At A Glance
Name: David Goodlett
Position: Director of proteomics and senior research scientist, Institute for Systems Biology
Prior Experience: Held positions in protein mass spectrometry for Johnson & Johnson and Bristol-Myers Squibb; postdoc in Richard Smith’s lab at Pacific Northwest National Laboratory
How did you get involved in protein mass spectrometry?
I was primarily interested in biochemistry, and I got a master’s degree in chemistry, but I was interested in the physical sciences as well as the biological sciences. So for my PhD I studied protein chemistry and mass spectrometry, really with no idea that the field would do what it’s done — with a really naïve understanding of what the future might hold. I was interested in instruments and technology — that was the mass spec part — and I was interested in protein structure and how that’s related directly to function.
That was in the mid ‘80s. I finished my PhD in 1991 — that was at North Carolina State, in the biochemistry lab of Frank Armstrong, and the mass spec lab of Richard Van Bremen. I was just intrinsically interested in the instrumentation and the technology, like any mechanic would be interested in taking a car apart and putting it back together. During my PhD, electrospray and MALDI ionization techniques were developed. I didn’t have access to them during my PhD, so I did a postdoc in Dick Smith’s lab at the Pacific Northwest National Laboratory, where I used electrospray to basically get non-volatile things like peptides directly into the gas phase, at a sensitivity greater than you could get with other techniques. Those two techniques really revolutionized the field, and they came along at the same time the genome was being sequenced.
What happened after that?
I finished my postdoc, decided I didn’t want to go into academics, and had a few interviews in the pharmaceutical industry. I got a nice offer from J&J to set up a new mass spec lab at this little division called the immunobiology research institute that was trying to understand a potential HIV therapeutic, which happened to be a peptide. In the course of being there — not as a result of my work — it became clear that the peptide really wasn’t as efficacious as it needed to be. It wasn’t toxic but it really wasn’t efficacious in treating HIV infections, so J&J basically changed the focus of that unit. Realizing that that was going to happen, and wanting to stay in basic research, I found a job with Bristol-Myers Squibb back in Washington State; this time in Seattle. There we were working to sequence major histocompatibility complex peptides. What we were trying to do is discover specific peptides presented on the surface of cells that we could associate with different types of cancer. We were hoping to find and sequence one peptide and then potentially make some sort of therapeutic to kill those virally infected or expressing cells. BMS eventually shut that site down, really for financial reasons, not scientific reasons — people were doing good science there but had failed to produce a new therapeutic. In ‘96 I knew this was coming, and in ‘97 it happened. I had a new house that I was trying to fix up and a new baby on the way — my second — and I really just didn’t want to move at that point. I just happened to meet Ruedi Aebersold on an interview trip that I did at Zymogenetics, where I was interviewing for a director position. I didn’t get that job, but later on [Ruedi] called me and we talked about the Institute for Systems Biology. That sounded like a good idea; it certainly sounded better than moving back to New Jersey to work for BMS. I took him up on the offer. This was in ‘97. It was supposed to only take about six months to get the institute started; and in the end it took two and a half years [during which time I was at the university]. In the spring of 2000 it finally happened, and I got hired by Lee Hood to set up the proteomics group at the institute.
How do you interact with the method development efforts that go on at the institute?
Our group is fairly small; there are five of us right now. Three of those people are in the lab running mass spectrometers. We do some of what you would call analytical method development, which basically revolves around decreasing the time from the beginning to the end of the experiment, or extracting more information from the same amount of sample. Ruedi’s group, which is a huge group by comparison — mostly postdocs — is involved in a lot of different projects to develop methodology and apply that methodology in a first pass to develop proof of principle. Once that methodology is developed and proven, then at some point we make a decision to transfer that into our high throughput facility.
How much of your work is geared toward research that you would call your own?
