At A Glance
Name: Tom Kerppola
Position: Associate Investigator, Howard Hughes Medical Institute; Associate Professor of Biological Chemistry, University of Michigan
Background: Assistant Professor, University of Michigan — 1994-2000; Research Fellow, Roche Institute of Molecular Biology — 1990-1994; PhD, biochemistry, University of California, Berkeley — 1985-1989
Tom Kerppola investigates protein interactions in mammalian cells and nucleoprotein complex architecture in an effort to elucidate mechanisms that contribute to the specificity of cellular regulatory networks. To this end, his laboratory at the University of Michigan uses a fluorescent assay technique that Kerppola helped develop called bimolecular fluorescence complementation analysis. The assay is also garnering interest from commercial drug discovery entities, many of whom Kerppola will address in a presentation at November’s Fluorescent Proteins in Drug Development conference in La Jolla, Calif. Kerppola took some time last week to discuss with Inside Bioassays the assay and its commercial potential.
How did you become interested in your current field of study?
Biologists in general have always been interested in protein interactions, and I think that protein interactions are sort of a standard currency in biological specificity. People use many different approaches to study protein interactions. GFP pulldowns, immunoprecipitation, and two-hybrid assays have been methods that have been used very widely and provided a lot of insight into how proteins form complexes and thereby acquire functions that are specific for a specific cell type or signaling pathway. We have certainly been interested in this for many years, and have studied these processes in vitro, in biochemical reactions. But most recently we felt the necessity to study these processes in the normal cell environment. So that’s where this got started.
What are your feelings on the importance of studying these interactions in a cellular environment?
I think that in order to understand specificity, we have to ask the questions in a relevant context. I think a simplistic view would be that if you have high enough concentrations of any two proteins, they will interact, so the question becomes: What are the relevant concentrations of the proteins, and what is the environment? What are the other interaction partners in that environment? Therefore, I think you can learn things about what can happen in vitro — I think that can be very powerful — but I think that only by asking the question in a cell, or preferentially, in an organism, do we get answers to what actually does happen in real life.
Tell me a little bit about the bimolecular fluorescence complementation analysis assay.
The concept here is very simple. This is something that was developed by Chang-Deng Hu in my laboratory. If you take a fluorescent protein and truncate it — basically take fragments of it — those fragments on their own will not be fluorescent. However, when brought together by an interaction between proteins that are fused to those fragments, it occurred to us that maybe association between those partners could bring together the fragments of the fluorescent protein, and thereby reconstitute a fluorescent complex. This would allow us to visualize, in the cell, the formation or interaction between those proteins through the formation of this fluorescent complex.
How did you become involved in the Fluorescent Proteins in Drug Discovery conference, and what are your thoughts on the use of this assay in drug discovery?
As far as the conference, I was simply invited to present. So in that sense, I guess there is interest on the part of the drug development community. In addition, diagnostics and other fields could have use of these strategies. As far as our interest in the lab, we are using it in basic research questions — trying to understand fundamental biology. But certainly other people are using these approaches for more product-directed purposes, and they are certainly reasonable to be applied in that context. At the same time I would say that understanding of the details of the process, in terms of how, in fact, does this assay work, and what are the critical constraints on the assay, those are issues that are critical to take into account when thinking about applying it in a more complex drug-development type of environment. We continue to work on improving the assay and adapting it for specific purposes, and it’s that improvement and adaptation that would be interesting to people in the pharmaceutical and biotechnology communities.
Can you talk about some of the improvements that need to be made to the assay, and how it might be adapted to a high-throughput approach?
I think that like every assay, this one has limitations. Among those is the fact the formation of these fluorescent complexes is — in contexts where it has been carefully studied — irreversible, which means that once formed, that complex no longer dissociates. So that can be a real limitation in studying a dynamic situation like biology often is in that you don’t get to study the dissociation of the complex. Other limitations include the fact that the formation of the actual fluorescence, or what’s called the maturation of this fluorescent protein, is a very slow process, and therefore, what you see is actually a complex that has formed up to an hour before they become visible, and that limits the time resolution of the assay. So those are issues that certainly need to be addressed. As far as the high-throughput application, I think whenever you go to larger and higher throughout situations, the level of false positives crops up, and the stringency of the assay in terms of what is an acceptable rate becomes an issue. As far as the background signal in those kind of assays, microscopy isn’t a very high-throughput type of analysis method. Fluorescence can certainly be adapted to many different modalities, like basic instruments that can screen large number of samples. But those types of applications tend to be less robust and more subject to fluctuations in signal and difficulty of actually quantifying what the level of signal is. So those are issues that will definitely need work. The potential is there. I think that fluorescence assays have a virtue in that they can be scaled up to large numbers of samples and a large scale of throughput, but work will need to be done to accomplish that.
