Multiplex Protein Quantitation using itraq Reagents in a Gel-Based Workflow

Multiplex Protein Quantitation using itraq Reagents in a Gel-Based Workflow Purpose Described herein is a workflow that combines the isobaric tagging reagents, itraq Reagents, with the separation power of SDS-PAGE in a global protein expression analysis. Overview itraq Reagents have proven to be a valuable tool in protein expression analysis. Post labeling with the itraq reagents, multiple samples or channels may be pooled and processed as a single sample. Upon MSMS fragmentation, each isobaric tag produces a unique reporter ion. This reporter ion not only distinguishes from which channel the peptides originated but also visualizes the relative quantitation of the peptide between the multiple channels 1. The current approach for protein quantitation using itraq Reagents requires enzymatic digestion followed by itraq Reagent tagging. This results in peptides containing the itraq Reagent on the lysine residues and the peptide N-termini. While this is a very useful workflow, it precludes the effective use of SDS-PAGE as a separation method for complex protein samples. Described herein is a labeling strategy that labels undigested proteins with the itraq Reagent. In this workflow, the itraq reagents label the lysine residues and the single N-terminal of the protein. Post labeling, samples are combined and separated with traditional protein separation techniques such as SDS-PAGE as illustrated in Figure 1. A B C D 114 115 116 Methods 31 30 29 28 Mix SDS-PAGE itraq Reagent Labeling In-Gel Digest LC-MSMS itraq Reagent Reporter Ions Figure 1. The Protein itraq Reagent Workflow Each protein sample is labeled with one of four itraq reagents and pooled. The pooled sample is then loaded onto a single SDS-PAGE lane for molecular weight separation. In gel digestion is performed and the subsequent peptides analyzed by LC-MSMS. Protein samples (25ug) were dissolved in of 0.5M TEAB (triethylammonium bicarbonate (ph 8.5), reduced with TCEP (tris(2 carboxyethylphosphine)) and alkylated with iodoacetamide. Proteins that precipitated during reduction and alkylation were dissolved by adding fresh urea. The four itraq Reagents (114,115,116 and )-(25ug protein) were prepared by adding 50 µl of ethanol to each vial. The resulting solution was then transferred to each sample tube and incubated at room temperature for 2 hours. Samples were then mixed and speed vacuumed to dryness. Linear Mass Analysis of Whole Protein Transferrin samples were reconstituted in 2% ACN, 0.1%TFA and desalted using a C4 ZipTip. Samples were spotted with sinnapinic acid onto a MALDI plate and analyzed on the Voyager Workstation remove TM after workstation. 114 115 116

SDS-PAGE of itraq Reagent Labeled Proteins The labeled proteins were reconstituted in SDS-PAGE sample buffer and loaded and run onto a single 4-20% gradient SDS- PAGE lane. The resultant protein bands were excised and subsequently digested with either trypsin or chymotrypsin in 100 mm ammonium bicarbonate. The extracted peptides were then injected (Agilent 1100 System) onto a 75 µ x 15 cm C18 PepMap column. Peptides were eluted from the column using a gradient and analyzed by on a QSTAR Elite-Hybrid LC/MS/MS System. Data was collected using Analyst 2.0 software and proteins identified with ProteinPilot Software. Validation of Protein itraq Reagent Protocol Two proteins, ovotransferrin and carbonic anhydrase were spiked at varying concentrations into four separate vials of a 6-protein mix (beta-galactosidase, serum albumin, serotransferrin, beta-lactoglobulin, alpha-lactalbumin and lysozyme.) One of the vials contained the 6-protein mix at a lower concentration. Each of the vials was labeled with a different itraq reagent. Post labeling, the samples were pooled and then run on a single SDS-PAGE lane. The lane was excised into ten slices and digested with trypsin. The subsequent peptides were analyzed by LC/MS/MS. Protein identification and quantitation was performed using ProteinPilot Software. Results and Discussion Analysis of intact transferrin before and after labeling with the itraq Reagents (Figure 2) showed an apparent 8499 m/z shift in the molecular weight of the protein. This increase corresponds to 59 itraq Reagent tags. Human transferrin is known to contain 58 lysine residues. This plus the addition of 1 itraq Reagent to the N-terminal of the protein suggest that the labeling reaction is complete. This is further demonstrated by A +2 B +2 Unlabeled Labeled +1 +1 Complete Shift Figure 2. Efficiency of Labeling Analysis of the unlabeled (A) and labeled (B) transferrin by linear MS exhibits a 8499 m/z difference. This is equivalent to 59 itraq Reagent tags. Human transferrin contains 58 lysine residues and 1 N-terminal for a total of 59 sites. 216,000 132,000 78,000 45,700 32,500 18,400 7,600 MW Stds. 6-Protein Mix 6-Protein Mix Labeled MW Shift Figure 3. Efficiency of Labeling on 6-Protein Mix Proteins tagged with the itraq reagent exhibit a clear shift in molecular weight when separated by SDS-PAGE. SDS-PAGE analysis of a 6-protein mix (Figure 3). A comparison between the unlabeled sample and the labeled sample shows a clear shift in the molecular weight of the proteins. In the itraq Reagent Application Kit - Protein protocol, the proteins are labeled prior to enzymatic digestion. This differs from the peptide itraq Reagent protocol that requires tryptic digestion of proteins prior to labeling. This difference in the order of digestion and labeling produces two key differences in the peptides generated in the workflow:

