Revealing the cellular degradome by mass spectrometry analysis of proteasome-cleaved peptides

Abstract

Cellular function is critically regulated through degradation of substrates by the proteasome. To enable direct analysis of naturally cleaved proteasomal peptides under physiological conditions, we developed mass spectrometry analysis of proteolytic peptides (MAPP), a method for proteasomal footprinting that allows for capture, isolation and analysis of proteasome-cleaved peptides. Application of MAPP to cancer cell lines as well as primary immune cells revealed dynamic modulation of the cellular degradome in response to various stimuli, such as proinflammatory signals. Further, we performed analysis of minute amounts of clinical samples by studying cells from the peripheral blood of patients with systemic lupus erythematosus (SLE). We found increased degradation of histones in patient immune cells, thereby suggesting a role of aberrant proteasomal degradation in the pathophysiology of SLE. Thus, MAPP offers a broadly applicable method to facilitate the study of the cellular-degradation landscape in various cellular conditions and diseases involving changes in proteasomal degradation, including protein aggregation diseases, autoimmunity and cancer.

Figure 1.. Establishing Proteasomal Profiling by Mass Spectrometry Analysis of Proteolytic Peptides (MAPP)

(A) MAPP involves immunoprecipitation of cellular proteasomes after cross-linking. Captured peptides are separated from the proteins and further subjected to mass spectrometry analysis. (B) Immunoprecipitated proteasomes were eluted (Pro. Eluate), separated from co-purified peptides (Pep. Eluate) and separated on SDS-PAGE or analyzed by mass spectrometry. Mock IP: Empty beads. FT: flow through. Wash: IP wash. (C) MAPP using antibodies against PSMA1, an isotype antibody control or GAPDH (Two tailed Welch’s corrected t-test, **P=0.0028, ***P=0.0003; Bars: mean ± standard deviation). (D + E) The fold changes in intensity of peptides identified by MAPP between untreated (UT) and Epoxomicin treated (1uM, 4 hours) cells were ranked. Orange: 2-fold or greater increase, Blue: 2-fold or greater decrease. Of the 4144 peptides identified by MAPP, 3302 peptides (D), or 80% the total (E), were reduced by at least 2 fold in intensity. (F) Schematic representation of the ZsGreen model. (G) Peptide intensity of ZsGreen peptides significantly changed under epoxomicin treatment. The graph shows area under the curve (intensity). Two-way ANOVA (p=0.0036, two degrees of freedom). (H) Thousands of identified peptides were filtered for identification in 2/3 replicates in one biological condition and at least 2-fold intensity above mock IP (beads). (I) High correlation of MAPP protein intensities between two biological replicates (n=1394 proteins, Log 10 transformed LFQ intensities; R²=0.978). (J) Poor correlation between the abundance of proteins identified by MAPP or standard proteomics (n = 979 proteins, Log 10 transformed LFQ intensities; R²=0.117). (K+L) Proteins identified by: solely MAPP ([A]), both standard proteomics and MAPP (shared [B]) and solely standard proteomics ([C]; Supplementary Figure 5A). Protein abundance was inferred from deep proteomics of HEK293 cells ^,. The average abundance for [A] is significantly lower than ([B]; Log 10 transformed LFQ intensities, Mann Whitney s****p <0.0001). Line: Mean error bars: 95% confidence intervals. n=65 or 779 proteins ([A] or ([B], respectively) (K). The amount of protein degraded per hour (Protein Turnover) was inferred for proteins identified by MAPP and standard proteomics from turnover rates ^, (Supplementary Figure 5B). The average protein turnover for [A + B] was significantly lower than ([C]; Log 10 transformed, Two tailed Mann Whitney test ****p <0.0001). Line: Mean error bars: 95% confidence intervals. n=785 or 2509 proteins (MAPP [A+B] and proteome specific [C], respectively (L).

Figure 2.. MAPP-identified peptides reveal properties of endogenous proteasome cleavage

(A+B) A sample MS/MS spectrum of a peptide from Pyruvate kinase (PKM) identified by standard proteomics following tryptic digest (A) or by MAPP (B). (C) The peptides identified by standard proteomics (top) or MAPP (bottom) are shown aligned to the PKM sequence (residues 1 – 151). Color scale indicates peptide abundance (Log10 transformed intensities). (D) Standard proteomics (right, n = 33,747 peptides) identifies peptides with arginine or lysine as the carboxy-terminal residue. By contrast, MAPP (left, n = 6,311 peptides) is not constrained by tryptic cleavage and captures peptides with almost every residue in the carboxy-terminal position. (E) Distribution of the lengths (number of residues) of peptides identified by MAPP (n = 6,311 peptides) and standard proteomics (n = 33,747 peptides). Peptides identified by MAPP are significantly longer than those identified by standard proteomics (Two tailed Mann-Whitney test, P<0.0001). (F) The amino acid motif surrounding the peptide cleavage site (P4 – P4’; n = 2396 unique cleavage sites). (G) Clustering of the peptide cleavage sites based on amino acid properties reveals three distinct motifs: a motif enriched for negatively charged residues (left, n = 701 unique cleavage sites), a motif enriched for hydrophobic residues (center, n = 581 unique cleavage sites), and a non-canonical motif (*) enriched for polar residues (right, n = 1085 unique cleavage sites).

Figure 3.. MAPP profiles distinguish between PBMCs from SLE patients and Healthy individuals.

PBMCs were purified from SLE patients and healthy individuals and analyzed by both MAPP and standard proteomics. (A) Matrix of the pairwise coefficient of determination in each sample of healthy individuals (n = 6) or SLE patients (n = 8). (B + C) Principal component analysis of protein intensities from PBMCs of healthy individuals or SLE patients, as identified by MAPP (B) or standard proteomics (C). (D) The abundance of proteasome subunits or actin (ACTB) in PBMCs was determined by standard proteomics (Log 10 transformed LFQ intensities; error bars: mean ±standard deviation). (E) The abundance of the indicated proteasome subunits in the protein fraction of the proteasome immunoprecipitate. Each subunit was normalized to the mean abundance (Log 2 normalized ratio of LFQ intensities; Bars: mean ± standard deviation). (F) Volcano plot of MAPP-identified proteins in PBMCs from healthy individuals or SLE patients (Log2 normalized ratio of LFQ intensities; -Log10 transformed P values). Two tailed t test, no correction for multiple comparisons. (G) Supervised clustering of proteins whose abundance changed by 2-fold or more between MAPP of healthy individuals and SLE patients (Log 10 transformed LFQ intensities standardized around 0, city block distance function). Identified histones are labeled. (H) Higher levels of histones are identified by MAPP of PBMCs from SLE patients, compared to healthy individuals (Histones lane, SLE/healthy). The ratio of histones in MAPP of SLE patients compared to healthy individuals is significantly higher than the ratio of other MAPP-identified proteins (Bootstrapped Mann-Whitney, P =0.0042; Log 2 transformed ratio of LFQ intensities), including other nuclear proteins (Two tailed Mann-Whitney, P < 0.0001, n = 13 (histones) n = 19 (nuclear) n = 202 (other) proteins. Bars: mean ± standard deviation.