N-Degradomic Analysis Reveals a Proteolytic Network Processing the Podocyte Cytoskeleton

BRIEF COMMUNICATION www.jasn.org

†‡ † | † Markus M. Rinschen,* § Ann-Kathrin Hoppe,* Florian Grahammer, ¶ Martin Kann,* † † † †‡ Linus A. Völker,* Eva-Maria Schurek,* Julie Binz,* Martin Höhne,* § Fatih Demir,** | †† † ‡‡ Milena Malisic,** Tobias B. Huber, ¶ Christine Kurschat,* Jayachandran N. Kizhakkedathu, †‡ †‡ Bernhard Schermer,* § Pitter F. Huesgen,** and Thomas Benzing* §

*Department II of Internal Medicine, †Center for Molecular Medicine Cologne (CMMC), ‡Cologne Excellence Cluster on Cellular Stress Responses in Aging-Associated Diseases (CECAD), and §Systems Biology of Ageing Cologne (Sybacol), BRIEF COMMUNICATION University of Cologne, Cologne, Germany; |Department of Medicine III, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; ¶Department of Medicine IV, Medical Center and Faculty of Medicine – University of Freiburg, Freiburg, Germany; **Central Institute for Engineering, Electronics and Analytics, ZEA-3, Forschungszentrum Jülich, Jülich, Germany; ††BIOSS Centre for Biological Signalling Studies and Center for Biological Systems Analysis (ZBSA), Albert-Ludwigs-University, Freiburg, Germany; and ‡‡Centre for Blood Research, Department of Pathology and Laboratory Medicine, Department of Chemistry, University of British Columbia, Vancouver, Canada

ABSTRACT Regulated intracellular proteostasis, controlled in part by proteolysis, is essential in of these genes leads to altered signaling maintaining the integrity of podocytes and the glomerular filtration barrier of the cascades in glomerular diseases and to kidney. We applied a novel proteomics technology that enables proteome-wide FSGS.2 Of critical importance are mem- identification, mapping, and quantification of protein N-termini to comprehensively brane proteins, such as podocin and characterize cleaved podocyte proteins in the glomerulus in vivo.Wefoundevi- nephrin3,4 and cytoskeletal proteins dence that defined proteolytic cleavage results in various proteoforms of important such as a-actinin-4 (ACTN4).5 podocyte proteins, including those of podocin, nephrin, neph1, a-actinin-4, and Proteolytic processing is an irrevers- vimentin. Quantitative mapping of N-termini demonstrated perturbation of protease ible, site-specificproteinmodification action during podocyte injury in vitro, including diminished proteolysis of a-actinin-4. that regulates essential physiologic Differentially regulated protease substrates comprised cytoskeletal proteins as well processes by generating protein isoforms as intermediate filaments. Determination of preferential protease motifs during (proteoforms) and controlling proteo- podocyte damage indicated activation of caspase proteases and inhibition of stasis.6 Deregulated proteolysis is a key arginine-specific proteases. Several proteolytic processes were clearly site-specific, driver for human diseases, and proteases were conserved across species, and could be confirmed by differential migration are promising targets for tailored thera- behavior of protein fragments in gel electrophoresis. Some of the proteolytic peutics.7 Recent studies indicated that changes discovered in vitro also occurred in two in vivo models of podocyte damage maintenance of the renal filtration (WT1 heterozygous knockout mice and puromycin aminonucleoside–treated rats). Thus, we provide direct and systems-level evidence that the slit diaphragm and Received October 21, 2016. Accepted May 8, podocyte cytoskeleton are regulated targets of proteolytic modification, which is 2017. altereduponpodocytedamage. P.F.H. and T.B. shared senior authorship.

J Am Soc Nephrol 28: 2867–2878, 2017. doi: https://doi.org/10.1681/ASN.2016101119 Published online ahead of print. Publication date available at www.jasn.org.

Correspondence: Dr. Markus M. Rinschen, University Hospital Cologne, Josef-Stelzmann-Str 26, 50931 Köln, 1 The majority of kidney diseases start in membrane, and podocytes. Podocytes Germany, Prof. Thomas Benzing, University Hospital the glomerulus due to the susceptibility are terminally differentiated, interdigi- Cologne, Kerpener Straße 62, 50937 Köln, Germany, of the renal filtration unit to metabolic, tating cells that engulf the glomerular or Dr. Pitter F. Huesgen, ZEA-3 Analytics, For- schungszentrum Jülich,Wilhelm-Johnen-Straße,52425 1 genetic, mechanic, and immunologic capillaries. Damage to these cells results Jülich, Germany. E-mail: [email protected], damage. The glomerular filtration unit in proteinuria.1,2 Genetic studies have [email protected] or [email protected] fi consists of three layers: fenestrated endo- identi ed essential podocyte proteins Copyright © 2017 by the American Society of thelial cells, the glomerular basement which maintain the barrier. Mutation Nephrology

J Am Soc Nephrol 28: 2867–2878, 2017 ISSN : 1046-6673/2810-2867 2867 BRIEF COMMUNICATION www.jasn.org barrier largely relies on proteases as key such as acetylation, or in vitro by Ala59 (Table 1, Figure 1D). The 34-kD regulators of podocyte function, includ- dimethylation (Figure 1A). Next, tryptic band was observed in both human and ing Cathepsin D, Cathepsin L, and ma- digest was performed and N-terminal mouse glomeruli (Figure 1E) and is trix metalloproteases.8–11 High copy peptides were enriched (see Concise therefore not the previously described numbers of intracellular proteases with Methods for details). Proteomic analysis short podocin isoform specific to hu- broad specificity are found in the podo- of two such enriched samples from mans (and not expressed in mice) cyte transcriptome and proteome, and mouse glomeruli identified a total of with a similar molecular size.20 Interest- they accumulate in disease states.8,9,12–14 3815 annotated N-termini, of which ingly, this podocin cleavage site is asso- Although essential proteases and a few 2976 were modified by dimethylation in ciated with a polymorphism in the of their targets have been identified, it vitro whereas 784 were endogenously human sequence with unknown patho- is largely unclear which particular po- acetylated (Figure 1B, Supplemental genicity21 (Figure 1F). Three of the docyte proteins are cleaved, which pro- Table 1; 55 were both acetylated and di- cleavage sites observed in nephrin and teoforms are generated at which site(s), methylated). We mapped these to their ACTN4 were located in close vicinity to and how proteolytic cleavage affects topology and classified them depending 12 described mutations leading to he- their physiologic function in health and on their positional annotation19: 1343 reditary nephrotic syndrome and/or disease. Distinction of processed and were “expected” termini, of which 720 FSGS (Figure 1F). precursorproteinsischallengingby started with intact or removed initiation Totest whether proteolytic processing standard proteomics because many methionine (P1, P2), six were termini is altered in podocyte damage, we chal- tryptic peptides are shared by both pro- mapping to known alternative transcrip- lenged cultured human podocytes with teoforms, whereas the informative pro- tion start sites, and 521 were termini puromycin aminonucleoside (PAN; tease-generated neo-terminal peptides matching to known or predicted sites 50 mg/ml, 24 hours), an artificial yet constitute only a very minor frac- for signal- or propetide removal (Figure well established model of podocyte in- tion that is rarely identified.15 Here, we 1B). Ninety-six termini were annotated jury,22,23 and performed quantitative overcome this limitation using the inno- as previously described cleavage sites TAILS analysis. Immunofluorescence vative Terminal Amine–based Iso- associated with known proteases. Ca- analysis demonstrated the expected typ- tope Labeling of Substrates (TAILS) thepsin D cleavage sites were markedly ical rearrangement of actin stress fibers technique, a strategy enabling site- overrepresented in the dataset as com- but also partial rearrangement of the in- specific mapping of protease-generated pared with all other known cleavage termediate filament vimentin (Supple- novel protein termini (“degrado- sites (Figure 1C). Two thousand four mental Figure 2A).24 Podocyte viability mics”).16–18 The aim of this first unbi- hundred forty-eight termini mapped as measured by 2,3-Bis-(2-Methoxy-4- ased, systems-level study of podocyte to “unexpected” positions, indicative of Nitro-5-Sulfophenyl)-2H-Tetrazolium- degradomics was to delineate proteolytic extensive proteolysis in podocytes in vivo. 5-Carboxanilide (XTT) assay was not protein modifications and demonstrate The vast majority (91%) of unexpected altered by PAN at this time point (Sup- dynamic proteolytic regulation of cyto- termini were localized outside known plemental Figure 2B). The PAN-treated skeletal proteins in injury. protein domains. and vehicle-treated proteomes were iso- To obtain an atlas of proteolytically We found novel termini in the slit di- lated and labeled with two distinct stable cleaved proteins in podocytes of living aphragm proteins podocin (Figure 1D), isotope variants of formaldehyde to allow animals we applied the TAILS technology nephrin, and neph1 (Kirrel) (Supple- quantification of protease-generated combined with mass spectrometry– mental Figure 1A), and sites in podocyte neo-termini (Supplemental Figure 2C, based protein identification. Technical cytoskeleton proteins (Supplemental n=4). After TAILS enrichment we precautions were taken to minimize ex Figure 1B). Immunoblot analysis with identified a total of 3085 N-termini (Fig- vivo proteolysis by harvesting freshly iso- two antibodies targeting two different ure 2A); 1665 termini were expected lated mouse glomeruli with a compre- epitopes validated the presence of a po- (P1,2), 235 of the termini resulted from hensive protease inhibitor cocktail and docin fragment with approximately 34 signal- or propeptide removal, and 1123 performing protein isolation rapidly at kD, in both glomerular preparations were unexpected (Figure 2A), the major- cold temperatures (see Concise Meth- and lysates of fresh, snap-frozen mouse ity occurring outside of protein do- ods). In the first steps of the TAILS pro- kidney (ex-vivo time ,10 seconds) (Fig- mains. Forty-five termini occurred cedure, proteins were denatured and ure 1D). The presence of podocin at the precisely at known cleavage sites which primary amines (e-amines of Lys side molecular mass of 34 kD was also con- are predominantly processed by granzymes chains; a-amines of native, unmodified firmed using in-gel digestion and mass and cathepsins (Supplemental Figure amino [N-] termini, and protease- spectrometry (Supplemental Figure 1C). 3A). Compared with homologous generated neo-N-termini) modified This is consistent with a stable podocin proteins observed in the glomeruli data- by reductive dimethylation18 (Figure proteoform comprising only amino set, most cleavages occurred at con- 1A). Hence, all N-termini were either acids 59–385 as indicated by the observed, identical protein residues or modified by endogenous modifications served neo-N-terminus starting with closely adjacent to them, highlighting

