DOI:.38/ncb97 P ( μm, hours) 1 2 4 P DMSO Figure S1 unningham et al. E 97 65 27 MFN1 GFP-Parkin Opa1 ctin GPDH HEK293 GFP-Parkin 19 115 97 65 27 Mitochondrial Fraction SH-SY5Y GFP-Parkin Mito DMSO Mito P yto DMSO SH-SY5Y ells DMSO 1 P 1 yto P and and and and D and and D F Relative bundance ytosol and 3 1 and and 5 15 25 35 45 55 65 75 85 9 95 5 DMSO Mitochondria and DMSO Supplementary Figure 1 ubiquitin chains are enriched in HEK293 and SH-SY5Y whole cell lysates and mitochondria upon P treatment. () SH-SY5Y Western blot confirmation of mitochondrial depolarization upon P treatment over time. MFN1 levels decrease and Opa1 processing increased over the course of 4hr with μm P treatment in SH-SY5Y cells. () ase peak chromatograms of immunoaffinity purified K-GG peptides from SH-SY5Y cells following Parkin overexpression (oe) and/ or μm P treatment for 2hr. hromatographic features showing ubiquitin signature peptides for,,, and 3-linked chains are denoted by arrows. Inset shows extracted ion chromatograms from peptides in the treated and untreated samples. () SDS-PGE gel regions excised for ubiquitin linkage analysis in Figure S1D. HEK293 cells stably expressing GFP-Parkin were treated with either DMSO or μm P for 3hr, fractionated into mitochondria, and lysed for trypsinization P 3 3 P 3 3 / 3 Time (min) and and D ytosol and Mitochondria and Parkin oe ontrol Parkin oe +P 2hr 3 3 3 3 and subsequent ubiquitin linkage analysis. Data represent one out of three independent experiments. (D) Ubiquitin linkage analysis on 4 gel regions ranging from 5-kDa. bundance (fmol) of each ubiquitin linkage is reported for DMSO (black circles) or P (red circles) treatment for each gel region. Data represent one out of three experiments. (E) Representative SDS-PGE gel showing excised regions for ubiquitin linkage analysis in Figure S1F. SH-SY5Y cells were treated with either DMSO or μm P for 3hr, fractionated into cytosolic and mitochondrial fractions, and lysed for trypsinization and subsequent ubiquitin linkage analysis. (F) Ubiquitin linkage analysis of cytosolic and mitochondrial fractions of 2 gel regions ranging from -kda. bundance (fmol) of each ubiquitin linkage is reported for DMSO (black circles) or P (red circles) treatment for each gel region from either the cytosolic or mitochondrial fractions of two replicate experiments in SH-SY5Y cells. WWW.NTURE.OM/NTUREELLIOLOGY 1 15 Macmillan Publishers Limited. ll rights reserved
OP1 Mfn1 GFP Tub ct GPDH Input + DMSO Input + P FT + P 185 115 65 25 25ug Input + P FT + P band 1 band fmol 4.5 4 3.5 3 2.5 2 1.5 1.5 3 1 1 2 3 4 5 6 7 8 9 and PU (a.u.) Parkin Input Flow-thru 1 2 3 4 5 6 7 8 9 and Supplementary Figure 2 Immunodepletion of GFP-Parkin confirms substrates other than Parkin carry -linked polyubiquitin chains. () HEK293 cells overexpressing GFP-Parkin were treated with um P or a DMSO vehicle control for 2hr. Lysates were immunodepleted with anti- GFP beads and input and flow through (FT) fractions were run on Western blot. MFN1 levels decreased and Opa1 processing increased confirming mitochondrial depolarization upon P treatment. This experiment was performed once. () Input and flow through (FT) fractions treated with P from () were run on SDS-PGE and excised into gel regions spanning 5-kDa. () Ubiquitin linkage analysis was used to measure the abundance of chains (fmol) for each gel region excised (left) and Parkin peptide abundance (U) was measured for each gel region excised (right). 2 WWW.NTURE.OM/NTUREELLIOLOGY 15 Macmillan Publishers Limited. ll rights reserved
