Probes Light the Way for Advances in FRET
Novel probes improve high-throughput, three-color and single-molecule measurements.
Gary Boas, News Editor
First described almost 60 years ago, Förster resonance energy transfer (FRET) traditionally uses two molecules — donor and acceptor — to visualize biological activity. Today the technique continues to evolve.
In FRET, if molecules are close enough to one another upon laser excitation, the donor transfers its energy to the receptor, which then emits light. Because this energy transfer occurs only when the molecules are in proximity, the emission can indicate ion binding or protein-protein interactions, for example. Therefore, the technique can contribute to a variety of applications, from basic science studies of measuring conformational changes to drug discovery and more.
Development of a three-color alternating laser excitation technique has helped to address some of the limitations of three-color FRET. The use of alternating laser beams — as opposed to the single excitation laser in three-color FRET — enables sorting of singly, doubly and triply labeled molecules in the same sample. (APD = avalanche photodiode, DM = dichroic mirror, EOM = electro-optical modulator, F = filter, MR = mirror, OBJ = objective, OF = optical fiber, PH = pinhole, P = polarizer.) Reprinted with permission of Biophysical Journal.
Recent developments have improved the efficacy of FRET by increasing its sensitivity. These developments include new types of probes that are optimized for FRET measurements, especially for biosensing applications. They also involve advances on the excitation side: for example, the use of alternating laser excitation and the addition of a third color to be probed. Improvements such as these have helped push the envelope of FRET measurement, leading to important new findings.
Using genetically encoded fluorescent proteins can enable FRET investigations of protein interactions in living cells in nearly real time. The proteins offer a relatively small dynamic range, however, and therefore cannot be used to probe anything other than known interaction partners. Applying them to screening libraries to identify new interaction partners can be problematic at best.
In the Dec. 5 issue of PNAS, researchers at the University of California, Santa Barbara, and Harvard University, in Cambridge, Mass., reported their use of genetically encoded proteins for high-throughput screening for interaction partners in the cytoplasm of Escherichia coli. The new FRET probe vastly improved efficiency, achieved by optimizing the FRET signal, thus enabling the mapping of protein interactions in the intracellular environment.
Researchers have reported the use of genetically encoded proteins for high-throughput screening of interaction partners. Libraries of target receptor and CyPet-peptide fusions were expressed, and cells showing FRET were detected and sorted with fluorescence-activated cell sorting. The researchers improved the sensitivity of the technique through four cycles of sorting. Reprinted with permission of PNAS.
“The idea of using living cells for [such] screening is not new,” said Patrick Daugherty, the principal investigator of the study, “but no one has pushed it, primarily because of the efficiency issue.” The bump in efficiency offered by the new probe knocks down this barrier and so opens up possibilities for high-throughput screening with FRET.
The quantitative screening method is based on intracellular coexpression of a target receptor and a library of ligands genetically fused with yellow and cyan fluorescent proteins optimized for FRET: YPet and CyPet, respectively.
The researchers performed screening using a B.D. Biosciences flow cytometer for fluorescence-activated cell sorting. Because of overlapping FRET signals between target and nontarget cells, they enriched the target cells through four cycles of sorting. After identifying interaction partners, they quantified the binding affinities using a Tecan fluorescence spectrophotometer.
They initially described the fluorescent proteins in a 2005 Nature Biotechnology letter. They evolved the pair through four FRET cycles, screening both yellow and cyan proteins to determine which offered the most improved maximum FRET-ratio change. Each of the cycles consisted of random mutagenesis to discover the amino acid residues that influenced FRET measurements. They subjected these proteins to further mutagenesis and screening until the pair used in the present study was identified.
Efforts by other groups to evolve FRET pairs have proved challenging because the investigators typically developed donors and acceptors independently, rather than evolving the proteins as a pair. According to Daugherty, the former approach improves the efficiency of the individual proteins. “But when you put them back together you find that the biology is more complicated than the physics. The physics of Förster theory for the individual proteins doesn’t capture everything that’s going on.”
In the PNAS study, use of the pair enabled the researchers to identify the ligand preferences of two types of protein-interaction modules: the C-terminal SH3 domain of Mona, a scaffolding protein linked to hematopoietic signaling, and the second PDZ domain of the postsynaptic protein PSD-95, which is associated with memory and learning. Thus they showed the potential of the pair for near-real-time detection of protein-ligand interactions, enabling high-throughput screening for optimal interaction partners.
Such application of these fluorescent proteins is potentially limited because the interaction partners fused with the proteins must be within about 5 to 10 nm and at an appropriate orientation for FRET to work. Flexible linkers may help avoid orientation issues, and similar strategies could help overcome distance constraints.