Certainly not more than 50 percent and probably more often no more than 25 or 30 percent on a typical day. For instance we have a paper that has just come out where we utilized gas phase fractionation to chop up the mass-to-charge scale and try to get at low intensity mass spec signals — essentially peptides — in a more efficient way. The other thing we’re working on is a technique that we’re referring to as shotgun CID [collision-induced dissociation]. This differs from the normal way we fragment peptides today, in that typically we take a single peptide that comes off an HPLC column in the presence of hundreds of other peptides, pick one, and fragment that one peptide. Then we generate information on the parent protein that the peptide came from by taking the fragment ion profile, and relating that information to the parent that identifies the protein. The shotgun approach, rather than being serial, is parallel. In that case we take everything that comes off the HPLC column simultaneously, we first measure all the parent ions, and then we fragment all of those ions together.
Computationally, it’s a little bit more interesting, you might say. It’s more difficult to get protein identifications out, but it’s not impossible, and basically it lets you get higher coverage in a single pass. In the normal serial manner in which ions are selected, there’s basically not enough time for the mass spectrometer to go through everything that’s co-eluting, pick [every peptide] individually, and fragment it. You don’t have enough duty cycle. You run out of duty cycle, and in a single pass there are peptides that are in there, usually the lower abundance ones, that just get missed. They aren’t identified because they don’t get selected; you don’t even know they’re present. With the shotgun CID approach, the idea is that now that you’re fragmenting everything together, you’re not going to lose any information. The single problem with it, however, is that you will occasionally get a degenerate answer, meaning that instead of telling the biologist, ‘It’s protein A,’ you tell the biologist, ‘Well, it’s protein A, B, C, or D.’ The point is that we think it’s a better situation than telling the biologist, ‘I only saw protein A; proteins B, C, or D might be in there but we don’t know because we didn’t see them.’ Some of them we’ll get absolutely correct, and some of them we won’t, and there’ll be a degenerate list. We’ll give that to the biologist, and the biologist can go on to do something like PCR or a Western blot to confirm which proteins are present. The point is that we think in this case it’s actually a cheaper approach [to exhaustive efforts to identify every protein], because it better uses the sample the biologist might have spent six months generating. In one pass we hopefully extract more information that the biologist can use.
What kind of hardware do you have in the lab?
For hardware we have a couple of chromatography stations; they’re both from Applied Biosystems. One is called a Vision and the other is called an Integral — they both let us do multidimensional chromatography. The difference is that the Vision can collect fractions and then reinject them onto a second column, so the separation is a little more complex. In the high-throughput facility, we currently have three mass spectrometers that are working; two of them are ion traps from Thermo Finnigan. One has an Agilent pump, and the other has a Surveyor pump from Finnigan. They both have Famos autosamplers, which are sold by Dionex now in the US. Then we have a Micromass Q-TOF Ultima that we got earlier in the year, and that has a Waters CapLC system on it. It actually has the guts of a Famos autosampler [as well], along with a Waters syringe pump system. We’re still getting the Q-TOF up and running, but the traps are surprisingly 24-7. It’s an amazing thing to me that we get as much output out of them as we do. There’s usually a problem with the chromatography and the sample will plug the plumbing somehow. But even that seems to be pretty robust with the current setup. We can get down to a few attomoles of peptide loaded on the column. The setup is both rugged and sensitive. One of the things we’re looking forward to is something that Kratos, the Shimadzu subsidiary, is beginning to sell. It’s called a MALDI-ion trap-TOF instrument, and that we hope will be used a lot by the folks here at the institute who still run gels. They might want to cut a band out of gel, put it on a MALDI plate, and do traditional mass mapping followed by some MS/MS on a select few of the peptides that don’t fit the mass map profile. We’re supposed to get that this month.
What is it like working at ISB?
It’s somewhere between academics and biotech or pharmaceutical life. It’s definitely more on the academic side. For the first two and a half years we’ve been in startup mode, and if you’ve ever been at a startup, it’s just insane. No day is the same. Some of us like that, and it’s a lot of fun. There’s a certain subset of society who aren’t good at doing the same thing, day in and day out! So far we haven’t had that situation. I expect that we won’t for a while. Even in our group, where, say, 70 percent of what we do is routine, the biology is never routine. The fact that we’re designed to transfer technology in from Ruedi’s group means that in a year we’ll probably increase our capacity, and have some other functionalities within the group.