A variety of stable isotope reagents have been developed for relative quantification in proteomics, including ICAT, SILAC, 18O, AQUA, and iTRAQ1,2,3. Most methods enable quantification in the full MS scan and peptide identification based on subsequent fragmentation (MS/MS) of precursor ions, with the exception of iTRAQ where both the identification and quantification is performed in the MS/MS scan.
This requirement was thought to make the iTRAQ reagent incompatible with ion traps which, upon MS/MS analysis of large m/z iTRAQ-labeled peptides, would suffer from the "1/3 Rule" and be unable to effectively analyze the iTRAQ reporter ions at m/z 114117.
However, the Thermo Electron (www.thermo.com) Finnigan LTQ performs additional scans beyond just MS/MS, which allows quantification of iTRAQ reporter ions in a subsequent scan. This new method utilizes MSn for direct analysis of iTRAQ labels using the Finnigan LTQ, with peptide identification in the MS/MS scan and quantification of the tag in a subsequent, MS3 scan.
The iTRAQ reagent preferentially labels the epsilon amino group from lysine, followed by other free amino groups in a peptide digest. Digestion with common proteases that cleave at lysine and arginine residues results in a peptide with a C-terminal lysine and will always yield a lysine y-1 ion in MS/MS scans.
The addition of the iTRAQ reagent to lysine increases the mass of the singly charged y-1 ion to 291. Selecting the 291 ion for MS3 results in fragmentation of the labile iTRAQ tag and efficient conversion to iTRAQ reporter ions. These ions can be easily resolved in the MS3 scan and used for quantification with sensitivity and precision comparable to MS/MS methods.
The goal is to develop a workflow on the Finnigan LTQ that supports quantitative proteomics with iTRAQ reagents.
iTRAQ Labeling of Peptides
A mixture of six enzymatically digested proteins was obtained from Michrom Bioresources (Auburn, CA). The proteins included horse cytochrome-C, horse myoglobin, bovine beta-lactoglobulin, bovine carboxypeptidase A, bovine carbonic anhydrase, and bovine serum albumin. The protein digest was diluted to 1 pmol/L with water containing 0.1% formic acid.
Fifty microliters of this solution was vacuum dried in a centrifuge. The dried peptide digest was labeled with the iTRAQ kit according to the manufacturer's instructions. Briefly, the digest was dissolved with 30 L of the dissolution buffer in the iTRAQ kit. Four tubes were prepared in the same manner and each tube was labeled with one of the four isobaric iTRAQ reagents. The first tube was labeled with the 114 mass tag, the second with 115, the third with 116, and the fourth with the 117 mass tag.
The contents of the four labeled tubes were combined into one tube, and the excess reagent was removed using a cation exchange cartridge provided in the iTRAQ kit. The labeled peptide digest was eluted from the cation exchange column using 500 L of elution buffer from the kit.
The resulting sample mixture contained peptides at a concentration of 100 fmol/L, which were labeled in a 1:1:1:1 ratio with one of the four iTRAQ tags. Five microliter injections of this sample were made for LC/MSn analysis with a Finnigan LTQ.
LC Separation and MS Analysis
The Finnigan LTQ was configured in the high-throughput mode using two C18 traps and a 10 cm x 75 m PicoFrit column. Five microliters of the iTRAQ-labeled digest were injected using the Finnigan Micro AS Autosampler.
Peptides were eluted with a 90-minute gradient from 0.1% formic acid to 50% acetonitrile with 0.1% formic acid at a nominal post-split flow rate of 300 nL/min. The Finnigan LTQ was run in positive ion mode using the nanospray source.
