Name: Julie Leary
Position: Adjunct professor of chemistry, University of California, Berkeley, since 1985.
Director of analytical facilities, University of California, Berkeley.
Background: Post-doc, Woods Hole Oceanographic Institute, Woods Hole, Mass.,1984-85.
PhD in chemistry, MIT, 1984.
BS in chemistry, Lowell Technological Institute (now part of University of Massachusetts, Lowell), 1977.
How did you get involved with proteomics?
I noticed about a year and a half ago that there was a lot of interest in getting information on inhibitors for enzymes. There were some people who were synthesizing combina-torial libraries of enzyme ligands —whether they were to be used as substrates, or as inhibitors. There was clearly a need for screening large numbers of these compounds, and I thought mass spectrometry really was the best way to do that — since you detect based on mass-to-charge, you could in fact do some sort of before-and-after analysis of your library, before it was first synthesized, and then after it was incubated with your enzyme or protein, to find out if it was a binder. So we published an article in [Proceedings of the National Academy of Sciences] that showed you could immobilize enzymes onto solid supports, and then incubate the library with the enzyme, and if something bound to the enzyme, it would stay behind with the solid support enzyme, and those things that didn’t bind would be left behind in the supernatant, and you would be able to tell with a ‘before and after’ experiment what bound.
The thing that was catchy with this was, we didn’t know if these ligands were binding specifically or nonspecifically. So that’s when I came up with the idea of ‘let’s develop a mass spec-based kinetic assay to get at the kinetics.’ Looking at the literature, it became apparent that many of the UV assays that are out there had a huge range of results for the same system. So different investigators were getting numbers that were rather different. A lot of that has to do with the fact that many enzyme systems —- either the substrates or the products —- don’t have chromophores that absorb in the UV, so investigators have to use coupling enzymes and work backwards. Whenever you introduce another variable like that, you introduce the possibility of more error. So that combination of wanting to screen combinatorial libraries of possible inhibitors and substrates, and seeing how much variation there was in the literature with regards to kinetic constants, told me that this was definitely a field that needed to be investigated more, and that mass spec might be the best way of doing this.
I’ve always been involved in mass spectrometry, but our main focus of research was structural elucidation of carbohydrates, which it still is. But I [also] now have another NIH grant to pursue this enzyme project — it’s about a year and a half old.
How does doing enzyme kinetics with a mass spec work?
When I decided I wanted to develop a kinetics-based assay by mass spec, the first goal was that it could not be any more complex or difficult than a standard UV assay, because if it was, people aren’t going to use it. Every enzymologist has a UV spectrometer in their lab, and for 60 years they’ve been using coupling enzymes to get at this data. So to try to convince them of new technology you have to make it very straightforward and easy to use. I also didn’t want to have to run calibration curves unless it was really necessary, because that adds another level of experiments you have to run in order to generate a calibration curve and then get the data from it. So what we developed is a single-point normalization factor. The way this works, is you first start off at some particular substrate concentration, which is in the linear portion of the Michaelis-Menton curve. You then incubate your substrate with your enzyme, and you monitor the formation of your product over time. At the point at which all of the substrate has been converted to product, so all you see in the mass spectrum is product and no substrate, you know that the concentration of your product must equal what the original substrate concentration was.
That’s crucial, because eventually what we want to do in order to set up our Lineweaver-Burke plot is get at velocity and product concentration. So now, if I have this ‘R-factor’, we put an internal standard into the quench solution, so we have an internal standard that has a structure very similar to the product: we know its concentration, we can measure the intensity of the internal standard in the mass spectrum, and set up a ratio of that information with the intensity of the product that we actually measure in the reaction in the mass spectro-meter. We measure the intensity, measure the intensity of the product, and we now know what the concentration of the product is because all of that substrate’s been converted to product-that equals our R-factor. So now we set up our tubes just as you would for a regular UV assay — different tubes with different concentrations of substrate — we put our enzyme in and start the reaction, quench it with methanol that has internal standard in it, then we can calculate all of the product concentrations from all of those different vials of different substrates with this R-factor. Then once we get product concentrations, we can calculate velocity, because velocity is just product concentration divided by the quench time.