2868 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2867–2878, 2017 www.jasn.org BRIEF COMMUNICATION

the highly regulated, site-specificnature of the proteolytic events (Supplemental Figure 3B, Table 1). We could reproducibly quantify 666 N-termini by the log2 normalized ratios of PAN-treated versus control podocytes, as determined from the stable isotope labels. We ranked them according to the ratio and divided them into ten percentiles (Figure 2B). Gene-ontology enrichment analysis of the most strongly increased(Q10)andmoststrongly decreased (Q1) N-termini showed decreased generation of N-termini on proteins associated with cytoskeletal organization (Figure 2C), whereas N-termini of proteins with structural activity and general cellular functions were increased upon PAN treatment. To test whether different classes of the cytoskeletal proteins are differentially regulated, we visualized the ratios of neo-termini in different classes of cytoskeletal proteins in cumulative histograms (Figure 2D, Supplemental Figure 4). Surprisingly, unexpected termini of actin-binding proteins were significantly decreased as compared with all other termini (Figure 2D). Also, neo-termini of tubulin and adhesion proteins were decreased, whereas cleavage of mitochondrial proteins was increased (Supplemental Fig- fi Figure 1. N degradomics identi es processed proteins in mouse glomeruli. (A) Scheme of the ure 4). For ACTN4, an actin-binding fl fi TAILS work ow (see main text and Concise Methods for details). (B) Classi cation of protein protein, we observed a decrease of cleav- N-termini identified by TAILS experiments (see Concise Methods for details). Expected termini age between C-terminal EF domains and are distinguished in termini with intact (P1) or removed (P2) initiating Met mapping to the start of thelongestknownisoform,terministarting at annotated signal peptide cleavage sites, previously the Spectrin domains (Figure 2E). identified protease-generated N-termini, and termini explained by alternative splicing sites. Using a C-terminal polyclonal antibody, Unexpected N-termini are neither annotated nor observed previously and occurred either within we detected two bands corresponding to or outside annotated protein domains. Blue, dimethylated termini modified in vivo representing the apparent molecular mass of ACTN4 protease-generated neo-N-termini by proteolytic cleavage; black, acetylated termini modified proteoforms which were, consistent with endogenously; and red, termini found in both acetylated and dimethylated forms (stably en- our TAILS data, decreased by PAN treat- dogenous), are distinguished. (C) Known cleavage sites as annotated by the TOPFINDer al- ment (compared with the P0 form, Fig- gorithm. Cathepsins (CATE, CATD) were the most abundant cleavage sites in glomeruli. ure 2F). The large cytoskeletal protein Cleavage sites of CATD were statistically overrepresented in the dataset as compared with the MYH9 exhibited a ladder-like staining , previously published terminome (cyan color; Bonferroni corrected Fisher exact test, P 0.05). (D) indicative of the complex proteolytic pat- Immunoblot for podocin from isolated glomeruli (left) and snap-frozen mouse kidney using two tern with many termini observed for antibodies against two different epitopes. *Denotes a band at 34 kD. (E) Immunoblot of mouse and human glomerular proteins for podocin detects a conserved lower mol wt band (*). (F) this protein (Supplemental Figure 5, A Overview of novel cleavage sites on nephrin (NPHS1), podocin (NPHS2), and ACTN4 and their and B). vicinity to known point mutations causing hereditary nephrotic syndrome and/or FSGS. The To infer quantitative protease hierar- mutations are as follows: NPHS1: Leu96Val, Asp105Asn, Ala107Thr, Ala107Glu, Ala107Val47–49 chies from the data, we generated motif and Arg299Pro, Arg299Cys, and Asp310Asn48,50; NPHS2: A61V (a polymorphism with yet un- logos25 of the top (Q10) and bottom known pathogenicity21); ACTN4: Lys255Glu, Thr259Ile, Ser262Pro, and Ser262Phe.5,51 Muta- (Q1) percentiles to determine whether tions are depicted in red, and polymorphism in magenta, and the most adjacent neo-termini specific proteolytic activities are affected peptide is depicted in yellow. All detected cleavage sites are depicted as black arrows. Glom. by PAN challenge (Figure 3A). The algo- Kid., glomeruli whole kidney; IB, immunoblot. rithm generates position-weighted

J Am Soc Nephrol 28: 2867–2878, 2017 Podocyte Proteomics and Degradomics 2869 BRIEF COMMUNICATION www.jasn.org

