Jeffrey Perkel has been a scientific writer and editor since 2000. He holds a PhD in Cell and Molecular Biology from the University of Pennsylvania, and did postdoctoral work at the University of Pennsylvania and at Harvard Medical School.
The defining tool of proteomics research is mass spectrometry, and it has been for years. But the way that tool is used has changed.
Early proteomics studies tended to focus on cataloguing proteins—they were qualitative analyses. More recently, researchers have insisted upon having quantitative information. They don’t want to know simply which proteins are present in a cell, tissue or compartment; they want to know in what abundances, as well. Such information is required, for instance, to identify protein biomarkers of disease.
In mass spec, explains Joe Fox, senior director of the pharmaceutical business unit at AB SCIEX, a protein or peptide must make the leap from liquid phase to ionized gas and then make its way to the detector. Not all molecules make that transition with equal efficiency, in part because of molecule-to-molecule variation and in part due to interference from competing molecular species and the surrounding matrix. Peptides A and B may thus be present in equal abundance in the sample and yet yield peaks of different heights in the mass spectra, complicating comparisons.
“There are areas of variability there,” Fox says. “Peptides can have large chemical differences, and as a result, the ionization of these compounds can vary.”
Researchers have devised tools and workflows to overcome these issues.
Quantitation strategies
There are two levels of quantitative mass spec, Fox says: relative and absolute. The former indicates that there is twice as much peptide A as B, but not the actual number of molecules to which that corresponds. Absolute quantitation fills in that missing data through the use of internal standards, which provide a benchmark against which to measure a given analyte’s response in the instrument.
Depending on the instrument type—whether it measures nominal or accurate mass—there also are two strategies for quantitation: global (or discovery-mode) and targeted, says Christie Hunter, director of omics applications at AB SCIEX.
“The trend,” says Fox, “has been in the area of proteomics to move more towards targeted quantitation.”
For relative quantitation, researchers have multiple options. One, called SILAC (stable isotope labeling by amino acids in cell culture), uses heavy or light isotope-containing amino acids to label proteins in two populations of live cells (say, drug-treated and control cells) [1]. After the proteins from both populations are harvested, they can be mixed and analyzed together in the mass spectrometer, where a given peptide will yield slightly different masses depending on which sample it came from. By comparing the intensities of the two peaks, researchers can directly compare the analyte’s abundance in both samples.
Another option uses isobaric labels, such as AB SCIEX’s iTRAQ® (isobaric tags for relative and absolute quantitation) and Thermo Scientific’s TMT™ (tandem mass tag) reagents. In this case, proteins in different samples (up to eight in the case of iTRAQ and up to 10 for TMT) are labeled in vitro with chemical tags of identical size, each of which contains a chemical fault line in a different place. The samples are then mixed and placed in the mass spectrometer. Because the tags are isobar—that is, they all weigh the same—the same peptide from multiple samples will behave identically in the mass spectrometer. But when fragmented in tandem mass spectrometry (MS/MS), they shatter to produce a fragment ladder from which the peptide can be identified as well as a series of different sized reporter ions, each indicating the molecules’ abundance in one of the starting samples.
Researchers may also use label-free methods, says Christie Hunter, director of 'omics applications at AB SCIEX. In this case, samples are not mixed. Instead, they are run sequentially, assessing peptide concentration by comparing spectral peak areas from run to run.
According to Hunter, labeling and label-free methods are both popular among proteomics practitioners. “The market divides maybe 50/50 between these labeling techniques and those that measure peptides in unlabeled forms.”
Instrumentation
After you have your sample, whether labeled or not, you need an instrument to analyze it.
One popular option for quantitative mass spectrometry is the triple-quadrupole mass spec, which is ideal for the selected-reaction monitoring workflow (SRM). In this case, the instrument is programmed to look for and quantify specific analytes and to ignore everything else. Thus, it is best suited to workflows in which researchers wish to quantify one or more peptides previously identified as interesting.
