ ID NO:

[0032] In some embodiments, the peptide mix further contains peptides contained in the peptide mix represent various key proteins involved in different stages of autophagy.

[0033] In a preferred embodiment, the peptides contained in the peptide mix according to the invention are derived from proteins of one of seven classes, namely proteins involved in autophagy regulation, proteins involved in autophagosome initiation, proteins involved in membrane nucleation, proteins involved in membrane expansion, proteins involved in pore closure and / or proteins involved in fusion of the autophagosome with lysosomes.

[0034] As used herein, the term “peptide derived from a protein” refers to a peptide sequence that corresponds to a peptide generated from a particular protein by proteolysis.

[0035] Proteases used for generating such peptides are LysC, AspN, GluC, chymotrypsin, proalanase, ArgC, elastase, pepsin, and trypsin. In a preferred embodiment, the peptides used in the peptide mix according to the invention are tryptic peptides, i.e., a peptide sequence that can be generated by treating the protein of interest with the endoprotease trypsin, which cuts amino acid chains after lysine and arginine residues if they are not followed by proline.

[0036] The peptides used in the peptide mix according to the invention may be non-modified or posttranslationally modified peptides. In one embodiment, the peptides are modified in the same way as proteins are modified posttranslationally, e.g., by phosphorylation, acetylation, methylation or ubiquitination. Autophagy is regulated on a posttranslational level and protein activity is modulated by phosphorylation, acetylation, methylation, and ubiquitination. Therefore, by including modified peptides in the peptide mix, the peptide mix according to the invention will be better suited to detect different forms of the proteins involved in autophagy.

[0037] The peptide mix according to the invention can be used in mass spectrometry (MS)-based analyses which support absolute protein quantification via accurate quantification of a predefined number of given peptides potentially encompassing numerous proteins. Quantification can be performed in targeted and non-targeted MS approaches. Nontargeted approaches such as Data Dependent Acquisition (DDA) and Data Independent Acquisition (DIA) monitor potentially all peptides being present in a given sample. Mass spectrometers operated in targeted mode focus on a pre-established set of mass-to- charge ratios corresponding to peptides-of-interest. Targeted approaches can be inclusive or exclusive, i.e., focusing on peptides-of-interest AND other peptides being present, or focusing ONLY on peptides-of-interest, respectively. By ignoring other peptides eluting from the coupled liquid chromatography (LC) column (exclusive), targeted proteomics greatly enhances the sensitivity for its targets. Combining this approach with the peptide mix according to the invention as spiked-in peptides further enables confident identification of the targets and robust normalization across samples, thus making the method suitable for absolute quantification.

[0038] More specifically, the peptide mix according to the invention can be used in a targeted proteomics approach, parallel reaction monitoring (PRM) or multiple reaction monitoring (MRM) or single reaction monitoring (SRM), allowing to accurately quantify tryptic peptides derived from human autophagy-relevant proteins in single MS measurements. In combination with a simple in-solution digestion protocol that can be performed on multiple, low protein concentration samples (atto- to millimolar concentrations) in a (semi)automatic manner the assay supports screening approaches.

[0039] In the peptide mix, the peptides are dissolved in any regular mass spectrometry buffer known in the art, e.g., in 50 mM ammonium bicarbonate buffer or in 0.1% formic acid in water. According to the invention, proteins involved in autophagy regulation are MTOR / FRAP, RPTOR, RICTOR, MLST8, AMPK (PRKAA1, PRKAB1 , PRKAG1), ribosomal S6 kinase (RPS6KA1 , RPS6KB1), TFE3 and TFEB. Therefore, in one embodiment, the peptide mix comprises or consists of peptides derived from ribosomal S6 kinase (RPS6KB1) and / or TFEB.

[0040] According to the invention, proteins involved in autophagosome initiation comprise ATG9A, AMBRA1 , WBP2, ULK1, ULK2, RB1CC1 / FIP200, ATG13, ATG101, PIK3C3 / VPS34, PIK3R4 / VPS15, BECN1 / Beclin-1 , and ATG14. Therefore, in one embodiment, the peptide mix comprises or consists of peptides derived from ATG9A, AMBRA1, WBP2, ULK1, ULK2, RB1CC1 / FIP200 and / or ATG13.

[0041] According to the invention, proteins involved in membrane expansion comprise MAP1 LC3A, MAP1LC3B, MAP1 LC3C, GABARAP, GABARAPL1, GABARAPL2, ATG2A, WIPI1, WIPI2, WIPI3, WIPI4, ATG3, ATG7, ATG12, ATG5 and ATG4A. Therefore, in one embodiment, the peptide mix comprises or consists of peptides derived from ATG2A, WIPI2, ATG3, ATG7, ATG12, ATG5 and / or ATG4A.