Spray voltage: 2.0 kV
Capillary temperature: 160C
Capillary voltage: 46.0
Tube lens (V): 110
Several scan modes were used including:
Full MS: 4001,200 m/z
Data Dependent MS/MS on top four ions with CID
Data Dependent MS3 of fragment ion m/z 291, the iTRAQ-labeled y-1 ion from MS/MS
Results and Discussion
The iTRAQ-labeled peptide digest of six proteins showed many intense peaks in the full MS, MS/MS, and MS3 modes (Figure 1). The latter mode was only triggered when the iTRAQ-labeled lysine y-1 ion at m/z 291 was detected. Many iTRAQ peptides were found in the Data Dependent MS3 scans. The data was searched using TurboSEQUEST in BioWorks 3.2, selecting two differential modifications with an increase of 144 Da at the N-terminus of the peptide and an increase of 144 Da at any lysine residue.
Using these parameters, BioWorks identified all six of the expected proteins. Coverage was typically excellent, e.g., 36% coverage for bovine serum albumin from 13 peptides distributed throughout the protein. The iTRAQ label confers 144 Da to the N-terminal residue of the peptide so that the b-1 ion in the MS/MS scan will also be increased by this mass.
The spectrum in Figure 2 of the labeled peptide (LVNELTEFAK) from BSA shows a b-1 ion at 258 due to the iTRAQ label instead of the expected 114 ion for leucine. Likewise, lysine (y-1 ion) has been increased to 291 instead of the normal 147 for the lysine y-1 ion. The MS3 spectrum of m/z 291 from the peptide LVNELTEFAK revealed all four iTRAQ labels with equal intensity, as expected.
The relative standard deviation from theoretical was 2.4% based on the expected ratio of 1:1:1:1. These ratios were determined by normalizing the intensity of the 115, 116, and 117 mass tags to the intensity for the 114 tag. Figure 3 shows the results of the MS3 scan for a peptide where iTRAQ labels are at the low end of the intensity range.
The carboxypeptidase peptide revealed nearly equivalent labeling by iTRAQ with all the ratios falling within 10% of the theoretical ratios of 1:1:1:1. Producing a significant y-1 lysine ion was not a problem with this large, 15 residue peptide.
The results for many iTRAQ labeled peptides in the digest of the six proteins in the mixture is shown in the Table. Peptides were selected that had iTRAQ label intensities of at least 200, that had no missed cleavages, and that terminated in lysine. The intensities of the iTRAQ labels for all peptides varied by about 100-fold, and the length of the peptides varied from 7 to 22 residues. For a given peptide, the raw intensity of the iTRAQ labels was relatively consistent for all four tags.
The standard deviation of the raw intensities of the four tags from each peptide ranged from 317%, relative standard deviation. Four proteins had multiple peptides that met the selection criteria, and the results for each peptide were averaged and listed in red below the last peptide belonging to that protein in the Table.
The iTRAQ ratios for all four proteins with multiple peptides were closer to unity and had a lower standard deviation than the two proteins with just one peptide that qualified. The average ratio for all 24 peptides in the Table is 1.026 with a relative standard deviation of 6.0%.
The method proposed by Carr et al., for relative quantification with iTRAQ works well on the Finnigan LTQ linear ion trap. Segregating peptide identification in one scan mode (MS/MS) and iTRAQ quantification in another scan mode (MS3) allows both functions to be performed well.
The C-terminal lysine peptides in the six protein digest ranged in size from 7 to 22 residues with varying degrees of basicity. No problems were observed in the generation of y-1 ions from these peptides, nor were any problems observed with the release of the labile iTRAQ tags. The precision of the LTQ for both low and high intensity iTRAQ ions suggests that this method can be used for complex peptide digests with a wide dynamic range.
The iTRAQ ion intensities varied over more than two orders of magnitude, suggesting that excellent detection is possible at low femtomole concentrations.
Because this technique separates the MS/MS scans from the quantitative MS3 scan, the approach maintains the sensitivity of the LTQ for diving deep into a proteome for protein identification. Only when significant ions are present for quantification does the LTQ spend a fraction of second to generate and quantify the iTRAQ labels in the MS3 scan.