Agilent Technologies launched a new high-end triple-quad at this summer’s American Society for Mass Spectrometry (ASMS) national meeting. The Agilent 6495 (PDF) is a second-generation, iFunnel-enabled triple-quad featuring improved ion optics, detection and collision-cell design to yield enhanced sensitivity, says Na Pi Parra, Agilent’s triple-quadrupole LC/MS product manager. “This new instrument is much more sensitive and precise [than previous-generation instruments], especially at lower [analyte] levels,” she says.
For discovery (i.e., untargeted) proteomics, researchers need a different class of instrument and workflow. Often, these are variants on the triple-quad configuation. For instance, users of AB SCIEX’s new TripleTOF® 6600—essentially a triple-quad in which the Q3 is replaced with a time-of-flight analyzer—can quantify ions using a workflow called SWATH™.
Proteomics workflows traditionally have been run in data-dependent modes, Hunter explains. As a peak elutes off the upfront liquid chromatography column, the instrument automatically selects, say, the 10 to 30 most abundant peptides for MS/MS fragmentation, then the next 10, and so on, until the peak passes. “It’s a stochastic sampling method,” she says, but by focusing on the most abundant molecules it can sometimes miss important features.
SWATH, in contrast, is data-independent [2]. The system collects all ions in a narrow m/z window (say, 5 Da wide), fragments everything, and then slides along the m/z range to the next window and repeats. In this way, all peptides over a given mass range can be fragmented and collected in just a few seconds, providing “a digital archive” of the sample. Thus, Hunter explains, data can be reanalyzed later to see if an interesting molecule is present that wasn’t known or considered when the data were collected. “If you need to go back and look for something, the data should be there,” she says.
According to Fox, the 6600 enables SWATH thanks to its high acquisition rate (100 Hz), reproducibility and wide dynamic range (five orders-of-magnitude). “You can now see deeper into your sample with that same level of quantitative reproducibility previously only achievable by targeted MRM,” Fox says.
Software solutions
The final piece of the puzzle for quantitative proteomics is software to make sense of the data. Mass spec vendors have their own in-house solutions, of course, but there are free alternatives, too. SWATH data, for instance, can be handled with OpenSWATH, an open-source software platform integrated with OpenMS. An alternative, and more broadly applicable tool, is Skyline.
According to Michael MacCoss, professor of genome sciences at the University of Washington, whose lab developed the popular, no-cost tool, Skyline takes a “protein- and peptide-centric” approach to proteomics data analysis—an uncommon approach when the project began.
Historically, MacCoss explains, most proteomics packages have taken a “spectrum-centric” point of view—they try to match a given spectrum to a particular peptide in a sequence database. "In contrast, we have been very interested in flipping the problem and asking if there is any evidence for a given peptide in the mass spectrometry data."
The approach is only valid for data-independent acquisition workflows, he says. But that seemingly subtle difference allows researchers to extract low-abundance signals within the presence of much more abundant ones, MacCoss says. “As a colleague of mine says, finding a needle in a haystack isn’t that hard, if you have matches and a good magnet,” he explains. “And that’s basically how things work in this peptide-centric analysis: You’re basically lighting the haystack on fire; all you’re doing is looking for things that stick to a magnet. In our case, we ignore everything that doesn't have the accurate precursor and product ion m/z for the target peptide.”
Skyline facilitates experimental design, optimization, SRM scheduling and data analysis.
“It’s definitely pretty powerful,” MacCoss says, and it is compatible with (and financially supported by) all major mass spec vendors. “It’s got about 30,000 installations now, and it gets booted up more than 6,000 times per week.”
Now at version 2.6, Skyline includes support for SRM, targeted MS/MS, MS1 filtering from data-dependent acquisition, and data-independent acquisition analysis, MacCoss says, making it a useful tool regardless of the proteomics strategy you choose.
References
[1] Ong, SE, et al., “Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics,” Mol Cell Proteomics, 1:376–86, 2002. [PubMed ID: 12118079]
[2] Gillet, LC, et al., “Targeted data extraction of the MS/MS spectra generated by data-independent acquisition: A new concept for consistent and accurate proteome analysis,” Mol Cell Proteomics, 11:O111.016717, 2012. [PubMed ID: 22261725]
Image: AB SCIEX's TripleTOF system