David Lubman on 2D Liquid Separations and Building Your Own MS

At A Glance

Name: David Lubman

Age: 49

Position: Professor of chemistry, University of Michigan, since 1983.

Background: Weizmann Institute of Science fellow, Rehovot, Israel, 1981-82.

Worked at Quanta-Ray/Spectraphysics, 1980-1983.

PhD in physical chemistry, Stanford, 1979.

Bachelor’s in chemistry, Cornell, 1975.

How did you get involved with proteomics?

We were basically instrument developers, and it looked like there were some very interesting applications of mass spec toward proteins and DNA at the time. This goes back to about 1991-92, and some of the problems that they wanted us to work on were a little too hard for the technology available. They were interested in phosphorylation site searching, looking for low-level proteins off of gels, looking at protein-protein and protein-DNA complexes and trying to sequence them and match them up with DNA sequences. So we were trying to work on these things, but the technology really wasn’t quite there. MALDI was in existence but hadn’t really been commercialized at that point — it was very crude. Electrospray was around but it wasn’t very popular yet. It wasn’t until the late 1990s that these machines started to find their way into labs.

We used to build our own [machines]. Actually we built a high-voltage MALDI instrument for detecting large DNAmers back in the 1990s.

Do you still build your own instruments?

We do, because we can service them ourselves and we can make them do things that commercial instruments can’t. But we use commercial instruments, too. We use a Micromass MALDI-TOF, MALDI Q-TOF and ESI-TOF. Then we have our own homebuilt machines — an ion trap TOF for doing MS/MS on online separations.

What can home-built machines do that commercial ones can’t?

Actually, the stuff that is being home-built now is being commercialized by various companies. We have an ion trap TOF that can rapidly sample, do MS/MS, then pulse the ions onto time-of-flight. So the major advantage there is speed of online separation. We do very rapid capillary electrophoresis separations of protein digests. What’s really nice about that is that CE provides very high resolutions of digests in a very short amount of time, so we can separate out 30, 60 peaks from a digest in 30 seconds or less with very high resolution, and then we can do online monitoring with MS/MS, which would be very difficult to do in the commercial machines.

Not many scientists use capillary electrophoresis — why is that?

Most people still want to do their separations by 2D gels and then you have a lot of garbage and it’s hard to purify sufficiently for CE separation. We use [CE separation] because we developed a new method, with Eprogen of Darien, Ill., for separation of proteins — a 2D liquid separation of proteins that is now being marketed by Beckman-Coulter. We use a column — a technique called electrofocusing — to separate proteins according to their pI in one dimension, then we use nonporous RP-HPLC to separate proteins in a second dimension, so we can get a 2D map similar to a 2D gel but all in a liquid phase. Since it’s not in a gel, it’s very easy to interface that to CE so you can take advantage of the very fast separations and the high resolution of the capillary electrophoresis.

Will liquid separations eventually replace 2D gels?

That’s the idea. The proteins [in liquid separations] can be relatively clean and purified. So all the background garbage doesn’t interfere with the CE. We don’t even have to clean them up, so we can easily digest them and do CE on them. The thing that’s nice about the 2D liquid separations is that you can collect the fractions or send them right into mass spec, so it’s very easy to do mass spec on the products, whereas with a gel it’s not a natural fit. The 2D liquid separations are a natural interface, so you can get very accurate molecular weights. Then we can collect it and do MALDI MS/MS and get the identification. So we know the identification and we know the molecular weight. And if the molecular weight is off from the database molecular weight, we know that [the protein] has been modified. But then we go into online CE followed by ESI MS/MS in our home-built instrument. And, from that, we can get very high coverage. We can actually get most of the sequence of the protein and we can look at the post-translational modifications, and with the molecular weight, we can match up to see if we have the whole protein. So once we do detailed sequencing with CE ESI MS/MS, it allows us to tell if we’ve covered the whole protein and where the modifications are. There are actually three steps. We can get the molecular weight, most of the sequence, and we can get the detailed post-translational modifications. Even with large proteins, up to 70 to 80 kDa, we can generally get 90 percent-plus coverage of the protein.

