At A Glance
Name: Michel Bouvier
Position: Professor and Chairman, Department of Biochemistry, Université de Montréal
Background: Associate professor/assistant professor, Department of Biochemistry, Université de Montréal — 1989-1997; Postdoc, Duke University — 1985-1999; PhD, neurological sciences, Université de Montréal — 1979-1984
SAN DIEGO — Michel Bouvier and colleagues at the Université de Montréal are at the forefront of academic drug discovery research, having recently published several notable papers on the use of bioluminescence resonance energy transfer to explore cellular signaling pathways. These include a December paper published in Nature Methods in which they used BRET to monitor real-time ubiquitination in live cells (see Inside Bioassays, 11/30/2004), and several papers this year describing the use of BRET to monitor GPCR receptor activity and GPCR dimerization. At ICB’s ScreenTech conference held here last week, Bouvier gave a keynote presentation on the latter topic, and sat down with Inside Bioassays for a more in-depth discussion.
Why is dimerization such an important concept in GPCR drug discovery?
In fact, we did not aim to study dimerization, we sort of came across that. By serendipity, we became convinced that dimerization of GPCRs did exist, so it was something that we could not avoid studying. There are several roles for dimerization. One very important role in terms of cell biology is the folding of the receptors themselves — dimerization is often used for proteins as a quality control system, allowing only the properly folded pieces of a protein to get to the cell surface; and, it is part of this process for GPCRs. It is required for the receptor to be completely assembled and be trafficked to the plasma membrane. Once it is at the plasma membrane, the homodimerization — it’s not clear yet what its role is in terms of signaling. Most likely, from the data available, we think that you need a dimer of a receptor to activate one heterotrimeric G-protein — although the proof of that is not completely there yet. Probably the most relevant portion of the dimerization story for drug discovery is heterodimerization, where clearly, some receptors not only homodimerize, but can also associate with other subtypes — sometimes closely related, sometimes not so closely related — to form structures that are composed of two different receptor subunits. In that case, what may happen is that you have an expanded diversity both at the binding selectivity level and the signaling level. And so far, the larger number of examples comes from the signaling level, where you have heterodimerization between two receptors, which appears to be directing the response of the receptor to a different signaling pathway. That becomes very important when you’re looking for drugs or compounds that can selectively activate these signaling pathways. And now there are a couple of examples coming out in literature where people have either designed or come across ligands or compounds that can selectively activate or block the heterodimers versus receptors expressed alone. That has two consequences. First of all, it means that you can target heterodimers pharmacologically, and secondly, now we’ll start to have the tools to really dissect what the roles of these heterodimers are in vivo. And I would say that the in vivo question is the hottest question right now — finding out what is really the role of these heterodimers in living animals and humans. Right now, most of the data comes from either an overexpression system; or in some cases, normally expressing cells; or in the best case scenario, a cell system. There is very little data in living animals. I think this is where the field needs to go now, demonstrating that heterodimerization may have very important functional consequences in living animals.
Although, as you said, there have been several papers on this idea, do you think the concept is entering commercial drug discovery?
I think the concept is slowly entering the drug-discovery enterprise. Industry is relatively conservative in general, which I think is normally a good thing, because you don’t want to rely on trends and uncertain data. But I think that in the last year or two, they have slowly become more interested in at least looking at the possibilities. And when you talk with people in the drug industry, many of them are starting to look at it. They realize that it is going to be difficult — it’s already difficult to get a compound against one receptor, but now you’re thinking that you might have to get one against two receptors, so this complicates the issue. And this was part of the resistance. I have a number of friends in the industry that are in power positions, and when they see me they say ‘Michel, you are complicating our life.’ And I say ‘I’m not complicating your life; life is complicated, I’m just trying to make a description of it!’ But I think that slowly it is happening. At this point it is still a largely academic enterprise. I predict that in the next few years, we’ll start seeing people getting actively involved in trying to identify bifunctional ligands that could selectively bind to some heterodimers, which may have therapeutic indication in the opioid receptor field, or the cannaboid receptor field, or other very actively investigated targets.
The assay that your lab developed to investigate this deals with the beta-arrestin pathway?
There are a number of assay, but one of them is indeed looking at beta-arrestin interaction with the receptor using the BRET technology. In contrast to some of the other technologies, this is not imaging. We’re not imaging the translocation of the beta-arrestin receptor; rather, we’re looking directly at the interactions between the receptor and the beta-arrrestin, looking in a spectrometer at the transfer of energy between the receptor and the beta-arrestin. So we can follow this in real time, in living cells, quantitatively, and without having a very complicated algorithm. It’s simply measuring the luminescence and the fluorescence and having a ratiometric measurement that allows us to determine the energy transfer between the two molecules, and thus allowing us to monitor [their] interactions. So this was based on the transfer of energy between a luciferase, which is on the beta-arrestin, and a GFP, which is on the receptor. And we have established a number of stable cell lines that express a given receptor with beta-arrestin. Then each of these cell lines becomes a reagent to screen for this receptor target. But recently, we wanted to get away from the need to modify the receptor. Many people in drug industry feel that they’d rather screen against a native receptor, or one that has not been modified; or, even better, receptors that are expressed endogenously in certain cell types. So we’ve constructed a beta-arrestin that we call ‘double brilliance’, in which we have tagged on both ends — on the C-terminus and the N-terminus — with luciferase and GFP. So now you have an intramolecular BRET signal that is constitutive, because you have fused these two things together. And once the receptor is activated, and beta-arrestin comes to the receptor and is engaged by the receptor, it changes conformation, which will lead to a change in distance and orientation between the energy donor and acceptor. That is reflected in the change in the BRET signal. So now you can monitor the engagement of beta-arrestin by a receptor by following the change in intramolecular BRET in cells that express receptors that are not modified.
