Abstract

Universal proteomics sample preparation is challenging because of the high heterogeneity of biological samples. Here we describe a novel mechanism that exploits the inherent instability of denatured proteins for nonspecific immobilization on microparticles by protein aggregation capture. To demonstrate the general applicability of this mechanism, we analyzed phosphoproteomes, tissue proteomes, and interaction proteomes as well as dilute secretomes. The findings present a practical, sensitive and cost-effective proteomics sample preparation method. Universal proteomics sample preparation is challenging because of the high heterogeneity of biological samples. Here we describe a novel mechanism that exploits the inherent instability of denatured proteins for nonspecific immobilization on microparticles by protein aggregation capture. To demonstrate the general applicability of this mechanism, we analyzed phosphoproteomes, tissue proteomes, and interaction proteomes as well as dilute secretomes. The findings present a practical, sensitive and cost-effective proteomics sample preparation method. Dedicated sample preparation for shotgun proteomics is essential for removing impurities and interfering species which may affect peptide chromatography, ionization during the electrospray process, and sequencing by mass spectrometers. To represent the in vivo state of the global proteome including membrane-bound proteins, it is of high importance to ensure complete lysis of cells and tissues before protease digestion. This typically requires strong detergents that are difficult to remove afterward, however crucial to avoid signal interference during MS analysis. Considerable developments have been made based on a variety of different biochemical principles which use filters, traps, or protein precipitation techniques which address different sample types (1.Wiśniewski J.R. Zougman A. Nagaraj N. Mann M. Universal sample preparation method for proteome analysis.Nat. Methods. 2009; 6: 259Crossref Scopus (5043) Google Scholar, 2.Shevchenko A. Tomas H. Havli J. Olsen J.V. Mann M. In-gel digestion for mass spectrometric characterization of proteins and proteomes.Nat. Protoc. 2007; 1: 2856-2860Crossref Scopus (3531) Google Scholar, 3.Zougman A. Selby P.J. Banks R.E. Suspension trapping (STrap) sample preparation method for bottom-up proteomics analysis.Proteomics. 2014; 14: 1000-1006Crossref Scopus (173) Google Scholar). However, a primary challenge remaining is the development of a universal sample preparation method that has the potential to scale across different sample amounts, which typically range from ng to mg of starting material. Moreover, such a method needs to be compatible with different lysis buffers, biological material (i.e. cell lines, tissues), robust, reproducible, cost effective, and perhaps above all; practical. Although several methods have been developed to individually address different proteomics sample preparation challenges, a simple solution spanning all sample types remains elusive. Here we report a mechanism, termed protein aggregation capture (PAC) 1, which uses the phenomenon of nonspecifically immobilizing precipitated and aggregated proteins on any type of sub-micron particles irrespective of their surface chemistry. We explore the fundamental process underlying this phenomenon behind methods such as SP3 and determine the optimal parameters leading to effective sample preparation for shotgun proteomics analysis by mass spectrometry of different sample types. Our developments demonstrate the potential for low cost, simple, robust and sensitive sample preparation procedures for proteomics analysis, which can be easily implemented in any setting with great potential for full automation. Chemicals were purchased from Sigma-Aldrich (Søborg, Denmark) unless otherwise specified. 1 μm diameter Sera-mag carboxyl magnetic beads (cas # 45152105050250 and cas # 65152105050250) were purchased from GE-Healthcare (Brøndby, Denmark). 0.5 μm diameter SIMAG-Sulfon (cas # 1202), SiMAG-Q (cas # 1206), and SiMAG-Octadecyl (cas # 1301) magnetic beads were all purchased from Chemicell GmbH (Berlin, Germany). 5–10 μm average diameter HILIC, TiO2, and Ti-IMAC magnetic beads were purchased from ReSyn Biosciences (Edenvale, Gauteng, South Africa). Carbonyl-iron powder with 5–9 μm diameter grain size was purchased from Sigma-Aldrich (cas # 44890). Human bone osteosarcoma epithelial (U2OS) and human epithelial cervix carcinoma (HeLa) adherent cells were grown in DMEM media (Gibco, Thermo Fisher Scientific, Waltham, Massachusetts) supplemented with fetal bovine serum (Gibco) at 10% final. The media also contained penicillin (Invitrogen, Thermo Fisher Scientific) at 50 U/ml and streptomycin (Invitrogen) at 100 μg/ml. Cells were grown in a humidified incubator at 37 °C with 5% CO2. In all cases, cells were grown to 80–90% confluency before harvesting with different lysis buffers in Nunc