Director of Proteomics,
Cold Spring Harbor Laboratory
At A Glance
Name: Michael Myers
Position: Director of Proteomics, Assistant Professor, Cold Spring Harbor Laboratory, since 2001.
Background: Senior research fellow, Cold Spring Harbor Laboratory, 1999-2001.
Postdoc in Nicholas Tonk's laboratory, Cold Spring Harbor Laboratory, 1996-1998.
PhD in neuroscience, Case Western Reserve University, 1996.
Mike Myers is scheduled to give a talk on the improved detection of phosphopeptides by MALDI-TOF mass spectrometry at next week's Association of Biomolecular Resource Facilities conference in Long Beach, Calif. ProteoMonitor spoke with Myers to find out more about his work, and about his opinions on mass spectrometry.
How did you get into studying post-translational modifications?
I sort of started in that field. In my graduate work, I worked on kinases and kinase signaling that's involved in neuronal differentiation and survival. And then for my postdoc I was working on tyrosine phosphatases, and that's sort of what led into the proteomics work.
What kind of technologies were you using for those studies?
For the kinases, it was mostly biochemical purification, and then just assays with exogenous substrates and hot ATP, and then the typical Western blot and that kind of stuff. And then it was sort of more of the same for my postdoc, except for at the end. The lab that I was in developed a certain kind of mutation that would make the enzyme trap substrates. And so the way most people were doing it was they would make one of these trapping mutants, express it, immunoprecipitate it, and then just look to see what's coming down in the IP, and then based on the molecular weight, make a guess, and buy some antibodies from whichever vendor has something of that molecular weight, and hope that it was right.
But I started using mass spectrometry to try to identify those things.
Did that make things considerably easier?
Well, I wouldn't say easier, because if you guessed right, there was nothing easier than that. It was just a question of guessing correctly or not. So I wouldn't say it made it easier, but I guess it's a less biased approach, and I think it gave more unique information. Because with the guessing, you had to know something already.
Were you using tyrosine phosphatases to study pathways?
With tyrosine phosphatases, you can find them in the genome quite easily, because there's a catalytic signature site. So identifying new phosphatases was pretty easy, but figuring out what they did was quite difficult. So if you know the substrate, then you can figure out what the phosphatase does. The idea was to come up with a method of taking that trapping mutant to another level.
What did you do after you finished your postdoc?
So now, my lab is mostly interested in post-translational modifications, centered around signaling, DNA damage, and DNA damage repair. So that includes phosphorylation and ubiquitination.
How do you go about studying the phosphorylation related to DNA damage?
That's sort of been a difficult process, because getting the phosphorylation sites by mass spectrometry has typically been quite difficult. So you usually end up having to do enrichment and/or chemical modification in conjunction with the enrichment. Those sort of things work, but they don't seem to be very sensitive technologies, and they have a lot of caveats that go along with them.
Often times, if you're doing an enrichment from a complex mixture, you're relying on your mass spec results being based on one or two peptides, which is always super dangerous.
So part of my lab does nothing other than technology development, and they've been doing some stuff to make it more sensitive to look at phosphorylation by mass spec. I mean, we haven't tried to do this, but we're planning on doing a large-scale screen for things that get phosphorylated in response to DNA damage. But we're not at that point yet.
Which kind of mass spec do you find is best for doing this kind of work?
Oh, I don't know. Whichever one you have. I have a MALDI 4700 from ABI — that's their original TOF/TOF — and then I have an LTQ from Thermo.
For the phospho stuff, I'm sort of still up in the air about which is a better platform. The nice thing about the MALDI is you can collect the data from the spot, but the spot doesn't go anywhere. So you can collect data, analyze it, and then come back to that spot and collect more data if you want. Whereas with the LC/MS/MS going into the Finnigan instrument, you only have that one shot to collect data. If you want to go back and collect more data from that sample, you'd better have more sample to reinject.
So at this point, I'm not convinced that there's one instrument platform that's better than anything else.
What are the most important traits for a mass spectrometer to have for doing the post-translational work that you do?
For me, I think sensitivity is probably the most important thing. I guess mass accuracy and resolution are kind of related to that, but not directly.
When you say you're doing technology development, what aspects of technology are you developing?
Mostly, it's sample prep. There's a little bit of instrument development — the settings and stuff like that on the instrument — but most of it is just how to prepare the sample to get it in the mass spec.
We've found a matrix where the phosphopeptides fly very well in it. Traditionally, the problem with MALDI and phosphopeptides is that the phosphopeptides tend to be highly suppressed when you use MALDI as the ionization force. That seems in our hands to be a matrix-dependent event, rather than a property inherent to the peptide.
Is this new matrix something that you're looking to patent?
No. It's something that's been used as a matrix before for looking at plastic polymers. I'm not sure it can be patented as a matrix. You might be able to get away with it as a use patent, but I don't see much to gain by patenting that, for me, at this point.
How far have you gotten in your work with phosphorylation and DNA damage?
We haven't done a global search specifically for the phosphorylation sites, but we've done a global search for the ubiquitination sites. I can tell you that we also found a lot of phosphorylation sites when we did that work. We do find stuff without doing IMAC enrichment.
