At A Glance:
Name: Forest White
Position: Assistant professor, Biological Engineering Division, Massachusetts Institute of Technology, since July 2003.
Prior Experience: Senior research scientist, MDS Proteomics, Charlottesville, Va., 2001-2003
Institute for Proteome Research, University of Virginia, 1999-2000
Postdoc, University of Virginia, 1997-1999. Worked with Don Hunt.
PhD, Analytical Chemistry, Florida State University, 1993-97. Studied with Alan Marshall.
How did you become interested in mass spectrometry?
[I did my PhD] at Florida State University, [working with Alan Marshall] at the National High Magnetic Field Laboratory. I actually went to Florida State to do organic synthetic chemistry and did that for a semester, and then talked to Alan about switching labs. He gave me a couple of papers that he had published recently on FTMS. I read the papers and I knew: This is what I wanted to do. Mass spec was it.
My PhD was basically on designing and constructing a novel FTMS instrument that would have multiple ion sources. What Alan wanted [me] to do is construct a system where you could do MALDI and then, if you weren’t getting the right results with MALDI, or you wanted to switch projects completely and go to electrospray, you would have only 15 or 20 minutes of downtime.
What led you to biological applications?
I wanted to learn a little more biology because instrument design is pretty cool, but it’s much better to design an instrument with an application in mind. The person who really has pushed the field of bioanalytical mass spectrometry is Don Hunt, so I went and postdoc-ed for him. When I got there, he was still very focused on doing immunology and looking at MHC class I and class II peptides. During my time there, he switched to bringing in a large component of proteomics into his lab, and that’s actually how I got started in proteomics.
One of my main projects [in Don’s lab] was developing phosphoproteomics, [together] with a first year graduate student called Scott Ficarro. (see PM 1-20-03). Don came into the lab one day and said, “I want you to analyze protein phosphorylation sites on blood platelets.” We went and read the literature, and there are a lot of papers published on IMAC. We applied the standard methodology and saw instantly what a lot of other people had seen, that there is [a lot of] nonspecific binding to an IMAC column. Don [suggested] that if you block the carboxylate groups, you [could] get rid of the nonspecific binding. Scott and I went back into the lab and spent the next several months implementing Don’s suggestion. A couple of years of development later, we applied it to yeast, and published a paper in Nature Biotechnology last year (March 2002;20(3):301-5).
After my postdoc, I spent two years in Don’s Institute for Proteome Research, which was funded by Novartis Agricultural. They presented a broad variety of very interesting problems, including posttranslational modifications, [and] phosphorylation was one of the main projects there. It was actually all in plants, so it was a completely different ballgame. The overall level of phosphorylation is much lower [and] there are a lot more issues with dynamic range. Overall, it was much more difficult looking at the phosphoproteome of a given plant than it is to look at the phosphoproteome of yeast. We looked at several other projects, and developed a technique for mapping novel sites of ADP ribosylation on proteins.
Does the institute still exist?
It was around for about two years, and then MDS Proteomics came in. They hired several postdocs from Don’s lab, and several of us who were in the institute, and that pretty much was the end of it. [The institute] was funded very well by Novartis Agricultural, but that was the only partner. So at any point they could have pulled funding, and it wasn’t a stable atmosphere.
I think [MDSP’s] main goal was to get Don Hunt’s technology for phosphoproteomics that we had developed, and also the differential analysis technology, and to be able to tap into Don as a mass spec consultant.
What did you do at MDSP?
We were pushing quantitative phosphoproteomics forward in human cells. We worked on several pilot projects for a variety of pharmaceutical companies. We would take a given kinase inhibitor and look at the phosphoproteome either before or after its addition and look at changes. If the kinase inhibitor was very specific, you would inhibit a given pathway. It’s a really nice technique for figuring out what a specific signal transduction pathway is.
How did you improve the technology?
Our mapping of a given phosphoproteome [became] much more comprehensive. [In the] yeast [study, which] was actually done at MDSP, we mapped 270 or so phosphorylation sites. At the end, right before I left the company, we were mapping on the order of 1,000-1,200 phosphorylation sites from a given sample.
The quantitation is the other thing we pushed forward, being able to do reproducible quantitation from cell state A to cell state B. [For this], we use isotope-coded derivatization. During the esterification procedure, you add a methyl group to the carboxylate. We can use either a non-deuterated, or a deuterated form of methanol and put on three deuteriums on each of the carboxyl residues. Not only does this give you quantitation, but it also gives you a check on your sequencing, because now you know that you have a certain number of aspartic acids or glutamic acids in your sequence.
You have just become an assistant professor at MIT. Why did you decide to go back to academia?
The main reason was that I really want to develop the phosphoproteomics technology beyond what we can do in a company. We had pushed the technology to a point where we were able to apply it to many different problems, but we still are not [able to deconvolute] the signal transduction pathways in the cell. To do that, there needs to be quite a bit of technology development, and the nice thing in academia is that you don’t necessarily have a timeline for that, you don’t have to do it in two months or six months. Plus, I tend to be excited by a broad variety of different projects.
What are your plans for the next few years?
One of the initial projects will be looking at the yeast cell cycle. If we arrest yeast at some given point in the cell cycle and then release them and take samples at discrete time points and look at the phosphoproteome, we should be able to figure out the signal transduction that’s going on at each of these time points. A lot of people have obviously looked at the yeast cell cycle and identified phosphorylation sites on given proteins, but no one has really taken a system approach, trying to figure out exactly how the different signal transduction networks are talking to each other. The nice thing about being at MIT is, there are some excellent biologists here who are experts in looking at the cell cycle, and there are also some very good modelers who can use the data to model signal transduction pathways. There is an effort here called CSBI, which is the Computational and Systems Biology Initiative, [and] I am going to be part of that. It’s basically a tie-in of a lot of different approaches to systems biology, including RNA expression biology, proteomics, and some very high-powered computing.
What I like to do initially is some technology development, so we can go in and map a larger fraction of the phosphoproteome. When we published that paper with a couple of hundred phosphorylation sites, we were looking at maybe one percent, or less, of the yeast phosphoproteome. What we would like to be able to do is look at maybe 20 percent of the yeast phosphoproteome and then track that 20 percent throughout the cell cycle and see how it changes with this quantitative approach.
The [ultimate] goal is to start looking at cancer, to figure out the signal transduction that occurs as you go from carcinogenesis and then progress through the various stages of cancer to metastasis. That should yield a broad variety of drug targets, so industry would be interested in this. Once we work on the yeast cell cycle, we will be moving into humans quite a bit, looking at breast cancer to begin with.
What challenges remain in proteomics?
There are three big problems: Sensitivity, so we need to develop more sensitive instruments and more sensitive sample handling techniques on the front end. There is dynamic range, which again can be addressed at the instrument side, and at the sample handling side by fractionation to simplify the mixture. The last one, that’s huge right now, is going to higher throughput [analysis]. That was one of the main issues within the company. After a while, you realized that your throughput is not enough to do what a lot of pharmaceutical companies wanted you to do, which was to analyze 30 samples or 50 samples and do it ten times over for each sample, so that you had real statistical numbers, and go down deep into the phosphoproteome on each sample. I think with the right instrumentation developments, and [advancements] on the software side, we can start addressing a lot of these problems. FTMS is the way that it has to go right now. Your peak capacity in FTMS far exceeds peak capacity in any other instrument, so you can look at more species with higher dynamic range and more accurate mass than you can on other instruments. This allows you to probe deeper into the proteome, and you are able to quantitate from one sample to another. If you can go deep into the proteome in a single analysis, now your throughput has increased quite a bit.