Table 1. Selected Neo-termini of important podocyte proteins Homologous Sequence of Terminus MaxQuant Cleavage Site Gene Symbol Protein (Uniprot) Residue Terminus in Human Found in Glomerulusa Score Annotation Cultured Podocyteb Nphs2 Podocin (Q91X5) APAATATVVDVDEVR 103.88 A59 Unexpected N/A AATATVVDVDEVR 99.94 A61 Unexpected N/A ATATVVDVDEVR 78.96 A62 Unexpected N/A TATVVDVDEVR 94.77 T63 Unexpected N/A ATVVDVDEVR 150.04 A64 Unexpected N/A TVVDVDEVR 82.01 T65 Unexpected N/A Nphs1 Nephrin (Q9QZS7) IEACDLSDDAEYECQVGR 160.55 I112 Unexpected, N/A VMTVRPEDHGAR 55.04 V317 Extracellular N/A Kirrel Neph1 (Q80W68) TRFSQEPADQTVVAGQR 137.91 T53 Signal peptide Yes, T21 Actn4 Actinin-4 (P57780) TQIENIDEDFR 170.95 T69 Unexpected Yes, T70 Podxl Podocalyxin (Q9R0M4) HVAPILDNQAVAVKR 77.78 H343 Unexpected N/A VAPILDNQAVAVKR 89.23 V344 Unexpected N/A Myh9 Myosin IIa (Q8VDD5) TQISELQEDLESER 93.42 T1111 Unexpected Yes, S1111 TLDSTAAQQELR 71.59 T1151 Unexpected Yes, T1151 TQELLQEENR 66.02 T1313 Unexpected Yes, T1313 Arhgdia Arhgdia (Q99PT1) SIQEIQELDKDDESLR 54.00 S34 Unexpected Yes, S34 Itgb1 Integrin beta-1 (P09055) TDEVNSEDMDAYCR 124.25 T504 extracellular Yes, T504 Calm Calmodulin (P62204) DKDGDGTITTKELGTVMR 82.01 D21 Cathepsin D/E Yes, F20c cleavage aAll N-termini are dimethyl-labeled. For more mass spectrometry details, please see Supplemental Material or deposited raw data (see Concise Methods for accessions). bHomologous, conserved cleavage sites were determined by the CPHOS software.44 cIn the human podocytes, this terminus occurs one amino acid N-terminal of the homologous amino acid. matrices of overrepresented amino acids in native and cultured podocyte pro- (e.g., complement) and the extracellular surrounding aligned cleavage sites infer- teins and transcripts.12 Degradomics matrix compartment (Figure 4A). How- red from the neo-termini. In Q1 (de- demonstrated that vimentin was cleaved ever, we also found that identical, creased N-termini), an R-T cleavage at D85 upon PAN treatment (Figure 3E). conserved protease cleavage sites on in- motif was overrepresented as compared Consistently, a C-terminal, but not an tracellular proteins are regulated in both with the nonchanged N-termini popula- N-terminal antibody identified a band in vivo and in vitro PAN damage models. tion, indicating that Arg-specific prote- corresponding to the molecular mass For instance, the proteolytic cleavage of ases such as trypsin-like serine proteases of 42 kD (P1, Figure 3F), which was re- vimentin was detectable in all three spe- might be inhibited (Figure 3A). This class producibly increased by PAN treat- cies and reduced in both in vitro and in is expressed in the podocyte transcrip- ment. Conversely, an N-terminal, but vivo PAN models (Figure 4B). This could tome but poorly studied.12 In the not a C-terminal antibody detected a be corroborated by immunoblotting: strongly induced quantile Q10 we presumably N-terminal protein frag- proteolytically processed vimentin observed an overrepresentation of an ment decreased by PAN treatment (P2, (stained by the N-terminal antibody) aspartate-rich A/D-X-X-D motif preced- Figure 3F). was reduced using an N-terminal anti- ing the cleavage site (Figure 3B). This is Finally, we asked whether the proteo- body in PAN rats, similar to cell culture the typical motif of the caspases. Consis- lytic changes defined by degradomic (Figure 4C). We also found that ACTN4 tently with this motif analysis, caspase 3/6 analysis in vitro could be also found in proteolytic processing was decreased, activity was transiently induced by PAN vivo. To this end, we analyzed proteolytic similar to the effect in human cells (Figure 3C) in agreement with previous cleavage in an in vivo model of rat PAN (Figure 4D). To further strengthen the results.23,26 Interestingly, protease net- injury by initial degradomics studies. We data, we analyzed a nonchemical model work analysis27 of observed cleavage identified 1334 N-termini, and quanti- of FSGS: in isolated glomeruli from sites revealed known downstream fied 552 termini in at least two biologic proteinuric WT1 heterozygous mice, a targets of caspase 3 including further replicates. We observed a large number known FSGS model,28 we found decrease proteases, but also vimentin and mito- of proteolytic cleavage products (di- of ACTN4 cleavage compared with chondrial proteins (Figure 3D, Supple- methylated peptides) belong to proteins wildtype littermates (Figure 4E). Im- mental Figure 4D). Vimentin, a key annotated as “extracellular” or “secreted.” munoblotting for vimentin showed a structural component of intermediate This can be explained by action of “ladder-like” staining in immunoblot- filaments, is very abundantly expressed proteases on both the serum proteins ting of lysates from WT1 heterozygous

2870 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2867–2878, 2017 www.jasn.org BRIEF COMMUNICATION

mouse glomeruli (Supplemental Figure 6). These experiments demonstrate the relevance and transferability of the TAILS analyses to the in vivo situation. Conceptually, the method used here is the counterpart to studying kinase substrates and signaling via phospho- proteomics, a method that has provided considerable insights into both glomerular and tubular pathophysiology.29–33 On a cell biologic level, protease processing appears dominant over alternative splicing or transcription start sites (Figures 1A and 2A). It is intriguing tospeculatethatsomeofthestable fragments might exhibit a cellular signaling function as described e.g., for the polycystin-1 C-terminal tail.34 The results of the quantitative “proof of con- cept” study are consistent with the current paradigm that cytoskeletal rearrangement is essential for podocyte injury.22,24 Yet, the ﬁndings could be partially explained by the fact that cytoskeletal proteins are extremely highly abundant in cultured podocytes and particularly amenable to mass spectrometry analysis.35,36 In conclusion, this study proves the existence of high- stoichiometric protein termini and spe- ciﬁc proteolytic machineries which mediate stress-induced processing of crucial components of podocyte structure and biology. These comprehensive results are also available now for further studies.

The color background denotes cutoffs for increased (magenta) and decreased (green) termini. In ACTN4 protein, five termini were identified by TAILS. Of these, two were significantly regulated. An antibody with a Figure 2. Quantitative mass spectrometry based proteomics reveals protein N-termini in C-terminal epitope (indicated) detected five cultured podocytes. (A) Positional classification of N-termini in cultured podocytes. (B) of the resulting proteoforms. All of the reg- Quantile analysis of quantified neo N-termini (n=666). The log2 ratio (PAN/vehicle) is ulated sites were conserved in mouse glo- plotted against the log2 rank. Termini reproducibly quantified in the first (decreased, first meruli. (F) ACTN4 proteoforms visualized by quantile, green) and last quantile (increased, 10th quantile, magenta) were further ana- immunoblotting as low–molecular mass lyzed. (C) Gene ontology analysis of N-termini increased and decreased by PAN. GO-BP bands. Quantification of intensity of the Slim terms significantly overrepresented (Fisher exact test, P,0.01 and FDR,0.01) are proteoform bands (P2–5) over full-length depicted in the bar graph. The enrichment factor versus the unchanged protein population protein (P0) by densitometry demonstrated is indicated at the x axis. (D) Cumulative histogram of actin-binding proteins and all other significant regulation of proteolytic process- proteins. On a global level, the distribution of ratios of actin-binding proteins was different ing during PAN injury, consistent with the as compared with the other proteins (Kolmogorov–Smirnoff test, P,0.001). (E) Overview of TAILS data. *P,0.05 in a two-tailed t test. ACTN4 cleavage sites. The x axis denotes the residue number of the protein, and the y axis a.u., arbitrary units; FL, full length protein; IB, the log2 fold change of PAN/veh treatment. Each terminus is denoted with its regulation. immunoblot; veh, vehicle.

