Markus Kalkum on Proteomics Using a Homemade MALDI Ion Trap

At A Glance

Name: Markus Kalkum

Position: Assistant professor, Beckman Research Institute, City of Hope, Duarte, Calif., since June 2003.

Background: Post-doc in mass spectrometry and gaseous ion chemistry with Brian Chait, Rockefeller University, 2000-03.

Post-doc in mass spectrometry, Max-Planck Institute for Molecular Genetics with Hans Lehrach, 1999-2000.

PhD in chemistry, biology, and pharmacy, Freie Universitaet Berlin and Max-Planck Institute for Molecular Genetics, 1999.

MS equivalent in chemistry, University of Constance, 1995.

How did you first get involved in proteomics?

I first got started on mass spectrometry during my thesis at the University of Constance in Germany at Michael Przybylski’s lab. This is where we had an interesting project about a cytokine called M-CSF — macrophage colony stimulating factor. We were very interested in characterizing the receptor binding site on that particular protein. We did this with chemical modification of histidine residues. So that was my first introduction to proteins and mass spectrometry.

Then I went to Berlin to the Max Planck Institute for Molecular Genetics, and did my PhD work there. But since the Max Planck Institute cannot offer the PhD title, I got the title from the Freie University of Berlin. That research was on mass spectrometric methods for proteome research. It involved bacterial conjugation and a collaboration with Dr. Eric Lanka. I am still collaborating with Eric — [the week before last] we just finished writing a review article for Current Protein and Peptide Science. He was working for many years on bacterial conjugation mediated by a plasmid called RP4. The funny thing is that it encodes for about 20 different genes that are responsible for transferring antibiotic resistance from one bacterium to another. The plasmid can be transmitted among many different bacteria species — so it’s not limited to one certain class of bacteria. We found that this particular plasmid encodes for a gene that makes up the main component of the conjugative pilus, or sex pilus. With mass spectrometry and a collaboration with one of [Eric’s] PhD students, Ralf Eisenbrandt, I found that the main subunit of the pilus is a circular protein — it has 79 amino acid residues and a head-to-tail peptide bond.

So after my PhD work, I left Berlin and went to New York and joined Brian Chait’s lab at Rockefeller University. This is when I really got started on proteomics. I did 2D gels before in Berlin, but in Brian Chait’s lab there are no 2D gels; there were very good collaborations with other biochemists and molecular biologists like Mike Rout and Robert Roeder at Rockefeller University. I’m still continuing a collaboration with Jinsong Zhang, a post-doc in Bob Roeder’s group. So I was working on histone deacetylase complexes in yeast and in humans, and I am doing that now. The publication is presently in preparation.

So your goal was to identify complexes?

We wanted to identify new components — unknown components — of these histone deacetylase complexes, to find out what their function is. My project with Brian Chait’s group is to systematically analyze how the proteins in that complex are connected to each other. We already know it’s not one complex — it’s different forms of the complex that you can get when you purify them from yeast cells, and they might all have slightly different functions. We focus on the core complex, which probably has 17 to 18 components. I’m trying to partition that complex by clipping parts of it. This is ongoing research.

What I learned in Brian Chait’s group is how important it is to make really high sensitivity measurements. It’s not so much that you need a very high resolution in mass spec to do proteomics — I think 11,000 resolution is probably okay for most peptide identification. But if you have high sensitivity and can look beyond the chemical noise — which gives you a limit in the femtomole range — then you can do really interesting things. The problem is, you cannot detect those peptides, due to the chemical noise, by normal mass spectrometry, but if you do MS/MS like on the MALDI ion trap which was built by Andrew Krutchinsky in Brian’s lab, then you are able to overcome this limit. This gives you a certain increase in sensitivity — I don’t know how much the factor is, but it’s at least around 10. And there’s lots of room for improvement.

So one of my new focuses here at the Beckman Institute is to further improve the MALDI ion trap and apply it to very specific biological problems.

Why build a MALDI ion trap — what are the advantages of this particular hybrid?

