| Metadata |
| datasetIdentifier | PASS00554 |
| datasetType | SRM |
| submitter | Christopher Colangelo <christopher.colangelo@yale.edu> |
| submitter_organization | Yale University |
| lab_head_full_name | Christopher Colangelo |
| lab_head_email | christopher.colangelo@yale.edu |
| lab_head_organization | Yale University |
| lab_head_country | United States |
| datasetTag | Rat_Brain_PSD |
| datasetTitle | 90 min LC-MRM/SRM Rat/Mouse Brain PSD assay that relatively or absolutely quantifies 112 targeted proteins (337 peptides) by quantifying 5 MS/MS transitions from each of 3 peptides/protein. |
| publicReleaseDate | 2014-12-31 00:00:00 |
| finalizedDate | 2014-08-27 07:33:50 |
| summary | We present a comprehensive workflow for large scale (>1000 transitions/run) label-free LC-MRM proteome assays. Innovations include automated MRM transition selection, intelligent retention time scheduling (xMRM) that improves Signal/Noise by >2-fold, and automatic peak modeling. Improvements to data analysis, including a novel Q/C metric, Normalized Group Area Ratio (NGAR), MLR normalization, weighted regression analysis, and data dissemination through the Yale Protein Expression Database. As a proof of principal we developed a robust 90 minute LC-MRM assay for Mouse/Rat Post-Synaptic Density (PSD) fractions which resulted in the routine quantification of 337 peptides from 112 proteins based on 15 observations per protein. Parallel analyses with stable isotope dilution peptide standards (SIS), demonstrate very high correlation in retention time (0.997) and protein fold change (0.94) between the label-free and SIS analyses. Overall, our first method achieved a technical CV of 11.4% with >97.5% of the 1697 transitions being quantified without user intervention, resulting in a highly efficient, robust, and single injection LC-MRM assay. |
| contributors | Christopher M. Colangelo, Gordana Ivosev, Lisa Chung, Thomas Abbott, Mark Shifman, Fumika Sakaue, David Cox, Rob Kitchen, Lyle Burton, Stephen A Tate, Erol Gulcicek, Ron Bonner, Jesse Rinehart, Angus C. Nairn, Kenneth R. Williams |
| publication | Colangelo, CM et. al. Development of a Highly Automated and Multiplexed Targeted Proteome Pipeline and Assay for 112 Rat Brain Synaptic Proteins. Proteomics, Submitted |
| growth | |
| treatment | |
| extraction | Rat cortex was homogenized using a Dounce tissue grinder in homogenization buffer (0.32 M sucrose, 20 mM HEPES, pH 7.4) with protease and phosphatase inhibitors (Sigma-Aldrich, St. Louis, MO). Nuclear and unhomogenized cell contaminants were removed by low-speed centrifugation at 2,000 x g for 10 min, followed by a high-speed centrifugation at 15,000 x g for 10 min to obtain the pellet containing synaptosome/synaptoneurosomes . The pellet was resuspended in the homogenization buffer and then applied to a Percoll (GE Healthcare, Waukesha, WI) gradient (3%, 10%, 15%, 23%) and centrifuged at 25,000 x g for 12 min. The interfaces between 10% and 15% (F3, fraction 3, crude synaptosome) and 15% and 23% (F4, fraction 4, pure synaptosome) were collected and subjected to hypotonic lysis (20 mM HEPES, pH 7.4, 1.0 mM DTT). For downstream protein identification, either F3 or F4 fractions were used, whereas for protein quantitation studies we combined F3 and F4 fractions. Subsequent purification of F3 and/or F4 was performed by centrifugation at 25,000 x g for 30 min and a detergent treatment (0.32 M sucrose, 20 mM HEPES, pH 7.4, 0.75% Triton X-100). The PSD fraction was then collected by centrifugation at 63,000 x g for 30 min. PSD samples were then washed with PBS and centrifuged again at 63,000 x g for 30 min. Final PSD samples were dissolved by sonication in urea buffer (7 M urea, 0.1 M Tris-HCl pH 7.5). |
| separation | |
| digestion | Cysteines were reduced with 4.1 mM DTT (10 min at 60 °C) and reactions were quenched on ice. Carbamidomethylation of cysteines was carried out for 30 min at room temperature in the dark using 8.3 mM IAA. Excess IAA was quenched with DTT and the reaction was diluted to ~1 M urea with 200 µl of 90 mM Tris-HCl buffer pH 8.0 (23°C) containing 2 mM CaCl2. Digestion with trypsin was carried out at 37°C for 15 h using a trypsin to protein ratio of 1:15. The reaction was quenched with 12 µl 20% TFA. Peptides were desalted by reversed phase chromatography using C18 UltraMicroSpin columns. Peptides were dried in a vacuum centrifuge, dissolved in 20 µl 3:8 v/v 70 % FA/0.1% TFA and protein quantitation was determined by hydrolysis and amino acid analysis (AAA). The peptide concentration was adjusted to 0.6 µg/µl for LC-MS/MS and 0.2 µg/µl for LC-MRM using 0.1% TFA. |