[0042] According to the invention, proteins involved in pore closure and fusion with lysosomes comprise LIVRAG, GABARAPL1 and GABARAPL2. Therefore, in one embodiment, the peptide mix comprises or consists of peptides derived from LIVRAG, GABARAPL1 and / or GABARAPL2.

[0043] According to the invention, selective autophagy receptors (SAR) proteins comprise SQSTM1 / p62, NBR1 , OPTN / Optineurin, CALCOCO1, CALCOCO2 / NDP52, TAX1BP1, WDFY3 / Alfy, CCPG1, FAM134B, RTN3, NIX / BNIP3L, TOLLIP, TNIP1, BNIP3, PHB2, NCOA4, NUFIP1, STBD1 , BNIP1 , PHBP2, TRIM16, FUNDC1, FKBP8, NLRX1, ATL3, TEX264, SEC62, PNPLA2, BCL2L13, TEX264, SEC62, FAM134A, FAM134C, CDK5RAP3, ATL3, SPART, YIPF3, YIPF4, GOLPH3, and LMNB1. Therefore, in one embodiment, the peptide mix comprises or consists of peptides derived from SQSTM1, NBR1, Optineurin, CALCOCO2, TAX1BP1 , Alfy, CCPG1, FAM134B, RTN3, NIX / BNIP3L, TOLLIP, TNIP1 , FKBP8, BNIP3, PHB2, NCOA4, NUFIP1, STBD1 , BNIP1, PHBP2.PNPLA2, BCL2L13, TEX264, SEC62, FAM134A, FAM134C, CDK5RAP3, ATL3, SPART, YIPF3, YIPF4, GOLPH3, and / or LMNB1. According to the invention, the peptides are derived from source proteins by in silica proteolytic digestion using a suitable endoprotease such as trypsin. Not all peptides from a given source protein are suited for targeted MS as some sequences might not fit the requirements of LC (highly hydrophilic or hydrophobic peptides), chemical stability (e.g. post-lysis methionine oxidation), or distribution of cleavage sites (lysine or arginine positions in the sequence prevent digestion into suitable peptides yielding either too large or too small peptides). Selected peptides need also be screened for their ability to be detected in MS. In particular, the peptides have to be able to be ionized and be stable after ionization. Furthermore, it must be possible to retain the peptide on columns commonly used with MS, e.g., LC C18 column coupled to the mass spectrometer. Lastly, it is important that MS signal generated by the peptides has a broad linear concentration range.

[0044] The peptides to be used in the peptide mix are then synthesized by methods known in the art, e.g., solid-phase synthesis. In a preferred embodiment, the peptide mix comprises at least 7 peptides selected from the group consisting of peptides having SEQ ID NO: 1 to 213.

[0045] Table 1: Validated peptides and their analytical figures of merit.

[0046] In other preferred embodiments, the peptide mix comprises at least 10, 20, 30, 40, 50, 60, 70, 80, 90, 200, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210 or all of the peptides having SEQ ID NO: 1 to 213. In another preferred embodiment, the peptide mix comprises at least one peptide derived from a protein involved in autophagy regulation, at least one peptide derived from a protein involved in autophagosome initiation, at least one peptide derived from a protein involved in membrane nucleation, at least one peptide derived from a protein involved in membrane expansion, at least one peptide derived from a protein involved in pore closure, at least one peptide derived from a protein involved in fusion with lysosomes and at least one peptide derived from a selective autophagy receptors (SAR) protein. Thus, the peptide mix comprises at least 7 peptides, preferably selected from the group consisting of peptides having SEQ ID NO: 1 to 213.

[0047] In another embodiment, the peptide mix comprises at least two peptides derived from a protein involved in autophagy regulation, at least two peptides derived from a protein involved in autophagosome initiation, at least two peptides derived from a protein involved in membrane nucleation, at least two peptides derived from a protein involved in membrane expansion, at least two peptides derived from a protein involved in pore closure, at least two peptides derived from a protein involved in fusion with lysosomes and at least two peptides derived from a selective autophagy receptors (SAR) protein. In one embodiment, the two peptides in each class are derived from the same protein. In another embodiment, the two peptides are derived from different proteins of that class. The peptides are preferably selected from the group consisting of peptides having SEQ ID NO: 1 to 213.

[0048] In one embodiment, the peptide mix consists of one peptide derived from a protein involved in autophagy regulation, one peptide derived from a protein involved in autophagosome initiation, one peptide derived from a protein involved in membrane nucleation, one peptide derived from a protein involved in membrane expansion, one peptide derived from a protein involved in pore closure, one peptide derived from a protein involved in fusion with lysosomes and one peptide derived from a selective autophagy receptors (SAR) protein. The peptides are preferably selected from the group consisting of peptides having SEQ ID NO: 1 to 213.

[0049] The peptides are preferably selected from the group consisting of peptides having SEQ ID NO: 1 to 213.