J Am Soc Nephrol 28: 2867–2878, 2017 Podocyte Proteomics and Degradomics 2871 BRIEF COMMUNICATION www.jasn.org

CONCISE METHODS TAILS at a flow rate of 300 nl/min. The nanoLC TAILS was performed essentially as de- system was on-line coupled to an Impact 18 Glomerular Preparation and Mice scribed usingca1mg/channelproteome II high resolution Q-TOF (Bruker) via a Cap- Male mice (age 8 weeks, 100% BL6N back- as start material. Briefly, proteome samples tiveSpray nano ESI source (Bruker) essen- ground) were euthanized by cervical disloca- were thawed, purified by chloroform/metha- tially as described.41 Nitrogen gas flow in the tion and kidneys were perfused with magnetic nol precipitation to remove potentially inter- emitter device was saturated with acetoni- beads to isolate glomeruli as previously de- fering amine-containing compounds, trile as dopant using a nanoBooster device scribed with modifications.29 Briefly, a reduced with 10 mM DTT at 65°C for 30 (Bruker). MS1 spectra with a mass range solution containing 200 mltosyl-activated min, alkylated with 25 mM iodoacetamide from m/z 200–1600 were acquired at 4 Hz, Dynabeads dissolved in 10 ml ice-cold for 1 hour at RT in the dark, and amine mod- with the Top 15 most intense peaks selected HBSS with 13 protease inhibitor (Roche) ified by reductive dimethylation38 using for MS/MS analysis using an intensity- was generated. Mice were decapitated, and 20 mM sodium cyanoborohydride and dependent spectra acquisition time between kidneys were immediately removed and per- 20 mM formaldehyde at 37°C overnight. 2 and 20 Hz and stepped collision energy fused with the solution. Glomeruli were For the PAN-treated proteome, heavy form- (5.6 and 8.4 eV for 50% of the acquisition 13 minced and pressed through a sieve (mesh aldehyde ( CD2O) was used. Reagent addi- time). size 100 mm) twice and collected by a magnet, tion was repeated and the reaction incubated and spun down. Tissue was snap-frozen im- for another 2 hours. For the quantitative ex- Analysis of nLC-MS/MS Data mediately after pelleting (1 minute, 3000 3 g) periments, PAN and control proteomes were Spectra were matched to peptide sequences and removal of the supernatant. The mouse combined at this point. Samples were then at a false discovery rate of 0.01 using the An- holding was done in the University of Cologne purified by chloroform/methanol precipita- dromeda algorithm42 as implemented in the animal facility, according to standardized tion and digested with trypsin (proteome/ MaxQuant Software package v.1.5.3.30. Glo- specific pathogen-free conditions. The exper- enzyme ratio, 100:1 w/w) at 37°C overnight. meruli datasets were searched against the imental protocol was examined and approved Trypsin addition was repeated at 1000:1 UniProt mouse proteome (release 2015_10) by the LANUV NRW (Landesamt für Natur, (w/w) and digest efficacy controlled using a with enzyme specificity set as semispecific Umwelt und Verbraucherschutz Nordrhein- silver-stained SDS-PAGE gel. A small aliquot (free N-terminus) ArgC, Cys carbamidome- Westfalen/State Agency for Nature, Environ- of 20 mg proteome was saved as “pre-TAILS thylation, and Lys dimethylation (+28.031300) ment and Consumer Protection North control.” Next, dialyzed HPG-ALD (ratio, 1:5 as fixed modifications, and var N-terminal di- Rhine-Westphalia). proteome/polymer, w/w) and 20 mM sodium methylation (+28.031300), N-terminal acety- cyanoborohydride were added to C-termini lation (+42.010565), or N-terminal pyroGlu Cell Culture and internal peptides. The reaction was al- formation from Glu (218.010565) or Gln Human podocytes generated by Saleem lowed to proceed for 2 hours at 37°C, then (217.026549). et al.37 were cultured at 33°C in RPMI an additional 20 mM sodium cyanobrohy- For PAN-challenge experiment data, sim- medium containing 10% FBS and 13 insulin- dride was added, followed by incuba- ilar searches against the Uniprot human pro- transferrin supplement (Gibco). Differen- tion overnight. The polymer was removed teome (release 2015_10). However, we noted tiation was induced by culturing podocytes by ultra-filtration using spin filter devices that in the employed version of MaxQuant, at 37 degrees for 10 days. Cells were treated (10 MW cut-off, Merck) and the flow-through endogenously N-terminally modified pep- withPANdissolvedinH2O(final concen- containing protein N-terminal peptides tides were only unreliably quantified if an tration 50 mg/ml) and appropriate vehicle collected. N-terminal dimethyl label was defined. As a controls were performed. No crossconta- workaround, datafiles were duplicated and mination of this cell line with other cell nLC-MS/MS defined as two groups that were searched us- lines is currently reported. The cell cul- Samples were desalted using home-made ing mutually exclusive search parameters. tures were regularly tested negative for C18-Stop-And-Go solid phase Extraction Both parameter sets used semispecific (free mycoplasma. tips as described.39 Peptides were separated N-terminus) ArgC as digestion enzyme spec- on a nano-LC- (2.5-hour gradient), run on a ificity, Cys carbamidomethylation as fixed and Proteome Preparation Q Exactive MS/MS (Thermo) coupled to a Met oxidation as variable modifications, and Podocytes were scratched from the dishes and nano-LC-MS/MS exactly as previously de- isotope labeling by light (+28.031300) or heavy proteins were dissolved in urea. Acetone pre- scribed.40 For the in vivo TAILS, an estimated (+34.063117) dimethylation of Lys residues. cipitation was performed and pellets were 1 mgofdesaltedenrichedN-terminipep- The first parameter set additionally required washed three times with ice-cold methanol. tides were separated on an Ultimate 3000 light (+28.031300) or heavy (+34.063117) Then, pellets were dried and resuspended RSLCDnano HPLC (Thermo) operated in N-terminal dimethyl labeling for the identifi- with 1 ml 6M GHCl and HEPES 200 mM, a two-column setup (Acclaim PepMap 100 cation of natural and protease-generated pH 7.5 with a protease inhibitor cocktail, C18, particle size 3 mm, ID 75 mm, trap termini with unblocked a amines. The second snap-frozeninliquidnitrogen,andstored column length 2 cm, analytic column length parameter set considered variable N-terminal at 280°C until further processing. Glomeruli 25 cm, Thermo) using a 60-minute gradi- acetylation (+42.010565) or N-terminal py- were processed in an identical fashion. ent (2% Acetonitrile to 30% Acetonitrile) roGlu formation from Glu (218.010565) or

2872 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2867–2878, 2017 www.jasn.org BRIEF COMMUNICATION

Gln (217.026549), covering endogenously blocked termini and internal peptides car- ried over by incomplete coupling to the polymer (present as unmodified peptides) or pyro-glu formation. For the dataset from PAN treated rats, these two parameter groups defined above were applied in two independent searches against the Uniprot rat proteome (release 2015_10) using Max- Quant v 1.5.6.5. In all searches, the PSM FDR was set to 0.01 and the “requantify” option was engaged. All search results were additionally filtered by removing potential contami- nants, reverse entries, and peptides not ending with Arg as expected from digestion of dimethylated proteins with trypsin, after which the peptide FDR as estimated from the number of reverse entries retained was ,0.01. For quantitative analysis, reverse hits were removed as well as peptides that were not quantified in at least two biologic replicates.