v (nm/sec) 1.5 1..5 USP2 USP. 2 4 6 UbM (μm) USP7 R eq (nm).5.4.3.2.1. 1 oncentration ( μm) n.s. Monomer K6 K11 K27 K29 K48 K33 K63 Linear D kda 37 * 25 Ub 4 : 6 48 63 44299.22 ounts vs. Deconvoluted Mass (amu) E kda 1 75 F Fraction Top and Fraction ottom and Ratio (ottom/top) K6.988.982.99 K11.4.2. K48.1.1 N/ 37 K63.7.15 2.16 25 Top and ottom and Supplementary Figure 3 USP is an ubiquitin specific deubiquitinase that prefers compact chain types. () Michaelis-Menten plot of USP2 (blue), USP (orange), and USP7 (red) catalytic domain activity. Data were fit using the equation v = V max [S] /K M +[S]. () USP s preference for compact ubiquitin linkages is not driven by a hyper affinity for dimeric ubiquitin chain types in a biolayer interferometry experiment. Equilibrium response is shown as a function of di-ubiquitin concentration. ompact chains (,, and ) are shown in colors. Greater overall signal was observed with dimers, but individual kinetic curves displayed some non-specific adherence to the tip and fitting showed no significant (n.s.) decrease in dissociation constant (all K D s > 5μM). () Gel of,, and 3 purified tetramers. The lower band of (*) migrates most closely to 3 tetramers. (D) Intact mass spectrometry of tetramer shows a single molecular weight peak at the expected MW of tetrameric ubiquitin. (E) SDS-PGE gel of tetramers showing regions excised for ubiquitin linkage analysis. (F) Relative amounts of detected chain types for each excised band in (E). 3 make up a larger fraction of linkages in the lower band indicating this chain type as an off-target product during the enzymatic formation of K6 tetramers. WWW.NTURE.OM/NTUREELLIOLOGY 3 15 Macmillan Publishers Limited. ll rights reserved
E Figure S4 unningham et al. U Ratio (KO/WT) 2. 1.5 1..5. Relative bundance Relative bundance +P Usp (um) 1 15 kda 1 75 37 25 15 HSPD1 WT TP5 USP_1 USP_2 USP_3 34. 34.5 35. 35.5 KO Time (min) 34. 34.5 35. 35.5 Time (min) Excised for linkage analysis Supplementary Figure 4 Mass spectrometry validation and ubiquitin linkage analysis of USP knockout cell line. (, ) Mass spectrometry confirmation of USP RISPR knockout. Ratios of area under the curve (U) between USP knockout HEK293 cells expressing GFP-Parkin and a wild type USP isogenic cell line for 2 mitochondrial matrix proteins (HSPD1, TP5) and three independent USP peptides (; see Table S2 for raw data). Ratios are not shown for the three USP signature peptides since no signal was detected in the USP knockout cell line, as represented in (). Error bars represent standard deviation from n=3 independent experiments (Table S2). () Representative SDS-PGE gel from D 19 115 97 65 27 and and and and D HEK293 GFP-Parkin WT DMSO KO P WT P KO DMSO Mitochondrial Fraction WT P KO P and and and and D 3 3 3 3 three independent experiments showing excision locations used for ubiquitin linkage analysis in Figures 3 and S3D. USP knockout (KO) or WT USP (WT) HEK293 cells stably expressing GFP-Parkin were treated with either DMSO or P (μm, 3hr), enriched for mitochondria, and lysed for trypsinization and subsequent ubiquitin linkage analysis. (D) Ubiquitin linkage analysis on 4 gel regions ranging from 5-kDa. fmol of each ubiquitin linkage is reported for P treated WT (black circles) or USP knockout (red circles) mitochondrial enriched fractions for each gel region. Data represent one out of two replicate experiments. (E) Representative SDS- PGE gel showing regions excised for ubiquitin linkage analysis in Figure 3. 4 WWW.NTURE.OM/NTUREELLIOLOGY 15 Macmillan Publishers Limited. ll rights reserved
S U P P L E M E N T R Y I N F O R M T I O N P (2 hours, 5 μm) Doxycycline - - + + - + + MFN1 GFP-Parkin Usp-3xFLG Opa1 ctin GPDH p-value 1e 1e 5 Experiment 1 MUL1 FKP8 TOM USP TDRKH MTX1 GDP1 VD3 VD2.1.1 3 2 ES 1 1 2 Log2(USP + P vs. P alone) 3 Experiment 2 1e 1e 5 VD3 PS2 H12 TOM H31 ISD1 RL34 PN p-value.1.1 MID49 PSMG2 F188 SRSF9 USP MID49 VD3 PS2 H12 PSMG2 TOM F188 SRSF9 H31 ISD1 RL34 PN 3 2 1 1 2 3 Log2(USP + P vs. P alone) Supplementary Figure 5 USP deubiquitinates outer membrane mitochondrial substrates. () Western blot confirmation of USP induction following doxycycline administration and mitochondrial depolarization upon P treatment. USP showed strong induction, while reduced MFN1 levels and increase in Opa1 processing were observed. Representative Western blot shown from experiment 1. (-) Volcano plots of ubiquitinated Figure S5 unningham et al. proteins identified in Experiments 1 and 2, respectively. The fold change of ubiquitination for each protein between ombo (USP overexpression and P) versus P is denoted on the x-axis with significance reported on the y-axis. oth values were determined based on the aggregate values of individual KGG peptides using LiME analysis. Proteins with p-value <.1 are labeled; mitochondria-associated proteins are colored red. WWW.NTURE.OM/NTUREELLIOLOGY 5 15 Macmillan Publishers Limited. ll rights reserved