Now that I’ve made you address the limitations of the technique, what are the major advantages of this assay?
I think the major strength of this assay as compared to many other methods, including other methods that allow visualization of interactions in living cells, is that the assay is exquisitely sensitive, so it can be used to study interactions between proteins with only modest over-expression, or ideally, at levels of normal expression. It can be adapted to a variety of different cell types and environments, and in that sense, you can study a problem in its normal context — you’re not limited to a particular cell line or a particular engineered context. You can actually ask the question in the environment in which you wish to study it, which is an important ability because much of what we hope to study is the normal physiology of an organism, rather than something we recreate in a model system.
Have you had interest from commercial entities, and is this something you would like to see through to commercialization?
My lab is really focused on the basic research side of things, but there are companies that are very interested in using this type of technology for a variety of purposes, and I think that it’s up to them to develop the capacity and the extensions of the assay that will make it commercially applicable. There is a large number of applications, and I’m sure that different entities will come up with different ways of using it.
How does this technique differ from the fluorescence-based assay developed by Stephen Michnick at the Universite de Montreal? This assay is the basis of the drug discovery program at Odyssey Thera.
There’s a whole family of techniques, and I would say they’re all interrelated. All of these are really descendants of original work by Jacob and Monod on the beta-galactosidase system, and then Hartley on the barnase or ribonuclease system. They were the first to demonstrate this complementation phenomenon where fragments of proteins can reassociate to form a functional complex. This work has been extended by many people. The first was called conditional complementation, where the association between the fragments required that those fragments be fused to interaction partners was done by Johnson and Varshavsky at [the California Institute of Technology]. They used that to study protein interactions in yeast. And as you mentioned, Michnick’s group showed complementation by dihydrofolate reductase in both bacterial and mammalian cells. And Helen Blau at Stanford demonstrated complementation by fragments of beta-galactosidase, again, by mutant variants in mammalian cells that could be used to visualize the sub-cellular locations of interactions. And Lynne Regan and her group demonstrated that fragments of fluorescent proteins could be used to detect interactions in E. coli. What our contribution initially was, was to identify fragments of fluorescent protein that could be used to visualize protein interactions in mammalian cells, and subsequently the demonstration that differently engineered variants of those proteins could be used in a multi-color approach to visualize several different interactions in the same cell simultaneously. And certainly Michnick’s group and Odyssey have used the complementation concept to study many interactions of proteins in cells, so these are really tools, and we are all borrowing ideas from each other, and improving our toolkit by basically taking pieces of each other’s work and adapting it for our purposes. I think it’s exciting when this information is shared, and the field can make greater progress when many minds contribute to the solution.
Have you sought patents on the method?
A provisional application was filed.
Are there any licensing issues with using the GFP variants?
These are issues best dealt [with] by lawyers. The whole intellectual property process is unfortunately rather inefficient and often gets in the way of scientific progress. It may in fact be partly to blame for the rather slow implementation of new technology by many biotechnology and pharmaceutical companies.
What is your lab interested in using this for going forward?
I think there are two directions that are exciting to us. One is the adaptation of these methods to still more physiological and more relevant environments, such as using this in intact animals, plants — basically normal living organisms. And I think the possibility of actually studying developmental regulation of tissue-specific interactions and related questions is very exciting. And the other thing is the adaptation of these methods to screen — looking for new interaction partners, modifiers for interactions, inhibitors of interactions, and signal-dependent interactions. Those are the directions that I think are most exciting, but again, there are many things that can be done using this tool, so we’re pleased to see other people are interested in using them as well. We are making these reagents available to non-commercial parties — basic researchers — and there is now a fairly large body of literature of people who have used the assay for different purposes, and we’re happy to see that it seems to be a successful tool.