Advances in Post-Translational Modification Analysis

by Jeffrey M. Perkel

Phosphorylation may be the biologist’s darling, but it is just one of 300 or so known post-translational modifications (PTM), according to a recent review.¹ Each serves to tweak protein activity in some way, whether by altering protein stability, localization, binding affinity, or enzymatic efficiency. Yet for all such changes, from acetylation to ubiquitination, the analytical tool of choice is the same. "Mass spectrometry (MS) is a central technology in the protein chemist's toolkit, enabling site mapping and quantification of chemical modifications on proteins, as well as detection of new types of structures," wrote Witze et al.

Natalie Ahn of the University of Colorado, who was corresponding author, has two mass spectrometers in her lab, an LTQ Orbitrap from Thermo Fisher Scientific and a 4000 Q TRAP from Applied Biosystems /MDS Sciex, the latter of which she uses in her search for phosphorylated peptides.

According to Ahn, a significant challenge to MS-based PTM analysis is sample preparation. Most peptides are unmodified, and only a fraction of those that can be are actually modified in any given protein. As a result, researchers (and their instruments) can be overwhelmed.

"I think one of the biggest challenges right now is how to simplify the sample by selecting for a modification of interest," she says. "It's important because very complex samples cannot be analyzed in depth."

Recently scientists have become increasingly sophisticated in purifying modified "subproteomes." Forest White, associate professor of biological engineering at the Massachusetts Institute of Technology, for instance, selectively isolates phosphotyrosine-containing peptides via a combination of immunoprecipitation and immobilized metal ion affinity chromatography.

"It allows us to identify hundreds of phosphotyrosine sites from small amounts of samples, say 10e6 cells," says White. By coupling that strategy with Applied Biosystems' iTRAQ chemistry, White's team can compile a relative quantitative assessment of phosphorylation changes in four samples at once. "It allows us to watch signaling events evolve," he adds.

White's lab uses a QSTAR XL from Applied Biosystems, a buying decision he says was biased by the lab's extensive use of iTRAQ chemistry. "The QSTAR works really well with iTRAQ," says White, who calls the reagent "addictive": "once you start using it you cannot get off it."

But prepping the sample is only half the battle; the other half is analysis.

PTM analysis requires tandem mass spectrometers, which first select eluting peptides and then fragment them for sequencing. The typical fragmentation strategy involves colliding the peptide ion with a gas to produce vibrational energy that shatters the peptide along its backbone, a process called collision-induced dissociation (CID, or collision-activated dissociation, CAD).

CID works well for short, unmodified peptides. But when it comes to PTMs, especially labile ones like phosphorylation, glycosylation, and sulfonation, CID often fails, says Joshua Coon, assistant professor of chemistry and biomolecular chemistry at the University of Wisconsin, Madison.

Consider what happens to a phosphopeptide, for example. Ideally, the peptide should randomly fragment along its backbone to produce a series of ions, each separated by the mass of a single amino acid, whether modified or not. Instead, Coon says, often CID produces just one predominant, and thoroughly unhelpful peak: the dephosphorylated peptide cation.

As a postdoctoral fellow, Coon helped develop a new fragmentation approach that largely circumvents this problem. That approach is called electron transfer dissociation, or ETD, and it is one he has continued to develop since joining the faculty at Wisconsin.

"Instead of making the peptide have collisions," Coon explains, "we react it with [negatively charged] fluoranthene, which transfers an electron to the [positively charged] peptide, and the peptide falls apart between peptide bonds while modifications are preserved."

Coon's lab has two Thermo LTQ Orbitrap instruments specially outfitted to perform ETD (the commercial instrument does not support this process yet, he says), which his team uses in its analysis of post-translational modification of histone tails.

Histones can be subjected to a variety of modifications, including phosphorylation, acetylation, and methylation. The latter two modifications are not particularly labile, Coon says, but they are found on rather long peptides.

And that exposes another advantage of ETD, he says. For fragmentation to be useful, the peptide must break randomly, so that all possible peptide fragments are generated. But as peptide length increases, he says, CID fragmentation becomes less random, making analysis difficult. That is not true of ETD, however.

"ETD breaks all the bonds more or less randomly, so if you have a 24-residue histone tail, you can map exactly which residues have modifications," Coon says.

