Principal investigator, National Institutes
of Health Mass Spectrometry Research Resource
At A Glance
Name: Michael Gross
Position: Principal investigator, National Institutes of Health Mass Spectrometry Research Resource at Washington University, since 1994. Professor of chemistry, medicine, and immunology, since 1994.
Background: Director, National Science Foundation Midwest Center for Mass Spectrometry, 1978-1994; Professor of chemistry, University of Nebraska, 1978-1994; Assistant, associate professor of chemistry, University of Nebraska, 1968-1978.
Postdoc, Fred McLafferty's lab, Purdue University, 1967-1968.
Postdoc, Edward Thornton's lab, University of Pennsylvania, 1966-1967.
PhD in organic chemistry, University of Minnesota, 1966.
The Mass Spectrometry Research Resource at Washington University aims to use mass spectrometry as a tool for biomedical research. Established in 1978, the facility has gone from focusing on gas chromatography mass spectrometry and isotope-ratio mass spectrometry to concentrating on Fourier transform mass spectrometry and tandem mass spectrometry.
ProteoMonitor spoke with Michael Gross, who has been a principal investigator at the facility for 12 years, to find out about the history of the resource center, and the proteomics work going on there.
When did you become involved in Washington University's Mass Spectrometry Research Resource, and what had you been doing before you were hired to work there?
I was a professor of chemistry at the University of Nebraska in Lincoln, and there I directed for 16 years a National Science Foundation Center for Mass Spectrometry. In 1978, the NSF founded six regional/national facilities in NMR, mass spectrometry, lasers, and surface science. They did this for a period of two years. The second year they did, maybe, eight or nine. Some of the facilities lasted for 16 years — ours did. Others dissolved or lost their funding.
That NSF facility in Nebraska was not very biomedically oriented. It was instrumentation oriented, and oriented towards a subdiscipline of mass spectrometry that we call ion chemistry — that's about the mechanism of how things come apart.
So I came to Washington University in 1994 because I was interested in expanding mass spectrometry in biomedical things. I brought along my interest in tandem mass spectrometry and Fourier transform mass spectrometry.
You had worked with tandem mass spec and FTMS at the NSF facility?
Yes. We built the second Fourier transform instrument 30 years ago in the mid- to late-70s. This has become a reasonably important high-performance instrument in proteomics.
My collaborator at the time, Charles Wilkins, who is now at the University of Arkansas, he and I teamed up and built the second instrument. Our interest was analytical applications. A lot of the dreams we had 30 years ago have come true.
A big thing that happened in FTMS is not only the development of computers, but the actively shielded superconductive magnets, which makes these instruments available in many laboratories now. Five years ago, I'd say this instrument was principally reserved for research laboratories, instrumentation laboratories, people who were at the forefront of instrumentation and proteomics. But now it's a relatively common instrument.
In the early 80s we sequenced an unknown peptide. It was one of the first examples of doing tandem mass spectrometry, and solving a peptide structure. It was a small one — a cyclic peptide with an unusual amino acid. That was, I think, 1982.
So we were developing instruments, but we were spurred on by possible applications which we were conducting then.
Of course, there were no databases for identifying peptides then. None of those resources were available then. So it was a pure and simple chemical problem, sequencing peptides then. I guess it's what you'd call de novo sequencing today.
I came to Washington University 12 years ago to assume the principal investigator position. My mandate and goal was to expand the resource, which, say 15 years ago, was a resource that had done extremely nice work in GCMS and trace analysis and isotope ratio mass spectrometry.
So now the resource became one that emphasized tandem mass spec and Fourier transform mass spec, and began to move more into the areas with the modern ionization methods, like electrospray.
There was here a small electrospray instrument. There was no MALDI, no tandem mass spec.
We brought those things here, and continued to do instrument development, but also to move more actively into application areas.
Indeed, the university needed the mass spec capability, and this program at the NIH called the National Centers for Research Resources is the one that was funding the resource. And because of the commitment of Wash U to update and invest in [the resource], that was convincing to the NIH that they should continue funding the resource here.
What were your first projects at Washington University?
The first years we worked mostly on DNA. Not on trying to have a competitive method for sequencing, but principally working on modified DNA where mass spectrometry can play a role. We worked on method development, and then in collaboration with my colleague here, John Taylor, and a colleague at the Eply Cancer Institute in Omaha, we worked on identifying modified DNA — modified either by radiation, or by carcinogens — involved in lung and breast cancer.
Most of this was done to support basic biomedical research, in trying to find the origins of disease, and also, as a spinoff, to try to identify potential biomarkers.
When did you get into protein work?
The protein work began almost immediately with an interesting application that involves proteomics. It has to do with antigenic peptides — major histocompatibility molecules — that are involved in immune response. These are proteins that ultimately end up on cell surfaces. They present peptides from all the proteins in the cell that are undergoing proteolysis. So what you get in the analysis is a partial sequence of a peptide that could be from a self protein, in the case of an immune response, or autoimmune response, a peptide that is from a non-self protein.
So the name of the game is what are these MHC molecules? What's the biochemistry of this peptide presentation? How does it happen? What are the families, their properties and motifs for recognition and binding to the peptides? That has turned into a proteomics project because we are interested, for example, in those peptides that are presented by a mouse that gets diabetes, verses a mouse that does not. The diabetes project is being done in collaboration with Emil Unanue in the immunology department here at Wash U.