Tell me about the projects you are working on now.

Using 2D liquid separation, you get pretty high reproducibility. So it’s good for inter-lysate comparisons — comparing a sample at one time versus later on. So we’re looking at problems where there are changes induced by cancer cells, and we’re trying to understand some of the processes that may occur. For example, we’re interested in breast cancer — looking at how a specially-developed panel of cancer cells develops from normal to intermediate stages into a fully malignant metastatic cell. We’re looking at four to five stages, and using this liquid separation technique, we can do highly reproducible comparisons between the protein expression profiles. We can look at proteins that are different markers at different stages of cancer, which is important for both early diagnosis and prognosis of the cancer. That is, can we find proteins in the malignant cells that are highly upregulated compared to normal cells. We’re looking at ovarian cancers — there are four different subtypes. We’re trying to find for each different subtype whether there are markers particular to that subtype. In that case, we’re looking for patterns but also specific proteins. We’re also looking for proteins that have very specific changes, for example phosphorylation patterns that might be directly related to the progression of cancer and could be related to the prognosis. So we’re looking for detailed changes to proteins that might be early markers or a predictive of the cancer itself.

We’re using 2D liquid separations followed by molecular weight mapping to do that. We can watch how the molecular weight changes as a function of cancer progression, and we can do detailed sequencing using MALDI and CE ESI-TOF mass spec. We’ve published some initial work on ovarian cancer, in which we talked about some of the possible markers. That came out in Analytical Chem[istry] recently. But now, we’re writing a much bigger paper on detailed analyses of markers for two types of ovarian cancer.

We are also doing protein chips. One of the big advantages of the liquid fractionation method is that from the typical prostate cancer cell we can see 2,500 proteins easily. We can collect them from the liquid phase and spot them on a chip. So then you take human serum, and spread them over these chips and you can see which proteins react to the antibodies in the serum. Certain kinds of cancers leak proteins into the blood. There’s an antibody response and that’s what we’re looking at on these protein chips. We’ve looked at about 50 different sera samples with our collaborators at the medical school. And we’ve been able to detect possible markers of prostate cancer.

So the two things [we are doing] are mapping the protein content of the cell, and looking for patterns of specific markers. The chips are a spinoff from that. And both are complementary because the chips tell you what proteins are circling the blood and what might be inducing antibodies. But they don’t tell you about all the proteins in the cells that might be important.

What other collaborations do you have?

We also have a big breast cancer project that we’re about to publish a paper on in Proteomics. That’s where we’re looking at a panel of breast cancer cells that develop from normal to metastatic, with the Karmanos Cancer Center Institute in Detroit. Then [we have a partnership with] a company called Cyagen that’s actually working on the CE mass spec.

What major obstacles remain in techniques and tools that you use?

I think that the instrument companies have to do more work. They’re going to have to make advances, especially if we want to look at detailed changes in proteins as a function of cancer. The instrument companies have to develop instruments that are less expensive and can be more widely disseminated. It obviously isn’t [a problem] for these pharmaceutical companies, or for a big facility to get an instrument, but if more researchers are to use mass specs themselves, then they need instruments that are more user-friendly and less expensive.

The other scientific problems that remain are dealing with things like fractionation and getting to membrane proteins. We’re doing work on that, trying to develop techniques for pre-fractionation and solubilization.

The other place that everybody talks about are low-level proteins. I’ve heard that there are enzymes expressed at 10 copy numbers per cell that catalyze a lot of reactions. But I don’t think that’s been proven yet. People are speculating and we need to get there. But it may be that we get there and find out they are not important.

Unfortunately, there seems to be less of a push these days among funding agencies to fund instrument development, mainly for academics. I think academics really need to lead the way on development of new instrumentation technology, and there isn’t a lot of funding out there for that.