At A Glance
Name: Michael Snyder
Position: Professor and chairman of molecular, cellular and developmental biology, Yale University.
Background: Postdoc, Ronald Davis’ laboratory, Department of Biochemistry, Stanford University School of Medicine, 1982 to 1986.
Ph.D., California Institute of Technology, 1983.
B.A. in chemistry and biology, University of Rochester, 1977.
How did you get into proteomics and developing the yeast protein array?
Back in the late ‘80’s, we knew they were going to sequence the yeast genome and we came up with a scheme for tagging all the genes and proteins of yeast using transposons. It was actually the first functional genomics project. From that, we got very interested in the large-scale biochemical characterizations of proteins. That was sort of late ‘80’s, early ‘90’s, not long after I first came to Yale as an assistant professor. I think that’s probably where some of the interest stemmed from for creating a yeast array. I was never terribly interested in sequencing genomes, but I was fairly interested in trying to understand the functions of genes and proteins. This tiny project we launched was the first to get at those sorts of things. Basically from there, it certainly seemed logical at the time, the late ‘90’s, to try to characterize proteins in a very high-throughput fashion using array technology.
The word at the time was that you can’t make protein arrays because if you put proteins down on a surface, they denature, they’re inactive, there’s no way to make lots of proteins, blah, blah, blah. And I guess we kind of just ignored that. It was the case of having a very talented postdoc named Heng Zhu. He basically took on the task of seeing if he could do this, and sure enough a lot of what was out there was myth — you can put proteins down on a surface and keep them active. He explored a variety of different means and came up ways that we liked that worked pretty well. People just weren’t thinking of things in a large sense. The very first set of things we did were the protein kinases of yeast. So there were 122 of these, and he cloned 121 of them and got them active and expressed and such. That was a big deal at the time. From there, we decided that that was rather inspiring, we should do the whole yeast proteome. That was what we did next.
What did you do with your yeast proteome array?
We realized the thing to do was to set up a company because people were clamoring for these sorts of chips and things, so then we launched Protometrix, which was basically to be able to commercialize these things, get them out. It was fairly logical to take this into the industrial center, so we launched the company, and they got bought out by Invitrogen (see related story, p.3). But they certainly are really terrific at making lots of proteins. As you probably know, they launched the yeast array this summer, and now this substantial human array just got launched last week. The human array foundation work was done by Protometrix.
Going back to when you first developed the yeast kinase chip with 122 protein kinases — what were some of the challenges of developing the first protein microarray? Were there other arrays out there then?
Around the time we were doing our stuff, people, including Stuart Schreiber’s lab, had just done a proof of principle assay, just spotting a few proteins down. Their paper came out at about the same time as our 122 kinases. So people weren’t really doing anything, and even today the biggest problem is being able to get large numbers of proteins expressed. I don’t think the challenge was ever the array technology itself. There are ways of optimizing the arrays and getting them better and better, but at the end of the day the biggest challenge is being able to get large numbers of proteins expressed and active, and that’s just a matter quite frankly of doing it — learning how to make lots of clones and setting up ways to make proteins 96 at a time. And streamlining the procedures so they stay pretty active. It’s learning how to make all these things that’s the biggest challenge.
What kind of assays have you done using the yeast chip?
We’ve done all kinds of assays — protein-protein are the obvious ones. One of the very first things we published was interactions with phospholipids. We’ve also done probing for small molecules. We just had a paper come out in Science probing with nucleic acids — finding new DNA-binding proteins. So you can use them for many, many things. There is some assay development associated with all of this — there’s no question that takes some effort — but at the end of the day, the biggest challenge was learning how to make lots and lots of proteins.
How long were you involved with Protometrix?
The company physically launched in August of 2001 and it was bought out by Invitrogen this April.
Aside from the yeast chip, were there other products that were produced?
Well, yeast and human — those were the two main focuses of the company. There’ll be a whole series of human chips coming out.
What were you studying before that motivated you to make a chip?
That’s a good question. One of the things that definitely was riding heavily on our minds was that a lot of yeast proteins are phosphorylated. That’s true of mammalian cells actually — people say that one third of mammalian proteins are phosphorylation, and that’s probably an underestimate. So there’s lots and lots of phosphorylation, and it’s always been really quite challenging to figure out the kinases that are adding the phosphates to those proteins. So one of the reasons we started with the kinases was in fact to see if we could starting with the substrates see which kinases like to phosphorylate particular substrates. That’s actually what we did in our very first protein chip study. We did it on a fairly small scale — we had like 20 substrates at the time and we ran all the kinases through them. It was successful. It gives you candidate kinase substrate targets, and using that information you can go back and validate which was the right one.
Aside from the studies on kinases and phosphorylation, what other studies did you do using the yeast chip?
Probably one of the best stories I can tell you about the power of this approach was when we looked for new DNA-binding activities. The power of chips is that when you’re looking for biochemical activities, you’re doing it in a very unbiased fashion. You’re serving all the proteins at once. So what we wanted to do was see if we could find a lot of the DNA-binding activities of yeast. So we just labeled up total genomic DNA and probed the proteome chip and looked for which proteins would bind DNA. We made up a list and half of the things made a lot of sense. The other half were kind of weird stuff like metabolic enzymes, nuclear pore complexes and things like that. What we did was we went into the nuclear pore complex and metabolic enzyme guys and we set up secondary assays to see if any of them were associated with DNA in vivo. And sure enough, it turns out a metabolic enzyme that you never would’ve guessed in a million years is associated with DNA actually is both in vivo and in vitro associated with specific sequences, and when you knock it out, the genes that it binds near are altered in their regulation.
So by doing this sort of unbiased probing, you can learn new things that you wouldn’t think to look for — that’s the power of unbiased screens.
What direction is your lab going in for the future?
Half of the lab does do proteomics — most of those people are doing protein chips. Definitely a big area for us is this phosphorylome stuff I was telling you about. We’re almost done finding all of the substrates for every kinase in vitro. We want to do some mapping of the sites and see if they map up in vivo. A big project is to work out this whole map. We’ll probably extend that to other post-translational modifications as well. These are extremely unexplored areas. For the case of yeast, there’s less than 160 known kinase-substrate interactions, and there’s probably thousands or tens of thousands of phosphorylations out there, which tells you how little we know. And so I’d really like to get into that in a much bigger way.
We’re also going to get into metabolomics, and the other half of the lab does genomics, where we set up trip chips, which is a method of finding targets for transcription factors. Imagine you have a transcription factor and you want to know all the targets. You can actually pull down that transcription factor along with its associated DNA in vivo. Then you label that DNA and you probe a microarray that carries all the regulatory sequences of yeast, and in one swoop you learn all the targets of that factor. It works great.
We’re also trying to map out all the transcribed regions of both humans and yeast. We just had a paper that came out in Science Express where we tiled the whole human genome. So we have oligonucleotide arrays that cover all of the human genome. We probed with liver polyplus to see where all the transcribed regions were and we discovered 10,585 new transcribed regions throughout the human genome. There’s a lot of transcribed regions people don’t know anything about.
Are you working on any new chips?
We are. We’re building an Aribidopsis chip with Dinesh Kumar, also of Yale. We’ve got funding from NSF to do that. And we do have another chip that I can’t say much about until we do filing on it — stay tuned.