[0050] In one aspect of the invention, the peptide mix according to the invention is a peptide standard for mass spectrometry (MS). Preferably, the MS is liquid chromatography tandem-mass spectrometry (LC-MS / MS). When the peptide mix according to the invention is used as MS standard, it is spiked into, i.e., added at a suitable amount (e.g., 0.01-250 fmol) to a sample of interest, e.g., a tryptic digest of a lysed tissue or cell sample. The peptide mix is used to determine the LC retention times and MS signals of respective endogenous peptides. The extracted ion currents of the peptides of the peptide mix are used to generate mass chromatograms for quantification of peptides of the peptide mix and the endogenous peptides. For quantification either the area under the curve or the intensity maximum of respective mass chromatograms, series of mass chromatograms, or mass spectra are used.

[0051] In a further embodiment, the methods according to the invention comprise a step of extracting the proteins from a biological sample via lysis, a step of digesting the tissue or cell lysate with trypsin or another suitable endopeptidase / endoprotease, a step of purifying the obtained peptides, a step of spiking the obtained peptide sample with the peptide mix (e.g., 0.01-250 fmol) according to the invention, and / or a step of introducing the spiked peptide sample into a mass spectrometer.

[0052] Examples

[0053] Example 1

[0054] Cell culture

[0055] A549 lung carcinoma cells were grown in Dulbecco’s Modified Eagle Medium (DMEM, PAN biotech) supplemented with 10% Fetal Bovine Serum (FBS, BioWest) and Penicilin Streptamycin (PAN Biotech). 2.2x106cells were seeded into 10 cm dishes to reach about 80% confluency on the next day. Treatments were performed by washing the plates twice with PBS before changing to the treatment medium for 2 h. Treatment media were Hank’s Balanced Salt Solution (HBSS, Thermo Fisher) for amino-acid starvation and DMEM glucose-free (PAN biotech) supplemented with 10% dialyzed FBS (BioWest), penicillin / streptomycin (PAN Biotech) and stable glutamine (GlutaMAX, Thermo Fisher). Concanamycin A (Sigma-Aldrich) was used at a final concentration of 10 pM.

[0056] Cells were washed twice with ice-cold PBS before harvesting by scrapping on ice. Cell pellets were snap frozen in liquid nitrogen.

[0057] MS samples preparation

[0058] Pellet were lysed in 5% sodium deoxycholate buffer in 50 mM ammonium bicarbonate pH 8.5. Samples were sonicated (Bioruptor, diagenode) at 8°C for 10 min alternating 30 s pulses on medium intensity and 30 s breaks. Protein concentration was measured and adjusted using BCA assay and samples were spiked with 10 or 100 fmol of heavy-labeled peptides. Samples were subsequently diluted to reduce sodium deoxycholate below 1 %. Samples were reduced by adding 1 mM DTT and incubated at 37°C for 30 min, then alkylated with IAA for 15 min in the dark at room temperature for 15 min.

[0059] Samples were digested with trypsin to a 1 :100 trypsin (Promega) to protein ratio for 15 h at 37°C with constant agitation. Trypsin was inhibited and sodium deoxycholate was precipitated adding 50% TFA to a final concentration of 2%. Samples were desalted using AssayMAP C18 cartridges (Agilent) on a Bravo liquid handling platform (Agilent), and eluted in 50 pl 80% acetonitrile and 0.1% formic acid. Solvents were lyophilized to remove organic solvents. Peptides were resuspended in 0.1% formic acid to a 1 pg / pl.

[0060] Synthetic peptides

[0061] Heavy-labelled synthetic standards for all peptides were acquired from SpikeTides (J PT) or custom synthesis (GenScript) with the following chemical modifications: Carbamoylmethylated cysteine, Carboxy-terminal13Ce-15N2 Lysines,13Ce-15N4 Arginines. Peptides were not purified but isotope label purity was high. Crude peptides were used and the purity of peptides was not taken into account for calculating concentrations.

[0062] LC-MS

[0063] LC-MS / MS measurements were performed on an EASY-nLC 1000 nano-flow LIHPLC system (Thermo Fisher Scientific) coupled to a Q Exactive HF-X hybrid quadrupole- Orbitrap mass spectrometer (Thermo Fisher Scientific). 5 pl of solubilized peptides in solvent A (0.1% formic acid in water) were separated on a fused silica HPLC column (75 pm internal diameter column Fused-silica PicoTip® emitter: SilicaTip™, New Objective) self-packed with Waters Acquity CSH C18, 1.7 pm (Waters) to a length of 20 cm) using a linear gradient of solvent A and solvent B (0.1% formic acid in 80% acetonitrile in water) from 4% solvent B to 30% over 85 min, followed by an increase to 100% buffer B over 8 min and 7 min at 100% buffer B at a 250 nl / min flow rate. The spray voltage was set to 2.3 kV with a capillary temperature of 250°C.

