by Jeffrey M. Perkel
In the world of protein mass spectrometry, there are, fundamentally, two types of people. One group focuses on intact proteins, the other on peptides. The advantage of the former strategy is that it exposes nuances of posttranslational protein modification – which modifications occur simultaneously, for instance, and which are mutually exclusive – that peptide work overlooks. Yet intact proteins are tough to handle, separate, and analyze.
“The general problem … is that the tools to separate intact proteins prior to the mass spectrometer are not that great,” says John Yates of the Scripps Research Institute in La Jolla, Calif., “they don’t work that well.”
Many researchers prefer to work with peptides instead.
“I’m not sure there’s a clear distinction [between peptides and proteins] because most people use peptides as a surrogate for detecting proteins,” says Andreas Huhmer, director of proteomics at Thermo Fisher Scientific.
Among the most common peptide MS applications, says Gary Kruppa, vice president of business development at Bruker Daltonics, is “bottom up” proteomics, in which complex protein mixtures are digested with protease into even more complex peptide mixtures, which are then separated and analyzed.
In such cases, fractionation is key—the simpler the sample going into the MS, the richer the data coming out. Pre-MS HPLC is de rigueur, whether at normal or nanoLC flow rates. Sometimes, though, it isn’t enough.
“One of the things that MS has to do is data-dependent data acquisition—the computer tells the instrument to select this peak and perform MS/MS on it,” Yates explains. “So if you have lower resolving chromatography, you have instances where many peptides co-elute, entering the MS at the same time, and the system cannot necessarily scan fast enough to collect MS/MS data on as many peptides as you like.”
One approach to this problem is to increase MS sensitivity by improving chromatographic resolution with sharper peaks, increased peak heights, and overall better signal-to-noise courtesy of ultra-performance LC setups, such as Waters’ ACQUITY and low-flow-rate nanoACQUITY UPLC systems.
According to James Langridge, director of proteomics at Waters, traditional HPLC uses columns packed with 3- or 5-micron beads. ACQUITY UPLC, in contrast, uses 1.7-micron beads. “These small particles are tightly packed,” he says. “That raises the pressure of the system”—from 3,000-5,000 psi up to 10,000-15,000 psi—as well as its performance. “Peak capacity or resolution is twice that of a traditional HPLC system while run times are 10-times faster,” he says
Sharper peaks, in turn, put less strain on the MS system, enabling it to more effectively probe complex mixtures.
Another increasingly popular pre-MS fractionation option, Yates says, is isoelectric focusing, a classic protein biochemistry technique (in fact, it is one of the two dimensions in 2D gel electrophoresis) that fractionates peptide mixtures by their isoelectric point, or pI, prior to LC. Commercial systems include BD’s Free Flow Electrophoresis System, Bio-Rad Laboratories’ PROTEAN IEF, and the dPC Fractionator from Protein Forest.
Yates’ lab uses Agilent’s 3100 OFFGEL Fractionator. According to Keith Waddell, Agilent’s LC-MS Applications Solution Manager, the OFFGEL resolves as many as eight samples into 24 pH-delimited fractions simultaneously, finally pushing the proteins out of the gel and into a liquid buffer that sits above it.
“This is actually pretty important,” Waddell says, “because one of the problems in classical gels is, sure you can cut the bands out, but how do you get the peptides out of the gel?”
Several companies provide hardware in support of another strategy, called LC-MALDI. In LC-MALDI, mixtures are separated via liquid chromatography, but then spotted onto a MALDI target plate for off-line analysis and archiving.
According to Aaron Hudson, senior marketing manager of biomarker and proteomics at Life Technologies, the LC-MALDI workflow (enabled with the company’s new AB SCIEX TOF/TOF 5800 System and TEMPO LC MALDI SPOTTER) decouples the MS from the time and throughput constraints of LC, giving researchers the ability to study peptide mixtures in depth and at leisure.
With an online ESI strategy, “Even if the separation is good, at any one point you have a lot of peptides eluting [from the LC] at once,” he explains. “The MS has to make intelligent decisions of what to look at.” But with a limit of “probably 5 to 10 MS/MS per second,” he continues, “you can miss things because the logic of the MS is not fast enough to keep up.”
