Professor and associate head
University of Georgia
At A Glance
Name: Jon Amster
Position: Professor and associate head of chemistry, University of Georgia, since 2003. Faculty since 1988.
Background: Postdoc in chemistry, University of California Irvine, 1986-1987.
PhD in analytical chemistry, Cornell University, 1986.
Fourier Transform mass specs are known for their high resolution and high mass accuracy. But despite these qualities, the instruments have some disadvantages over other types of mass specs, such as tandem mass spectrometry.
Jon Amster recently published a paper in Analytical Chemistry describing a method for improving the specificity of Fourier Transform mass spectrometry. He discussed the advantages and disadvantages of the instrument with ProteoMonitor this week.
Why do you choose to use Fourier Transform mass specs, and to develop improvements for this instrument?
Fourier Transform mass spectrometers are high-resolution and accurate mass instruments. We like that, and also, it's a very versatile instrument. You can do so many kinds of experiments with it. You can come up with all sorts of interesting tandem mass spectrometry experiments, in addition to mass analysis.
Its reputation is for being high-end and high-complexity. But I think that the barrier is now down for using these sorts of instruments, and there are certainly a lot of FTMS instruments going out into the field to people who are not trained as FTMS experts.
I've been using FTMS since I was a graduate student, which was 25 years ago. It's been a long time. I was originally attracted by its high resolution, high mass accuracy capabilities, and its flexibility in doing a variety of complex and sophisticated measurements. Over the years, I've been interested in different aspects of the technology.
When I first got into academics, I was using it to study ion molecule chemistry, with the goal of trying to come up with an ion molecule reaction that would allow us to get the amino acid sequence of peptides. That was a lot of people's dream. It never quite worked out that way, but other techniques came along that accomplished that goal fairly soon.
More recently, in our proteomics phase of development, we've been interested in applying the accurate mass capabilities of the instrument to try to increase the throughput of proteomic analysis.
When I first started thinking about this, the only comparison was really gel electrophoresis-based proteomics. So I first started thinking about using accurate mass measurement as a means to achieve a no-gel type of proteomics measurement in 1997. John Yates was working on MS/MS shotgun approaches at about that time.
I did some thought experiments and convinced myself that if one did a batch proteolytic digest of a protein mixture, and measured the masses of all the peptides that were in that mixture with enough mass accuracy, we could identify a high enough fraction of them to actually identify all the proteins that were present.
What are the main advantages of FTMS over other kinds of mass specs?
The main advantages are resolution and mass accuracy. There are other instruments now that claim low ppm mass accuracy. Time of flight instruments, for instance. Whether or not they can actually achieve that on practical samples is still a debate — there's a difference between what manufacturers specify and what users actually achieve in the field. But the resolution is important because if one is looking at batch digest mixture, even if you do some chromatography to reduce the complexity of the mixture, you're still going to have many components and many mass spectra that you obtain.
What we do to try to improve the throughput is do chromatography to generate a relatively small number of fractions to analyze — perhaps 100 fractions — and then use MALDI mass spectrometry to measure as many components in that fraction as possible. So for instance, if you had 100 components in a fraction, and 100 fractions, conceivably you'd be examining 10,000 peptides.
The competing methodology is using tandem mass spectrometry to identify those thousands of peptides that are present in a batch digest, and the advantage of our technique is since we're only measuring molecular weight of these peptides, we can examine several of them — let's say 100 of them at a time — 100 per mass spectrum. Whereas if you're doing tandem mass spectrometry, you have to select these one at a time, and in serial fashion obtain a tandem mass spectrum for each peptide. So that approach is slower, because you have to examine things one at a time, and it's more data intensive because you have to have a tandem mass spectrum for each component.
By our approach, we're just getting molecular weight information for each component. If you have a high-resolution mass spectrometer, you can get 100 of those at a time. Some of those might overlap in mass at low resolution, but at the kind of resolution that we get, we can separate these things. We can distinguish peptides that differ by 20 or 30 millimass units.
What are some of the disadvantages of using FTMS?
The disadvantage is that molecular weight by itself doesn't have nearly as much specific information for identifying the peptide and the protein that peptide came from as a tandem mass spectrum does. So a lot of our development work is directed toward making that identification based on molecular weight more specific. We try to develop methods that provide extra specificity in this identification step.
In the identification step, you start off with a database of potential proteins for the organism that you're studying, and you can calculate all the peptides that would be generated through a batch digestion of all those proteins. And you can calculate the molecular weights of all those peptides. That list of molecular weights is what you search with the masses that you measure. If you have enough mass accuracy, then hopefully a significant fraction of those peptides will uniquely identify a peptide in your list, and the protein that it came from.
But some of the drawbacks are that that only works well for relatively simple organisms — ones that don't have a large number of proteins in their database of proteins. So we apply this mostly for prokaryotic organisms.
It works best for organisms in which there is very little post-translational modification, because if you have to anticipate modifications of the peptides, then you have to include those in the list that you search against. And if you start putting too many things in that list, then the specificity of molecular weight goes down.
[This method] would be a poor choice for looking at human samples, or any type of mammalian samples. It's better suited for exploring fundamental biology, and there is a lot of work to be done there. Amongst prokaryotic organisms, it turns out that probably two-thirds of prokaryotic organisms are of a size that's suitable for this type of analysis.
Are you trying to make improvements so that you can study mammalian samples?
Well, potentially that will happen one day. Right now we're working on developments to make it work even better for larger bacterial genomes.
The recent Analytical Chemistry paper describes a technique to improve specificity by creating a mass label for a subset of peptides in our batch digests — ones that contain the amino acid cysteine. It doubles the identification specificity of those peptides.
In other words, if you look it up in a table, and it turns out that there's two peptides in your list that have the same mass, this technique would allow you to pick out the right one.
How much time does the new labeling technique take to perform?
In the sample preparation, it adds perhaps an extra 30 minutes. In the data acquisition and data interpretation, it adds no extra time at all.
Are you planning on filing any patents for this new method?
There are other technologies that patent aspects of this method. We've made a patent disclosure with our university. There's a company that sells some complementary technology. They've approached us, and they're interested in joint development efforts.
How are you going to apply this new method?
We're now working with a microbiologist at our university who has an archeal organism that he's been studying for a long time. He has produced some interesting mutant strains of the organism to probe certain key biochemical pathways. We're going to apply some of these techniques to study this organism from a proteomic point of view.
This organism produces methane from carbon dioxide, so it has the potential to be a source of cheap energy. We're examining some of the key biochemical pathways, which if you could control, could potentially turn this organism into an industrial-scale process for turning carbon dioxide into methane.
Where do you get your FTMS instruments from?
My instruments come from Bruker Daltonics.
What other FTMS improvements are you working on?
We're working on some methods to improve mass accuracy, because if you improve mass accuracy, that's the best way to get better identification specificity for any kind of proteomics measurement. We've recently submitted a paper on achieving sub-part-per-million mass accuracy using this technology for examining peptide mixtures. This paper describes a way to do it by external calibration, which is very important.