[0064] Mass spectrometer was operated in PRM mode at 60’000 resolution with an AGC target set as 1e6 and a maximum injection time of 118 ms. Isolation window was set at 1.5 m / z, normalized collision energy was set at 27. Data was acquired in centroid mode. Targeted precursors were monitored during 6 min windows in the calibration curve experiments and 3.5 min windows in the later experiments.

[0065] Spectral libraries were acquired using a mixture of the synthetic peptides of reference in Data Dependent acquisition with the same resolution and collision energies.

[0066] Data Analysis

[0067] Raw data was analyzed using Skyline (27, 28). MS2 spectra were matched to libraries generated using the synthetic peptides. Extracted ion chromatograms were manually inspected, signal from interfering ions were removed and integration peak boundaries were modified if needed. Precursors with less than 3 valid transitions were excluded from further considerations. To normalize the intensity across runs, two additional peptides from the background matrix INVYYNEATGGK, QSVENDIHGLR from Tubulin beta chain 4b and Type I cytoskeletal keratin 18, respectively, were monitored for the calibration curve. For experiments, precursor intensities were normalized on their respective heavy- labelled peptides.

[0068] Further data analysis was performed using in-house R code. Protein-level intensities were obtained by geometrically averaging peptide data.

[0069] Example 2

[0070] Target selection

[0071] In order to cover bulk and selective autophagy subtypes, respective literature was screened for relevant proteins (H. Yamamoto, S. Zhang, N. Mizushima, Autophagy genes in biology and disease. Nat Rev Genet 10.1038 / s41576-022-00562 -w, 1-19 (2023). T. Lamark, T. Johansen, Mechanisms of Selective Autophagy. Annu Rev Cell Dev Biol 37, 143-169 (2021). M. Bordi, R. De Cegli, B. Testa, R. A. Nixon, A. Ballabio, F. Cecconi, A gene toolbox for monitoring autophagy transcription. Cell Death Dis 12, 1044 (2021). D. Siva Sankar, J. Dengjel, Protein complexes and neighborhoods driving autophagy. Autophagy 17, 2689-2705 (2021)). In addition, selected, autophagy-relevant proteins from ongoing research projects were included (J. Zhou, N. L. Rasmussen, H. L. Olsvik, V. Akimov, Z. Hu, G. Evjen, S. Kaeser-Pebernard, D. S. Sankar, C. Roubaty, P. Verlhac, N. van de Beck, F. Reggiori, Y. P. Abudu, B. Blagoev, T. Lamark, T. Johansen, J. Dengjel, TBK1 phosphorylation activates LI R-dependent degradation of the inflammation repressor TNIP1. J Cell Biol 222 (2023)). The chosen targets include proteins relevant to autophagy regulation (ribosomal S6 Kinase and TFEB), the initiation complex (LILK1, RB1CC1 / FIP200, ATG13), autophagosome biogenesis (ATG2A, ATG9A, WIPI2, BNIP1, WBP2), the ATG8 lipidation machinery (ATG3, ATG7, ATG12, ATG5, ATG4A) and mammalian ATG8s (GABARAPL1 , GABARAPL2), as well as SARs (SQSTM1 , NBR1 , Optineurin, NDP52, TAX1 BP1, Alfy / WDFY3, CCPG1 , FAM234B, RTN3, NIX, TOLLIP, FKBP8, BNIP3, AMBRA1, PHB2, NCOA4, STBD1, ATGL) (Table 1).

[0072] Table 2: Target proteins monitored by PRM

[0073]

[0074] Peptides of proteins-of-interest for targeted proteomics are commonly generated by proteolytic digestion using the endoprotease trypsin. Not all tryptic peptides from a given source protein are suited for targeted MS as some sequences might not fit the requirements of LC (highly hydrophilic or hydrophobic peptides), chemical stability (e.g. post-lysis methionine oxidation), or distribution of trypsin cleavage sites (lysine or arginine positions in the sequence prevent digestion into suitable peptides yielding either too large or too small peptides). For these reasons, only peptide sequences between 7 and 25 amino-acids in lengths were considered. Due to the sensitivity to oxidations, methionine- containing sequences were also excluded. Unfortunately, these restrictions led to the exclusion of some important protein candidates. As an example, human ATG8 proteins are very short proteins that only generate a limited number of tryptic peptides. Among this small set, some peptides needed to be removed as their sequences did not match our quality criteria. As a result, only three peptides from GABARAPL1 and GABARAPL2 were further considered and all LC3s (MAP1LC3A, B, B2 C) did not make it to the final list. As a next step, shortlisted peptides were screened on their detection in previous MS experiments by our group as well as in publicly available MS datasets from PeptideAtlas (17). This led to a selection of 114 peptides sequences from 41 source proteins (Table 1). These were synthesized following the AQUA principle using isotopically labelled arginine or lysine variants (18), supporting their use as synthetic spike-ins in order to aid identifications of endogenous peptides. To assess the usability of these peptides, we first determined their linear concentration range, i.e. the range in which changes in quantities generate a proportional change in the recorded MS signal.

