Supporting Information for. Top down Proteomics of Large Proteins up to 223 kda Enabled by. Serial Size Exclusion Chromatography Strategy

Transcription

1 Supporting Information for Top down Proteomics of Large Proteins up to 223 kda Enabled by Serial Size Exclusion Chromatography Strategy Wenxuan Cai a,b,1, Trisha Tucholski c,1, Bifan Chen c, Andrew J. Alpert a,d, Sean Mcilwain e,f, Takushi Kohmoto g, Song Jin c, Ying Ge* a,b,c a Department of Cell and Regenerative Biology, University of Wisconsin-Madison, Madison, Wisconsin 53705, USA b Molecular and Cellular Pharmacology Training Program, University of Wisconsin-Madison, Madison, Wisconsin 53705, USA c Department of Chemistry, University of Wisconsin-Madison, Madison, Wisconsin 53706, USA d PolyLC Inc., Columbia, Maryland 21045, USA e Department of Biostatistics and Medical Informatics, University of Wisconsin-Madison, Madison, Wisconsin 53705, USA f UW Carbone Cancer Center, School of Medicine and Public Health, University of Wisconsin- Madison, Madison, Wisconsin 53705, USA. g Department of Surgery, School of Medicine and Public Health, University of Wisconsin- Madison, Madison, Wisconsin 53705, USA. 1 These two authors contributed equally to this work. *To whom correspondence may be addressed: Ying Ge, Ph.D., 8546 WIMR II, 1111 Highland Ave., Madison, WI, ying.ge@wisc.edu; Tel: ; Fax: S-1

2 SUPPLEMENTAL METHODS Preparation of Cardiac Protein Extract The donor heart was maintained in cardioplegic solution before dissection. Dissected tissue was snap-frozen immediately in liquid nitrogen and stored at - 80 C for subsequent analysis. All procedures involving sample handling were performed at or below 4 C at all times to reduce oxidation and degradation of samples. Extraction of the sarcomeric proteins was performed as follows. Cardiac tissue was homogenized using a Polytron homogenizer in 10 vol (ml/g tissue) of wash buffer (5 mm NaH 2 PO 4, 5 mm Na 2 HPO 4, 5 mm MgCl 2, 0.5 mm EGTA, 0.1 M NaCl, 1% Triton X-100, 5 mm DTT, 1 mm PMSF, and 1 mm Na 3 VO 4 containing protease inhibitor cocktail). Homogenate was centrifuged at 17,000 rcf for 3 min at 4 C and the supernatant was discarded. The washing step was repeated twice and the resulting pellet was homogenized in 5 vol (ml/g tissue) of LiCl extraction buffer (5 mm EGTA, 0.1 mm CaCl 2, 0.7 M LiCl, 25 mm TrisHCl (ph 7.5), 5 mm DTT, 1 mm PMSF, 1 mm Na 3 VO 4 and protease inhibitor cocktail) and incubated at 4 C for 10 min to extract sarcomeric subproteome. The homogenate was centrifuged at 17,000 rcf and the supernatant containing the sarcomeric proteome was collected and further centrifuged at 17,000 rcf for 30 min to remove residual pellet. MS-SAFE protease and phosphatase inhibitor cocktail was used in the extraction of sarcomeric proteome to be analyzed by MS and tandem MS (MS/MS). Prior to 1D RPC and 2D ssec-rpc analysis, the sarcomeric protein extract was desalted by washing 200 µl of the aliquot through 10 kda MWCO filters using 0.1% formic acid (FA) in H 2 O twice, 1% FA in H 2 O four times, followed by 3 final washes with 0.1 % FA in H 2 O. The volume of the protein extract recovered after washing was normalized to the starting volume (200 S-2

