At A Glance
Name: Roman Zubarev
Age: 40
Position: Professor of ion physics and biological mass spectrometry, Uppsala University, Sweden, since 2002.
Background: Associate professor of biological and chemical mass spectrometry, University of Southern Denmark, Odense, 1998-2002.
Post-doc in chemistry with Fred McLafferty, Cornell University, 1997-98.
PhD in ion physics, Uppsala University, 1997.
Worked at Institute for Technology and Automation, Sumy, Ukraine, 1986-92.
MS in engineering and applied physics, Moscow Institute of Engineering Physics, 1986.
How did you get involved with proteomics?
I come from the physics side. When I got my PhD at Uppsala University in ion physics I was doing desorption studies. Then I did a post-doc with Fred McLafferty at Cornell, and there I got involved with protein studies, and that’s where the electron capture dissociation technique was discovered.
What got you interested in proteins?
The diversity of the protein world. Genomics is more about information and proteins are much closer to the level of real biological systems. Also, that’s what we could do. When I came to this field, it was already clear that the future of mass spectrometry is in proteins. There was a group of people who said in the 1980s that large masses are good. A community was created that tried to ionize heavy masses, and this community was mostly interested in proteins. This laser desorption community is still alive and well.
Proteins seemed to be good objects to [provide] these large masses in the gas phase. So everyone was using proteins. DNA and RNA popped up in mass spectrometry, but they were never considered a big field. Somehow, proteins are better ionized — they are kind of a natural object to study by mass spectrometry. And what we later came to realize was that in fact, it was a very good choice to study proteins because of the bio-logical applications.
You are originally from Russia — how did you end up in Sweden?
I was working on the development of a time of flight mass spectrometer with [Cf-252] as the ionization source. This was a so-called plasma desorption instrument. This was the second half of the 1980s. That’s how I learned about this, first by publications and then by meeting in person some people involved in the field, such as Bo Sundqvist from Uppsala and Peter Roepstorff from Odense and other people, who visited Russia. When the Soviet Union collapsed, I got an offer from Bo Sundqvist to come to Sweden to be a PhD student. I received the offer in 1991, so in 1992 I went to Sweden. I felt at home almost immediately.
How does ECD work?
First, as much as there was a community that wanted to have these large masses in the gas phase in the 1980s, Fred McLafferty had had this idea of achieving non-ergodic fragmentation of large molecules for a long time. He tried different things — UV lasers, collisions with surfaces, and nothing seemed to work. There were a number of post-docs involved in the project, but the post-docs would come and go. When I came, I started to work with photo dissociation. The hypothesis was that UV light should fragment the molecules, but serendipitously it was discovered that the UV light from the laser was desorbing electrons from some surfaces. These secondary electrons would interact with positively charged ions. The positively charged ions produced by electrospray are multiply charged, so they like [the electrons], and they are attracted to each other, and they capture them. When an electron is captured, some en-ergy is released — between 4 and 7 eV — it’s not much but it’s larger than the bond strength. So a chemical bond gets broken. One can even say that the electron cannot be captured unless some bond is broken, because there must be a pathway for this energy to be released, and this fragmentation is a natural pathway for that.
It turns out that Fred McLafferty had predicted this seven years before it was experimentally discovered, and then forgot about it. I found some reference about it, and one reference led to another, and finally I found this idea published. He tried it, but he tried different things [than what I found]. He thought one should collide the multiply charged ion with neutral sodium or potassium atoms that have low ionization energy, and that that’s how electrons should be captured. He never tried this particular experiment. Interestingly, now people have done exactly that, and guess what — it works.
So then you applied this technology to combining top-down and bottom-up approaches …
The credit for this suggestion should go to Neil Kelleher of University of Illinois at Urbana-Champaign. We were looking only at backbone cleavage. But he suggested that if this cleavage is so fast, then the labile modifications, if they’re present, should be intact. So what he did was prepare peptides with sulfated cysteine. And he saw that yes, it worked: The sulfate group that is normally very labile — you attach it and it flies away — it stayed on the fragments during ECD. That was a hint that other things should work in the same way, and that caused us to try other things. We found that post-translational modifications were indeed surviving this process.
How do post-translational modifications survive ECD?
People are still debating [that question]. It [may be] because the process happens so fast that the energy does not have time to distribute over the whole molecule to reach the points of where the stable modifications are. Alternatively, [maybe] it doesn’t happen that fast but it happens at a very specific place far away from the modifications. So the important thing is that when the electron is captured, then the molecule turns into a radical. And certain bond strengths in radicals are very different than the bond strengths in molecules. As a result, certain bonds become much weaker, and so they break.
So when it was discovered that post-translational modifications survived, it was now clear that that is probably the most important application field for ECD. That is how I became really interested in proteomics. My view of proteomics is that it had a path with several stages. One was identifying the proteins. To do that, you only need one peptide, and you don’t have to have full information about the protein. This stage has been tremendously successful, and the database search and identification approach is very sensitive and quite reliable these days. The next stage is quantification. And yet another stage is determination of post-translational modifications. And here it’s a much more complicated task than the previous two. That’s because you need basically 100 percent sequence coverage.
You can either do a top-down approach, and there you have 100 percent coverage automatically because you’re working with intact molecules. However, you may not receive 100 percent information about the position of the modifications. So if you do bottom-up, you receive 100 percent information for pieces, but then you might have holes — uncovered parts of the sequence. So [my approach is that] you put the two approaches together.
So where are you going next with this technology?
One wants to find an application field where it can really be important. So post-translational modifications seems to be a really nice field. This next stage of proteomics seems to be specially created for ECD. That’s the focus now.
Are you starting in one particular area?
People say that proteomics is about mining data — like gold mining. And you know what people say about gold rushes: Who has benefited from gold rushes most? They say spade merchants — those who provided spades and other equipment for those who did actual gold mining. So that means that small companies and researchers providing instrumentation for proteomics are probably in a good position. Therefore, I see the task of my group not so much as mining proteins themselves, but providing tools for mining.
So you want to improve the ECD technique and make it available for others to use?
That’s right. We need to make it something that one can take from a shelf and use — kind of a kit, or easy-to-use technology. We’re not there yet. But this is an important task. We need to work on ECD efficiency — ECD only works currently in Fourier-transform mass spectrometers. Currently there are three manufacturing companies that provide these instruments, and maybe a fourth one soon. All three are now offering ECD as an option. But they don’t guarantee anything. There’s no standardization, no generally accepted approach on how to evaluate how well ECD works. We need to provide a gold standard so everyone can refer to it.
Where do you get your funding?
One of the biggest sources is [the] Wallenberg foundation. They provided us with over €1 million ($1.2 million) for buying the latest FT-MS. And they provide us with money for research as well. We also interact with manufacturing companies to provide the ECD technology. We encourage them to commercialize it.
What are the big challenges that still face proteomics research?
There are many, but I think the biggest challenge will be protein separations. I think that if there were a robust and very sensitive technique for separation of proteins that differ by one post-translational modification — if they could provide us some sort of device to do that, the rest is easy.