[0075] Peptide characterization

[0076] As with all analytical techniques, quantitative accuracy can only be achieved within a certain analyte concentration range. Quantitative accuracy of targeted proteomics assays is achieved by determining the linear range of detection, the limit of detection and the limit of quantification of respective analytes, i.e. the analyte concentrations at which the method generates a signal clearly separated from the noise and a reproducible signal, respectively. These values are peptide-specific and depend in part on the way peptides ionize; hence, they have to be determined experimentally for all of the target analytes by measuring calibration curves using concentration gradients.

[0077] The calibration curves for the targeted peptides were measured in five replicates across 11 different concentrations from 3-fold serial dilutions (0.15 fmol to 10'000 fmol on column) and five blank samples. Known amounts of heavy-labeled peptides were spiked into the same tryptic whole-cell-digest from an A549 human lung carcinoma cell line. These spiked digests were then separated into five different samples. Peptide desalting and MS sample preparation was performed separately on each of these samples to reproduce the variation introduced by these steps.

[0078] Only peptide precursors for which a minimum of three fragment ions were detected were used for further analyses. As expected, some peptides reached their upper limits of quantification at high concentrations resulting in a loss of linearity of the measured signal or exhibited poor chromatographic results such as tailing. These datapoints were manually removed from further analysis and the last concentration in the linear range was considered the upper limit of quantification for these targets. The remaining points were used to fit a calibration curve using the MSStats LOB / LOD package. Briefly, this method enables to fit a nonlinear regression model to the intensity-concentration response curve. Limits of quantification (LOQ) and detection (LOD) can then be derived from this model. LODs were defined as the concentrations corresponding to the intersection of the upper bound of the intensity prediction interval for the blank samples (corresponding to the noise level) and the mean predicted intensity of spiked amounts. LOQs were set as the concentrations corresponding to the intersection between the upper bound of the predicted noise intensity and the lower bound of the confidence interval of predicted intensities for given concentrations (see Fig. 1 as example). Out of the 1413 considered peptides, 1200 could either not be reproducibly detected or accurately quantified and were removed from further considerations, leading to a final list of 213 peptides with quantified LODs and LOQs (Table 2).

[0079] Robustness of the assay in a biological context

[0080] To assess the reproducibility of our approach, a first setup experiment was performed comparing samples under normal growth conditions (DMEM) to samples treated for 2 h by Concanamycin A (DMEM + ConA), which inhibits the V-ATPase and thereby, lysosomal hydrolases (Figure 3). This assay is commonly used to measure autophagy fluxes as proteins which are degraded in lysosomes will accumulate under ConA treatment. The ratio of (DMEM + ConA) / (DMEM) reflects autophagy-dependent protein degradation, i.e. the autophagy flux. Biological triplicates of each condition were prepared and the amount of spike-in peptide was adjusted for each peptide based on the calibration curves, to ensure a stable signal from all standards reflecting respective peptide concentrations.

[0081] From each biological replicate, three technical replicates were performed to assess technical variation. Peak intensities of light peptides were normalized to their respective heavy-labeled standards. Technical coefficients of variations (CVs) were calculated for each technical triplicate and averaged across all biological replicates. Technical CVs ranged from 0.98% to 12.9% (Figure 4a) depending on the peptides. Following treatment with ConA, the lysosomal hydrolases are inhibited resulting in the accumulation of material which was targeted for recycling through autophagy. This effect could be clearly observed for a variety of well characterized autophagy-related proteins, such as GABARAPL2, SQSTM1, TAX1BP1 and CALCOCO2 but was also noticed for many other ATG proteins for which this effect was not previously described (Fig. 2b, 2c). Thus, proteins involved in autophagy initiation like the LILK1 complex members LILK1 , FIP200 and ATG13, as well as proteins involved in ATG8 lipidation like ATG3, ATG4, and ATG5 were also stabilized by ConA, indicating that also autophagy regulators are turned over by lysosomal degradation under growth conditions.

[0082] Assessment of stimulus-dependent protein degradation by autophagy

[0083] Having established the technical principles for our assay, it was next determined if it was possible to identify differences in protein regulation of two classical and well described autophagy inducing stimuli, i.e. amino acid and glucose starvation. Both treatments were performed for a duration of 2 h and coupled to the inhibition of lysosomal degradation with ConA to study autophagy flux.