petridishes (100 or 150 mm diameter). To generate stably expressing GFP-TTP cells under a doxycycline inducible promoter, ZFP36/TTP was gateway cloned into a pCDNA4/TO/GFP expression vector by gateway cloning (Thermo Fisher Scientific), and co-transfected with pcDNA6/TR (Thermo Fisher Scientific) into U2OS cells. Cells were selected with zeocin and blasticidin for 14 days, after which individual clones were picked and screened for GFP-TTP expression. For SILAC labeling, cells were cultured in media containing either l-arginine and l-lysine (Light), l-arginine [13C6] and l-lysine [2H4] (Medium) or l-arginine [13C6-15N4] and l-lysine [13C6-15N2] (Heavy; Cambridge Isotope Laboratories, Tewksbury, Massachusetts). RAW264.7 macrophage cells were derived from Mus musculus and grown in 10% in DMEM media with 10% FBS in 150 mm diameter Nunc petridishes. The media was removed, and cells were washed with PBS before addition of phenol-red free DMEM media without serum, penicillin, and streptomycin. Cells were stimulated with lipopolysaccharids (LPS) with 1 μg/ml for 4 h. Four hundred microliters of the media was removed and processed for secretome analysis and filtered through 0.22 μm filter (Sartorius #16532) before further processing. Cells lysis as presented in this study was performed with either one of the three buffers: (1) 6 m guanidine hydrochloride in 100 mm tris hydrochloride (Life technologies, Carlsbad, California) at pH 8.5, (2) 1% SDS in 100 mm 100 mm Tris Hydrochloride (pH 8.5) or (3) 0.1% NP-40 in 1× phosphate buffered saline solution (pH 7.4) containing β-glycerol phosphate (50 mm), sodium orthovanadate (10 mm), and protease inhibitor mixture (Roche, Basel, Switzerland). In all cases, supernatant from adherent cell plates was removed and the cells were rinsed with ice cold 1× PBS before the addition of the lysis buffer. Guanidine hydrochloride buffer was pre-heated to 99 °C before the addition to the cell plates. After the addition of guanidine or SDS buffer, cells were manually collected and heated at 99 °C for 10 min followed by sonication using a probe to shear RNA and DNA. Proteins were immediately reduced and alkylated with the 10 mm tris(2-carboxyethyl)phosphine (TCEP) and 11 mm 2-chloroacetamide (CAA) for samples lysed with guanidine hydrochloride or SDS lysis buffer before further processing. For cells lysed using 0.1% NP-40 buffer, 1 μl of benzonase (≥250U/ul) was added to the lysis solution for 1 h on ice. The lysis solution was centrifuged at 5000 × g for 10 min and the supernatant was transferred to a new tube. GFP-TTP immunoprecipitations were performed using GFP-Trap magnetic agarose beads (Chromotek, Martinsried, Germany) according to manufacturer's instructions. Cell lysis and immunoprecipitations were carried out using low salt EBC lysis buffer (150 mm NaCl; 50 mm TRIS pH 7.5; 1 mm EDTA; 0,5% NP-40) for 1 h followed by 5 washes with the same buffer. Proteins were eluted by boiling in 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.1 m Tris-HCl (pH 6.8). Aggregation was induced by the addition of acetonitrile (unless stated otherwise) and magnetic microparticles were added to solution followed by mixing to uniformly mix the bead solution. The solution was allowed to settle for 10 min and beads were separated using a magnet for 60 s. Magnetic microspheres were retained by magnet and the supernatant was removed by vacuum suction. In the case of analysis by protein gel electrophoresis (SDS-PAGE), the supernatant was transferred to new tubes. Beads were washed using acetonitrile once followed by one wash with 70% ethanol. See extended methods for experiment specific protocols. Samples and washes were prepared for analysis by protein gel electrophoresis (SDS-PAGE) by the addition of 4× LDS sample buffer (Thermo Fisher Scientific) to 1× final, and DTT (100 mm). Samples were heated for 10 min at 80 °C. For eluting bead bound protein aggregates, LDS buffer (containing DTT) was added to bead containing solutions and the mixture was heated for 10 min at 80 °C. Heated beads in LDS buffer were separated by magnet and the supernatant was analyzed by SDS-PAGE or transferred to a new tube and stored at −20 °C until SDS-PAGE analysis. Samples were loaded on NuPAGE 4–12% Bis-Tris protein gel (Thermo Fisher Scientific) and ran with 200 volts for 40 min. Gels were stained for 15 min using instant Blue (Expedeon, San Diego, California) and destained overnight with Milli-Q water and scanned on EPSON V750 PRO. Proteins were aggregated on microspheres and washed as described above. For on-bead digestion, 50 mm HEPES buffer (pH 8.5) was added to submerge microspheres. Proteins were reduced and alkylated with the 5 mm tris(2-carboxyethyl)phosphine (TCEP) and 5.5 mm 2-chloroacetamide (CAA) for 30 min if not treated immediately after lysis. Lys-c (Wako Chemicals) was added at ratio of 1:200 (to protein) and allowed to react for 1 h at 37 °C followed by the addition of trypsin at a ratio