Further Bioinformatics Analysis of TAILS Dataset For all quantified peptides, the median of the normalized log2-transformed ratios was calculated. The distribution of median log2 ratios was further divided in ten quantiles. N-terminal pyro-Glu formation from Gln andGlu is a known side reaction that can occur nonenzymatically at physiologic pH and in Figure 3. Protease class analysis reveals caspase 3/6 activation as an overarching signaling mechanism in PAN injury. (A) Position-weighted matrix of Neo-termini in the decreased vitro during the sample preparation. Hence termini population. The one-letter code demonstrates the frequency of each amino acid as all peptides with N-terminal pyro-Glu were compared with the unchanged termini population. The arrow indicates the cleavage site of not considered as native N termini, similar to the protease. The motif of proteases cleaving after an Arginine, “R,” is overrepresented in internal peptides with unmodified N termini the decreased termini population (as compared with nonchanged termini). This matrix that escaped polymer capture. corresponds to a Serine protease motif. (B) Position-weighted matrix of Neo-termini in the Information on the termini sequences, increased termini population. The arrow indicates the cleavage site of the protease. The matched proteins, and positional informa- “ ” motif of proteases cleaving after an Aspartate, D, is overrepresented in the decreased tion were obtained from the MaxQuant out- termini population. This motif corresponds to a caspase motif. (C) Immunoblot for cleaved put files. Termini were further annotated caspase 3–6 (Asp175). Protein lysates of human podocytes treated with PAN (50 mg/ml) for using the TopFIND Explorer software as im- the indicated time points. Immunoblots were stained with a cleaved-caspase specifican- plemented in the web portal (clipserve.clip. tibody, b tubulin indicated a loading control. *Delineates significance in a two-way ANOVA fi fi 19,27 with Dunett post-test versus veh (0h PAN treatment), P,0.05, n=5. (D) Protease web27 of ubc.ca/top nd/top nder). Abinaryde- fi activated caspase 3 in the dataset. Dark circles correspond to hits from the dataset. The cision tree was used for classi cation: Termi- protease of interest (caspase 3) is connected with downstream substrates inherent to this nus maps to position 1 (P1) or 2 (P2) in the dataset. Caspase activates a variety of other proteases, including caspase 6, 7, and 8, and the mitochondrial proteases PMPCB2, MTAP2, and cathepsinB. Substrates include mainly structural and metabolic targets of caspases, as well as the intermediate filament protein proteoforms (P1, 2) can be visualized by vimentin (D85). (E) Overview of vimentin cleavage sites. The x axis denotes the residue immunoblot using specific antibodies with number of the protein, and the y axis the log2 fold change of PAN/veh treatment. Each definedepitopes.Quantification of intensityof terminus is depicted with its regulation. The color background denotes statistical cutoffs the proteoform bands (P1+2) over full-length for increased (magenta) and decreased (green) termini. In the vimentin protein, many protein (P0) was performed using densitome- cleavages can be found by TAILS analysis. One of the increased cleavage sites (D85) is a try. *P,0.05 in a two-tailed t test. a.u., arbitrary known caspase cleavage site. All of the regulated sites are also present in native mouse units; cl. casp, cleaved caspase; IB, immuno- glomeruli. (F) Immunoblot of PAN-treated podocytes for vimentin. Two of the detected blot; IDed, identified; veh., vehicle.

J Am Soc Nephrol 28: 2867–2878, 2017 Podocyte Proteomics and Degradomics 2873 BRIEF COMMUNICATION www.jasn.org

the respective key word and all other proteins. Homologous sites (between mouse and human) were determined by a modiﬁed CPHOS software using the termini as an in- put.44 This software uses the NCBI homolo- gene groups for alignment across species. A binary decision tree was to determine extracellular proteins in the rat in vivo dataset. First proteins were ﬁltered for uniprot keyword “secreted” (YES/NO). If NO: Pro- tein is annotated with GO-CC term “extracellular region, extracellular region part, extracellular space, extracellular matrix, extracellular matrix part.” These proteins were determined as “extracellular.” All re- maining proteins were determined as “other proteins.”

Raw Data Deposition The raw data and maxquant output associated with this study were made publicly available at PRIDE/ProteomeExchange (http://www.ebi.ac.uk/pride).45 Datasets in- Figure 4. Proteolytic events are partially conserved in in vivo podocyte injury models. (A) clude: (1) PXD005139 (2) PXD005140 (3) Classification of protein N-termini identified by TAILS experiments in PAN-treated rats. A binary decision tree was used to classify proteins. A majority of proteolytic N-termini (in- PXD006356. dicated by dimethylated termini) localizes to extracellular compartments. (B) Structure of vimentin and localization of a proteolytic cleavage site determined in mouse glomeruli, Human Samples human podocyte cells, and rat glomeruli. The cleavage occurs between residue R196 and Unaffected kidney regions from male patients E197 in all three species. The cleavage process is reduced and in the top regulated quantile at the age of 60–80 were obtained from ne- in both in vivo and in vitro PAN injury models. (C) Immunoblot of lysates from isolated phrectomy surgeries. Written informed con- glomeruli of rats treated with PAN for 2 days versus vehicle control. Immunoblot was sent was obtained from the patients the day probed for vimentin using an N-terminal antibody. One cleaved proteoform can be visu- before the surgery. All procedures were ac- fi alized by the antibody in both human cells and rat glomeruli. Quanti cation of intensity of cording to the declaration of Helsinki. The the cleaved proteoform bands (Pcl) over full-length protein (P0) was performed using procedure was approved by the local ethics densitometry (only for rat glomeruli, n=3). *P,0.05 in a two-tailed t test. (D) Immunoblot of committee. Glomeruli were isolated by siev- glomeruli of rats treated with PAN for 2 days versus vehicle control. Immunoblot was probed for a-actinin. Several cleaved proteoforms (Pcl) can be visualized by immunoblot ing with differently sized meshes. Glomeruli using specificantibodies.Quantification of intensity of the cleaved proteoform bands (Pcl) were put on ice and then directly snap-frozen over full-length protein (P0) was performed using densitometry (only for rat glomeruli, n=3). in liquid nitrogen. *P,0.05 in a two-tailed t test. (E) Immunoblot of glomeruli from WT1 heterozygous knockout mice (WT1KO/WT) and control (wildtype) littermates. Immunoblot was probed Immunoblot for a-actinin. Quantification of intensity of the cleaved proteoform bands (Pcl) over full- Samples from mouse and human glomeruli length protein (P0) was performed using densitometry. *P,0.05 in a two-tailed t test. a.u., and from podocyte cells were lysed in 2% SDS arbitrary units; ctrl., control; glom., glomeruli; i.p., intraperitoneal; Pcl, cleaved proteoform; buffer and denatured by DNA shredding veh., vehicle; S.E., short exposure; WT1KO, gene deletion of WT1. by sonication and boiling. Samples were re- ducedwithDTTandrunona9%,10%,or corresponding protein sequence (YES/NO). (PAN/veh) and GO terms were enriched us- 12% SDS gel by gel electrophoresis, and If NO: Terminus starts after cleavage of Signal ing the Perseus software using a Fisher blotted onto a PVDF membrane by semidry peptide (YES/NO). If NO: Known cleavage exacttest(P values were controlled at transfer. Membranes were blocked with 13 (YES/NO). If NO: Alternative transcription FDR,0.05).43 Proteins were classified in “ac- Rotiblock (Roth) for 1 hour at room tem- start site (YES/NO) If NO: unexpected termi- tin binding,”“cytoskeleton,”“microtubule,” perature and incubated with the primary nus, inside domain (either P1-P1’ or P1’ to “cell adhesion,” and “Mitochondrion” antibody at a 1:1000 dilution overnight END) (YES/NO). If NO: unexpected termi- using the respective uniprot keywords. A Kol- (Table 2). After washing with PBST buffer, nus, outside domain. Further bioinformatics mogorov–Smirnoff test (two-way) was blots for mouse and rabbit antibodies were analyses were performed using Perseus v. performed to delineate differences in distri- decorated using fluorescence-labeled anti- 1.5.43 Data were ranked according their ratio butions between the proteins annotated with bodies (IR Dyes, Licor) 1:25,000 and

2874 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2867–2878, 2017 www.jasn.org BRIEF COMMUNICATION