[0084] Peptides derived from 38 proteins were detected and accurately quantified. To identify potentially different responses, the results were filtered for proteins showing a significant difference in abundance between the respective treatments (p-value < 0.01 in a one-way ANOVA) yielding a total of 25 differentially regulated proteins. Upon clustering of the results three clear clusters formed (Fig. 3). The top cluster comprised mostly the well- described p62-like autophagy receptors, i.e. ubiquitin-dependent soluble SARs, showing a clear flux under all of the treatments, also under basal conditions. This cluster also contained the ATG8-homolog GABARAPL2, a member of the core autophagy machinery functioning as SAR docking site in the autophagosomal membrane. These data imply that under all tested conditions selective targeting of cargo proteins to autophagosomes takes place. As a mammalian ATG8 and defining component of autophagosomes, GABARAPL2 is constitutively turned over, suggesting that receptors following the same pattern are also turned over in a constitutive way. While most of the other proteins assigned to this cluster are known to bind a wide range of ubiquitylated substrates, NCOA4 a selective ferritinophagy receptor was also assigned to this cluster. This could suggest a basal turnover of this receptor probably in an unbound state in conditions in which iron is not limiting, and turnover of the ligand-bound receptor when ferritinophagy is required by the cell.

[0085] The second cluster comprises known ubiquitin-independent SARs, such as the mitophagy receptors BNIP3 and NIX, and the reticulophagy receptors CCPG1 and FAM134B. While the overall abundance of these receptors does seem to change upon different treatments, it does not seem to be regulated by autophagy as inhibition of lysosomal degradation does not generate significant changes in their abundances.

[0086] The last cluster mostly comprises core proteins involved in autophagy regulation, such as proteins involved in initiation of autophagy, autophagosome biogenesis, and ATG8 lipidation. The proteins assigned to this cluster all show a lower abundance under basal conditions, which seems to be caused by autophagy as this amount increases upon ConA treatment, similarly as observed in the setup experiment (Fig. 2). Upon starvation, ConA treatment ceases to significantly impact the abundance of these proteins. This implies that upon stress, these proteins are spared from degradation to ensure a proper autophagy response.

[0087] Example 3

[0088] Robustness of the assay compared to standard MS workflows and in a biological context

[0089] Amino acid starvation is a strong and commonly used autophagy inducing stimulus. Coupling it to Bafilomycin A1 (BafA1) treatment, which inhibits V-type proton-ATPases blocking lysosomal degradation (5), enables to measure the accumulation of autophagosomal cargo that would be degraded in the absence of BafA1 treatment. To assess the reproducibility of our approach and compare it to standard shotgun proteomics approaches, A549 cells were starved for amino acids in Hank's Balanced Salt Solution (HBSS) with and without BafA1 for 2 h, samples processed as outlined above, and resulting peptide mixtures were measured on the same LC-MS / MS system comparing PRM-based targeted proteomics to Data-Dependent Acquisition (DDA)-, and Data- independent Acquisition (DIA)-based discovery proteomics (n=3 technical replicates per treatment).

[0090] Out of the 37 protein targets of the assay, 31 could be identified and quantified with a signal above the LOQ in a minimum of one sample using PRM. Standard DIA measurements identified 17 of these proteins and DDA identified 9 (Figure 4a). When it comes to quantification, DDA measurements resulted in large coefficients of variation (CVs). Normalized DDA measurements using the MaxLFQ algorithm (21) performed surprisingly well with an average CV of 8.5%; however, only 7 proteins were retained and depending on the protein the CV might be significantly higher (max. value of 44%). Not surprisingly, targeted proteomics developed with specific targets in mind outperformed the other approaches when measuring those targets. In terms of reproducibility, PRM measurements produced an average CV across conditions of 4.8% with most proteins being below 10% (Figure 4b).

[0091] We then tested the reproducibility of the PRM assay on a larger sample number, using cells harvested in growth conditions (DMEM) in the absence and presence of BafA1 for 2 h. Biological triplicates were prepared for these conditions and all of them were measured 3 times to clearly separate the noise originating from biological differences and sample preparation from the purely technical MS-related noise. For these quantifications, 60 peptides which were identified without interferences in all samples and which gave rise to a signal above the LOQ were considered for CV calculations. CVs calculated for all technical triplicates and averaged across samples ranged from 0.5% to 14.8% with a majority below 5% showing high technical reproducibility of the assay with variations lower or in the range of the variation observed across biological replicates (Figure 4c-f).

[0092] Example 4

[0093] Assessment of stimulus-dependent protein degradation by autophagy

[0094] Having established the technical principles for our assay, we next asked if we could identify differences in autophagy-dependent protein regulation comparing 2 h amino acid starvation and 2 h glucose starvation to growth conditions in complete media (DMEM). These treatments were coupled to BafA1 treatments for 2 h to study the autophagy flux in all three conditions.

