At A Glance
Name: Marvin Vestal
Age: 68
Position: Scientific fellow and vice president of mass spectrometry platform R&D, Applied Biosystems
Prior Experience: Developed the first commercial MALDI-TOF mass spectrometer, as well as the Applied Biosystems MALDI-TOF/TOF
How did you get involved in proteomics?
It’s kind of a long story, but prior to about 20 years ago there wasn’t any such thing as protein mass spectrometry. Mass spectrometry was limited to small molecules and volatile systems. As new ionization techniques came along it became more accessible. When MALDI was first described in 1988, Vestec was in operation and we decided to build a commercial system, thinking there was a future for this technology. I really got involved in it from the instrument-building side, both for the MALDI-TOF and electrospray.
How did Vestec come about in the first place?
Vestec was founded on the basis of a development that my research group at the University of Houston and I made called thermospray, which was a technique for interfacing liquid chromatography with mass spectrometry. It had some applications [in proteomics].
Was it relatively simple to switch over from thermospray to MALDI time-of-flight instruments?
Your question implies that we thought of coupling the two; it wasn’t our idea actually. I was doing some work on laser desorption with mass spectrometry, but Franz Hillenkamp and Michael Karas in Germany were the ones who developed the idea of using the matrix with the laser desorption to get ionization of large molecules. In 1988, at the Bordeaux mass spectrometry meeting, they presented a spectrum of ß-galactosidase showing an intact molecular ion of mass 116,000, and that was truly revolutionary.
So how did you get into building MALDI-TOFs?
As a result of that, there was a group at Rockefeller — Brian Chait and Ron Beavis — who very quickly started applying the technique to proteins. They had lots of good collaborators at Rockefeller of course, and they developed a somewhat simpler instrument for doing it. We collaborated with them and, in fact, licensed some of their technology in developing the first commercial MALDI-TOF instrument. Because the laser is pulsed in MALDI, it lends itself naturally to a time-of-flight analyzer. But at the time the development of time-of-flight analyzers had been dormant because there weren’t many applications for them. Those analyzers really hadn’t progressed very far from the early ‘50s when they were first introduced.
What were the sort of challenges you had to overcome?
I suppose the biggest jump was this technique we called delayed extraction, which actually had been anticipated or described in a general sense in the very early papers on time-of-flight. That’s what actually made MALDI practical, because without that, one could not get high resolution and routine sensitivity. By producing the pulse of ions, and then waiting a short time to do the acceleration, it solved a whole bunch of problems in a practical sense. We introduced that commercially in ‘95 after some early work by Bob Brown at Colorado State.
How did the TOF/TOF come about?
In one sense it came out of a conversation I had with Al Burlingame, drinking a beer by the swimming pool at a meeting in Palm Springs about five years ago. It’s clear that for many of the applications in protein identification one really needs MS/MS. What people were doing at that point was using MALDI to get the molecular weight of the peptides. But if they couldn’t identify it from the peptide mass fingerprinting, then they would take another aliquot of the sample — usually the majority of it — and do MS/MS to get the fragmentation and hence the more detailed structural information.
It seemed to me it would make sense, since people were doing MALDI as a first pass anyhow, to go ahead and do the MS/MS on the sample as well, which would be more efficient, faster, and more sensitive. Also, back in the earlier days, there were four sector mass spectro meters that did high energy CID and gave very high performance, but were very expensive and didn’t lend themselves to ionization techniques like MALDI and electrospray. The other MS/MS techniques were all low energy collisions, which has its advantages and disadvantages as does the high energy. But there was no high energy technique, and there was no MALDI technique for doing MS/MS, so it seemed like a reasonable thing to do.
I had been thinking about the prospects, and I talked to Al Burlingame, who’s an old friend, and he said, you guys really ought to figure out a way to do MS/MS with MALDI. I started designing it on my way home, and building it a few months after that.
What are you working on now? Can TOF/TOF technology still be improved?
We do have an active program in continuing to develop the TOF/TOF technology. We’re also looking at other [aspects of protein analysis]. For a lot of the applications, the mass spectrometer alone is not enough. You need the sample prep and the separation on the front end, and the bioinformatics on the back end, and the software to interpret the data. Mass spectrometrists are used to doing a lot of this stuff themselves, but as you try to approach a more general market, you need more integrated systems.
How much of a role does mass spec hardware play in advancing proteomics technology?
There’s still some improvements we can make in that area, and we’re actively working on it. The more general area which limits all of the techniques is what we call chemical noise, where for some level of concentration there’s a lot of other endogenous materials that give signals that are not the things you’re interested in looking at. This becomes a question of purification. It’s a fairly complicated issue but ultimately in most cases now the limitation is the background or chemical noise rather than the basic sensitivity of the instrument.
What role will proteomics play in understanding biology?
It’s essential! The genome gives you the floor plan, but the interpretation and the application has turned out to be somewhat more distant than some people might have thought. I think the way one gets there, to a great extent, is through the proteins, that is, you really do have to understand these protein machines that make things happen in the cell. Genomic data is absolutely vital; without it we couldn’t do what we can do now, but we still have to do an awful lot of work to do on the proteins to really understand these mechanisms in biology. Not only proteins but metabolites, too. I think the emphasis on doing this kind of research with proteins, using mass spectrometry, is going to be a very rapidly growing field for the next ten years at least.