3 µl) and the sample was collected. Protein concentration after the desalting procedure was approximately 2 mg/ml. Hexafluoroisopropanol was added to the desalted sample to a final concentration of 100 mm and the sample was centrifuged at 16,100 rcf for 30 min at 4 C prior to 1D RPC-MS and 2D ssec-rpc-ms analysis. ssec Column Conditioning and Protocol Brand-new PolyHEA columns (PolyLC, Inc.) were first flushed with 15 column volumes of H 2 O to remove methanol. To equilibrate the stationary phase residues and ensure reproducible retention time, the column was flushed overnight using 1% FA in H 2 O as the mobile phase. Before injection of the samples/analytes, three blank runs injected with H 2 O were performed using the same flow rate and gradient as the sample (100% isocratic elution with 1% FA in H 2 O) to ensure stable baseline and pressure. A flat baseline is most desirable for achieving reproducible separation and UV signal. For ssec experiments, columns were connected with capillary tubing (Valco, PEEK tubing, 100 µm i.d., 2 cm) using two bolts and two PEEK ferrules. The fitting was secured to prevent leakage, but care was taken not to over-tight the fittings. Connection was monitored to check for solvent leaks prior to blank runs and sample injections. A 50 µl injection loop was used for single column SEC, while a 250 µl injection loop was used for 2sSEC and 3sSEC due to larger injection volume. For SEC, 2sSEC, and 3sSEC, 50 µl, 100 µl, and 150 µl protein extract (2 mg/ml) were injected, respectively (corresponding to 100 µg, 200 µg, and 300 µg). SEC/sSEC separation was performed at a flow rate of 0.5 ml/min and fractions were collected at a rate of 1 fraction/min using an automatic fraction collector. The S-3

4 number of fractions to be collected was determined based on the test run prior to the runs with fraction collection. The columns were flushed with water following each use and stored at 4 C. Fractions were collected in 1-minute intervals (0.5 ml per fraction) and subsequently concentrated to 100 µl using 10 kda MWCO centrifugal filters for SDS- PAGE and LC-MS/MS analysis. Tris(2-carboxyethyl)phosphine (TCEP) was added to the ssec fractions to be analyzed by MS and MS/MS to a final concentration of 1 mm to minimize protein oxidation. The expected pressures for SEC, 2sSEC, and 3sSEC for this experimental setup and flow rate (0.5 ml/min) are approximately 1000 psi, 1500 psi, and 2000 psi, respectively. The columns were packed at approximately 7000 psi. Data Analysis and Protein Identification All total ion chromatograms (TIC) were smoothed by the Gauss algorithm with a 3.45 smoothing width. Mass spectra were deconvoluted using the Maximum Entropy algorithm 1 incorporated in the DataAnalysis software. The resolving power was set to 80,000 for proteins that were isotopically resolved online, and 10,000 for proteins that were not isotopically resolved. Tandem mass spectra were exported as.ascii files from the DataAnalysis software and analyzed by the MASH Suite Pro software 2 developed inhouse using an S/N threshold of 3 and minimum fit score of 60%. All fragment ions were manually validated and the mass lists were output for protein identification using the MS- Align+ search algorithm (version 0.7.1). 3 The search was performed against the Uniprot human database (June 16, 2015). Protein identifications were considered good candidates if p-values were below All protein identifications were manually validated following S-4

5 database output using MASH Suite Pro software. Mass tolerance for both precursor ions and fragment ions was set to 10 ppm. Proteoform Profiling Mass spectra were averaged every 0.5 min from 11 min to 45.5 min in the chromatogram ( min, , etc.). The resulting averaged mass spectra were treated as compound mass spectra, and were deconvoluted using the Maximum Entropy algorithm. 1 High-resolution deconvolution using a resolving power of 80,000 was performed for the proteoforms that were isotopically resolved (3-50 kda) in the 1D analysis and 2D analysis (ssec Fractions 7-12). The SNAP algorithm was used to generate the mass lists containing the mono-isotopic masses of the proteoforms for the compound mass spectra that were deconvoluted using high-resolution setting. Retention time window for each proteoform was determined based on the time segments in which the species was found. Parameters for the SNAP algorithm intensity threshold set to 0.01 % with an absolute intensity of 500, and the Quality factor and S/N was set to 0.6 and 3, respectively. Low-resolution deconvolution with a resolving power of 10,000 was performed for the proteoforms that were not well isotopically resolved by the instrument (40, ,000 Da) for both the 1D analysis and 2D analysis (ssec Fraction 1-7). For all low-resolution deconvoluted spectra, the Sum Peak algorithm was used to generate the mass lists. The intensity threshold for 1D and 2D analysis using the Sum Peak algorithm was set to 1000 and 800, respectively, based on relative signal intensity. A 0.01% intensity threshold and S/N of 3 were used for the Sum Peak parameters for both 1D and 2D data processing. It is to be noted that the S-5