[0095] Some conditions like amino-acid starvation (HBSS) considerably reduced protein levels of autophagy cargoes making it difficult to detect and / or quantify these peptides. To calculate protein quantities, we used the following strategy: if a set of peptides was identified in all samples and resulted in a signal above the LOQ, only this set was used for inferring the protein amount. If no valid peptide was found across the whole experiment the protein was considered missing and not quantified. Lastly, if a protein was identified with a given valid peptide set in some samples but none of these peptides were valid in other samples, the protein abundance was imputed with the same standard deviation as the one observed for measured samples and a downshifted mean of 20%. We regard this as a rather conservative estimation of protein abundance in the case of missing values and it can easily be adjusted based on the biological setting of the experiment. This strategy led to the quantification of 34 proteins with peptides above the LOQ.

[0096] To identify potential patterns of differential regulation of these proteins across the tested conditions, we filtered the results for proteins showing a significant difference in abundance between treatments (p-value < 0.01 in one-way ANOVA). This filtering yielded a total of 24 proteins. Upon clustering of the results, experimental conditions could be clearly separated, and three major clusters emerged (Figure 5a). Cluster 1 is mostly comprised of components of the core autophagy machinery (Figure 5a-b). These proteins do not show degradation patterns in any of the tested conditions but seem to accumulate under amino-acid starvation. The levels of these proteins consistently and significantly increase by about 20% once cells are starved for amino acids, regardless of lysosomal activity (Figure 5b, p<0.05, T test).

[0097] Cluster 2 constitutes of ubiquitin-dependent, soluble SARs and the ATG8 homolog GABARAPL2 (Figure 5a, c). SARs like p62 / SQSTM 1 are being turned over to some extent under basal conditions and amino-acid starvation drastically increases their flux, corroborating their central role in adaptation to amino acid starvation.

[0098] In cluster 3 membrane-bound, ubiquitin-independent autophagy receptors BNIP3, NIX / BNIP3L, PHB2 and CCPG1 are not being degraded within lysosomes under any conditions but interestingly seem to be at their most abundant levels under growth conditions (Figure 4a, d). The mechanism leading to their decrease under stress conditions is not known. For some proteins, the trends observed by targeted proteomics were also observed by immunoblotting, although the noise levels were higher (Figure 4e- f). Interestingly, glucose starvation decreases the overall level of SARs compared to growth conditions. This decrease is independent of lysosomal activity as BafA1 treatment has no effect (Figure 4c). The block of turnover is also reflected in the poor clustering of the samples starved for glucose with and without BafA1 , as this drug should underscore the autophagy-related cargo degradation. The small but significant stabilization of GABARAPL2 by BafA1 indicates that lysosomes are principally still active, questioning the role of the monitored proteins in autophagy-dependent adaptation to glucose starvation.

[0099] Example 5

[0100] Determination of protein regulation during mitophagy

[0101] After demonstrating that our method detects differences in protein regulation within 2 h of amino acid starvation, we wanted to assess whether we could monitor the flux of specific SARs after inducing selective autophagy. We used ARPE-19 cells, a retinal pigment epithelia cell line, treated with the iron chelator DFP to induce selective BNIP3 / BNIP3L- dependent and PINK1-Parkin-independent mitophagy (22). These cells expressed the mito-QC reporter, a FIS1 mCherry-GFP fusion protein enabling to monitor mitochondrial turnover in the lysosome using fluorescence-based approaches (22). Briefly, this fusion protein emits both red and green fluorescence signals until it is brought to the lysosome, where the acidic conditions quench GFP resulting in a red-only signal that can be detected by flow cytometry or microscopy. We treated cells with DFP for 24 h with and without concanamycin A (ConA), another V-type H+-ATPase inhibitor used to monitor accumulation of autophagosomal cargoes (5). After 24 h of DFP treatment, we observed a significant increase in red mito-lysosomes compared to non-treated control cells by IF microscopy (Figure 6a-b). GFP-quenching in lysosomes could be blocked by ConA. Also, flow cytometry analyses confirmed these results (Figure 6c).

[0102] Having established that the mito-QC reporter cells behave as anticipated, we performed replicate experiments and subjected them to targeted proteomics as outlined above. Protein abundance data were again clustered to gain insights into differential protein regulation. Soluble SARs like p62 / SQSTM 1 , TAX1BP1 , NBR1 and CALCOCO2 show a clear autophagy-dependent turnover under growth conditions and increased turnover after DFP treatment (Figure 6d-e). Interestingly, OPTN which is linked to PINK1-Parkin- dependent mitophagy, did also slightly respond to DFP treatment (23, 24). These trends were also observed by immunoblotting, although the data was noisier (Figure 6f-g).