| acquisition | For label-free samples, 1 μg (5 μl of 0.2 μg/μl) of digested peptides were injected. For SIS “gold-standard” samples, the SIS peptide mixture was added to resuspended digests for a final concentration of 100 fmol/μl and 0.2 μg/μl PSD digest. Liquid chromatography was identical to sMRM. For extended MRM (xMRM) data collection, label-free samples used an xMRM method containing 1697 transitions with peak windows of 6 min and a cycle time of 2.5 sec. To trigger acquisition of additional transitions we used a threshold of 200 counts which was 4-fold above the highest observed noise level of about 50 counts. We note that the xMRM approach can be used on AB Sciex triple quad (e.g. 5500, 6500) and triple quad/Q-TRAP (e.g. 5500 QTRAP, 6500 QTRAP) instruments “Gold-standard” SIS spiked samples used an xMRM assay with 1371 transitions with peak windows of 5 min and a cycle time of 2.5 sec. The resulting LC-xMRM data was processed and quantitated using MultiQuant™ 2.1 software (research version). The SignalFinder™ 2 algorithm (research version) was used to integrate and score peak groups. Data results from MultiQuant were exported and uploaded into YPED [3, 4].The data was also processed via MultiQuant 2.1 MQ4 and Skyline (1.4) using default parameters for each. The Skyline data results have been uploaded to Panorama at https://www.panoramaweb.org/labkey/project/Yale%20Keck/PSD%20MRM/begin.view. |
| informatics | We developed an R package entitled MRMandSWATHgraphics, which performs data quality assessment and creates associated graphical plots from either MultiQuant or AB Sciex Peakview™ result tables. We also created a MATLAB package entitled Protein Quant Data Analysis that provides normalization and fold-change analysis. The software includes standard normalization techniques, such as quantile and total-area sum normalization (TAS), and also includes a new algorithm called Most-Likely Ratio Normalization (MLR) [19]. The fold-change calculation was based on a weighted fold-change analysis [20] with modifications for label-free analysis. All of the raw data collected on the integrated transition peak areas were used as input for the fold-change determinations. Following MLR normalization the protein level fold changes were calculated as follows: (i) For each set of biological replicates, a table of weighted average areas was determined using the transition reproducibility values determined during normalization—'weighted area response'. (ii) Weighted analysis of variance was used to determine the likelihood of difference between the experimental test conditions. The output was in essence a P value as if a t-test had been performed between the different experiments, but the use of weighted values better accounts for poor-quality values in the data being processed. (iii) The ”signal-quality” values and the analysis of variance values (from step ii) were used to determine the peptide fold-change values via a weighted-average fold-change calculation. (iv) A peptide signal-quality table was determined from the transition signal-quality table by calculating the median signal quality for each set of transitions. (v) The peptide variance was then determined by calculating the summed weighted average for the transition analysis of variance using the signal-quality table as the weighting factor for the individual transitions. (vi) A protein fold-change value was determined as previously described [19] except that the protein value was again a weighted average using both the peptide signal quality (step iv) and the peptide variance (step v) values as weights. The output was a fold change up or down for each protein and a confidence value as determined from the peptide variance and the peptide signal-quality values. All calculations were carried out in Matlab by executing custom-coded algorithms and data exported either in figure format or as text as previously described [19]. R scripts were then used to generate fold change graphics such as bar charts and scatter plots. All R scripts used in the manuscript are attached as a zip file (“Colangelo_2013_ManuscriptRscript.zip”). |
| instruments | AB Sciex 5500 QTRAP
AB Sciex TripleTOF 5600 |
| species | Rattus norvegicus |
| massModifications | static: C+57.021464
variable: K+8.014199, R+10.008269 |