Table 2. Primary antibodies Protein Company Species Application Dilution Cat. No. Epitope Vimentin Cell Signaling Technology Rabbit polyclonal antibody Western blot 1:1000 #5741 Oligopeptide IF 1:1000 Surrounding Arg 45 Vimentin Santa Cruz Biotechnology Goat polyclonal antibody Western blot 1: 1000 sc-7558 Oligopeptide at the C terminus ACTN4 BD Mouse mAb Western blot 1:1000 612576 AA 629–825 (crossreactivity ACTN1) Podocin Sigma-Aldrich Mouse mAb Western blot 1:1000 P0372 AA 367–383 Podocin Santa Cruz Biotechnology Rabbit polyclonal AB Western blot 1:1000 sc-21009 AA 1–130. Filamin Cell Signaling Technology Rabbit polyclonal AB Western blot 1:1000 #4762 Unknown Myh9 Sigma-Aldrich Rabbit polyclonal antibody Western blot 1:1000 M8064 AA 1949–1960 Cat. No., catalogue number; IF, immunoflurescence; AB, antibody. developed using a Licor Odyssey imaging XTT Cell Viability Assay dilution in PBS for overnight incubation. system. Blots for goat antibodies were XTT cell viability assay (Biotium, Hayward, Next, slides were washed using PBS three decorated using a horseradish-coupled CA) was performed according to the man- times and were decorated with phalloidin antibody and developed using enhanced ufacturer’s instructions. Two hours of in- (1:100) coupled to Cy5 and a Cy3 or Alexa chemoluminiscence in a Fusion imaging cubation with XTT assay solution was 488 labeled secondary antibody (Jackson system. performed in living cells before measurement ImmunoResearch) (Table 3). Samples were of XTT uptake using a plate reader (Perqin washed with PBS and DAPI and samples ’ Podocin Immunoprecipitation and Elmer) set up according to the manufacturer s were mounted using Prolong Gold (Thermo In-Gel Digestion instructions. Scientific). For immunoprecipitation, human glomeruli were lysed in ice-cold IP buffer and homog- Reagents Rats, PAN Treatment, and enized using a glass douncer. The IP buffer PAN was obtained from Santa Cruz Biotech- Glomerular Isolation used contained 1% Triton X-100, 20 mM Tris nology (CAS Number 58–60–6), and was Rats were housed in a specific pathogen–free pH 7.5, 25 mM NaCl, 50 mM NaF, 15 mM dissolved in water. Cytochalasin CAS facility with free access to chow and water Na4P2O7, 1 mM EDTA, 0.25 mM PMSF, and 22144–77–0 was obtained from SIGMA and and a 12-hour-day/12-hour-night cycle. An- 5 mM Na3VO4. Lysates were precleared with dissolved in DMSO (1M). Appropriate vehi- imal procedures were approved by local au- protein G beads. Supernatants containing cle controls were used. The HPG-ALD poly- thorities (Regierungspräsidium Freiburg equal amounts of total proteins were incu- mer is commercially distributed by Flintbox G15–134). Male CD IGS rats (Crl:CD[SD]) bated for 1 hour at 4°C with anti-Podocin (flintbox.com/public/project/1948/). (body wt, 101–125 g) were obtained from antibody (Sigma-Aldrich) and 40 ml of Pro- Charles River. Animals were randomly tein G beads. The beads were washed three Immunofluorescence assigned to treatment groups. Animals were times with IP-buffer and bound proteins Podocytes grown on coverslips (50% conflu- treated by intraperitoneal injection of either were resolved by 10% SDS-PAGE. One half ency) were washed with 13PBS, fixed using PAN (15 mg/100 g body wt) (Santa Cruz of the sample was analyzed by in-gel tryptic 4% PFA at room temperature for 10 minutes, Biotechnology) dissolved in 0.9% NaCl or digestion, Stage-tip, and nLC-MS/MS analy- and immunostaining was performed as solvent alone (26). Kidneys were removed in sis exactly as previously described.20 The described previously. Samples were blocked deep anesthesia (ketamine [100 mg/kg] and other half of the sample was subjected to im- using 5% donkey serum dissolved in 0.1% xylazine [5 mg/kg] in 0.9% NaCl [10 ml/g]) munoblotting. Triton in PBS. Antibodies were used in a 1:400 and immediately perfused with ice-cold

Table 3. Secondary antibodies and other dyes Antibody Company Cat. No. Application Dilution Rabbit anti-goat, HRP coupled Dako E046601–2 Immunoblot, ECL 1:25,000 Donkey anti-rabbit, IRDye 680RD Licor 926–68073 Immunoblot, Licor 1:25,000 Donkey anti–mouse 800 IRDye 800 RD Licor 926–32280 Immunoblot, Licor 1:25,000 Donkey anti-rabbit, A488 Jackson ImmunoResearch 711–546–152 Immunofluorescence 1:400 Donkey anti–mouse Cy 3 Jackson ImmunoResearch 715–165–150 Immunofluorescence 1:400 Donkey anti–mouse A488 Jackson ImmunoResearch 715–545–150 Immunofluorescence 1:400 Phalloidin, 647 Dyomics 647P1–33 Immunofluorescence 1:400 Dapi Invitrogen D1306 Immunofluorescence 1:10,000 Cat. No., catalogue number; HRP, horseradish peroxidase; ECL, enhanced chemoluminescence.

J Am Soc Nephrol 28: 2867–2878, 2017 Podocyte Proteomics and Degradomics 2875 BRIEF COMMUNICATION www.jasn.org

