There’s no shortage of scientists delving into studies of post-translational modifications. A new technology called electron transfer dissociation, or ETD, has gotten plenty of attention -- but is it ready for prime time? GT spoke with experts in the field to get their views on the most efficient and effective ways to detect post-translational modifications (PTMs), as well as their take on ETD and other relevant technologies. Following are excerpts of conversations with the PTM pros.
Assistant professor, chemistry
University of Wisconsin
You were one of the electron transfer dissociation pioneers. Why is the technology important?
So ETD is a technology that I helped develop during my postdoc in the [Don] Hunt lab. The idea going in was that we needed a way to sequence phosphopeptides; that was really the inspiration to work on the technology.
The typical way people had been doing peptide sequencing was that once you had the peptide in the mass spectrometer, you make it fall apart with collisions. Then you measure the mass of all the pieces [and] from the pieces you can reconstruct what the sequence was. But the problem was that, for phosphorylation sites, for example, that particular bond is very labile. So when you do that collisional excitation process, what happens many times is that the phosphate group gets displaced. You break that off preferentially, because it’s the weakest bond -- and then when you mass analyze all the products, you have a peak that shows you lost the phosphate, but it doesn’t have the backbone bond breakages that you’d need to deduce the sequence. That was the main limitation.
When I started my postdoc we needed a way to do this. It turns out that Fred McLafferty’s group at Cornell University in the late ’90s had developed a technology called electron capture dissociation. If you have a magnetic field, they discovered, you could have low energy electrons in the same space as the peptide ions, and then the peptide ions would capture the ions and then fall apart in this very informative way. Even if you had something like phosphate on the peptide, it didn’t matter -- the phosphate wouldn’t fall off, [so] you’d only break the peptide backbone.
So we asked, what if you use our ion trap device [and] what if we put in a negatively charged molecule, an anion. If we choose the right anion, then that anion would react with the peptide and an electron would come off of the anion and jump onto the peptide. If that could happen, you’d get the exact same chemistry as the McLafferty group described. It took us a year to build the instrument to find out, but it does work. And it turns out it works very efficiently and is very fast.
Associate professor, biological chemistry and oncology
Institute of Genetic Medicine
Johns Hopkins School of Medicine
Did you work with electron capture dissociation (ECD) or did you jump straight into ETD?
You know, we don’t have a Fourier transform mass spectrometer [to do ECD], so that is our weakness. I know groups that are using electron capture dissociation, they have instrumentation that is capable of it, but they have not made such great inroads in using [it] for studying post-translational modifications in the same way -- which basically speaks to certain difficulties which still exist. We basically short-circuited this as soon as we saw that ETD was on the horizon. We had a feeling that this basically would be amenable to rapid and yet robust analysis in terms of phosphorylation, but also other labile PTMs like O-GlcNAC-sylation [the addition of O-GlcNAc or O-linked N-acetylglucosamine].
Is ETD ready for prime time?
This has generally been the issue in the field, because the claims are usually enormous. There’s virtually no new mass spectrometer that comes in the market where you don’t hear, “It’s 10 times more sensitive than the previous one.” And you just wonder, “How many orders of magnitude are we getting better?” because surely in the past five years the same claim has been made by the same people, sometimes serially by the same manufacturer, so we should be 1,000 [times] better than we are now.
So maybe it’s not unreasonable for people to say “we need to see the data” because there is a big difference between some proof of principle. [We] started out talking about ECD -- I can tell you the proof of principle papers for ECD have been published, they’ve been glorified, they’ve been talked about. It essentially gives you the same kinds of results that ETD does, and one could argue that the reason ETD was developed [was] because it was going to be a simpler way of doing what ECD does. The moral was that in the practical world, ECD was not delivering in the global sense. Of course it had the technical capability, but people don’t have FT-MS instruments. For us it was important to be able to show [ETD] in a global scale. That’s when people start believing you. I think that ETD, there is no catch here -- we don’t have any special tricks.
What did you have to do in that study to get the global proteomic profile using ETD and collision-induced dissociation?
We took a human embryonic kidney cell line -- basically a well-characterized model where we said we want to find a source of sites of phosphorylation where we are more likely to find known things. In a way it’s a system that’s biased against us because it’s a system that’s been well studied.
We were interested in figuring out what can we identify in terms of what is known and what is unknown, also to figure out if something was known, how did this compare to the literature on those proteins or sites or peptides? On both sides I think we got some surprising answers. When we started out we thought we’d identify mostly the known sites and also a few unknown ones.
