Mutant Chronicles: The quest for a better red-fluorescent protein

Red-fluorescent protein labeling cellular structure in Zebra finch rotundus. The picture is one of many examples of the remarkable insight into cellular structure and dynamics obtained by using fluorescent proteins. The development of new red-fluorescent proteins resistant to inactivation by laser light and oxygen is a goal of the Jimenez group. Credit: Photo courtesy of Dr. Richard Mooney, Duke University School of Medicine
Microfluidics system used in the Jimenez lab to select the top-performing cells during a directed evolution experiment. Lasers measure the fluorescent properties of mutated fluorescent proteins in each cell passing through the system, and another laser provides a “tractor beam” to grab onto the best mutant cells for further investigation. This system allows the group to rapidly investigate tens of thousands of different proteins. Credit: The Jimenez group and Steve Burrows, JILA

Because red fluorescent proteins are important tools for cellular imaging, the Jimenez group is tailoring them to its own biophysics research. The group’s quest for a better red-fluorescent protein began with a computer simulation of a protein called mCherry that fluoresces red light after laser illumination. The simulation identified a floppy (i.e., less stable) portion of the protein “barrel” enclosing the red-light emitting compound, or chromophore. The thought was that when the barrel flopped open, it would allow oxygen in to degrade the chromophore, thus destroying its ability to fluoresce.

The group decided that its next step(s) would be to tweak the natural protein to make it more stable. Tweaking proteins is a huge challenge because most combinations of mutations result in a complete loss of the necessary structure to maintain fluorescence. Even so, the group succeeded in developing a new approach to real-world protein improvement that employs a laboratory strategy for directed evolution.

Directing evolution is challenging. The first step requires creating a library of hundreds of thousands of cells containing different mutations of a single protein. This step is now relatively easy, thanks to the tools of molecular biology. The second step requires screening the fluorescence properties of each cell to select only those few that contain top-performing mutant proteins.

To accomplish the selection process, the group uses microfluidics combined with several laser beams. Its microfluidics system contains micron-sized three-dimensional transparent channels that carry small streams of liquid and allow cells to flow through them one at a time. As the mutant cells pass through the microfluidics channel, lasers measure the fluorescent properties of each mutant cell to assess how well the cells maintain their fluorescence when repeatedly excited by the series of laser beams. Another laser acts as an optical trap that works like a tractor beam to grab onto the best mutant cells for further investigation. The microfluidics setup itself readily removes the cells that are poor performers by simply allowing them flow out of the device.

To make matters more challenging, directed evolution requires repeating the two steps described above multiple times. The Jimenez group is currently in the middle of round three of its quest to evolve a better red-fluorescent protein.

Although the group has already shown that the specific improvements suggested by the computer simulation don’t work, the first round of the directed evolution experiment has come up with an improved red-fluorescent protein with a less floppy barrel that is 2–4 times more stable than mCherry. The combination of mutations that resulted in this improvement has not been previously observed in nature and was completely unexpected.

The group named its new mutant protein Kriek, after a Belgian beer made via the fermentation of cherries. Clearly, the researchers are adept at doing more than biophysics. They include JILA Ph.D. Jennifer Lubbeck (2013) and Fellow Ralph Jimenez, Kevin Dean and Amy Palmer of CU’s Department of Chemistry and Biochemistry, as well as colleagues from the University of Tennessee Space Institute and Florida International University.

Improving the Photophysics of Red Fluorescent Proteins for Biological Imaging

A microfluidic cytometer for rapidly sorting FP-expressing mammalian cells based on the selective measurement of photostability, magnitude of dark-state conversion, or the combined effects of both.

Microfluidic design for measuring and sorting on photostability of fluorescent cells in flow.

Microfluidic design for measuring and sorting on photostability and fluorescence lifetimes of fluorescent cells in flow. Cells are hydrofocused to a narrow stream and their photophysical properties are investigated using an array of ten beams. When cells shows desirable properties, an optical trap is activated to deflect them from their initial path and isolate them from all the cells that have not been selected.

Although they have become important tools in biological microscopy, fluorescent proteins (FPs) are dimmer and less photostable than small-molecule fluorophores.  These limitations thus limit the signal output needed for next-generation single molecule measurements.  The goal of this project is to generate new red fluorescent proteins (RFPs) with high photostability and brightness as well as a reduced tendency to convert to transient dark states.  Our approach to developing new RFPs involves constructing targeted genetic libraries based on existing RFPs, expressing these libraries in yeast cells and screening the photophysical properties of each cell at high rates in a microfluidic flow cytometer.

To this end, we have developed an innovative microfluidic cytometer for rapidly sorting FP-expressing yeast cells based on the selective measurement of photostability, fluorescence lifetime, or the combined effects of both.  In the cytometer (shown schematically in the figure), ten elliptically shaped laser beams intersecting a hydrofocused stream in a microfluidic channel interrogate each cell. Fluorescence lifetime and the initial fluorescence intensity of each cell is measured in the first beam.  Cells then traverse eight beams in which the cells are repeatedly excited to introduce low photostability due to photobleaching.  In the last beam, the cell’s fluorescence intensity is measured again to assess the degree of photodamage, measured as the ratio of pre- and post-bleaching fluorescence intensities, that occurred to the FP molecules in the cell.  Subsequently, a piezo-steered infrared trapping beam deflects cells with desirable properties using optical forces.  In our measurements, the multi-beam sequence provides millisecond timescales of excitation and dark intervals to separate the effects of reversible dark-state conversion and irreversible photobleaching.  We have also added single-cell excited-state lifetime measurements to this instrument to enable screening and sorting on the basis of multiple parameters, including fluorescence lifetime and photobleaching.  This technology is uniquely capable of sorting genetic libraries of cell-based FPs on the basis of their excited-state dynamics.  Its ability to investigate the various processes that limit the photon output of FPs position this method as a versatile and powerful new tool for characterizing FP photophysics and engineering new FPs.