DU-FLG HSP Transfection: TOM-MLS- TOM-MLS- TOM-MLS- TOM-MLS- I: USP-FLG USP-FLG USP2-FLG USP5-FLG TRID-FLG Transfection: GFP-Parkin + USP-FLG I: GFP-Parkin DU-FLG HSP TOM DU-FLG TOM-MLS USP-FLG HSP I: Transfection: GFP-Parkin + ß-Gal H-Ub wild type H-Ub K6R H-Ub K11R H-Ub K6R/K11R H-Ub K48R/K63R H-Ub KallR TOM-MLS USP2-FLG GFP- Parkin TOM-MLS TRID-FLG TOM TOM-MLS USP5-FLG H-Ub D * ß-Gal % of cells without mito (TOM staining) 24hr P ß-Gal WT Figure S6 unningham et al. Supplementary Figure 6 typical ubiquitin chain linkages signal for mitophagy. () Immunostaining of HeLa cells co-transfected with GFP- Parkin and mitochondrial-localized, FLG-tagged DUs. 24 hours after transfection cells were immunostained for FLG (green) and endogenous HSP (red). Each DU co-localizes with HSP indicating proper mitochondrial targeting (see merge). In addition, expression levels appear uniform across all DUs transfected. Representative cells are shown for each transfection. Scale bar, μm. () Large field images from Figure 5. Solid yellow outlines highlight cells with GFP-Parkin and DU-Flag expressing cells. Dashed yellow outlines highlight cells expressing only GFP- Parkin. Scale bar, μm. () Immunostaining of HeLa cells co-transfected with GFP-Parkin and H-tagged ubiquitin variants as listed in Figure 5. Only GFP-Parkin expressing cells are outlined. Solid lines indicate TOM positive staining cells; dashed lines indicate TOM negative staining cells. Scale bar, μm. (D) Quantification of TOM in HeLa cells treated with P (15μM, 24hr) and transfected with either beta-gal or H-ubiquitin. significant difference in TOM was observed with a 24 hour P treatment preventing analysis of individual H-ubiquitin mutants tested in Figure 5. Each point represents a single experiment with a mean intensity from 25- cells measured; the line represents the mean from n= 5 experiments and error bars represent S.E.M. Significance is reported as one-way NOV with Holm-Sidak s multiple comparison test (* p <.5). 6 WWW.NTURE.OM/NTUREELLIOLOGY 15 Macmillan Publishers Limited. ll rights reserved
Supplementary Figure 6 Full scans WWW.NTURE.OM/NTUREELLIOLOGY 7 15 Macmillan Publishers Limited. ll rights reserved
Supplementary Table 1 Total peptide counts of selected proteins from isolated cytosolic (orange highlights) and mitochondrial fractions of HEK293T cells stably expressing GFP- Parkin confirm proper enrichment within each fraction. Upon P treatment Parkin is recruited to mitochondria and MFN1 and MFN2 are degraded. bbreviations are defined as: IMS inner membrane space; IM inner membrane; OM outer membrane. Supplementary Table 2 rea under the curve (U) measurements for a USP knockout HEK293 cell line expressing GFP-Parkin and a wild type USP isogenic cell line. Data are reported for two mitochondrial matrix proteins (HSPD1, STP5) and three independent USP peptides for three replicate experiments. Supplementary Table 3 Ubiquitin chain type analysis calculations for each chain type in four separate replicate experiments. ll amounts have been normalized by a panel of mitochondrial specific peptides to account for sample to sample variation. Supplementary Table 4 Mass spectrometry identified proteins after a K-GG immunoprecipitation (see methods) for two biological replicates. For each protein identified we are reporting the log2 fold change in ubiquitination during P treatment in the presence or absence of USP overexpression with the significance of the fold change (see Methods). Supplementary Table 5 Peptide abundance levels for proteins identified in the proteomic screen described in Figures 4, S4, and Table S4. 8 WWW.NTURE.OM/NTUREELLIOLOGY 15 Macmillan Publishers Limited. ll rights reserved