Is there any compelling reason why you would use bioluminescence, and not fluorescence resonance energy transfer [FRET]?
As anything, there are advantages and disadvantages to both techniques. The reason why FRET is more used right now is because it is older — it has been around for a long time. Of course, people have much more experience with it. But let’s start with the advantaged of FRET, because it’s well-known. You can do imaging with FRET, so you can see where the interaction is taking place within the cell. At this point, with BRET, this is a very difficult endeavor. Why? Because in contrast to FRET, where you are exciting your energy donor with light and putting a lot of energy in the system; in BRET, you don’t do that. All the energy comes from the luminescence enzymatic reaction of luciferase, so the level of energy is much lower. So yes, you can detect it under a microscope with a good camera, but at this point, we don’t have spatial resolution. So I would be able to distinguish between a nuclear interaction and a membrane interaction, but I would have a hard time saying whether it is in the endoplasmic reticulum or the Golgi. So this is one disadvantage of BRET. Now, as for the advantages there are a couple. Because you are not shining light in the system, there is no chance that you will non-selectively excite the acceptor, which is one of the problems with FRET and requires very good controls. Unfortunately, in many of the excitation FRET papers, these controls are not done, and that is a real concern and issue. Here we don’t have to care about it. The second thing is that empirically, we had better success of good energy transfer without having to play too much with the length of the spacer we put between the donor and acceptor, compared to FRET, where there is more playing around. We think that this has to do with the half-life of the excited species. When you’re doing these measurements, because you need to have a close distance between the donor and acceptor, and also parallel orientation between the two species, the relative position of the donor and acceptor is very important. Since proteins are flexible and moving, you’re really sampling different intermediaries from the overall conformational possibilities. And because the luminescent species has a longer half life, you’re sampling more intermediates, and you have a better chance to detect the energy transfer than with FRET.
Right now you just need a simple plate reader to conduct these assays, since imaging is not necessary?
There are a couple of characteristics that you need to perform these assays well. You need to have a sensitive plate reader, because you need to detect a relatively low level of light output. There is such equipment on the market, so this is not a real problem; but, some of them that are out there do not have the required sensitivity. The other thing is that you need to be able to read, almost simultaneously, two different wavelengths — the light that comes from the luciferase and from the GFP. So either you have equipment that is filter-based where the filter can change very, very rapidly from one to the other; or, you can have equipment that can read both at the same time because it has a beam splitter, or other technology.
Are people working on imaging capabilities for the BRET assays?
We are working on that. I know that there is another group — Dave Piston’s group at Vanderbilt. There are a couple of issues, but the key issues are trying to get luciferase that will give higher light output; having GFPs that are brighter and have a higher emission coefficient; and also having more sensitive cameras.
Can this be adapted to a high-throughput format for industrial-scale drug screening?
Yes, in a plate reader platform, not doubt about that. We have a new university group on therapeutic drugs at Montreal, and we are equipped with robotized and completely integrated screening platforms. One that we did recently was a screen against the CCR5 receptor, looking for an antagonist. But we are a university, so we did not screen one million compounds — but we did screen 26,000 compounds, which for a university is pretty good. And we came up with 19 hits, three of which we are pursuing right now because they were in the sub-micromolar range, and the chemists tell us we can do a little chemistry on these. So within the university group, we are trying to improve these hits, maybe make them leads, and see if people are interested. And we’re also going to screen for other receptors, with a major focus on orphan disease, which we think is a mission for universities. We’re not there to compete with the drug industry — we’re there to complement them, and in some cases, to help them.
There has been a noticeable trend in the US of academic labs taking more of a drug-screening angle. Do you see this same trend in Canada?
The same thing is happening. There are a number of initiatives, like the one at the University of Montreal, and there’s one very exciting one, which is called the Canadian Chemical Biology Network, where we’re trying to get the biologists, biochemists, pharmacologists, pharmacists together so we can have a centralized bank of compounds coming from the labs of a number of chemists. They could put their resources together, make sure the quality control is there, make sure we know what is in each of the wells, that the bank is well-curated and we can have a high level of confidence in these, and then make it available to a number of screening facilities across the nation. People can then apply it to their specific targets. I think one very important thing is having diversity in the type of screening that one applies. I’m absolutely convinced that the choice of a screening platform is not a trivial choice — it clearly determines the types of molecules that you will get. A screening pathway determines the pathway that you’re going to be looking for. And too often, in my opinion, people are trying to activate or block a receptor, and so the target is the receptor, without taking too much time to think about which pathways this modulates. We all know now that each of these receptors can engage a number of different pathways, and that the receptor conformations that will activate one specific pathway are different, and can be stabilized differently by different compounds. So if you don’t know ahead of time which of these pathways you really want to modulate, then the chances are better if you’re screening against different screening pathways, so you will have a compound that will correspond to your therapeutic applications. The other avenue is to really first do the biology, find out which pathway you want to modulate, and then take a screening assay that corresponds to this pathway directly. So diversity in screening is going to be a major issue in the coming few years. Academia is well-positioned to do that by focusing on smaller, more diverse, chemical libraries, rather than focusing on huge libraries without a lot of diversity.