Name: Fred McLafferty
Position: Professor Emeritus, Cornell University, since 1992
Background: Professor of Chemistry, Cornell University, 1968-92; Professor of Chemistry, Purdue University, 1964-68; Dow Chemical, 1950-64 (from 1956: Director, Eastern Research Laboratory for basic research, Framingham, Mass.); Postdoctoral fellow, University of Iowa, 1949-50; PhD in Organic Chemistry, Cornell University, 1950; BS in Chemistry and Mathematics, University of Nebraska, 1943.
Fred McLafferty leads a Cornell research team that has developed a new technique it claims can increase by four times the size of proteins that can be analyzed, from those that contain about 500 amino acids to those containing more than 2,000 amino acids.
Using a top-down approach, the researchers break up the protein into pieces and weigh the masses of the entire protein and of the individual pieces, then match them against those of known protein sequences in a database to identify the protein. Their work is described in a paper published in the Oct. 6 issue of Science.
In 2003 McLafferty spoke with ProteoMonitor about his work in protein analysis and the top-down approach [see ProteoMonitor, 6/13/2004]. Recently we caught up with him to talk about the new technique reported in Science.
Can you describe your new technique?
This top-down technique is where we get the whole protein into the mass spectrometer. And that gives us the molecular weight so we know the size of the whole protein. And then if we break it up and it’s a linear protein, then the pieces that we get can tell us about the sequence and the about post-translational modifications. For example, if we break a bond 100 residues in from the N-terminus and get that mass, and then we get another mass that’s 57 … daltons, greater than that, then we know that that indicates that the next amino acid’s glycine because glycine weighs 57.
Of course the DNA predicts sequence so that you can look and see if the masses come out the way the sequence predicts. And if there’s a jump in the sequence, let’s say by 14 mass units, that’s like putting on a methyl group, and so it indicates a methylation has taken place. So, this top-down technique has been valuable, except we haven’t been able to break up proteins when the proteins get as big as, let’s say, 50 kilodaltons. We have a great deal of trouble breaking them up inside the mass spectrometer because they form tertiary structures just like they fold up in solution, only they fold up even tighter in the gas phase.
What we’ve done is find ways to break these up by adding energy. The first way we found is that we can put things in the solution; we electrospray solutions of the protein into the gas phase. And by putting in additives that we didn’t expect, they break up more easily.
And then as they enter the mass spectrometer through a capillary — and this is still at high pressure, this is where the solution is evaporating and the protein is first going into the gas phase — if we heat the capillary, it slows down this folding into this very tight structure. You then have a collision region that’s still at relatively high pressure, but these low-energy collisions keep them from folding up, and then we have yet another collision region at lower pressures where we have much more energy and this then can break up the protein.
And so with this sort of combined crash-and-bash method, we now, instead of [seeing] relatively few cleavages for 50 kilodalton proteins, for ones that are even greater than 200 kilodaltons, we’re able to get cleavages and find out about sequence errors and post-translational modifications.
What kind of advantage does this technique offer over the bottom-up approach?
The first thing, [with] the bottom-up approach, you break up the protein ahead of time. So that destroys your molecular weight information. And then you have to get your sequence and your post-translational modification information from just the masses of the small peptides. Of course, if you don’t know that there’s been a modification, you don’t know what mass to look for. If it’s a totally unknown protein, then the particular peptides can be changed in mass by any sort of modification.
If you’re doing top-down, you can look and see how much the molecular weight has been modified and that tells you what the total numbers of modifications are. And then you look at the pieces and see how much each piece has been modified or hasn’t been modified.
The bottom-up approach is actually very good for identifying the protein. You have the sequences predicted and as soon as you see half-a-dozen peptides of the predicted masses, then you can have certain confidence in the identification. But it’s very difficult unless you’re looking for a very specific modification in a protein that you already understand.
Does the bottom-up approach offer any advantages over your technique?
The bottom-up approach is cheaper. We used a Fourier transform mass spectrometer, which is the most expensive one and it takes more training to run the mass spectrometer, shall we say. And there are not that many around. So all of these are restrictions.
The bottom-up approach is well established in hundreds of laboratories, so there’s no reason they should stop doing it. The easy way to say it is if they can’t solve their problem with the bottom-up approach, then the top-down approach stands ready and waiting.
But before that, we couldn’t go to large proteins and, of course, the bottom-up approach can break up any protein. They can digest any size of protein but they don’t do too well with large proteins because the mixture they get is so complicated.
How did you go about creating this technique?
For some years, we were the main players that did this top-down approach. Everybody else did the bottom-up. And that was a great way to get everybody in the business, so to speak.
We ran into the problem of how to break these things up, and part of my research here has been on these conformations, the way the protein folds up in the gas phase and how it folds up differently than in solution. Of course, protein folding in solution is a marvelous field of research all by itself. The native conformational structures are the key to enzymatic success, and so there’s a lot of research like that at Cornell, so doing it in the gas phase was just a complementary way to see how they fold when you take away the aqueous solution.
And so that research helped us understand why we were having these troubles and helped us try things that would get us around these troubles. I think the main thing [is] since we were doing more top-down than anybody else, we were having more trouble than anybody else.
What is this going to do for proteomic research?
The top-down approach for unknown proteins gives you far more information on structure, so if you’re isolating proteins of high biological interest, certainly you want to get all the information you can, and this then is a technique that gives you far more information. You never know what specific information is going to be the key to things, but post-translational modifications are now becoming a very interesting part of things.
This kind of research has become far more popular in just the last few years. The bottom-up approach really just can’t do this kind of thing. It’s being appreciated now that top-down with smaller proteins is a very powerful tool. Now that we can go to larger proteins, I think that people will start applying that to larger proteins.
How easy will this be for other researchers to do?
Of course, nobody is doing it yet, but I’d think there shouldn’t be any real trouble in pushing this. I do have former students in most of these instrument companies. For example, our laboratory came up with a technique called electron capture dissociation that in my prejudiced view, everybody does believe it’s been a revolutionary technique for both bottom-up and top-down methodologies. And when we first did it on the Fourier transform instrument, people thought, ‘Well, you’ll never use it.’
But now it can be used on most any instrument and everybody’s doing it, and all the instrument companies offer it, so this’ll be used by other people. It’s useable with other types of mass specs. The real problem is that when you go to a much larger protein, you get a much more complex mixture and the Fourier transform mass spec has, by far, the best resolving power and is, by far, the best at doing mixtures and things like that.