Förster resonance energy transfer (FRET) imaging can show the location of protein-protein interactions in cells, in tissues or throughout an organism. However, it requires exciting a fluorophore with light, which can induce autofluorescence, damage sensitive tissue and cause photobleaching. Furthermore, because the technique involves the transfer of energy from a donor to an acceptor fluorophore with overlapping spectra, stimulating the donor often inadvertently excites the acceptor. In contrast, bioluminescence resonance energy transfer (BRET) does not have problems associated with photoexcitation because it uses an enzymatic reaction that produces bioluminescence to donate energy to a fluorophore. It has proved useful for high-content screening, which does not require high spatial resolution. Nevertheless, the method has not produced sufficient resolution for imaging protein-protein interactions because bioluminescent enzyme substrates produce less intense light than fluorophores.Researchers performed bioluminescence resonance energy transfer (BRET) imaging in tobacco (shown) and in Arabidopsis seedlings (not shown). Images reprinted with permission of PNAS.Now a new CCD camera and a device that separates the bioluminescence and fluorescence BRET signals have enabled scientists at Vanderbilt University in Nashville, Tenn., and at the University of Tennessee, Knoxville, to detect more of the bioluminescent signal. As a result, they have successfully performed BRET imaging with subcellular resolution in both plant and animal cells. Principal investigator Carl Hirschie Johnson said that their success means that biologists can now use BRET imaging when photoexcitation is undesirable. In their experiments, the researchers aimed to show that BRET has achieved sufficient resolution for imaging. For that reason, they demonstrated the technique in commonly used specimens, including mammalian cells, tobacco seedlings and tissues, as well as in Arabidopsis seedlings, tissues and cells. Tobacco is agriculturally important, and Arabidopsis is the most commonly used plant for genetic, molecular biology and biochemistry experiments, while mammalian cells often are employed in biomedical research. Shown are BRET images of plant tissues (left and middle) and a cell (right) from Arabidopsis seedlings. The nucleus is the slight bulge in the top left of the cell.For the BRET donor and acceptor in plants, the researchers used Renilla luciferase and enhanced yellow fluorescent protein, respectively. In mammalian cells, they employed Renilla luciferase optimized for human codon bias and Venus, a yellow fluorescent protein variant. For both plants and mammalian cells, BRET was measured as the ratio of the blue luminescence of the luciferase substrate to the yellow fluorescence of yellow fluorescent protein. In plants, they used the natural luciferase substrate, coelenterazine. However, they found that the serum proteins in mammalian cell media could trigger luminescence by oxidizing coelenterazine. In other words, autoluminescence was occurring. To solve this problem, the researchers employed ViviRen, modified coelenterazine from Promega Corp. containing ester groups that inactivate it until it encounters natural enzymes in the cells that cleave the ester groups. It exhibited less autoluminescence than natural coelenterazine — even in serum-free media — which could be the result of the presence of dissolved oxygen.Advanced imagingThe scientists monitored BRET with an Olympus microscope contained entirely in a wooden box that excluded light to increase the resolution of the images. Inside the box they placed metal plates that could be heated to control the temperature of the microscope. A Dual-View imager from Optical Insights LLC of Tucson, Ariz., was attached to the bottom port of the microscope, and the camera was connected to the imager. Both the camera and the imager were outside of the box, connected via wires that ran through a hole in the box.The BRET ratio of luminescence to fluorescence signal can change over time because luminescence sometimes fluctuates. To measure the contribution of the blue versus the yellow signal, the scientists separated the signals with the imager, which includes a beamsplitter and filters that further refine the signal, reducing background noise and enabling them to look at the images pixel by pixel. Johnson noted that they could not have used spinning disk confocal microscopy to separate the signals because they were using long exposures, ranging from 100 ms to 10 min. He added that the imager is commonly used for FRET, so familiarity with the device made it an easy choice.For the camera, the researchers used a Hamamatsu electron-bombardment CCD. They tested various cameras and empirically found this device was most suitable to their application because it is highly sensitive. The company then made special modifications for the researchers, including removing an aluminum mask and adding a GaAsP photocathode and Peltier cooling to –25 °C using circulating water. Removing the mask and adding the photocathode reduced a source of noise called ion feedback, and cooling reduced another type of noise called dark current. Johnson said that liquid nitrogen cooling might further reduce the dark current but that it requires a large tank that needs frequent refilling with liquid nitrogen — a great inconvenience. Mission accomplishedUsing the advanced imaging equipment, the scientists produced BRET images that showed the location of the dimerization of COP1, a protein involved in light-regulated development in plants. They successfully visualized the process in whole tobacco seedlings and tissues, as well as in whole Arabidopsis seedlings, tissues and cells — down to the nucleus. Besides achieving subcellular imaging, they discovered that dimerization of this protein occurs in the roots as well as in other tissues, surprising because roots are not exposed to much light, yet the protein is involved in light-regulated development. The results of the experiments highlight the importance of using BRET instead of FRET in plants, where photoexcitation is undesirable. Photoexcitation would have affected the process that they were trying to study because the protein is involved in light-regulated development. Also, plants have a high level of autofluorescence because of the presence of chlorophyll and other pigments, and this autofluorescence interferes more with FRET measurements than with BRET measurements.Nevertheless, the natural plant pigments increased the contribution of the yellow fluorescent signal to the BRET ratio. Although plants need light to grow, they can be germinated in the dark, resulting in seedlings without pigment. The researchers used nonpigmented tobacco seedlings to normalize for the pigments in tobacco. While performing BRET imaging with mammalian cells, they tracked the nuclear entry of a protein and its subsequent interaction with another protein. These are BRET images of mammalian cells.They reported the experiments with seedlings, plant tissues, and plant and mammalian cells in the June 12 issue of PNAS.Johnson said that it should be possible to image nuclei, mitochondria and perhaps the Golgi in both plant and animal cells. However, he noted that the limits of BRET imaging have yet to be fully explored. In other words, the technique is likely to have even more potential than has been demonstrated thus far.Although BRET has advantages over FRET, he said that the former may never replace the latter completely and that the techniques will remain complementary. He noted that fluorescence imaging still has a higher resolution than bioluminescence imaging and that techniques such as lifetime imaging are not yet possible with bioluminescence.He said that BRET likely will improve and gain functionality with the appearance of imaging equipment that further increases photon throughput and lowers the background. Fluorophores with longer-wavelength fluorescence and luminescent substrates with shorter wavelengths would further refine the spectral distinction, and brighter substrates are desirable because they would allow shorter exposure times. For mammalian imaging, Johnson would like to have fluorophores that emit above the wavelength of hemoglobin.