At A Glance
Name: Peter James
Position: Professor of Protein Technology, Lund University, Lund, Sweden
Prior Experience: Worked with Al Burlingame at the UCSF mass spectrometry facility, and at Thermo Finnigan in the early days of the ion trap mass spectrometer
At the recent 5th Siena Meeting on Functional Proteomics, Peter James told a few jokes at the expense of his fellow Europeans (Heaven is where Italians are lovers, Hell is where they organize everything...) and presented some of his recent advances in proteomics technology. Here he talks about his background and explains a few details of his methods:
How did you get into proteomics initially?
I came from Oxford, where I was a biochemist during the ‘80s, and where I was building a protein sequencer. I have always been in protein sequencing and membrane proteins. We were working on calcium signaling, but unfortunately the guy I was working with [R.J.P. Williams] was only interested in three proteins. The ATPase that pumps calcium out of the cell, the channel that lets calcium into the cell, and the calcium-binding protein in the cell which transforms the signal. Therefore our world was very small, and it was clear there were only three proteins in the world; [Williams’] motto was ‘calcium ber alles.’ We started doing knockout experiments, and we noticed that nothing changed. Some in the group said, ‘Well, our protein’s not very important.’ Quite the opposite, because when you knock it out, you see that about 30 new proteins are induced on a gel. The thing is so important that there’s about three or four backup systems.
That’s when I realized I couldn’t go through life looking at one protein, and that 2D gels were the way. That was ‘87. That was when we started doing 2D gels [and then identifying the proteins] by protein sequencing. But soon you realize, ‘I’ve done two or three 2D gels and I’ve got 200 proteins,’ which is 12 hours a shot with the Edman sequencer. That’s when I went to UCSF, and spent some time with Al Burlingame, and then went down to San Jose and worked with Finnigan, the mass spec manufacturer. I spent a couple years there, and then in ‘93 five almost identical papers appeared practically simultaneously with the idea of protein fingerprinting. That’s when we suddenly realized, ‘Wow, now we can do 200 proteins a day.’ That’s how I got into it, a real slow slide into depravity.
What were you doing for Finnigan?
I had a grant from EMBL, where I could choose to go where I wanted. No one believed in ‘90 that [mass spectrometry] was the way to go, but I had spent some time with Don Hunt and was really absolutely sure. I went to Finnigan to watch the instruments. That was when they had the ion trap under development; it was hidden at the back and no one was allowed to see it. I thought [the grant] was a good way of getting into the firm and getting to play with the ion trap. That was a brilliant place to be at that time.
I was working with Finnigan but also for EMBL but I basically brought the biological problems in, because at that stage they didn’t have a lot of biology. There was a bunch of mass speckies and they’re going, ‘trypsin, uhh…is it in the toolbox?’ It was that early on. So it was fun because they would let me do the opposite. They would also let me rip the mass specs apart. ‘What’s this? Rip! No you don’t do that, you idiot!’ They were really sweet.
What made you move to Lund?
I went back to Switzerland, but at that particular time the University of Z rich and the ETH, which is the Swiss Federal institute of Technology, were like Harvard and MIT. Two things, same place, both hate each other. I wanted to work with the medics, but the medics were at the university, and I was at the ETH – the big brother with all the money. So [the university] had no mass specs but all the medicine, and I wanted to get into medicine. At that point, in ‘98-’99, I decided to leave to go somewhere where I could do medical mass spec.
[At the same time] I was also doing some advising in Sweden. There’s a family called the Wallenbergs, who used to basically own every single company in Sweden. They have a fund that kicks up [about $100 million] in interest per year. The terms of the foundation require [that the interest is] reinvested into science. The Swedes didn’t really get involved in the human genome project so they really wanted to fix that and get into post-genomics. They set up two large systems, in the north of Sweden and in the south of Sweden, with four platforms: bioinformatics, DNA-based sequencing, phenotype profiling, which is RNA analysis, metabolite analysis, protein analysis – that was [where I was advising them] – and there’s structural biology. While I was advising, I suddenly realized I might as well stay in Europe [instead of taking a position in the US.] The opportunity just presented itself [to take a faculty position at Lund University].