Hanks balanced salt solution (HBSS) via the Wilbrand-Hennes (Cologne), and Nadine T, Nissinen M, Herva R, Kashtan CE, Peltonen L, renal arteries. The renal cortex was removed Wettengl (Juelich). The authors acknowledge Holmberg C, Olsen A, Tryggvason K: Posi- tionally cloned gene for a novel glomerular and minced, and the glomeruli were isolated the proteomics core facility of the Cologne protein–Nephrin–is mutated in congenital using the sieving method, which results in Excellence Cluster on Cellular Stress Re- nephrotic syndrome. Mol Cell 1: 575–582, 95% purity of glomeruli in the preparation. sponses in Aging-Associated Diseases. The 1998 The following sieve sizes were used: 180, 106, WT1 heterozygous knockout mice were a 5. Kaplan JM, Kim SH, North KN, Rennke H, and 53 mm. The tissues were constantly rinsed kind gift of Jordan Kreidberg. Correia LA, Tong HQ, Mathis BJ, Rodríguez- Pérez JC, Allen PG, Beggs AH, Pollak MR: with 4°C cold HBSS during glomeruli isola- This work was supported by the University Mutations in ACTN4, encoding alpha- tion. The glomeruli were collected on the low- of Cologne (UoC PostDoc grant to M.M.R.), actinin-4, cause familial focal segmental est sieve. Purity of renal glomeruli was verified Deutsche Forschungsgemeinschaft (CRC glomerulosclerosis. Nat Genet 24: 251–256, using light microscopy. The final glomerular 1140 to F.G. and T.B.H., CRC 992 and HU 2000 pellet was obtained by centrifugation at 2300 1016/8-1 to T.B.H., KA 3217/4-1 to M.K., BE 6. Lange PF, Overall CM: Protein TAILS: When termini tell tales of proteolysis and rpm for 7 minutes and was snap-frozen in 2212 to T.B.) and an intramural Koeln- function. Curr Opin Chem Biol 17: 73–82, liquid nitrogen. Fortune Grant (to M.M.R.). T.B.H. was 2013 The rats used for immunoblotting were supported by the BMBF (01GM1518C), by 7. Vandenbroucke RE, Libert C: Is there new hope phenotyped for proteinuria in the previous the European Research Council (ERC grant for therapeutic matrix metalloproteinase – study,46 and an untouched aliquot of glomer- 616891 and H2020-IMI2 consortium BEAt- inhibition? Nat Rev Drug Discov 13: 904 927, 2014 uli was used. At day 2, rats were euthanized DKD), and by the Excellence Initiative of 8. Kistler AD, Peev V, Forst A-L, El Hindi S, and glomeruli fractions were obtained by siev- the German Federal and State Governments Altintas MM, Reiser J: Enzymatic disease of ing of the perfused kidneys. An aliquot of glo- (BIOSS). P.F.H.is in part supported by an ERC the podocyte. Pediatr Nephrol 25: 1017– meruli was snap-frozen and stored at 280°. starting grant (ID 639905). J.N.K. is the re- 1023, 2010 Then, glomeruli were lysed in 4% SDS, and cipient of a Career Investigator Scholar 9. Yaddanapudi S, Altintas MM, Kistler AD, Fernandez I, Möller CC, Wei C, Peev V, lysates were sonicated, boiled, and subjected award from the Michael Smith Foundation Flesche JB, Forst A-L, Li J, Patrakka J, Xiao Z, to immunoblotting as described above. No of Health Research. J.N.K. acknowledges fund- Grahammer F, Schiffer M, Lohmüller T, further freeze-thaw cycles were introduced. ing from the Natural Science and Engineering Reinheckel T, Gu C, Huber TB, Ju W, Bitzer The rats used for degradomic studies were Research Council of Canada. M, Rastaldi MP, Ruiz P, Tryggvason K, Shaw euthanized at day 4 and glomeruli were lysed The author contributions were as follows: AS, Faul C, Sever S, Reiser J: CD2AP in mouse and human podocytes controls a and processed as described in the “Proteome Designed study: M.M.R., P.F.H., T.B. Wrote proteolytic program that regulates cytoskel- ” “ ” Preparation and TAILS sections. Paper: M.M.R., P.F.H., T.B. Performed ex- etal structure and cellular survival. JClinIn- periments: M.M.R., A.K.H., F.G., J.B., L.A.V., vest 121: 3965–3980, 2011 E.M.S., M.H., M.M., F.D. Analyzed Data: 10. Yamamoto-Nonaka K, Koike M, Asanuma K, WT1 Knockout Mouse Model M.M.R., P.F.H., T.B. Interpreted data: M.M.R., Takagi M, Trejo JAO, Seki T, Hidaka T, Male proteinuric WT1 heterozygous mice Ichimura K, Sakai T, Tada N, Ueno T, P.F.H., T.B. Contributed new tools: C.K., (FVB background) at the age of 14 weeks Uchiyama Y, Tomino Y: Cathepsin D in po- J.N.K., M.K., T.B.H. were compared with wildtype male control docytes is important in the pathogenesis of proteinuria and CKD. J Am Soc Nephrol 27: animals (littermates). Proteinuria was vali- 2685–2700, 2016 dated by gel electrophoresis and Coomassie 11. Li S-Y, Huang P-H, Yang A-H, Tarng D-C, staining. The mouse model was previously DISCLOSURES Yang W-C, Lin C-C, Chen J-W, Schmid- Schönbein G, Lin S-J: Matrix metal- described and is an established model of None. FSGS.28 Glomerular preparation was per- loproteinase-9 deficiency attenuates diabetic nephropathy by modulation of podocyte formed as described above. Glomeruli were functions and dedifferentiation. Kidney Int lysed in 4% SDS buffer and then subjected to REFERENCES 86: 358–369, 2014 immunoblotting as described above. The ex- 12. Rinschen MM, Schroeter CB, Koehler S, perimental protocol was examined and 1. Pavenstädt H, Kriz W, Kretzler M: Cell biology Ising C, Schermer B, Kann M, Benzing T, approved by the LANUV NRW (Landesamt of the glomerular podocyte. Physiol Rev 83: Brinkkoetter PT: Quantitative deep-mapping 253–307, 2003 of the cultured podocyte proteome uncovers für Natur, Umwelt und Verbraucherschutz 2. Brinkkoetter PT, Ising C, Benzing T: The role shifts in proteostatic mechanisms during dif- Nordrhein-Westfalen/State Agency for Na- ofthepodocyteinalbuminfiltration. Nat Rev ferentiation. Am J Physiol Cell Physiol 311: ture, Environment and Consumer Protection Nephrol 9: 328–336, 2013 C404–C417, 2016 North Rhine-Westphalia). 3. Boute N, Gribouval O, Roselli S, Benessy F, 13. Altintas MM, Moriwaki K, Wei C, Möller CC, Lee H, Fuchshuber A, Dahan K, Gubler MC, Flesche J, Li J, Yaddanapudi S, Faridi MH, Niaudet P, Antignac C: NPHS2, encoding the Gödel M, Huber TB, Preston RA, Jiang JX, glomerular protein podocin, is mutated in Kerjaschki D, Sever S, Reiser J: Reduction of ACKNOWLEDGMENTS autosomal recessive steroid-resistant ne- proteinuria through podocyte alkalinization. phrotic syndrome. Nat Genet 24: 349–354, JBiolChem289: 17454–17467, 2014 2000 14. Sever S, Altintas MM, Nankoe SR, Möller CC, The authors acknowledge excellent technical 4. Kestilä M, Lenkkeri U, Männikkö M, Lamerdin J, Ko D, Wei C, Henderson J, del Re EC, Hsing L, support by Ruth Herzog (Cologne), Astrid McCready P, Putaala H, Ruotsalainen V, Morita Erickson A, Cohen CD, Kretzler M, Kerjaschki

2876 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2867–2878, 2017 www.jasn.org BRIEF COMMUNICATION