6 mass range for the high-resolution and low-resolution deconvolution was overlapped for the inclusive counting of the proteoforms. A mass list of proteoforms found in 1D analysis in mass range 3-50 kda by highresolution deconvolution was exported to Microsoft Excel (Table S1.1). Similarly, mass lists from 2D (ssec Fractions 7-12) of proteoforms in mass range 3-50 kda (high-resolution deconvolution method) were generated for 2D analysis. Molecular mass (M + H + ) for these lists were rounded to the nearest integer to account for cases in which proteoforms were double counted due to the difference in the mass determination using the SNAP algorithm. Following rounding, the mass lists of 2D Fractions 7 12 were combined for a list of total proteoforms in the mass range of 3-50 kda (Table S1.2). The 1D and 2D mass lists described above were further processed in Microsoft Excel to remove redundant proteoforms. There were two cases for having redundant proteoforms within the mass lists. In the first case, a proteoform eluted over a minute and was counted in two adjacent 0.5 min retention windows. To account for these cases, Remove Duplicates function was used for the mass lists generated from 1D and 2D data processing. In the second case, a proteoform could elute in two or more ssec fractions in the same retention time window (only applicable to the 2D analysis). In order to account for these redundant proteoforms, the mass lists from 2D analysis were combined and Remove Duplicates function was applied to remove redundancies between different ssec fractions. Absolute intensity cutoffs of 1,000 and 500 were applied for 1D and 2D mass lists, respectively, due to a relative low intensity (2-fold reduction) of the 2D data compared to 1D. S-6

7 Manual processing of the 1D and 2D compound mass spectra for Fractions 1 7 was performed to generate a list of proteoforms greater than 50 kda (Table S1.3). First, the mass spectra of the 2D analysis (ssec Fraction 1-7) were manually analyzed to identify proteoforms > 50 kda, and determine their retention time. Based on the proteoforms > 50 kda found in the 2D analysis and their retention time, mass spectra of the 1D analysis were analyzed for visible charge states of the proteoforms detected in 2D analysis. Cutoffs for including the potential high MW proteins found in 1D were determined based on having higher than 0.01% intensity relative to the m/z peak of the highest intensity, within the same compound mass spectra. For a comparison of proteoforms detected in 1D and 2D analysis, non-redundant mass lists containing proteins of 3 50 kda were combined for a total count of proteoforms detected between the two. Duplicates were removed from this combined list to reveal the number of species detected by both 1D and 2D analysis. The proteoforms > 50 kda that were found in both 1D and 2D analysis was determined manually. To determine the total proteoform count for 1D and 2D analysis, the mass lists containing proteoforms in the mass range 3 50 kda were combined with the lists of those > 50 kda that were generated manually. Table S1: Mass list of proteoforms detected with 1D RPC and 2D ssec-rpc analysis Mass lists of proteoforms in the mass range of 3 50 kda detected with 1D RPC (Table S1.1) and 2D ssec-rpc (Table S1.2) analysis and proteoforms with MW larger than 50 kda detected with 1D and 2D analysis (Table S1.3). Table S1 is in excel format and attached separately. S-7

8 Online CAD fragmentation Table S2: Targeted MS/MS of high MW proteins in ssec Fraction 5 ssec Fraction 5 - MS/MS MS+ Segment RT (min) Mr (Da) Isolation range (m/z) CAD energy (ev) i ii iii iv Table S3: Targeted MS/MS of high MW proteins in ssec Fraction 7 ssec Fraction 7 MS/MS MS+ Segment RT (min) Mr (Da) Isolation range (m/z) CAD energy i ii iii iv v vi vii viii ix x xi xii xiii xiv Though proteins greater than 50 kda were not identified using this method, online CAD of the kda protein (Figure 4) yielded a complex tandem mass spectrum with over 400 fragment ions (Table S4), which were all manually validated using the MASH Suite Pro. Additionally, large fragment ions with high charge states (9 +) were isotopically resolved in the S-8