One could say, 'Let's do the diabetes proteome,' and then do all the proteins. But in this case, the focus is on the antigen, hypothesizing that the person, or mouse, that gets diabetes is presenting as part of the MHCs a peptide from a protein that is mistakenly identified as antigen — as non-self. That causes the beta cells of the pancreas — the cells that make insulin — to be destroyed.
To me, it's kind of interesting, because our immune system has been doing proteomics for a long time. What the immune system does is it recognizes that a cell is infected, for example, because it has proteins that are non-self that are presented as peptides that are identified. So the message is in the peptides again, just as in proteomics. By recognizing the peptide, the immune system can tell if the cell is OK or not OK.
So we don't isolate the proteins. We analyze very complex mixtures of peptides from these MHCs for the purpose of trying to understand what kind of peptides are presented. Like in regular proteomics, we never sequence the whole peptide. We only sequence parts of it. We use accurate mass with Fourier transform. We need that because in database searching there are no restraints. We're not digesting with trypsin, or with any other enzyme. So the searches are very slow. We have to interrogate every possible peptide.
But we don't work with the proteins. The living system does. From the peptides, we identify the protein. We get the peptides directly from the cell system by isolating the MHC proteins that have peptides bound to it.
By sequencing the peptides, we can figure out what proteins are being proteolyzed in the cell, and we can understand, for example, what features, or motifs, a peptide has the allows it to bind to the MHC.
What have you found with the diabetes project?
We have found that there's a very significant difference between the mouse that gets diabetes and the one that does not. The MHCs themselves are different. Not significantly, only by two amino acids.
The range of peptides that are presented is amazingly different. It's completely orthogonal — that is, the peptides presented by one system, as far as we can tell, do not overlap with the other system.
And the other feature that's quite interesting is this: In the region of the peptide that binds to the protein — that region comprises nine amino acids — at position nine, there's almost always an acidic amino acid in the mouse that gets diabetes, verses the mouse that does not.
Have you analyzed the functional differences between the proteins in diabetic vs. non-diabetic mice?
No. We don't know what is the antigen that causes diabetes; that triggers T-cells to destroy the beta cells from the pancreas. But we know something about their properties now. We know that this small difference in the MHC produces an incredibly large difference in the peptides that are presented.
And we've extended this to an MHC system from humans, and we get incredibly similar results.
Do you think this could lead to some kind of therapy?
Well, it could, if we could find the antigen. What we can do is look as at many MHCs as we can to get a feeling of the landscape. The ultimate goal would be to discover what does the MHC present in the mouse that gets diabetes that causes the T-cell to trigger destruction of beta cells in the pancreas?
At the moment we've acknowledged that finding these one or two antigens in a very complex mixture will be difficult. But we're taking advantage of proteomics to understand what kinds of peptides bind. So we're getting some idea of what properties the antigen will have.
What other biomedical mass spec applications are you working on?
The other application of mass spec that we have going on builds on a long and ongoing interest at Washington University in muscle. There's a syndrome called muscle wasting that inflicts people with AIDS, and has implications in obesity and aging.
Part of our resource for many years was devoted to trying to understand the details of muscle wasting by doing amino acid analysis. And the reason that amino acid analysis was done is that one could inject into a human isotopically labeled amino acids, and use that, plus isotope-ratio mass spectrometry, to understand the rates of muscle metabolism. So it's a metabolism study.
It tells you about metabolism, but it doesn't say anything about the basic mechanism. What is it about the muscle from an AIDS patient versus a normal person that leads to muscle wasting?
So we're building on a decade of experience with isotope-ratio mass spectrometry in understanding metabolism of muscle in AIDS vs. non-AIDS patients. Now we're in a position to build on that with analytical proteomics, to try to understand what it is that causes wasting.
We're doing classical proteomics experiments centered on the mitochondria of muscle cells. It's a collaboration with two groups in the medical school that are affiliated with the Resource: Reid Townsend's group, and Kevin Yarasheski's group. Kevin's worked for years between the mass spectrometry lab and the clinical laboratory in understanding muscle metabolism.
We're using Fourier transform mass spectrometry to do classical control versus diseased state proteomics.
Are you also doing instrumentation development work?
Yes. We're developing a new design cell for the Fourier transform mass spectrometer. This is the heart of the instrument, where the ions are injected, and where the accurate mass measurements are made. We're trying to get more dynamic range out of the Fourier transform instrument, and better accurate mass measurement by cell design.
That's our number one instrument project.
The last idea that we're working on that's related to proteomics is we're using biophysical methods to try to understand protein-ligand interactions. We use both hydrogen-deuterium exchange and the new method, where we use hydroxy radicals to probe protein surfaces, and to watch how protein surfaces change as the result of binding to a ligand.
So I would call this an exploration of protein function and protein folding. You've already identified the protein. What you're trying to do now is learn something about its properties — namely, how does it bind ligands?
We're using tandem mass spectrometry as our measurement tool, coupled with an idea that might be known as chemical footprinting. We're footprinting proteins, with and without their ligands, by using two strategies that we're developing and implementing: one is using HD exchange, and the other one is reactivity with OH radicals.
The idea is that those proteins that are available, and not protected by being part of a protein-ligand interface, are available for HD exchange, or are available for reaction with OH radicals.