Besides, Hudson notes, ESI and MALDI tend to ionize different, but overlapping, sets of peptides, meaning that a two-instrument approach is more comprehensive. “The MALDI seems to find tryptic peptides with arginine termini more readily than does electrospray,” he says. “ESI detects peptides with lysine termini more readily. So there is a reason for this complementarity.”
According to Kruppa, Bruker Daltonics has devised a way to derive the maximum benefit from that complementarity—assuming users have the necessary hardware.
Researchers with both a MALDI and electrospray-based MS can run a combined workflow in which half of the eluent of an LC run is shunted to an ESI system (such as Bruker’s HCT Ultra ion trap) and half to the company’s PROTEINEER fc spotter. Then, once the LC-ESI run is finished, the MALDI MS (Kruppa recommends a TOF/TOF system, such as the company’s ultraflex) gets a crack at the data.
“You can tell the MALDI instrument what the LC instrument found so it doesn’t waste time doing MS/MS of peptides that were identified in the LC-ESI run,” he explains. “It focuses only on what’s new.”
Software also plays a significant role for another peptide MS application: peptidomics. Peptidomics is to peptides what proteomics is for whole proteins; that is, it seeks to enumerate, quantify, and characterize naturally occurring peptides, such as those found in serum, or bound to MHC receptors.
Such peptides are more difficult to study than those used in bottom-up analyses, says Hudson, because they were not all generated by the same protease. Many bottom-up studies, for instance, use trypsin to digest proteins. As a result, researchers know that each peptide should end with a basic amino acid (lysine or arginine). But that isn’t the case with natural peptides, Hudson says.
“Natural peptides are more difficult to identify with standard search engines because you can’t reduce complexity in the search space as you don’t know what the ends are,” he says.
Life Technologies’ ProteinPilot software can circumvent this problem, he says. The software’s probabilistic algorithm (called Paragon) analyzes MS spectra unconstrained by user guidance on terminal residues or possible post-translational modifications. Instead, it can determine on its own whether any of over 250 different modifications are present.
The software, Hudson says, “can quickly look for natural peptides in a complex mixture without increasing the number of false positives, because it doesn’t need to assume what the end looks like.”
On the hardware side, says Yates, “What’s in vogue now is high resolution, high mass accuracy [equipment].” His team, which focuses on bottom-up studies, uses Thermo Orbitrap mass analyzers—what Yates calls “the ‘it’ instrument.”
“I think Orbitrap is still the leading instrument to do large-scale peptide identification,” says Huhmer, “simply because you have a parallel detection capability, where the Orbitrap analyzer gives the accurate mass precursor information, and in parallel, you have the LTQ ion trap detector very quickly detecting the MS/MS spectra of those precursors that come through, which provides you with the peptide identification information.”
But there are other choices, including a new generation of high-performance quadrupole-time-of-flight (Q-Tof) instruments from Bruker Daltonics (the maXis), Agilent Technologies (the 6530 Accurate Mass Q-TOF), and Waters (the Xevo Q-Tof).
Q-Tofs, says Waddell, provide high mass accuracy and resolution in both MS and MS/MS modes. In contrast, Orbitraps (and other FT-MS systems) offer exceptional mass accuracy and resolution in MS mode, but generally only nominal mass accuracy in MS/MS.
High-performance instruments can’t do everything, however. Triple quadrupoles (such as Waters’ Xevo TQ MS, Thermo’s TSQ, and Agilent’s 6460) are preferred for targeted (as opposed to discovery), quantitative peptide analysis using so-called single reaction monitoring, or SRM.
“It’s very hip right now,” says Huhmer. “Everyone’s jumping on the bandwagon and trying to quantify peptides.”
And ion traps are ideal for MS(n) studies, for instance to determine both a peptide’s sequence and post-translational modifications, says Kruppa.
“I wouldn’t rule one or the other [instrument] in or out based on accurate mass,” he says.