9 tandem mass spectrum (Figure S10). However, the search algorithm did not match the tandem mass spectrum to any protein in the database with high confidence, suggesting that more advanced algorithms for top-down protein identification, independent of the precursor mass, is needed for confident identification of high MW proteins. Furthermore, other fragmentation methods, such as electron-based dissociation and ultraviolet photodissociation method, can be utilized in the 2D ssec-rpc-ms pipeline to improve protein identification and characterization. Table S4: Fragment ions generated by targeted MS/MS (CAD, 15 ev) of kda protein Mass list of fragment ions generated by online CAD (15 ev) fragmentation of a kda protein. All fragment ions listed have been manually validated using Mash Suite Pro software. Table S4 is in excel format and attached separately. S-9

10 SUPPLEMENTAL FIGURES Figure S1: Comparison of SEC with pore sizes A) 300 Å, B) 500 Å, and C) 1000 Å for the fractionation of sarcomeric protein extract. Top panel shows representative UV chromatograms and bottom panel shows corresponding SDS-PAGE analysis of fractions collected and pooled from two technical replicates (12.5 %). S-10

11 Figure S2: Three technical replicates of ssec experiment demonstrated highly reproducible separation: A) UV chromatograms for three 100 µl (200 µg total protein) injection of the sarcomeric protein extract for ssec Å column series. B) SDS-PAGE analysis of ssec fractions collected from individual runs. TR: technical replicate. Red and blue marks to the right of each gel represent molecular weight markers. S-11

12 Figure S3: ssec prior to RPC-MS analysis did not alter relative abundance of proteoforms of ctni: A) TIC of 1D RPC-MS analysis (black trace) of the whole sarcomeric protein extract aligned with TIC of RPC-MS (purple trace) analysis of ssec Fraction 10. Highlighted region ( min) corresponds to the elution of ctni. Top-down MS of ctni for 1D and 2D are shown. B) Deconvoluted mass spectra of ctni proteoforms showing similar relative abundance between 1D and 2D analysis. S-12

13 Figure S4: ssec fractionation enabled detection of high MW proteins: Top-down mass spectra with insets showing the zoomed-in views of the individual charge states of the corresponding proteins with MW kda, 80.9 kda, 65.2 kda, 72.3 kda, 69.6 kda, 62.7 kda, 53.5 kda, and 47.6 kda. S-13

14 Figure S5: 3sSEC fractionation separates intermediate MW proteins, which were otherwise suppressed by coeluted low MW proteins. A) TIC of 1D RPC-MS (black trace) analysis for the whole sarcomeric protein mixture aligned with TICs of RPC-MS analysis for ssec Fractions 7 and 12 (green and blue traces, respectively) of the sarcomeric protein extract. Proteins with MW 47.4, 47.3, 47.2, and 50.9 kda were revealed in Fraction 7, which coeluted with abundant sarcomeric protein MLC1 (21.8 kda) within retention time window min in the 1D analysis. B) Zoom-in views of the mass spectra of the proteins found in 1D or 2D analysis showing ions from the MCL1 (pink highlight), 47.4 kda (indigo highlight), 47.3 kda (orange highlight), 47.2 kda (purple highlight), and 50.9 kda (violet highlight). Deconvoluted mass spectrum of the each proteoform is shown to the right of the raw spectrum. S-14

