At A Glance
Name: Min Dong
Position: Postdoc, Department of Physiology, University of Wisconsin
Background: PhD, Department of Physiology, University of Wisconsin; BS, Science and Technology University of China
Tell me a little about how you became interested in conducting assays for botulism toxin?
We started working on this toxin about three years ago. The first project we did was looking for a receptor for those toxins, so that’s actually the first paper we published last year, on the receptor for one of those toxins. So as we worked on those toxins, we found out that there’s really not good methods right now to detect those toxins or test for toxin activity. There are methods available, of course, but the most effective way right now is to actually just inject toxin samples into mice, and then you just watch as the mice die. The other way you can do it is to just run a gel. For example, because those toxins are just proteases that cleave the proteins, then you run a gel and see the cleaved fragment of the protein, and you know the toxin worked. Or you can run HPLC for the cleaved peptides. But those methods are tedious work, and takes a half-day or almost a day to get the results. So we want to find a way to assay for toxin activity that’s really fast and really simple and really cheap. Initially, the idea was just to provide a method that you can do in the lab to test for toxin activity. But we think the method can be applied for large-scale screening of toxin inhibitors — looking for some kind of small molecule. That’s how we started this project about a year ago. We’ve spent about one-and-a-half years on this project.
I see that this is a FRET-based assay, and based on GFP variants. So is this assay for live cells?
We actually have two assays. One assay is just in vitro, in a test tube. In that one, you purify some recombinant proteins — a chimera protein, where you put CFP on one end, YFP on the other, and in the middle, a protein fragment that can be cleaved by those toxins. Together, it’s a chimera protein that you can just purify using E. coli. Then you just dump it in a test tube, and for example, if you want to test a sample that contains toxin, you just mix it with this reporter, and let it incubate for a certain period of them. Then you measure the FRET signal using a variety of fluorescence detectors, like a spectrometer — a really sensitive spectrometer — or a very simple plate reader. We actually use a plate reader to do the screenings. It’s very simple — using a 96- or 384-well plate, you put your samples in and mix it with reporter protein, and you just put it in the plate reader.
In that case, you’re just getting an aggregate signal — it’s not single-molecule fluorescence?
No, we use a certain amount of the protein concentration. For example, right now, we use a 390-molar protein. How much protein you need depends on how sensitive your detector is. For the really sensitive [readings], there are some high-end spectrometers available that can detect single-molecule fluorescence, but we don’t have that. So this is the first real method we developed that you can use to detect the toxins in vitro. And then we want to develop some kind of method that can develop toxin in a cell-based assay. The biggest advantage of a cell-based assay is that we can screen not only the inhibitors that inhibit the toxin protease activity, but has the potential to reveal the inhibitors that can inhibit other toxin action steps — for example, small molecules that can inhibit toxin binding to cells, or inhibit the toxin internalization, or translocation out of the vesicles.
So you might stumble upon other mechanisms of action of a small molecule?
Exactly. That’s why we want to do this. We tried to do this using one of the cell models we have, which is PC12 cells. Then we just express this chimera protein in cells. Because it’s a CFP-YFP, this protein can just simply express itself. But that doesn’t work — or it doesn’t work well. So we had to solve this problem, and what we found out is that we can not just use a truncated version of the protein when linking CFP and YFP when expressing in cells, because that is essentially a soluble protein, and soluble proteins inside cells don’t work — it works, but not well. We found out we have to actually use the full-length protein, which is very big. And for that, we didn’t expect to see FRET inside cells, because it is such a long protein. However, when we expressed it, we did observe the FRET, so, that actually gives us an advantage to have an ideal reporter for toxin activity in cells, because it’s exactly the same as the endogenous target for those toxins. The reason for why the full-length protein works and why the soluble protein doesn’t is actually because the full-length protein is anchored to the plasma membrane of cells — it’s not floating around inside cells. We found out this anchor to the plasma membrane is actually required for the toxin cleavage to be efficient. So that’s one of the reporters. And then another reporter for cell-based assays — there are actually seven toxins — and those seven toxins attack three proteins in cells. We have to develop two reporters that can be cleaved by all seven of those toxins. The first one, as I said, that is anchored to the membrane can be cleaved by toxin A, toxin E, and toxin C, and the rest act on another reporter. That reporter actually is the same thing — we can not just use a soluble fragment. We have to anchor it on certain subcellular organelles: secretory vesicles. So we anchored it to these vesicles, and because of that, we could not use FRET so we developed a new method to detect the toxin action. Because we label the molecule using two colors, the CFP is inside the vesicles the entire time, and the YFP is anchored to the vesicles before we add toxin. But after we add toxin, the toxin cleaves the linker, and the YFP becomes released from the vesicle, and becomes a soluble fragment all over the cells. So instead of watching the FRET signal, we can watch the redistribution of YFP signal.
The convenient part of this is that before those proteins are cleaved, they’re actually in the cytosol, on the vesicles. There is virtually no fluorescence signal in the nuclear area. And after the toxin action, when the YFP becomes a soluble fragment, they actually spread out everywhere, and enter the nucleus, so the nucleus is a convenient area to watch this change from no fluorescent signal to a relatively high level of YFP signal.
Is this a commercially available assay?
I think there systems available — just an automated microscope where you put cells into your 96-well plate and then take pictures of every well, and there is software that automatically recognizes the nuclear area. Then you can just measure the fluorescence signal change. But we don’t have such equipment right now.
What do you use?
For the cell-based FRET, we just use a regular epifluorescence microscope. But to watch the FRET, you need a special filter set. So companies like Chroma design the filter sets to do this kind of stuff. Basically you just take three pictures: one for CFP specifically, one for YFP, and one in which you use excitation filters for CFP and an emission filter for YFP, and you calculate how much FRET signal you get from those three pictures. For the redistribution, you just take a picture. That is just proof-of-principle. We don’t actually have the equipment to develop the assay, so right now we just take a picture showing addition of toxin, and then afterwards take another picture that shows the YFP toxin becoming a smear throughout the cell.
Have you evaluated any higher-end equipment or thought about it to increase throughput?
We would like to, but we don’t have any resources to do that. Hopefully after the paper comes out, we can get some help from interested companies.
This seems like something that might draw interest from a biodefense standpoint, considering that it’s botulism toxin. Have you had any interest in this area?
Sure, there is a definite interest in this method. One of the very interesting projects we have right now is that we’ve tried to develop an assay that can be used to screen milk samples. We tested it, and it works to the extent that we can detect a 90-molar level of toxin in milk.
Are you partnering with someone on that?
We worked on it because of the concern that we got from the California health department, because they were worried about someone tampering with milk, so they tried to find a way to screen the samples before people put them onto market.
Does this have applicability to other types of infectious disease that work by way of toxin, such as anthrax?
I’m not an expert on toxins, but the principle should be the same, as long as the toxin functions as a protease, and many toxins actually are. If it can cleave a certain protein fragment, then you can definitely use this approach. I’m not sure of how many toxins would work with this approach?
Is this FRET-based linker protein assay something your group designed, or is the assay new because of its application to botulism?
It’s not a new method. People used it several years ago. But the initial application was actually used to detect native protease activities. So it’s actually a very mature method. But to use it to detect toxin action, this is a fairly new assay. But in principle, it is not.