[0103] In contrast to the starvation experiments summarized in Figure 5, DFP treatment led also to a turnover of membrane-bound SARs: the mitophagy receptors BNIP3 (25) and BNIP3L / NIX, which is described as a major mitophagy receptor involved in DFP- dependent mitophagy (26), showed low turnover under growth condition but strongly accumulated in DFP treatment and blocked lysosomal degradation (Figure 6d, h). AMBRA1, FKBP8 and PHB2 which have also all been linked to mitophagy did not respond to the outlined treatments (27-30). Surprisingly, other receptors also seemed to strongly respond to DFP treatment, like TOLLIP a ubiquitin receptor (31), NCOA4 a ferritinophagy receptor (32), and CCPG1 and FAM134B, two ER-phagy receptors (33, 34). The parallel accumulation of FAM134B and CCPG1 suggests an increased reticulophagy / ER-phagy upon DFP treatment which, to the best of our knowledge has not been described yet, despite DFP being a commonly used iron chelator. This emphasizes the need to comprehensively monitor autophagy receptors to fully grasp cellular responses to stresses.

Claims

Claims1. A method for monitoring autophagy selectivity, the method comprising probing a sample via mass spectrometry for the amount of at least two different selective autophagy receptor (SAR) proteins using a peptide mix as peptide standard, wherein the peptide mix comprises peptides derived from the at least two different SAR proteins.

2. The method according to claim 1, wherein the at least two different SAR proteins are specific for at least two different autophagy types selected from aggrephagy, mitophagy, pexophagy, lysophagy, zymophagy, ER- / reticulo-phagy, ferritinophagy, glycophagy, xenophagy, ribophagy, midbody autophagy, golgiphagy, nucleophagy and lipophagy.

3. The method according to claim 1 or 2, wherein the SAR proteins are selected from the group consisting of SQSTM1, NBR1, Optineurin, CALCOCO2, TAX1BP1 , Alfy, CCPG1, FAM134A, FAM134B, FAM134C, ATL3, BCL2L13, CDK5RAP3, GOLPH3, LMNB1 , SEC62, SPART, TEX264, YIPF3, YIPF4, RTN3, NIX / BNIP3L, TOLLIP, TNIP1, FKBP8, BNIP3, PHB2, NCOA4, NUFIP1 , STBD1, BNIP1 , PHBP2 and PNPLA2.

4. The method according to any of claims 1 to 3, wherein the peptide mix comprises one or more peptides having SEQ ID NO: 1, 2, 25, 26, 27, 28, 29, 30, 31 , 32, 33, 34, 35, 36, 37, 38, 39, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 62, 63, 64,65, 66, 67, 68, 69, 70, 71 , 72, 73, 78, 79, 80, 81, 82, 92, 93, 115, 116, 117, 118,119, 120, 121, 122, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140,148, 149, 150, 151 , 152, 153, 154, 168, 169, 174, 175, 176, 177, 190, 191 , 192,193, 194, 206, 207, 208.

5. The method according to any of claims 1 to 4, wherein the peptide mix further contains peptides derived from proteins that are involved in autophagy regulation, in autophagosome initiation, in membrane nucleation, in membrane expansion, in pore closure and / or in fusion of the autophagosome with lysosomes.

6. The method according to any of claims 1 to 5, wherein the proteins involved in autophagy regulation are selected from the group consisting of ribosomal S6 kinase (RPS6KB1) and TFEB.

7. The method according to any of claims 1 to 6, wherein the proteins involved in autophagosome initiation are selected from the group consisting of LILK1 , LILK2, RB1CC1 / FIP200 and ATG13.

8. The method according to any of claims 1 to 7, wherein the proteins involved in membrane nucleation are selected from the group consisting of ATG9A, AMBRA1 , and WBP2.

9. The method according to any of claims 1 to 8, wherein the proteins involved in membrane expansion are selected from the group consisting of ATG2A, WIPI1, WIPI2, WIPI3, WIPI4, ATG3, ATG7, ATG12, ATG5 and ATG4A.

10. The method according to any of claims 1 to 9, wherein the proteins involved in pore closure and fusion with lysosomes are GABARAPL1 or GABARAPL2.

11. The method according to any of claims 1 to 10, wherein the peptide mix comprises at least 7 peptides selected from the group consisting of peptides having SEQ ID NO: 1 to 213.

12. The method according to any of claims 1 to 11, wherein the peptide mix comprises at least one peptide derived from a protein involved in autophagy regulation, at least one peptide derived from a protein involved in autophagosome initiation, at least one peptide derived from a protein involved in membrane nucleation, at least one peptide derived from a protein involved in membrane expansion, at least one peptide derived from a protein involved in pore closure, at least one peptide derived from a protein involved in fusion with lysosomes and at least one peptide derived from a selective autophagy receptors (SAR) protein.

13. A mass spectrometry peptide standard comprising a peptide mix, wherein the peptide mix comprises peptides derived from at least two different SAR proteins, for use in a method according to any of claims 1 to 12.

14. The peptide standard according to claim 13, wherein the peptides are isotopically labelled heavy peptides.

15. The peptide standard according to claim 13 or 14, wherein the peptide mix comprises at least 7 peptides selected from the group consisting of peptides having SEQ ID NO: 1 to 213.

Images