1) Only the peptide corresponding to the n- terminal of the protein will contain an n- terminal itraq tag. This is unlike the peptide labeling workflow where all the n-termini are labeled. 2) Lysines tagged with the itraq Reagent are effectively blocked from cleavage by some digestion enzymes. In keeping with trypsin as a digestion enzyme, cleavage will only occur at the arginine residues generating the equivalence of an arg-c digest and produces less peptides. The particular benefit of After identification with ProteinPilot Software, the highest coverage attained for each protein was recorded (figure 4.) In many cases, higher coverage and therefore higher confident identifications were made under the chymotrypsin digest. Coverage for BSA for example, was only 12 % with the trypsin digest but improved to 41% by switching to chymotrypsin. A trypsin digest, (equivalent to arg-c,) of BSA generates 4 fragments greater than 7 kda with the largest over 11 kda. These peptides correspond to 51% of the BSA sequence that is inaccessible when digesting with Figure 4. Digestion Efficiency Two SDS-PAGE lanes containing the 20-protein mix were processed though ingel digestion with either trypsin or chymotrypsin. Extracted peptides were then analyzed by LC-MSMS. Trypsin 60% Chymotrypsin Percent Coverage 50% 40% 30% 20% 10% 0% Beta Galactosidase Phosphorylase B Lactoperoxidase Trypsin Average Chymotrypsin Average Conalbumin Apotransferrin Bovine Serum Albumin Amyloglucosidase Alpha Amylase Catalase Glutamic Dehydrogenase Peroxidase Lactic Dehydrogenase Proteins Glyceraldehyde 3P Carbonic Anhydrase Glutathione S-Transferase Beta-Lactoglobulin Myoglobin Hemoglobin Lysozyme Cytochrome C larger peptides is a notable reduction in complexity of the generated peptidome. To demonstrate the differences in protein coverage, a comparison between trypsin and chymotrypsin was performed. Two lanes were loaded with equivalent amounts of a 20-protein mixture (gel not shown). Upon staining, each lane was cut into 16 slices of equivalent area. One set of gel plugs was digested with trypsin while the second set was digested with chymotrypsin. The samples were analyzed by LC-MSMS on a QSTAR -Elite mass spectrometer. trypsin. This is not an issue with chymotrypsin, which has more cleavage sites (F, Y, W, L.) Upon further analysis of the digests, it was noted that the chymotrypsin digest produced a larger number of missed cleavages than trypsin (excluding lysine residues.) Chymotrypsin, with a greater number of cleavage sites, would be expected to generate a larger number of low molecular weight peptides (<1000 m/z.) However, the missed cleaves generated under the digestion conditions provided a larger molecular weight