We did several proof-of-principle studies for this. The first paper we published that showed proof of concept was in Analytical Chemistry in 2001. That was with glutathione S-transferase, so that was a model system where we knew all the km and Vmax values for all the substrates. We also did inhibitors with that — we knew what a good inhibitor was for that system so we calculated ki. Since then, we’ve done the kinetics for hexokinase, NodST, and we’re now working on an enzyme system with Chris Walsh [at Harvard] involving nobobiosic acid. The first few times we’ve done these measurements, we always dovetail them with the traditional method to make sure we’re somewhere in that same range of published results that other investigators have come up with. In several instances now we’ve seen that our data is actually more accurate.
The biggest advantage of our system is that you don’t need to have chromophores. So you can look at any system directly without having to couple it with an enzyme. The second advantage is, you don’t have to use radioactive labels. And you can do it on a very inexpensive ion trap instrument, so it’s not like you have to put it on a million-dollar mass spectrometer.
What are you working on now?
We just had a paper accepted in the Journal of the American Society of Mass Spectrometry on scanning multiple substrates at one time. So what you do is, you have your enzyme and you want to see which substrates are stronger and work better with the enzyme. We can put in three or four substrates at once and come up with specificity constants—-kcat/km.
You look at several substrates on one spectrum?
We [always] use single ion monitoring —- you’re not looking at the whole mass spectrum, and that’s why the quantitation turns out to be so precise and accurate-because you’re only monitoring one or two ions at a time instead of the whole mass spectrum. [With more than one substrate] you’re just monitoring the mass of each one of those substrates. So a typical one would have one peak for the internal standard, and then one peak for each one of your substrates.
How many substrates can you monitor at once?
It’s really going to depend on the enzyme system. For the proof-of-principle paper that’s coming out in JASMS, we did five. One of the things we’re working on now with Chris Walsh is looking at tandem enzyme systems, where you have two enzymes, and you’re monitoring the eventual product and the intermediates of that whole system from beginning to end. So you have, ‘substrate one reacts with enzyme one’; that gives you intermediate one; that reacts with enzyme two; that gives you product two, etc. This looks like it is actually going to work as well. The other thing that we’re doing is determining mechanisms of catalysis — you alter both the substrate concentrations at the same time if you have a bi-substrate system, and you actually see ‘how does this enzyme catalyze this reaction?’ Mass spectrometry can get at all this data very straightforwardly. So I’m really hoping to turn on the mass spec community and more importantly, the biochemistry community, to the idea that mass spec is very versatile and can get at a lot of kinetic parameters in one or two experiments.
Are there other labs working on mass spec kinetics?
Jack Henion several years ago showed that you could use mass spec to get at kinetic constants. But the procedures that he used are very different from mine — he again used calibration curves, which I don’t want to have to use and don’t use in any of our data. ... And also Don Douglas from Canada looked at pre-steady state kinetics using a dual flow system that he developed.
What else are you working on?
Another new project that we have that we’re really excited about is mapping the 40S human ribosome. This is in collaboration with Jennifer Doudna at Berkeley. Jennifer is interested in looking at post-translational modifications that occur to the 40S human ribosome during various disease states, in particular hepatitis C. We’ve recently been able to take the 40S ribosome, denature it into its component proteins, and then using FT-ICR, we can get the exact masses of all the proteins. Once we get the exact masses, we can use databases to search the proteins if they’re known proteins, and find out if any of them have been post-translationally modified. Then when we want to know where the modification is, we can do MS/MS. We haven’t published this yet, but we’ve been able to identify 24 of the 27 proteins, and all but one is post-translationally modified.