Another problem may arise when the researchers begin to look at full-length proteins (in the current study, they investigated only how the peptide interacts with small protein modules). The FRET signal will decrease as the size of the protein increases, Daugherty said.
Despite these possible limitations, the FRET hybrids show potential for a variety of applications. “We believe the power of the approach will be for domain-level mapping using protein fragments and for discovery of highly specific intracellular ligands for basic research and pharmaceutical target validation,” he said.
The applications will likely extend to a range of host cells. The researchers primarily used E. coli as an expression host because of its rapid growth rate, high transformation efficiency and ease of manipulation, but they also observed improvements in FRET dynamic range when using the hybrids in the cytoplasm of yeast and mammalian cells. A way to screen for protein interactions in different cell types will open up a number of applications. In fact, the investigators have already sent out more than 300 samples of the FRET pairs for a variety of studies, resulting in roughly 30 citations thus far.
Quantum dots have, in recent years, attracted a tremendous amount of interest because they present an exciting alternative to organic dyes for biological sensing and imaging applications. In particular, some of their unique spectroscopic properties make them powerful donors in developing a variety of FRET-based assays.
Because of the broad absorption spectra of the particles, researchers can excite them safely away (100 to 200 nm) from the absorption spectrum of the acceptor, thus reducing its direct excitation contribution. At the same time, the narrow emissions of quantum dots facilitate signal separation and simplify data analysis. Furthermore, being able to array multiple copies of acceptor-labeled proteins and peptides around a single quantum dot can enhance the FRET efficiencies in these systems.
Several groups have developed basic concepts involving quantum-dot-based FRET and have applied them to design assays for DNA and small analyte detection. In these studies, the scientists took advantage of the significant decrease in direct excitation of the acceptor to improve assay sensitivity.
Researchers have worked with such conjugates over the past several years but until recently did not have information about the distribution or heterogeneity of acceptors in the conjugates — understanding of which is crucial for interpreting data from highly efficient FRET systems.
In the Nov. 29 Journal of the American Chemical Society, the investigators — Thomas Pons, Igor L. Medintz and Hedi Mattoussi of the Naval Research?Laboratory in?Washington along with Douglas S. English and Xiang Wang of the University of Maryland, College Park — reported that they used single-molecule FRET to characterize solution-phase dye-labeled protein-quantum-dot conjugates. This let them determine the heterogeneity and distribution in the population composition as well as obtain structural information about the conjugates.
Investigators have used single-molecule FRET to characterize protein-quantum dot conjugates used as FRET probes. Information provided by the measurements likely will contribute to improved efficacy of the conjugates in FRET studies. Reprinted with permission of JACS.
They studied self-assembled conjugate sensors consisting of CdSe-ZnS core-shell quantum dot donors surrounded by dye-labeled protein acceptors. They performed single-particle FRET measurements using a Zeiss confocal microscope. Placing 100-μl samples on a glass coverslip about 10 μm above the 100×, 1.3-NA objective, they excited them with 488-nm light from an argon-ion laser.
The same objective collected fluorescence from the quantum dot donors and dye acceptors, which was then detected by two PerkinElmer single-photon-counting avalanche photodiodes. A National Instruments acquisition counting board recorded the corresponding time traces, which the scientists analyzed with custom software written in LabView, also from National Instruments.
The experiments first revealed a one-to-one correlation between single-molecule and ensemble-based FRET measurements, with respect to both derived FRET efficiencies and donor-acceptor separation distances. More importantly, they demonstrated that heterogeneity of self-assembled conjugate populations dominates these conjugates. They showed, for example, that for quantum-dot-protein conjugates with low ratios of dye-labeled protein to quantum dots (low valence), a mixture of fully bound/conjugated and fully unconjugated quantum dot substrates exist in the solution.
They confirmed a longstanding idea that the distribution of protein acceptors in the protein-quantum dot conjugates follows a Poisson distribution — a discrete probability distribution describing the number of events occurring in a given period of time if these events occur with a known average rate and are independent of the time since the last event. This finding is important, because it means that investigators can account for the heterogeneity in subsequent analyses and that it applies to a whole range of biological problems — in particular, sensor designs based on quantum dots and FRET.
The results of the study are especially significant because researchers often assume a balanced system when performing such measurements. “You would like to have a system that has one valence — that has a single acceptor attached to each donor,” Mattoussi said. “This study tells us that you have to understand the heterogeneity and use various analytical techniques (for example, column chromatography) to separate conjugates having distinct valencies.”