peptidome and improved the protein coverage. The less than optimal performance of chymotrypsin does not affect the quantitation results because the isobaric nature of the tag ensures that the tag does not influence the enzyme specificity and therefore missed cleavages occur at the same rate across the four tagged channels contained within the one sample. Careful adjustment of the chymotrypsin concentration may be necessary to produce a similar digest. The itraq Reagent Protein Protocol was validated by spiking in two proteins at various concentrations into four vials of the 6-protein mix, one of which contained the 6 proteins at a lower concentration (figure 5). The results of the 10 independent ProteinPilot Software searches, one for each trypsin digested gel slice, are summarized in figure 7. When multiple identifications of a protein were found, the identification with the 4700 MS/MS Precursor 2090.03 Spec #1 MC[BP = 116.1, 47129] Protein MW Expected Ratio Observed Ratio Beta-galactosidase 089 1:1:0.3 0.9:0.7:0.8 Ovotransferrin 79181 0.5:1.5:0.2 0.6:1.5:0.2 Serotransferrin 78894 1:1:0.3 1:1:0.4 BSA 70907 1:1:0.3 1.3:1.2:0.3 Carbonic Anhydrase 28983 2:0.4:3 2.1:0.6:3.3 Beta-lactoglobulin 20206 1:1:0.3 1.0:1.0:0.4 Lysozyme 16654 1:1:0.3 1.6:1.1:0.4 Alpha-lactalbumin 16615 1:1:0.3 - Figure5. Validation Sample 25 ug of an 8-protein mix labeled with itraq reagents are loaded onto 2 lanes of and SDS-PAGE gel. The sample contains two proteins in the mix that are artificially regulated. Post staining, the gel lane is excised into 10 plugs for digestion and analysis. highest coverage is displayed. The two spiked proteins, carbonic anhydrase and ovotransferrin exhibit ratios that are very close to the expected ratios of 2.0:0.4:3.0 and 0.5:1.5:2.0 respectively. The extremely low P-values (PVal) for these Figure 6. MSMS of an Ovotransferrin Peptide Ovotransferrin was spiked into the 6-protein mix at an expected ratio of 0.5:1.5:0.2. Analysis of the itraq reporter ions on peptide CVASSHEK*YFGYTGALR yields ratios that are very close.to the expected values. Excellent peptide coverage is maintained. 100 90 116.14 CVASSHEK*YFGYTGALR 4.7E+4 80 116.14 70 114.14 114.14 115:114 116:114 :114 Expected Ratio 0.50 1.50 0.20 60 115.14 Observed Ratio 0.52 1.39 0.23 % Intensity 50.14 114 Mass (m/z) 40 y10(+1) 30 H b8(+1) 20 y4(+1) 145.13 y7(+1) y9(+1) 232.14 y5(+1) y8(+1) b2(+1) 1933.99 b9(+1) y12(+1) 10 402.26 y17(+1) y6(+1) 112.11 b3(+1) b7(+1) 1015.59 8.63 y11(+1) b10(+1) y13(+1) V 301.23 b6(+1) Y 1325.75 b12(+1) 1802.93 1999.07 225.12 408.31 743.38 1018.73 1181.74 1431.84 1675.82 539.36624.31 A F 312.15 904.54 169.11 426.22 0 9 449 889 1329 1769 2209 Mass (m/z)

Figure 7. Summary of Proteins Identified by Trypsin In-Gel Digestion Digested peptides were analyzed by the QSTAR Elite and analyzed with ProteinPilot Software. Both carbonic anhydrase and ovotransferrin exhibit ratios very close to the expected ratios of 2.0:0.4:3.0 and 0.5:1.5:2.0 respectively. The proteins contained within the 6- protein mix exhibit ratios close to the expected ratio of 1.0:1.0:0.3. identifications provide a high degree of confidence that the observed ratios are correct. Additionally, the error factors (EF) denote that the standard deviation of the quantitation is below 20%. The relative ratio of these 6 proteins is close to 1.0:1.0:0.3. The variation of beta-galactosidase from the expected profile (1.0:1.0:0.3) may be due to a pipetting error. ProteinPilot Software identified two contaminant proteins, superoxide dismutase and ubiquitin, exhibiting quantitation profiles similar to carbonic anhydrase. This suggests that they are contaminants form the carbonic anhydrase source. Conclusions The itraq Reagent protein labeling protocol is an effective method of multiplexed global protein quantitation. This has introduced the itraq Reagent workflow to traditional methods of protein separation such as SDS-PAGE. The incorporation of the itraq Reagents onto the intact proteins has proven efficient and complete according to linear MS. Lysine residues, tagged with one of the itraq Reagents, are blocked from cleavage with digestion enzymes. However, switching to alternate enzymes that do not target lysine such as chymotrypsin may attain higher protein coverage and therefore more confident identifications. Proteins identified by ProteinPilot Software from the gel slices exhibited ratios close to the expected values. This data shows that the Protein itraq Reagent protocol is an effective method for global protein quantitation. References 1. Ross, P. L., et al. (2004) Multiplexed Protein Quantitation in Saccharomyces cerevisiae Using Amine-reactive Isobaric Tagging Reagents, Mol. Cell. P 2. S.Wiese, K.. Reidegeld, H. Meyer and B.Warscheid, A Quantitative Proteomic Approach by Isotope- Coded Protein Tagging (ICPT) and Mass Spectrometry, 53rd ASMS Conference, 2005 rot. 3(12), 1154-69 Applera Corporation is committed to providing the world s leading technology and information for life scientists. Applera Corporation consists of the Applied Biosystems and Celera Genomics businesses. Applied Biosystems/MDS SCIEX is a joint venture between Applera Corporation and MDS Inc. For Research Use Only. Not for use in diagnostic procedures. Applied Biosystems and AB (design) are registered trademarks and Applera and itraq are trademarks of Applera Corporation or its subsidiaries in the U.S. and/or certain other countries. MDS and SCIEX are registered trademarks of MDS Inc. Voyager and ProteinPilot are trademarks and QSTAR is a registered trademark of Applied Biosystems/MDS Sciex. All other trademarks are the sole property of their respective owners. 2006 Applera Corporation. All rights reserved. Information subject to change without notice.