What is your relationship with Amersham Biosciences? Are they commercializing your tags for quantitative analysis?
[Amersham] is commercializing the tags now. It’s slightly different from ICAT. In ICAT you have an affinity purification, in these you don’t. But there is a trick. In generation two [of these reagents] there’s a trick involving the hardware, software, and chemistry [that allows you] only to see those peptides that are either increasing or decreasing in expression level. We do 2D HPLC purification, but we don’t do any affinity purification. We’re not relying on one peptide per protein, which is essentially what you’re doing with ICAT. That’s dangerous [because] if there’s just one cysteine per protein you only get one chance at it if you are watching for the peptide to elute. [Although] we have [to deal] with all the peptides, we can ignore all those ones whose expression level is staying the same. That’s Mach 2 and Amersham is developing it.
What’s this integrated sample handling idea you’ve spoken about?
Basically you have a 96-well plate, and at the bottom of the 96-well plate you have a very fine aperture. You electrophorese all the proteins out of the gel pieces down through a membrane that only lets through peptides. So the proteins concentrate down, and we also add trypsin at the same time in a minute amount. The point is that if you deal with 5 picomoles of peptide, you’re just about in the range where trypsin works. But if you go down to 5 femtomoles then you’re way below [the amount] where trypsin is active. Trypsin will only digest at pathetically slow [rates], and mostly what you’ll see in the mass spec will be either keratins or tryptic fragments, or one or two peptides from your target protein, instead of 60.
But, if you take your 5 femtomol and concentrate that a 1,000 times, you get the same digest as you would with 5 picomol. It’s just makes the digestion much more effective, and because it’s in such a tiny volume, it goes incredibly quickly. The digests go in about five or 10 minutes. Then as the peptides are produced they go immediately through the membrane and stick to the MALDI target. There’s no sample handling of this at any point. We have one of these things in the lab, but we have to get this thing into a format other people can use. The agreement is for Amersham to get it together into a system they can put out.
Are there other similar projects in the more initial stages?
We’ve got a very large robotic system for chunking up 150 samples in a week. We’ve been collaborating with Advion Biosciences, because [although] MALDI gives you a certain number of hits, you’ve got to check the ones you’re not so sure about by MS/MS. Advion has come up with a nice little thing which we’ve had in the lab for the best part of the year in this horrible beta version, but now it’s working properly. [With it] you can analyze 1,500 samples in 24 hours. Basically it’s got a 96-well plate, and picks the solution up with a graphite-embedded pipette tip which then butts up into the back of a silicon plate, in other words one of these nano devices, and it butts up and electrosprays out of one tiny nozzle, and there’s 100 nozzles on the plate. For us that’s very important because we can’t afford any cross contamination; we can’t use the same tip again and again. We use these things once and then throw them away.
And you’re also developing non-gel based separation technologies?
Now we’re starting a new non-gel-based proteomics setup. In Sweden we’re doing all this SNP analysis, where you walk up to a patient, and say, ‘Hey you’ve got a 50 percent chance of getting breast cancer within the next five years.’ That makes them very nervous. So SNPs are nice, but it’s a hell of a gun to hold to someone’s head. On the other hand, once they’ve got the cancer, we can take the tumors out and say, ‘Well, the best treatment method will probably be this.’ But wouldn’t it be nice if you could have something a little earlier, like a a test? Together with [colleagues] who are working on protein chips, we’re working on a method for getting very low molecular weight stuff out of blood in a format that we can say, ‘Wait a minute, we see myosin, or myosin fragments, and that’s a fairly clear diagnostic of someone having had a heart attack.’
If you get cancer, we pick out the highly expressed proteins and then try and find them or bits of them in the blood as an early diagnostic. Let’s say it’s a very early stage tumor, when it’s not yet a tumor but it’s heading in that direction. Let’s do a blood test every three to six months. That’s where we want to go: very early stage diagnostics. Trouble is, many of these things are membrane proteins, and that’s why we’re looking to membrane protein analysis. Those are the buggers that are exposed to the surface. When a cell is falling to pieces, the membrane proteins get the extracellular stuff chopped off, so those are probably the best markers.