Examples of Mass-directed Discovery
An example of a mass-directed biomarker discovery is illustrated by the analysis of an E. coli lysate spiked with known amounts of bovine serum albumin and serotransferrin (Table). This model simulates the up- and down-regulation found in complex biological samples. E. coli lysate, bovine serum albumin (BSA), and serotransferrin were digested with trypsin and analyzed using an Agilent 1200 Series HPLC-Chip/MS system interfaced to an Agilent 6210 TOF mass spectrometer. The HPLC-Chip configuration included a 40-nL enrichment column and a 150 mm x 75-µm analytical column packed with ZORBAX 300SB-C18, 5-µm material. A 100-minute-long gradient method was used. The solvents employed were: (A) 0.1% formic acid in water and (B) 0.1% formic acid in 90% acetonitrile/water. After initial loading at 3% (B), the gradient stepped to 8% (B) at 0.5 minutes, then 45% (B) at 85 minutes, 80% (B) from 90 to 92 minutes and back to 3% (B) at 92.01 minutes. The analytical flow rate was 300 nL/min, and the sample was loaded at 4 nL/min.
Accurate mass data from the TOF mass spectrometer were extracted and evaluated using the Agilent MassHunter Profiling software. Targeted LC/MS/MS data was obtained using a HPLC-Chip/MS system interfaced to an Agilent 6330 Ion Trap LC/MS. Peptides were identified using the Agilent Spectrum Mill for MassHunter Workstation software to search the SwissProt protein database.
Figure 1a shows a total ion chromatogram for one of the spiked E. coli samples and demonstrates the complexity found in biologically based samples. Figure 1b depicts an extracted ion chromatogram (EIC) for m/z 504.2506 1.9 ppm, a singly charged ion from the serotransferrin digest.
The multiple peaks for this exact mass further demonstrate the sample complexity. An MS spectrum obtained at 9.2 minutes shows that the 504.2506 ion has a relatively low abundance (Figure 1c).
Most likely this ion would not have been chosen for MS/MS analysis using data-dependent logic in a typical MS/MS experiment, since many other ions in the spectrum have significantly higher abundances. This is the typical situation for many biomarkers and is the fundamental reason why many undirected MS/MS experiments fail to discover low-level protein biomarkers.
One of the major challenges in mass-directed biomarker discovery is the analysis of the TOF MS data. Thousands of ion profiles are collected from the control and sample during the LC/MS acquisition. To address the challenge of such a large data set analysis, algorithms were designed to automatically determine the features (sets of ions from the same peptide) at various retention times.
Next, the extracted features are compared between samples to determine the up- or down-regulated peptides in samples when compared to the control. Ultimately, a plot of up- and down-regulated features is produced (Figure 2). The MassHunter Profiling software identified 22 features (sets of ions) that were up- or down-regulated by 2x and 4x when compared to the control. This data demonstrates the ability of a mass-directed strategy to discover differential expression levels between controls and samples.
The next step in the mass-directed method for biomarker discovery is to use the retention time and mass information from the initial TOF analysis for target identification using a Trap MS. The TOF mass information was imported into the Trap MS control software as an "include" mass list for MS/MS acquisition. At the appropriate retention time windows, the ion trap obtained MS/MS spectra for the ions included in the imported mass list.
This retention time and mass-directed approach ensures that the Trap MS will collect MS/MS information for the up- and down-regulated peptides regardless of their relative abundance. In addition, this approach maximizes the acquisition duty-cycle and therefore the quality of the MS/MS spectra of the targeted peptides.
The MS/MS data were then processed using Spectrum Mill for MassHunter Workstation software to search the SwissProt protein database. The MS/MS data from this targeted identification strategy resulted in the correct identification of the BSA and serotransferrin peptides in Samples A and B (Figure 3).