D, Rudensky A, Nikolic B, Reiser J: Proteolytic 27. Fortelny N, Cox JH, Kappelhoff R, Starr AE, 37. Saleem MA, O’Hare MJ, Reiser J, Coward RJ, processing of dynamin by cytoplasmic cathepsin Lange PF, Pavlidis P, Overall CM: Network Inward CD, Farren T, Xing CY, Ni L, L is a mechanism for proteinuric kidney disease. analyses reveal pervasive functional regu- Mathieson PW, Mundel P: A conditionally J Clin Invest 117: 2095–2104, 2007 lation between proteases in the human immortalized human podocyte cell line 15. Lai ZW, Petrera A, Schilling O: Protein amino- protease web. PLoS Biol 12: e1001869, demonstrating nephrin and podocin ex- terminal modifications and proteomic ap- 2014 pression. J Am Soc Nephrol 13: 630–638, proaches for N-terminal profiling. Curr Opin 28. Schumacher VA, Schlötzer-Schrehardt U, 2002 Chem Biol 24: 71–79, 2015 Karumanchi SA, Shi X, Zaia J, Jeruschke S, 38. Boersema PJ, Raijmakers R, Lemeer S, 16. Huesgen PF, Lange PF, Rogers LD, Solis N, Zhang D, Pavenstädt H, Drenckhan A, Mohammed S, Heck AJR: Multiplex peptide Eckhard U, Kleifeld O, Goulas T, Gomis-Rüth Amann K, Ng C, Hartwig S, Ng KH, Ho J, stable isotope dimethyl labeling for quanti- FX, Overall CM: LysargiNase mirrors trypsin Kreidberg JA, Taglienti M, Royer-Pokora B, tative proteomics. Nat Protoc 4: 484–494, for protein C-terminal and methylation-site Ai X: WT1-dependent sulfatase expression 2009 identification. Nat Methods 12: 55–58, 2015 maintains the normal glomerular filtration 39. Rappsilber J, Ishihama Y, Mann M: Stop and 17. Huesgen PF, Lange PF, Overall CM: Ensem- barrier. JAmSocNephrol22: 1286–1296, go extraction tips for matrix-assisted laser bles of protein termini and specificpro- 2011 desorption/ionization, nanoelectrospray, and teolytic signatures as candidate biomarkers 29. Rinschen MM, Wu X, König T, Pisitkun T, LC/MS sample pretreatment in proteomics. of disease. Proteomics Clin Appl 8: 338–350, Hagmann H, Pahmeyer C, Lamkemeyer T, Anal Chem 75: 663–670, 2003 2014 Kohli P, Schnell N, Schermer B, Dryer S, 40. Rinschen MM, Bharill P, Wu X, Kohli P, 18. Kleifeld O, Doucet A, auf dem Keller U, Brooks BR, Beltrao P, Krueger M, Reinert MJ, Kretz O, Saez I, Schermer B, Prudova A, Schilling O, Kainthan RK, Starr Brinkkoetter PT, Benzing T: Phosphopro- Höhne M, Bartram MP, Aravamudhan S, AE, Foster LJ, Kizhakkedathu JN, Overall teomic analysis reveals regulatory mecha- Brooks BR, Vilchez D, Huber TB, Müller R-U, CM: Isotopic labeling of terminal amines nisms at the kidney filtration barrier. JAm Krüger M, Benzing T: The ubiquitin ligase in complex samples identifies protein Soc Nephrol 25: 1509–1522, 2014 Ubr4 controls stability of podocin/MEC-2 N-termini and protease cleavage products. 30. Hoffert JD, Pisitkun T, Knepper MA: Phos- supercomplexes. Hum Mol Genet 25: 1328– Nat Biotechnol 28: 281–288, 2010 phoproteomics of vasopressin signaling in 1344, 2016 19. Fortelny N, Yang S, Pavlidis P, Lange PF, the kidney. Expert Rev Proteomics 8: 157– 41. Beck S, Michalski A, Raether O, Lubeck M, Overall CM: Proteome TopFIND 3.0 with 163, 2011 Kaspar S, Goedecke N, Baessmann C, TopFINDer and PathFINDer: database and 31. Rinschen MM, Benzing T, Limbutara K, Hornburg D, Meier F, Paron I, Kulak NA, analysis tools for the association of protein Pisitkun T: Proteomic analysis of the kidney CoxJ,MannM:TheimpactII,avery termini to pre- and post-translational events. filtration barrier–Problems and perspectives. high-resolution Quadrupole Time-of-Flight Nucleic Acids Res 43: D290–D297, 2015 Proteomics Clin Appl 9: 1053–1068, 2015 Instrument (QTOF) for deep shotgun pro- 20. Völker LA, Schurek E-M, Rinschen MM, Tax J, 32. Rinschen MM, Pahmeyer C, Pisitkun T, teomics. Mol Cell Proteomics 14: 2014–2029, Schutte BA, Lamkemeyer T, Ungrue D, Schnell N, Wu X, Maaß M, Bartram MP, 2015 Schermer B, Benzing T, Höhne M: Charac- Lamkemeyer T, Schermer B, Benzing T, 42. Cox J, Mann M: MaxQuant enables high terization of a short isoform of the kidney Brinkkoetter PT: Comparative phosphopro- peptide identification rates, individualized p. protein podocin in human kidney. BMC teomic analysis of mammalian glomeruli re- p.b.-range mass accuracies and proteome- Nephrol 14:102,2013 veals conserved podocin C-terminal Wide protein quantification. Nat Biotechnol 21. Weber S, Gribouval O, Esquivel EL, Morinière phosphorylation as a determinant of slit di- 26: 1367–1372, 2008 V, Tête M-J, Legendre C, Niaudet P, Antignac aphragm complex architecture. Proteomics 43. Tyanova S, Temu T, Sinitcyn P, Carlson A, C: NPHS2 mutation analysis shows genetic 15: 1326–1331, 2015 Hein MY, Geiger T, Mann M, Cox J: The heterogeneity of steroid-resistant nephrotic 33. Rinschen MM, Schermer B, Benzing T: Perseus computational platform for com- syndrome and low post-transplant recur- Vasopressin-2 receptor signaling and auto- prehensive analysis of (prote)omics data. Nat rence. Kidney Int 66: 571–579, 2004 somal dominant polycystic kidney disease: Methods 13: 731–740, 2016 22. Shankland SJ: The podocyte’s response From bench to bedside and back again. J 44.ZhaoB,PisitkunT,HoffertJD,KnepperMA, to injury: Role in proteinuria and glomer- Am Soc Nephrol 25: 1140–1147, 2014 Saeed F: CPhos: A program to calculate and ulosclerosis. Kidney Int 69: 2131–2147, 34. Talbot JJ, Song X, Wang X, Rinschen MM, visualize evolutionarily conserved functional 2006 Doerr N, LaRiviere WB, Schermer B, Pei YP, phosphorylation sites. Proteomics 12: 3299– 23. Kanjanabuch T, Ma L-J, Chen J, Pozzi A, Torres VE, Weimbs T: The cleaved cytoplas- 3303, 2012 Guan Y, Mundel P, Fogo AB: PPAR-gamma mic tail of polycystin-1 regulates Src-de- 45. Vizcaíno JA, Csordas A, del-Toro N, Dianes agonist protects podocytes from injury. Kid- pendent STAT3 activation. JAmSoc JA, Griss J, Lavidas I, Mayer G, Perez-Riverol ney Int 71: 1232–1239, 2007 Nephrol 25: 1737–1748, 2014 Y, Reisinger F, Ternent T, Xu Q-W, Wang R, 24. Lee HW, Khan SQ, Faridi MH, Wei C, Tardi 35. Bartram MP, Habbig S, Pahmeyer C, Höhne Hermjakob H: 2016 update of the PRIDE NJ, Altintas MM, Elshabrawy HA, Mangos S, M, Weber LT, Thiele H, Altmüller J, Kottoor database and its related tools. Nucleic Acids Quick KL, Sever S, Reiser J, Gupta V: A po- N, Wenzel A, Krueger M, Schermer B, Res 44(D1): D447–D456, 2016 docyte-based automated screening assay Benzing T, Rinschen MM, Beck BB: Three- 46. Rinschen MM, Grahammer F, Hoppe A-K, identifies protective small molecules. JAm layered proteomic characterization of a novel Kohli P, Hagmann H, Kretz O, Bertsch S, Soc Nephrol 26: 2741–2752, 2015 ACTN4 mutation unravels its pathogenic Höhne M, Göbel H, Bartram MP, 25. Colaert N, Helsens K, Martens L, potential in FSGS. Hum Mol Genet 25: 1152– Gandhirajan RK, Krüger M, Brinkkoetter Vandekerckhove J, Gevaert K: Improved vi- 1164, 2016 P-T, Huber TB, Kann M, Wickström SA, sualization of protein consensus sequences 36. Rinschen MM, Schroeter CB, Koehler S, Ising Benzing T, Schermer B: YAP-mediated by iceLogo. Nat Methods 6: 786– 787, 2009 C, Schermer B, Kann M, Benzing T, mechanotransduction determines the po- 26. Sanwal V, Pandya M, Bhaskaran M, Franki N, Brinkkoetter PT: Quantitative deep mapping docyte’s response to damage. Sci Signal 10 Reddy K, Ding G, Kapasi A, Valderrama E, of the cultured podocyte proteome uncovers (474), 2017 Singhal PC: Puromycin aminonucleoside in- shifts in proteostatic mechanisms during 47. Büscher AK, Kranz B, Büscher R, Hildebrandt duces glomerular epithelial cell apoptosis. differentiation. Am J Physiol Cell Physiol 311: F, Dworniczak B, Pennekamp P, Kuwertz- Exp Mol Pathol 70: 54–64, 2001 C404–C417, 2016 Bröking E, Wingen A-M, John U, Kemper M,

J Am Soc Nephrol 28: 2867–2878, 2017 Podocyte Proteomics and Degradomics 2877 BRIEF COMMUNICATION www.jasn.org

MonnensL,HoyerPF,WeberS,KonradM: Nephrol Dial Transplant 25: 2970–2976, non-Finnish congenital nephrotic syndrome. Immunosuppression and renal outcome in 2010 J Am Soc Nephrol 21: 1209–1217, 2010 congenital and pediatric steroid-resistant 49. Philippe A, Nevo F, Esquivel EL, Reklaityte D, 51. Choi HJ, Lee BH, Cho HY, Moon KC, Ha IS, nephrotic syndrome. Clin J Am Soc Nephrol Gribouval O, Tête M-J, Loirat C, Dantal J, Nagata M, Choi Y, Cheong HI: Familial focal 5: 2075–2084, 2010 Fischbach M, Pouteil-Noble C, Decramer S, segmental glomerulosclerosis associated 48. Schoeb DS, Chernin G, Heeringa SF, Matejas Hoehne M, Benzing T, Charbit M, Niaudet P, with an ACTN4 mutation and paternal V, Held S, Vega-Warner V, Bockenhauer D, Antignac C: Nephrin mutations can cause germline mosaicism. Am J Kidney Dis 51: Vlangos CN, Moorani KN, Neuhaus TJ, Kari childhood-onset steroid-resistant nephrotic 834–838, 2008 JA, MacDonald J, Saisawat P, Ashraf S, syndrome. J Am Soc Nephrol 19: 1871–1878, Ovunc B, Zenker M, Hildebrandt F; 2008 Gesselschaft für Paediatrische Nephrologie 50. Machuca E, Benoit G, Nevo F, Tête M-J, (GPN) Study Group: Nineteen novel NPHS1 Gribouval O, Pawtowski A, Brandström P, This article contains supplemental material online mutations in a worldwide cohort of patients Loirat C, Niaudet P, Gubler M-C, Antignac at http://jasn.asnjournals.org/lookup/suppl/doi:10. with congenital nephrotic syndrome (CNS). C: Genotype-phenotype correlations in 1681/ASN.2016101119/-/DCSupplemental.

2878 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2867–2878, 2017