As a result, it is possible to apply ETD to longer peptides, such as those generated by the proteinase LysC, as opposed to shorter, tryptic fragments. That difference could decrease sample complexity (by reducing the number of peptides), while increasing the likelihood that any given peptide contains a modification. The technique does have drawbacks, though. "It doesn't work so well on smaller, less charged peptides," says White.

Andreas Huhmer, Thermo Fisher Scientific's marketing director for proteomics in San Jose, says, "ETD is so crucial that it will become a de facto standard to provide information on post-translational modifications." He adds that ETD-ready Orbitrap systems will be commercially available later this year. Bruker Daltonics' HCTultra PTM Discovery System already includes the functionality.

Matthias Mann, professor of proteomics and signal transduction at the Max-Planck Institute for Biochemistry in Martinsried, Germany, demonstrated another new PTM-friendly fragmentation technology last year.² Called higher energy C-trap dissociation (HCD), the technique takes advantage of the Orbitrap’s architecture.

The LTQ Orbitrap is a hybrid instrument. The ion trap component is very fast, and very sensitive, but has relatively poor mass accuracy (about +/- 0.5 Da); the Orbitrap provides excellent mass accuracy (a few parts-per-million) and resolution, but is relatively slow and insensitive. In a typical tandem application, the Orbitrap is used to measure the precursor ions, but the ion trap is used for tandem analysis. As a result, fragment ions are measured at relatively low resolution. In addition, the ion trap tends to lose very small ions, such as the small immonium ions that are characteristic of phosphotyrosines.

Mann demonstrated that by fragmenting ions at high collisional energy in a separate chamber of the instrument, called a C-trap, it is possible to both recover those low-molecular weight ions, and to harness the Orbitrap's accuracy and resolution to do MS/MS. (The LTQ Orbitrap XL, a next-generation instrument, uses a dedicated octapole chamber between the ion trap and the Orbitrap for this purpose.)

According to Mann, HCD spectra "are high-resolution and have much richer fragmentation, so they are much superior [to CID]. But that comes at a cost: to analyze in the Orbitrap, you need 10-times more ions, let's say. So you give up some sensitivity."

As a result, Mann recommends using the typical ion trap-based CID for a first-pass MS/MS analysis, followed by selective, subsequent analyses using HCD.

According to Neil Kelleher, professor of chemistry at the University of Illinois, Urbana-Champaign, the mass spectrometer is just one of three components required for successful PTM analysis; the others are the upstream sample processing and the downstream data analysis.

So even though high-resolution, high-mass accuracy instruments, like the Orbitrap and Kelleher’s lab's 12-Tesla Thermo LTQ FTMS (Kelleher is a consultant for Thermo Fisher Scientific), simplify analyses by providing better data, PTM work requires commitment.

"There's a lot of tricks on the front and back end to do it well," Kelleher says. "If you have a phosphopeptide, do you in fact know the exact site that is phosphorylated, or do you not know that?"

"And that's hard to do right," he adds. "Because it is hard to nail down 10,000 peptides and do it in a meaningful way.”

Kelleher’s advice: make friends with local experts and technicians with direct access to mass spectrometers, as a thorough analysis will take time. Still, he says, the mass spec itself is not to be underestimated.

"For mass accuracy, the high end is the FTMS, the Porsche if you will," he says. "The middle performer is time-of-flight—a Lexus. And then you have your standalone ion traps, which are like a Honda in terms of resolving power, but acquire data very quickly."

Each level of sophistication (and cost), Kelleher says, provides about a 10-fold improvement in typical mass accuracy, from 0.1 Da for an ion trap, to 0.001 Da for the FTMS—enough to distinguish modifications (such as phosphorylation vs. sulfonation or acetylation vs. trimethylation), which differ by just a few milliDaltons.

"You want to get the best mass accuracy you can afford," he says. "That's most of what you are buying."

References:
¹Witze ES et al., “Mapping protein post-translational modifications with mass spectrometry,” Nature Methods, 4(10):798–806, 2007.
²Olsen JV et al, “Higher-energy C-trap dissociation for peptide modification analysis,” Nature Methods, 4(9):709–12, 2007.