MALDI is a very robust technique in comparison with electrospray, and it can tolerate a certain amount of small molecules that interfere with the mass spectrometric analysis of peptides. The problem still is that due to the matrix and other small molecules, you get chemical noise. You also get chemical noise in electrospray mode, but it’s made of different components. It’s not really known what the noise is. There was a paper by [Krutchinsky and Chait] where they figured out that most of the noise in the MALDI analysis is clusters of the MALDI matrix with an unknown component. But if you focus on the selection and collection of ions in a trap within a certain mass range — preferably a small mass window — then you can remove those components that make up the noise. They fragment rather unspecifically and you get mainly unspecific signals and only a few specific signals from the matrix clusters. But when there is a peptide inside — hidden underneath the chemical noise —it will generate very specific fragments. And the beautiful thing is that the 3D ion trap uses resonant excitation, so that when you break your ion at certain positions and you get smaller pieces, they are out of resonance. This way you get only a few fragments, but these are very sensitive. So you’re not diluting your single [MS spectrum] out into many fragments; you get a small number of specific fragments. This is the reason I think it is a good technique. Unfortunately, you cannot buy [a commercial MALDI ion trap] yet.

So you started with a commercial ion trap and then added a MALDI source?

Yeah — it’s the same thing that Andrew did. You start out with a commercial ion trap and set up a new source for it. But I’m improving it in some ways so that it can be compatible with other machines — so that it can measure my MALDI plates first at very high resolution and sensitivity on a TOF device, and then afterwards using the ion trap for a very efficient fragmentation of the ions. I had a few other ideas for how one could improve it, but everything that I do in this regard will be to match needs that emerged from the projects of my biologic interests and my collaborators.

How long have you been at Beckman? Tell me about the Institute.

I started at the end of June last year. Ever since, I’ve been setting up my lab and trying to acquire external funding.

City of Hope is a national medical center and designated cancer institute. It has had a research institute since the ‘50s, and was endowed by the Arnold and Marbel Beckman Foundation in 1983 and renamed the Beckman Research Institute. Right now I think there are five different Beckman institutes in the US — at Caltech, Stanford, University of Illinois at Urbana-Champaign, City of Hope, and UC Irvine. The head and CEO of the institute is Arthur Riggs. There is a hospital here and the research institute collaborates well with the cancer institute.

There’s a well-established mass spectrometry group focused on proteomics that is directed by Terry Lee, and also a core facility here. They are all, like me, located in the immunology division. There’s good infrastructure, and there are experts on chromatography and separations and electrospray. They’re interested in the new technology I’m pursuing with MALDI and ion traps. This is why I got hired for that position.

Tell me about some of the applications of your MALDI ion trap work.

If you want to do real biochemical studies of certain protein complexes, and you want to find out what the functionality is, it’s good to combine this knockout strategy with the strategy to analyze complexes. So it turns out that you then have to analyze basically the same protein over and over again. If you had antibodies against all these proteins, you could just do a Western blot every time. But we don’t have antibodies against all these proteins, and it would also be rather expensive to do that. So we decided to set up a list of best-identifying peptides for each protein. I’m trying to use this technique now for a whole number of other complexes and variations.

The other idea is that I had this collaboration with Tom Muir’s lab at Rockefeller where we looked at peptides secreted from bacteria in very crude supernatants. What I have not tried yet is to combine some kind of affinity with enrichment of some secreted peptides, and then use the hypothesis-driven MALDI ion trap approach to see at what levels they can be detected. So I would like to create an instrument for very high sensitivity detection of bacterial peptides. I think this could be very useful to create a diagnostic tool in the future.

What would this diagnostic tool be used for?

We showed in a PNAS paper about auto-inducing peptides that you can characterize bacterial strains just because you find these peptides. These peptides are not highly expressed, so if you take other bacteria and find out what are highly expressed markers, the diagnostic tool might be useful to detect an infection correctly and diagnose it correctly right away. If you go to a doctor and they take a blood sample, they can tell you that you have an infection. But they cannot tell you what infection you have and if the infection still persists. And this is why I think that my appointment here at the immunology group would be useful, because I have lots of expertise here and I can just go to my colleagues and ask them.