15 Figure S6: ssec-rpc-ms reveals low abundant protein PTMs: A) TIC of 1D RPC-MS (black trace) analysis for the whole sarcomeric protein mixture aligned with TICs of RPC-MS analysis for ssec Fraction 6 (purple trace) of the sarcomeric protein extract. The corresponding top-down mass spectra of the proteins eluted are shown at the right panel. A protein with MW 79 kda was revealed in ssec Fraction 6, which regularly coeluted with smaller proteins the in 1D RPC-MS analysis and remained undetected. B) Zoom-in views of the top-down mass spectra and the corresponding deconvoluted spectra of the 79 kda protein detected in RPC-MS analysis of ssec Fraction 6 (2D ssec/f6-rpc). Deconvoluted mass spectrum (resolving power 10000) revealed a proteoform with 80 Da mass shift from the major peak, indicating phosphorylation of the 79 kda protein. S-15

16 Figure S7: 2D ssec-rpc enabled targeted MS/MS analysis of high MW proteins in 3sSEC Fraction 5. Detailed information regarding the molecular weights and retention windows for individual proteins (numbered i-iiii) analyzed by MS/MS is shown in Table S2. A) TIC of the first MS experiment (Run 1, orange trace) aligned with TICs of the second experiment (Run 2, purple and teal trace) for targeted MS/MS analysis (teal trace) on the proteins of interest in defined time segments. For the remaining time segments wherein no proteins of interest were found, MS data were collected (purple trace). B) Run 1 allowed for the determination of retention windows for individual proteins of interest, and the resulting top-down mass spectra and the deconvoluted spectra allowed for the determination of precursor MW. Online MS/MS analysis by CID yielded complex high-resolution tandem mass spectra for individual proteins of interest. Collisional energy for each protein of interest is highlighted in red font. S-16

17 Figure S8: 2D ssec-rpc enabled targeted MS/MS of high/intermediate MW proteins in 3sSEC Fraction 7. The detailed information regarding the molecular weights and retention windows for individual proteins (numbered i-xiv) analyzed by MS/MS is shown in Table S3. A) TIC of the first MS experiment (Run 1, blue trace) aligned with TICs of the second experiment (Run 2, purple and orange trace) for targeted MS/MS analysis (purple trace) on the proteins of interest in defined time segments. For the remaining time segments wherein no proteins of interest were found, MS data were collected (orange trace). B) Run 1 allowed for the determination of retention windows for individual proteins of interest, and the resulting top-down mass spectra and the deconvoluted spectra allowed for the determination of precursor MW. Online MS/MS analysis by CID yielded complex high-resolution tandem mass spectra for individual proteins of interest. Collisional energy for each protein of interest is highlighted in red font. S-17

18 Figure S9: A) Top-down mass spectrum and the deconvoluted spectrum for the 47.4 kda protein (ssec fraction 7, RT min) determined by first MS run (Run 1). Targeted MS/MS experiment was performed by isolating precursor ions isolated across m/z in Run 2, yielding complex high-resolution tandem mass spectrum. Insets show zoom-in views of the isotopically resolved fragments. 44 bonds of 440 were cleaved, yielding 10 % bond cleavage. B) A novel proteoform of trifunctional enzyme subunit β was identified with high confidence (2.4 E-14). C) Representative b and y fragment ions with high mass accuracy validated using MASH Suite Pro. S-18

19 Figure S10: Large fragment ions with high charge state are isotopically resolved in the tandem mass spectrum of the kda protein. A), B), C), and D) show zoomed-in portions of the tandem mass spectrum from m/z Ions with Da mass shift were also identified by manual interpretation of the spectra, indicating the presence of phosphorylation, which is consistent with the multiply-phosphorylated proteoforms seen in the mass spectrum of the kda protein. S-19

20 REFERENCES (1) Ferrige, A. G.; Seddon, M. J.; Jarvis, S.; Skilling, J.; Aplin, R. Rapid Comm Mass Spectrom 1991, 5, (2) Cai, W.; Guner, H.; Gregorich, Z. R.; Chen, A. J.; Ayaz-Guner, S.; Peng, Y.; Valeja, S. G.; Liu, X.; Ge, Y. Mol Cell Proteomics 2016, 15, (3) Liu, X.; Sirotkin, Y.; Shen, Y.; Anderson, G.; Tsai, Y. S.; Ting, Y. S.; Goodlett, D. R.; Smith, R. D.; Bafna, V.; Pevzner, P. A. Mol Cell Proteomics 2012, 11, M S-20