By Jeff Perkel
Suppose you want to quantify a particular metabolite—whether small molecule or peptide—in the blood of all participants in a clinical trial. How would you do it? For all its precision, mass spectrometry is primarily a qualitative technique. It takes effort, and in some cases dedicated hardware, to make it quantitative.
That is largely because all molecules do not ionize with equal efficiency, especially in mixtures, says Michael Gross, editor of the Journal of the American Society for Mass Spectrometry and principal investigator of the Washington University Mass Spectrometry Research Resource. "In electrospray and MALDI the response is very sample-dependent, and furthermore it is mixture-dependent," he says.
In other words, two peptides present at the same concentration in a sample—two tryptic fragments of the same protein, for instance—may not ionize equally, and thus, will appear to be present in different amounts. In addition, a single peptide may ionize at different rates in isolation and in serum. As a result, peak intensity and volume cannot typically be directly correlated to analyte concentration—at least not without help. That's where quantitative mass spec comes in.
There are two basic approaches to quantitative MS. Absolute quantification enables a researcher to know that a particular sample contains X moles of a particular compound—that is, the compound's absolute concentration. Relative quantification simply reveals the relative levels of compounds or peptides in two or more samples—that one sample contains twice as much of one thing as another, for instance.
Absolute Quantification
By all accounts, the preferred tool for absolute quantification is a liquid chromatography-coupled triple-quadrupole mass spec running in "multiple reaction monitoring" (MRM) mode. According to Gross, triple quads are ideal for this purpose because of their "high duty cycle." Instead of scanning over a range of mostly irrelevant mass-to-charge (m/z) ratios, they can be set to probe only specific preset values.
Further, says Jason Wood, a product manager at Varian, the filtration feature inherent in triple quads helps to reduce sample complexity and enhance the signal, however faint. "Triple-quads have great matrix rejection," he says. "They reduce background, so you can actually see what you are looking for."
Absolute quantification experiments are more rigorous than typical MS experiments, requiring the use both of standard curves and "spike-in" controls—isotopically labeled isoforms of the analyte you wish to measure, which can be distinguished from the experimental ions in the mass spec but which will otherwise behave identically. These provide a control for both efficiency of sample recovery during protein extraction steps, as well as ionization efficiency.
Though there are MALDI-based quantitative systems (such as Applied Biosystems' new Flashquant MALDI-triple quad system), MALDI should for the most part be avoided for quantification purposes, because of both shot-to-shot variability and the high concentration of ionized matrix it introduces into the experiment. "If you double your analyte in MALDI, will the signal double?" asks Gross. "I wouldn't count on it."
Instead, for biological sample quantification, the preferred method is electrospray ionization, which Gross recommends coupling to a good liquid chromatography system (providing sharp peaks and reproducible elution times) to improve the system's specificity—that is, to ensure you are in fact quantifying what you think you are.
MRM provides an additional level of specificity, says Chris Elicone, a senior marketing manager at Applied Biosystems. Suppose you were interested in quantifying a metabolite of m/z 500. Many such ions could exist, but only one of those ions would fragment in a collision cell to produce, say, an ion of m/z 200, and that's what MRM looks for.
In MRM, ions are filtered in the first quad, fragmented in the second, and specific fragmentation ions quantified in third. In this case, you would set the first quad to allow only ions of m/z 500 into the collision cell, fragment them, and then set the third quad to read only ions of m/z 200. Because that particular transition of parent to fragment ion is diagnostic of the parent ion, "that very unique process provides really great sensitivity," says Elicone.
Though Applied Biosystems does offer a triple quad system, the API 5000, for peptides Elicone recommends systems such as the company's 4000 Q TRAP (a hybrid quadrupole-linear ion trap), because peptides do not fall apart as predictably as small molecules do in a collision cell; the ion trap's iterative MS(n) capabilities provide confidence in ion identification. "To me, that's the safest way to go, because in peptides, the MRM transitions aren't as specific as in small molecules," says Elicone.
According to James Langridge, senior manager for Proteomics Business Development at Waters, features to consider when purchasing an instrument for absolute quantification (like Waters' Quattro Premier XE triple quad) include: a large linear dynamic range (shoot for three orders of magnitude or more), a low limit of detection (in the low femtomole range), and a low limit of quantification (tens of femtomoles).
Another characteristic Langridge stresses is accurate mass, for increasing confidence in the identity of the ion you are measuring. Most triple-quads have mass accuracy of plus-or-minus 50 ppm; QTofs, which can also be used for quantitative MS, have mass accuracy of 3 ppm or better.
Relative Quantification
Unlike absolute quantification experiments, relative quantification typically requires specific reagents, not hardware. The idea is to label two or more samples with different tags, mix and trypsinize them, and then run the mixture through a tandem mass spectrometer. A given peptide will then yield two (or more) peaks, one for each sample, whose sizes provide an indication of relative concentration between samples.
John Yates, professor of cell biology at the Scripps Research Institute, has used a number of "home-brewed" approaches to relative quantification, including labeling rats by feeding them food spiked with N15-containing algae for up to two months, and comparing those rats (or rather, their proteins) to unlabeled (N14) rat proteins. Invitrogen's SILAC growth media provide a similar sort of isotopic differentiation.
Other commercial systems support relative quantification based on addition to proteins of specific tags, such as Applied Biosystems iTRAQ reagent and the ExacTag reagents from PerkinElmer Life and Analytical Sciences. These tags attach to proteins in different ways (for instance, binding either to amine groups or to cysteine residues), and support different levels of multiplexing. ExacTag enables multiplexed detection of up to 10 samples simultaneously, while ICAT (also from Applied Biosystems) can be used on two samples, ICPL (from Bruker Daltonics) on two or three, and iTRAQ on up to eight.
ExacTag is an example of an isobaric label: each tag has the same nominal mass (about 972), so that a single peptide from a variety of different samples (each labeled with a different tag) will appear as a single peak in the first MS stage. But the mass of each tag is unevenly distributed on either side of a fragmentation site, so that in a collision cell, each releases a different-sized reporter ion, whose relative intensity reflects the total amount of that protein in the original sample.
According to Peter Banks, technology manager for biochemistry at PerkinElmer Life and Analytical Sciences, ExacTag reagents will work with any tandem MS, regardless of its design. "It is tandem-MS agnostic," he says.
If you think quantification is in your future and you need to consider a dedicated quantitative instrument, Elicone offers this final piece of advice for product testing: check for reliability. Company demos do not mirror the kind of workload an instrument will see once it enters the lab setting, so you need to see what kind of abuse the system can take.
"Keep running the sample sets over and over again," he says. "Take a look at how many samples you can run before you see a decay in signal intensity."