This is more a reflection of how uncharted the phosphoproteome is. At this point it’s a modest number that we found 1,435 phosphorylation sites of about 1,500. Then we went back to the peptides. When we went back to the databases, they had already been described in the literature — so we have already captured those peptides so we know that they are detectable in a mass spec experiment. We went back to all of the papers that had described phosphorylation on those sites [and] what was very nice to see was we found 80 percent of the sites that were ever described on those peptides in 103 separate publications.
So we have one experiment and we have [captured] 103 separate publications, and we can find 80 percent of observed events over maybe 20 years of research.
Staff scientist and manager of
the Proteomics Shared Resource
Fred Hutchinson Cancer Research Center
How do you go about detecting a PTM event, which is necessarily rapid and reversible?
We start off with a prayer. One of the hardest things, obviously, is stoichiometry. [Investigators] come down with all of these reasons of why they think something is phosphorylated. The first question I always have for them is, “Do you have any idea of the stoichiometry?” They say, “We think it’s 20 percent or 50 percent” -- and oftentimes the evidence is rather weak. You can look all day for phosphorylation, but if the stoichiometry isn’t high enough, you may never detect it.
The latest thing we do is called neutral loss scanning, that’s done with [an] ion trap instrument. There are two approaches: the first is a structural approach, single protein of interest, then you are trying to get as much information about the entire protein sequence as you can. We’ll also use MALDI-MS to get what we’ll call a peptide map. We’ll do neutral loss experiment as well -- that’s done on [an] electrospray instrument -- hopefully we can combine that information together to get a good picture of what’s going on with this one protein.
The other sort of experiment is that I pull down a protein complex, or that I’m looking at a yeast cell lysate or something of that nature. In that case it would be interesting to find phosphorylated proteins from that; that’s more what I’d call a proteomics sort of phosphorylation analysis.
Are you interested in getting an ETD accessory?
So we have ECD on our FT instrument. We do not have ETD yet -- it’s one of those things where I’m going to let my friends try it first. It’s a significant investment and I’m not sure how well the instrument manufacturers have engineered this just yet.
Professor, analytical chemistry
Are you looking into the ETD approach?
The electron transfer turns out to be one class of reactions that is particularly attractive for characterizing post-translational modifications that are otherwise fairly labile under conventional techniques. In terms of my work, ETD is a class of reactions that we’re studying. We’re also studying other kinds of ion-ion reactions, like proton transfer reactions and metal transfer reactions -- things that we think have relevance to protein structural characterization.
One example of something we’ve done recently: we’re interested in making these ETD reagents in different ways. The way people generally make an ETD reagent is with chemical ionization or some gas phase technique where you have to volatilize your neutral compound.
We wanted to make an ETD reagent by electrospray. That’s a little bit tricky because most things that yield strong signals by electrospray are either strong acids or strong bases -- so they really want to do proton transfer, they don’t want to engage in electron transfer.
We solved this problem by spraying carboxylic acids attached to PAHs, and it turns out that these things fragment by loss of CO2 when they’re in the mass spectrometer if you activate them. So what you end up with is a product which is a nice electron transfer reagent. We’re searching for more efficient reagents, that are a bit higher in mass than ones that are commonly used. The reason why we want that is in these electron dynamic ion traps that we use as reaction vessels, they have a limited dynamic range in terms of the range that can be stored. If you’re studying a high-mass biomolecule of interest, ideally you’d like to have a relatively high mass reagent so that you can store them in the same place at the same time.
Department of Clinical Medicine
University of Oxford
You don’t have an ETD in your own lab. Are you waiting to see how it fares?
ETD, especially last year in the meetings, was praised as something that could add something significantly new to our analysis of post-translational modifications. In some cases, like in phosphorylation, it helps because it encourages the phosphate group to stay on the site chain even during fragmentation. It also helps for glycosylation. For these kinds of modifications, I’ve seen examples where ETD actually was helpful -- but that’s about it.
There is a lot you can still do with more classical instruments. If you do your experiments right, with a good ion trap where it can do MS to the N experiments, or a Q-TOF where you can do exact neutral loss type experiments, you can do a lot.
How are you doing phosphorylation analysis if you’re not using ETD?
We do either a classical DVA experiment on the ion trap or Q-TOF, or exact neutral loss experiments using [a] label-free [method]. We have this Q-TOF label-free quantitation system, it’s actually called protein expression profiling, that can be used to quantitate phosphorylation sites as well. It’s not only a qualitative analysis, it’s also quantitative: you can determine ratios, different phosphorylation states, and so forth.