of 1:100 (unless specified otherwise). Trypsin digestion was allowed to occur overnight at 37 °C. Beads were separated by magnet and the supernatant was transferred to new tube and acidified. In-solution digestion with guanidine hydrochloride buffer was carried out under similar reduction and alkylation conditions. Lys-c was added to solution and allowed to react for 1 h at 37 °C. The concentration of guanidine hydrochloride concentration was reduced to >1 M before the addition of trypsin for overnight digestion. Solution was acidified by with 1% trifluoroacetic acid (TFA) and centrifuged for 5 min at 5000 × g and the supernatant transferred to new tubes. Peptide mixtures were clarified with solid phase extraction. Briefly, hydrophobic C18 sep-pak (Waters Corporation, Taastrup, Denmark) were prepared by washing with acetonitrile and 0.1% TFA, followed by loading of the acidified peptide mixtures by gravity. Sep-paks were washed with 0.1% TFA and peptides were eluted using 50% Acetonitrile (0.05% TFA). Organic solvent was evaporated and peptides concentrated using a speedvac before MS analysis. We used skeletal muscle which were isolated for previously published study (See extended methods Schönke et al.). Frozen gastrocnemius muscles were crushed using mortar and pestle. Powdered muscle was homogenized using Ultra Turrax T8 homogenizer (IKA Labortechnik, Staufen im Breisgau, Germany) in 4% SDS buffer (100 mm Tris-HCl, pH 7.4). Protein lysates were boiled at 100 °C for 5 mins. Lysates were sonicated using a tip and centrifuged at 16,000 × g for 10 mins followed by reduction and alkylation as described above, the supernatant was then processed using FASP or PAC or frozen until further analysis. Urea powder was added to 400 μl of filtered cell culture supernatant for a final 2 m concentration, and pH for digestion adjusted with 40 μl Tris 1 m pH 8.5. FASP protocol was adapted from as previously described (1.Wiśniewski J.R. Zougman A. Nagaraj N. Mann M. Universal sample preparation method for proteome analysis.Nat. Methods. 2009; 6: 259Crossref Scopus (5043) Google Scholar). Samples were heated for 10 min at 56 °C and centrifuged (7000 × g, 10 min). Following centrifugation steps were performed applying the same conditions. Ultracel-30 membrane filters (#MRCF0R030, Millipore, Burlington, Massachusetts) were cleaned with 10% acetonitrile and 15% methanol, filters were centrifuged and equilibrated with 200 μl urea buffer (2 m, 0.1 m Tris, pH8.5), and centrifuged again. Samples were added into the filters and the filters were centrifuged and washed two times with urea buffer. Reduction was performed by 1 μl of 0.5 m TCEP in 100 μl urea buffer. The device was centrifuged, and alkylation was performed by 1 μl of 550 mm CAA in 100 μl urea buffer for 30 min in the dark. Filters were centrifuged and 200 μl urea buffer was added before another centrifugation. Subsequently, 4 μl of 0.5 μg/μl lys-c in 40 μl urea buffer was added for 3h at 37 °C with gentle orbital shaking. 4 μl of 0.5 μg/μl trypsin was added for an overnight digestion in the wet chamber at 37 °C with gentle orbital shaking. 1.5 ml eppendorf tubes were cleaned with absolute methanol and air dried, before inserting the filter device, which was then centrifuged. Subsequently 40 μl of milli-Q water was added followed by centrifugation. The enzymatic digestion was stopped by acidifying the sample to pH 1000 more phosphorylated peptides and 779 localized sites were identified on average using microsphere protein aggregation followed by protease digestion compared with standard in-solution digestion. Moreover, high degree of overlap for localized sites was found between the two methods (Fig. 3C) and no bias in the phosphopeptide enrichment was observed between the two experiments as we achieved an enrichment efficiency >99% for all replicates (supplemental Fig. S2B). Surprisingly, we did not observe a major difference between missed cleavage rates for phosphopeptides between the two methods (supplemental Fig. S2C). We next assessed the potential of using magnetic microparticles for proteomics analysis of organs and tissues. This can be particularly challenging as it often requires harsh solubilization buffers for efficient protein extraction from hard and soft tissues. To test aggregation on microparticles we used skeletal muscle tissue samples from Mus musculus. After homogenization and solubilization in 4% SDS lysis buffer, we examined protein recovery and digestion by using either the established filter-aided sample preparation (FASP) protocol or via aggregation of proteins on magnetic microspheres with sulfonic acid surface chemistry. Significantly higher peptide recovery (>2 fold) after Lys-C/Trypsin digestion was observed using microspheres from initial starting material of 1.8 mg (as determined by tryptophan assay) (12.Wiśniewski J.R. Gaugaz F.Z. Fast and sensitive total protein and peptide assays for proteomic analysis.Anal. Chem.