On July 1, investigators at the University of Colorado, Boulder, received a $320,000, four-year grant from the National Institute of General Medical Sciences to develop targeted libraries of fluorescent proteins; express the libraries in mammalian cells; and screen the proteins for increased brightness, increased photostability, and decreased conversion to “dark states” using a microfluidic cell sorter that they recently implemented.
The researchers will also conduct multiparameter screens to identify mutations known to enhance multiple photophysical properties and combine synergistically to improve signal output.
Such information is important because protein engineering work based on a single selection scheme often optimizes one property at the expense of another.
The research has three specific goals: increase the fluorescent signal 20-fold with a multiphase screen that probes the dark-state relaxation rate; increase the photostability of red fluorescent proteins by at least 7-fold with a screen that probes photobleaching; and identify mutation combinations that have synergistic effects on fluorescent proteins’ photophysical properties.
Amy Palmer, an assistant professor of chemistry and biochemistry at the University of Colorado and a co-principal investigator on the grant, spoke with CBA News this week about the work and how improved fluorescent proteins can impact drug discovery.
Can you give me a little background on this work?
I am a relatively new faculty member here at the University of Colorado. I guess I have been here about three years. A lot of the research that my group has been involved in relates to developing improved tools that enable us to image specific proteins, ions, and other molecules inside living cells.
As part of this work, we use different kinds of fluorophores. One kind of fluorphore that is really useful is naturally occurring fluorescent proteins. I was introduced to these fluorophores when I did my postdoctoral work at the University of California, San Diego, with Roger Tsien, who has done a lot of pioneering work with these fluorescent proteins.
That is where I learned about them and how useful they could be to illuminate cell biology. Although these fluorescent proteins have revolutionized cell biology, they are not really as good as they could be, and some evidence of that is the fact that you can see in the scientific literature somewhat marginal improvements to fluorescent proteins on a regular basis.
One of the things that I was interested in is seeing if we could make really profound improvements to these proteins by trying to explicitly optimize some of their photophysical properties that are currently suboptimal.
To do this, I basically initiated a collaboration with Ralph Jimenez, another scientist here who is also a co-principal investigator on the grant. One of the advantages of bringing him on board is that I have a lot of expertise in protein design and optimization, essentially making protein libraries and using these types of proteins in cells. Ralph’s expertise is essentially as a physicist. He has expertise in building lasers and developing optical tools to make very specific measurements.
We decided that one way we could really make a dramatic improvement in these fluorescent proteins is if we could make large libraries, meaning make a lot of mutations in these proteins and then screen them for the specific properties that we were interested in, and then select out the ones that have those properties of interest.
To do that really effectively, we realized that we needed a relatively high-throughput way to look at these proteins and select what we wanted. The important point was that it needed to be high-throughput, and in addition, the types of measurements that we wanted to do were rather sophisticated.
Instead of just exciting the protein and shining a light on it, and looking at the light that comes out, we wanted to do things that were much more difficult to do. This is where working together has really helped, and we realized that Ralph knew how to conduct all of these measurements, if we could integrate the optics into microfluidic devices.
Then we had this idea that we could make libraries of proteins, and then we could put the proteins into cells, such that each cell would express one particular library member. We could then actually screen the cells themselves in these microfluidic devices. That enables us to determine if the properties of the protein that each cell is expressing are good or bad, or if we want to keep them or discard the cells that do not have optimal properties.
The idea would be to collect the cells that do, and iterate on that model. I would say that our ultimate goal in the project is twofold. The technology of the optically integrated microfluidic sorter is very useful and sophisticated, and we want to create these improved fluorescent proteins that can be used really as cell biology tools that we hope will enable a new round of experiments that are not currently possible.
How would this be applicable to drug discovery?
This technology could potentially be applied in drug discovery endeavors in several ways, one of which is to dramatically improve the sensitivity of any cell-based assay that relies on fluorescence and is being used for compound screening. Our goal is to increase the fluorescence signal by 80-fold, so you could imagine that likewise, the sensitivity of any assay using the proteins would be dramatically improved.
That is just one way these proteins cold be used in drug discovery, but it is probably the most broadly applicable way.
How do these fluorescent proteins work, and how do they improve what is currently available?
I guess that we are still limited in how bright naturally fluorescent proteins are, so you can imagine if you have an assay where you can’t have very many copies of a protein inside of a cell, or if, for example, you want to express something for imaging an organism, not just cells. Tissue penetration then becomes a problem, because if you have a protein deep within a tissue and it is not very bright, then you essentially cannot see it.
But if you can improve the signal dramatically, then it might enable deeper tissue imaging, not just cell-based imaging.
I guess the advantage we have is that in this device we are creating, we can optimize multiple properties simultaneously. This is a bottleneck problem in the field — a lot of the screening techniques result in proteins that have improved properties in one area, but decreased properties in another. So for example, it might improve fluorescence intensity, but decrease luminescence duration, or photostability.
We have multimodal selection, so we can select properties that are going to be synergistic instead of antagonistic. We think that kind of screen has not really been conducted, so we are hoping that this approach will really open the door to create more useful tools and to better understand how the fluorescence is controlled.
Once we gain a better handle on that, we can use that to our advantage in iterating and creating better tools.
What is the next step in this project?
I would say, for myself and my group, we actually take the proteins and use them to develop sensors. For example, we are developing sensors that enable us to look at specific molecules, ions, and signaling components within cells.
We would probably take these improved proteins and try to make dramatically improved sensors. Likewise, these improved sensors could be used in drug discovery and screening down the road.
We also want to try to push the limits of these proteins, so our ultimate goal is to be able to look at single copies or single molecules within cells. The aim, then, is to make the proteins bright enough so that you actually look at one protein expressed inside of a cell.
That would really open up a huge area of science where people are now trying to better understand how very small numbers of molecules interact in cells. So I think we want to make sensors and extend this work by enabling researchers to look at single molecules inside cells.
How are you using the grant money?
The grant was awarded this year, so we basically used part of the money to support graduate students who are working on the project. Some of the money is used to actually build the instrument. Some of it is also used to purchase standard supplies, such as consumables.