The technique, called photoactivated localiztion microscopy, or PALM, also used work done by Lippincott-Schwartz in developing photoactivated fluorescent proteins. In the PALM project the researchers were able to control ultraviolet light to activate only a few molecules at a time. Even though each molecule appears as a 200-nanometer-diameter spot, the center of the spot could be determined to within 2 to 25 nanometers.
By activating only a few fluorescent proteins at a time and repeating the process thousands of times, a computer image was created in which all the molecules could be determined, often with near-molecular accuracy.
ProteoMonitor caught up with Lippincott-Schwartz this week to discuss the implications of the research on the proteomic field.
Can you summarize the study, what you were looking for and what you actually achieved?
The goal of the study was to create a microscope system that was capable of nanoscale localization of molecules at high density. Right now, conventional light microscopy is limited by the diffraction limit of light which is around 200 nanometers, 300 nanometers, so you can’t resolve two objects that are closer than 200 to 300 nanometers in dimension. That’s a real problem if you’re trying to understand how molecules which are only 2 nanometers in size are localized with respect to each other. So this approach was aimed at trying to develop a technique that could overcome this barrier.
How and why did you become involved in this project?
My lab, we developed the photoactivatable GFP. That reagent is key for this type of methodology. It worked because you can turn on fluorescence and you can [turn] it on at a very low level so that only a few molecules at a time become fluorescent — a few molecules in a population of 10,000 molecules.
Each of these few molecules that you turn on is separated by more than 200 nanometers in diameter, so you have the capacity to pinpoint them with incredible resolving capacity. So you build a map of the distribution of all of these molecules with time using the photoactivatable proteins. And you can’t do that with any other fluorescent probe. You need photoactivatable proteins to do it. So my lab was really the first lab to develop photactivatable proteins. Since the development of photactivatable GFP, there’ve been many other photoactivatable fluorescent proteins that have come into the scene. My lab has been working with many of these.
So it was natural for Eric to come to me and say, ‘Would you like to help us?’ Because he essentially had no access to any of these photoactivatable proteins and [didn’t] know how to work with them [or] create any reagents with them. Plus, we were a natural site to build the microscope because NIH was able to provide funding on a moment’s notice. Within a week or two we got into motion all of the equipment.
How much did NIH fund for this project?
I think it’s on the order of about $70,000.
Can you talk about the proteomic applications of this technique?
People who are interested in knowing where their proteins are localized at the nanoscale level would be very interested. The next step in this technique is to perform double labeling. That’s going to be really, really important because people in the proteomics field, they’re interested in knowing how proteins interact with each other. But they’re interested in how proteins are localized in the subcellular level in order to relate the protein to particular functions.
There are two directions [in which] this technique is going. One is to be able to look at more than one type of molecule at this type of high resolution at the same time. The other approach is to relate the distribution of the molecules to an electron microscope image because then you can look at how the proteins are distributed at an ultra-structural level. You’re not just looking at dots. You’re looking at dots on a background of an electron microscope image. And that’s what we were able to accomplish in the paper, barely. We’re now trying to see if we can do this more routinely.
What’s preventing you from doing either or both of these things now?
The roadblock [to double labeling] is that the current photoactivatable proteins are not different enough in their absorption spectrums to have them be clearly separated when we’re doing this type of imaging. But there are other types of photoactivatable reagents which we know will be separatable. And these are caged fluorescent molecules. So you hit [them] with light and they get uncaged and the fluorescence appears. And they are not fluorescent proteins. They’re like fluorescent dyes. We’re collaborating with various chemists to try to generate these reagents.
What about the electron microscope, what’s preventing you from being able to do that more consistently?
Again, it’s really just a learning curve. We need to learn how to fix the specimens better, how to prevent the photoactivatable proteins from becoming inactive during the fixation process. And then also develop better ways to localize the specimens when we shift from the light to the electron microscope.
Once you’ve accomplished that, what will proteomic researchers be able to do with this technique?
They will be able to determine whether two proteins that have some type of interaction, how they are localized at a nanoscale level. Let’s say you have a structure, and you have two different types of proteins that working in that structure. You want to know how these proteins are arranged with respect to each other. It may be, for instance, that they’re arranged concentrically around themselves or there may be one type of molecule that’s only localized to only one part of an ultra-structural element.
People are very keen to understand how these molecules are arranged relative to each other. This technique will allow that to occur. This technique can also allow one to make observations whether proteins on the plasma membrane, for instance, …are segregated from each other into micro-domains which previous techniques really haven’t been able to resolve. Yet there are very many biochemical assays that suggest that there should be some type of micro-domain organization. I think this technique’s going to be able to show that.
Does this technique have the potential to be a kind of ‘Aha!’ moment in proteomics?
I think so. Right now, we essentially know the gene products. We know roughly how these gene products are localized. But we don’t have any understanding of how these molecules are distributed at high resolution. And the techniques for doing that are very crude at this point.
[The success of this technique in proteomics] is going depend on how people use it, how easy it is for people to make use of it.
How easy is it to use right now? Is it a technique that researchers can duplicate?
I think that’s one of the things that Eric and Harald are interested in trying to do over the next year, trying to generate a user-friendly microscope of this type. Right now, the system that is in use at NIH is pretty cumbersome.
Why is it cumbersome? Can you describe the system now at NIH?
It’s mainly the objective. Normally, when you’re looking through a microscope, you have two eyepieces that allow you to look at a field that is several hundred microns across. And the field that we’re looking at in this PALM [photoactivated localization microscopy] microscope is very small. It’s maybe two microns across. When you are scanning, it’s very hard to find the area of interest that you want to go in and image because you have a field of cells that are hundreds of microns in size and you’re trying to decide on what object that you want to analyze in your fixed specimen. You’re only seeing a small part of the whole specimen.
But if you had a microscope that would allow you to see the whole specimen immediately, you can much more quickly then go in and decide what area of interest you want to focus on and analyze.