In this project, which is a part of a collaboration with Amy Palmer's group at the CU BioFrontiers Institute, we have examined a wide range of genetic libraries based on multiple RFPs.  For example, one set of libraries contains sequence diversity vicinity of the chromophore to alter the steric bulk, hydrogen bonding, and local dielectric properties.  These libraries are being investigated to uncover relations between mutations in the chromophore pocket and the propensity for triplet state formation, cis-trans isomerization, and excited state proton transfer, all of which which lead to dark state formation.  Other libraries with sequence diversity more distant from the chromophore pocket will also be investigated in an attempt to improve barrel rigidity and reduce permeability to molecular oxygen, which is likely to be important in reducing irreversible photobleaching.  In general this work will provide insight into the triad of structure-function-dynamics which control the dynamics of excited electronic states in proteins.


Characterization of dark state conversion in fluorescent proteins

Fluorescent proteins are characterized with rich photo-dynamics due to the existence of several non-fluorescent or “dark” states.  Upon excitation, FPs get trapped in these dark states via a process called dark state conversion (DSC) and subsequently relax back to the ground states. Recovery of the molecules from these dark states to the ground state (Ground State Recovery, GSR) has profound impact on in-vivo imaging and super-resolution microscopies. We study the dynamics of DSC/GSR and molecular origin of the dark states using various spectroscopic techniques.

Genetically Encoded Biosensors

Figure 1. Microscope image of HeLa cells expressing FRET sensors in growth conditions.
Figure 2. Sensor response to metal chelating agents and added zinc ions.
Figure 3. Experimental workflow.
Figure 4. Single cell measurements immediately and several seconds after added zinc ions reveal sensor heterogeneity within the microfluidic device. Pair-matched response (inset).

Genetically encoded biosensors are important tools for real-time measurement of the distribution of metal-ions (e.g. Ca2+, Zn2+) and other analytes in living cells (Figure 1).  Among a handful of design platforms, the fluorescent protein (FP)-based ratiometric sensors have the distinctive advantage of being able to compensate for potential signal variations caused by cellular fluorophore concentration and excitation laser intensity. As shown in Figure 2, these sensors report the concentration change of intra- or extracellular metal ions by recruiting one (or a few) metal ion(s) to their metal sensing domains, inducing a conformational change which leads to a change in fluorescence resonance energy transfer (FRET) efficiency between the donor (CFP) and acceptor (YFP).

Over the years, most efforts in rational sensor design and optimization have encountered unpredictable and surprising results, due to poorly understood interactions between molecular components of the FRET constructs and the cellular environment. One goal of this project is to develop genetically encoded metal sensors with dramatically improved optical, physical, and chemical properties and to enable high speed event detection at low signal levels in living cells. Using a directed evolution approach to systematically explore the landscapes of the sensor constructs, this study will also provide insight into the rational design of biosensors.

The experimental method mimics the typical calibration experiment utilized by sensor designers to determine full response by measuring the minimum FRET signal and the FRET at later times reaching saturation. We design targeted libraries of the sensor constructs and express the library members in living cells (Figure 3). Using microfluidic technology developed in our lab, the library-encoded cells are loaded into a microfluidic system in which cellular responses are chemically induced and optically measured, and are screened and sorted at high-throughput. Micron-scale laminar flow and droplet generation techniques offered by microfluidics enable us to achieve rapid reaction initiation, accurate timing control, high-throughput detection of the sensor response, and sorting at the single cell level (Figure 4). Our interest extends from the fundamental molecular dynamics, such as the mechanism of metal-ion sensing, to the important cellular processes of metal homeostasis, such as the dynamics of metal-ion permeation through the plasma membrane, as well as the improvement of these sensors for measurements in a variety of cell organelles with differing chemical environments.


Microfluidic cytometer for characterizing photosynthetic efficiency and lipid accumulation in algae

Figure 1. Image of the diatom Phaeodactyl tricornutum, with red fluorescence indicating the presence of lipid droplets accumulated in the cell.

Unicellular photosynthetic organisms (e.g. - phytoplankton, algae, cyanobacteria) are the world’s dominant primary producers and account for ~ 45% of all carbon fixation. They are taxonomically diverse with over 10,000 species identified. These organisms have a strong potential to produce precursors for biofuels and high-value chemicals. However, production of these precursors varies greatly both among these species and within populations of a single species (often driven by growth conditions). The reasons for these variabilities are not fully understood since information on individual algal cells under natural conditions is currently inaccessible.  Our research focuses on novel technologies developed in our lab (using custom-designed, multifunction, high-through-put, analytical instruments) that simultaneously evaluate, in individual cells, photosynthetic efficiency (utilization of solar energy), oil/biomass production potential and other photophysical properties. In light of the critical importance of these photosynthetic microorganisms for human health, energy needs, and ecosystem sustainability – and in the absence of established technology for the characterization of microbial functional diversity – this research addresses an urgent need.

Components of this project include (i) studies of laboratory-grown cyanobacterial and algal cultures, (ii) development of new measurement techniques to monitoring parameters needed to differentiate individual cells in mixed natural communities, (iii) assessment of the functional diversity within communities and (iv) development of improved strains using screening and directed evolution techniques. The results of this research are expected to lead to new species identification, characterization of natural environments, changes in microbial oil/biomass content, and identification of photosynthetic microorganisms with superior properties leading to economically competitive biofuel and high-value chemical production.

Our collaborations with other laboratories bring expertise to our research in diverse areas that includes specialists in limnology, biology, biochemistry, biophysisics and photobiologists thus providing the necessary intersection of novel technology and ecological expertise.