IMAC is immobilized metal affinity chromatography. Phosphoro groups have an affinity to certain kinds of metal. So you can immobilize the metal on a bead, and then you can pass your mixture over the beads, and the phosphopeptide should interact with the metal on the bead and be retained, and the rest you can wash away. Typically the metals that are used are nickel, iron, gallium.
We have a couple of problems with IMAC. One is that things that are negatively charged tend to stick a lot. So if you have a lot of aspartic acids or glutamic acids in a peptide, that peptide will tend to stick. And then the other problem we have is that things just won't elute once they bind. Well, either they won't elute, or they get destroyed by the buffer that you use to elute.
There are ways to get around the problem with the acidic residues. Forest White, when he was in Don Hunt's lab [at the University of Virginia], used a kind of chemistry to put methyl esters onto those groups, so they were no longer negatively charged. That pretty much solves the problem of non-specific binding to the IMAC beads, but there's a lot of side reaction to that chemistry, and there are some issues of how complete the modifications are.
So you don't enrich for phosphopeptides?
In one strategy we do, but not by using IMAC. This one actually is being patented, so unfortunately, I won't tell you about it now. But we have a different way of enriching for them, yes.
Whether or not to enrich depends on what you're looking for. If you have a candidate protein and you want to know where it's phosphorylated, you don't have to enrich for the groups then. But if you want to take a normal cell and a tumor cell and find out what's the difference in the phosphoproteome for those two things, you have to do some kind of enrichment. Otherwise, there are just too many other things there to get in the way.
Do you have any comments on what instrument vendors could focus on to make it easier to do research on post-translational modifications?
Yes, you just hit on one of my sort of pet peeves with instrument vendors.
I don't come from a mass spec background. I'm sort of mostly self taught. I think it's time that instrument vendors start seriously thinking about instruments that they can sell to basically any laboratory. So any fairly large group can handle having a MALDI instrument in their lab. And a good MALDI instrument will really tell basically anybody who does any kind of biochemical research a lot about what they're doing.
My background is mostly in biochemistry and molecular biology. I started my training right when PCR came out. Part of my training was pre-PCR, and part of it was post-PCR. PCR really changed the way that everybody did molecular biology. My view is that mass spec is the same thing for biochemistry.
Now it's at the point where pretty much any lab can have a very good mass spec in their lab, and it will be like a PCR machine — it will revolutionize the way that they do work.
So I can't imagine if I left Cold Spring Harbor, I'd have to go somewhere that would buy me a mass spec. I couldn't do work without a mass spec anymore. So the vendors, right now, they're concentrating on bigger, more expensive instrumentation. I think they need to concentrate on having at least a couple really good — they don't have to be research grade — but really good, turn-key mass specs that they can start selling to labs that do a lot of biochemistry. And I think a MALDI instrument is perfect for this, because a MALDI instrument is pretty difficult to break. An LC-based instrument is much easier to break.
I think right now a STAR [Q-TOF] from Applied Biosystems is like $250,000. And nobody's going to want to spend $250,000 for that instrument unless they're a mass spec lab. They need to get that instrument probably a little below $100,000. Then a lot of labs will just buy it, because at that point it's not going to be a monstrously huge expense.
What's going to be sacrificed to achieve the lower cost for mass specs?
I think they're going to get to a point where the sensitivity is going to be a problem. They're going to get to a point where instead of looking at your sample, you're going to be looking at environmental contamination. And that's probably not too far away. Probably another 10- or 100-fold increase in sensitivity, and they're going to be at that level, where a lot of the peptides that you see are going to come from bacteria and spores in the air.
The idea is that some people want to be able to do proteomics from a single cell. And when you're at the point where you can look at the proteins coming out of a single cell, the problem you're going to have is how are you going to keep all of the environmental contamination out? So you're going to have a single cell, but how many bacteria cells are going to get in there, or how many mold spores are going to get in there, or how many pieces of dust are going to get in there?
They're going to get to that point, and it's going to be a nightmare. All of a sudden, every mass spec lab is going to have to have a big clean room, and we're going to have to don clothing to keep ourselves out of those samples.
So at some point, the issue is going to turn with sensitivity. Suddenly, you're going to be so sensitive that it's going to hurt you. You're going to have to start taking super precautions to stop environmental contamination from getting into stuff.
I used to have a problem where if people in my lab were doing Western blots on the same day that they were preparing samples for the mass spec, you would see a lot of casing coming from the milk that they would use to block. The problem with casing is when you're weighing it out, you're basically dumping out powdered milk onto a scale. Then you get this cloud of powdered milk that's floating over everything. We were picking up casing that was coming from that powder. It's probably not that far off that these things are really going to start affecting people.
I think the mass spec vendors should concentrate on making [mass spec] not a specialized discipline any more; on making it the sort of thing that everybody is capable of doing. I think the main thing right now is to bring down cost.
Now there must be 10 different software packages you can use to analyze data. At least the back end is getting a little bit easier to handle. There's no reason that my lab should be the only lab at Cold Spring Harbor that has mass specs in it.