In fact, Mattoussi and his colleagues previously described methods to control population composition in protein-quantum-dot conjugates. In the July 2006 issue of Nature Materials and in the October 2006 Journal of Physical Chemistry B, they reported studies in which they used histidine-driven self-assembly of peptides and proteins onto quantum dots. Their methods allowed them to control the number of peptides/proteins per quantum dot and, in some cases, even enabled them to control the protein orientation in the conjugates. The former ability in particular is essential for characterization of quantum-dot-based sensors.
Continued improvement of conjugation techniques and of molecular probes based upon quantum dots likely will lead to increased application in a number of areas, including intracellular sensing of protein interactions, ligand-receptor binding and monitoring of protein trafficking in living cells as well as across cell membranes.
Recent years have seen the introduction and development of three-color FRET methods in which molecules are labeled with three probes, each with its own emission spectrum, allowing monitoring of three-molecule interactions and up to three interprobe distances. (Most FRET experiments, of course, use a pair of probes — the donor and acceptor — to measure bimolecular interactions and/or interprobe distances.) Researchers have described both ensemble-based and single-molecule approaches to three-color FRET, and the latter has been used to explore complex dynamics of biomolecules with surface-immobilized and diffusing molecules.
Application of single-molecule three-color FRET can be limited, however, because all three probes must be within 10 nm of each other to report on their interactions and distances. If they are not, the measurements are, in effect, no different from those of singly or doubly labeled species that may also be present in the sample. Furthermore, the technique cannot detect species that are not appreciably excited by the single-laser excitation used.
To address these limitations, researchers at Seoul National University in South Korea, at University of Oxford in the United Kingdom and at University of California, Los Angeles, have introduced three-color alternating laser excitation of single molecules.
As reported in the January issue of Biophysical Journal, the technique — known as three-color ALEX — uses alternating laser excitation of single molecules labeled with up to three probes to measure molecular interactions and interprobe distances, as well as to observe and sort singly, doubly and triply labeled molecules that are present in the same sample.
To describe three-color ALEX, University of California professor Shimon Weiss, lead investigator of this project, draws an analogy to fluorescence-assisted cell sorting.
“What fluorescence-activated cell sorting does for cells, ALEX does for single molecules,” he said. “We are using time-division multiplexing of different excitations to look at coincidences, stoichiometries and distances of different molecules, allowing us to sort and identify each molecule in a mixture.”
The technique is a substantial extension of two-color ALEX, originally described in the June 15, 2004 issue of PNAS. With the two-color method, a donor-excitation laser measures FRET efficiency and tracks the donor presence while an acceptor-excitation laser directly monitors the acceptor presence. The ratios of the two photon counts (the FRET efficiency and probe stoichiometry) are calculated for each molecule and then are binned into a two-dimensional histogram. The additional laser for the three-color technique allows measurement of three FRET efficiencies and three different probe stoichiometries for a three-probe system. Information from each molecule is binned into a six-dimension histogram.
The three-color technique enabled analysis of molecular interactions and interprobe distances. A 477-nm laser excites B in the 0- to 36-μs domain, a 532-nm laser excites G in the 40- to 76-μs domain and a 633-nm laser excites R in the 80- to 116-μs domain, all generating photon counts (F). Reprinted with permission of Biophysical Journal.
To demonstrate the method, the researchers modified a previously developed two-color ALEX single-molecule fluorescence microscope. The instrument included three excitation sources: a Melles-Griot argon-ion laser at 477 nm, a World Star Tech solid-state laser at 532 nm and a Melles-Griot HeNe laser at 633 nm. A combination of polarizers and electro-optical modulators alternated the three sources. Light from the sources was collimated and sent to an Olympus inverted microscope, then focused on the sample through a 60×, 1.2-NA water-immersion objective. The same objective collected fluorescence, which was sent to Perkin-Elmer avalanche photodiode detectors.
Using singly, doubly and triply labeled DNA species, the team found that the technique enabled FRET-independent analysis of three-molecule interactions; monitoring and sorting of all seven types of labeled species; measurement of three interprobe distances, even when there is no substantial FRET among the probes; and investigation of conformational heterogeneity with much higher resolution than is available with conventional single-molecule FRET.
A number of applications could benefit from the three-color single-molecule technique, Weiss noted; e.g. genomic and proteomic analysis, early diagnosis, drug discovery and fundamental questions in molecular machines.
Also, in the Nov. 17 issue of Science, Weiss and colleagues reported a two-color ALEX study that resolved a 25-year-old question involving the initiation of transcription. By measuring distances within single molecules, they showed that initial transcription occurs through a sort of “scrunching” mechanism.
MORE FROM PHOTONICS MEDIA