On the Path Toward More Useful Fluorophores

Lynn M. Savage, Associate News Editor

Where would modern biomedical research be today without fluorescent probes? Some would argue that it would be blindly groping down a dark alley, with no efficient means of discovering the finely honed interplay of proteins and other substances in the human body, especially at the cellular and molecular levels.

But such probes do exist, whether as dyelike molecules, artificial metallic dots or a number of other brightly glowing forms. Still, there is room for improvement among the probes — more brightness and greater endurance, for example — and there remain biological applications that demand fluorescing particles that do not yet exist. Investigators across the globe are attempting to fill these gaps.

Don’t spare the rods

Zinc oxide has several desirable optical properties — including its high stability in cellular environments — but despite its advantages, it has received little attention as a potential biosensing material. However, researchers at Pennsylvania State University in University Park report in the April 26 ASAP edition of Langmuir that they have constructed substrates featuring nanoscale ZnO rods that greatly enhance biomolecular fluorescence.

The investigators — Adam Dorfman, Nitin Kumar and Jong-in Hahm — produced the ZnO nanorods on silicon wafers by chemical vapor deposition. Using a scanning electron microscope made by FEI Co. of Hillsboro, Ore., they measured the rods’ average diameter to be 677 ±65 nm. Likewise, they made substrates featuring silicon nanorods with diameters of 547 ±17 nm with which to compare the ZnO rods.

On various substrates, such as glass and quartz slides, silicon and polymethylmethacrylate (PMMA), the scientists placed antibovine immunoglobulin conjugated with fluorescein isothiocyanate (FITC-antiIgG). With a confocal laser scanning microscope from Olympus and a 488-nm argon-ion laser operating at 40 mW, they confirmed that FITC-antiIgG alone does not fluoresce. Replacing the substrates with those featuring silicon nanorods provided a negligible amount of fluorescence, but employing those with ZnO rods resulted in three orders of magnitude greater fluorescence at the same protein concentration (200 μg/ml) (Figure 1).

Figure 1. Zinc-oxide nanorods laid out on silicon scaffolds enhance fluorescence detection. Fluorescein conjugated with an antibody and placed on silicon wafers or other materials exhibits little fluorescence contrast (a). However, the same conjugate, when adsorbed onto single ZnO nanorods (b) or patterned arrays of ZnO nanoplatforms (c), exhibits strong fluorescence. Reprinted with permission of the American Chemical Society.

They also compared platforms with ZnO thin films and with ZnO nanorods. They found that, although the film was better than the other substrate materials, using the rods provided better fluorescence enhancement.

To evaluate how the ZnO rods might be used to detect biological agents, the researchers tested them with oligonucleotides specific to Bacillus anthracis, B. cereus and a modified form of B. anthracis that carried 6-carboxyfluorescein. No fluorescence was detected from samples with the B. cereus oligonucleotide, but intense fluorescence was detected from those with B. anthracis and complementary sequences. They report that the technique enables the detection of labeled DNA at concentrations as low as tens of attomoles per liter without amplification.

They note that they are investigating the exact mechanism behind the enhanced fluorescence.

Reducing the quench

Fluorescent dyes such as Cy5 are handy for cell and tissue imaging and can act as reliable sensors of protein interactions. However, you cannot increase the sensitivity of the imaging system or sensor just by adding an unlimited quantity of a dye onto an antibody or other carrier: Nonfluorescent dimers form, quenching the fluorescence of the system.

Now, though, investigators at the US Naval Research Laboratory in Washington, at Scripps Research Institute in La Jolla, Calif., and at Nova Research Inc. in Alexandria, Va., have devised a more effective approach for accruing dye particles onto a molecule.

They created a scaffold for Cy5 molecules from a variant of cowpea mosaic virus. The mutant includes inserted cysteine groups that enable dyes to readily attach to the virus’s protein shell. The location of the cysteine ensures that the dye particles are spaced far enough apart (6.5 nm) that nonfluorescent dimers cannot form, thus eliminating quenching.

The researchers report in the April 19 issue of Journal of the American Chemical Society that they fashioned several versions of the scaffold complexes, composed of different dye-per-virus ratios up to a maximum of 42 dye molecules per virus. They also added NeutrAvidin, a binding protein from Amersham Bio?sciences Corp. of Piscataway, N.J., that enables the complex to act as the recognition element of a sensor.

They used a spectrometer from Varian Inc. of Palo Alto, Calif., to measure the amount of dye and virus in their samples and a fluorometer from Horiba Jobin Yvon of Edison, N.J., to measure the fluorescence intensity of the scaffolded Cy5. They found that the fluorescence signal increased linearly as the amount of dye on each scaffold rose, indicating minimal quenching (Figure 2).

Figure 2. Using cowpea mosaic virus as a platform for Cy5 enhances fluorescence because the fluorophore’s molecules can be closely arranged without overpacking to the point that self-quenching occurs. In a fluorescence microarray application, nonscaffolded Cy5 (a, b) does not have the same response as Cy5 on the viral platform (c). Reprinted with permission of the American Chemical Society.

They also compared the complex with a mixture of Cy5 and nonmutated virus and with free Cy5. They confirmed that Cy5 attaches to the mutated form of the virus but not to the wild type, and that the complex provides enhanced fluorescence intensity over the dye alone.

A healthy glow into old age

The investigators noted that energy transfer between the virus and Cy5 may provide some of the fluorescence enhancement but that it does not account for all of it. Research is under way to elucidate the full cause.

Many commercial fluorophores are water-soluble, making them suitable for biological imaging applications. However, they tend not to be very stable, photobleaching after only a few seconds. A group of researchers has developed a novel dye based on terrylene diimide that is not only water-soluble, but also highly fluorescent in various environments for as long as minutes at a time.

Scientists at Ludwig Maximilians Universität in Munich and at Max Planck Institüt für Polymerforschung in Mainz, both in Germany, and at the University of Illinois at Urbana-Champaign synthesized and characterized WS-TDI, a molecule that has a hydrophobic core but also a number of hydrophilic sulfonic groups that enable water solubility.

They report in the April 19 issue of Journal of the American Chemical Society that WS-TDI, although water-soluble, forms self-quenching aggregates in polar solvents but remains in its monomeric form in such polymers as PMMA and polyvinyl alcohol. Also, adding a surfactant breaks up the aggregates into monomers, leading to an increase in fluorescence.

WS-TDI absorbs at ~680 nm and has a maximum emission of ~720 nm. Using an inverted microscope from Carl Zeiss of Göttingen, Germany, either a 635-nm laser from PicoQuant GmbH of Berlin or a 633-nm helium-neon laser and an avalanche photodiode from EG&G Optoelectronics of Vaudreuil, Quebec, Canada, the researchers investigated the photostability of WS-TDI, finding that it is more photostable than oxazine-1, sulforhodamine-B and other commonly used water-soluble dyes.

They also found that WS-TDI is less stable than — and emits about one-third the number of photons of — the water-insoluble form of TDI. They note, however, that the lower stability and emissions are the price to pay for achieving water solubility.

Comparing dye abilities

The investigators compared the ability of WS-TDI, Alexa 647 and a styryl dye to image living cells. Where the latter fluorophores began to photobleach three seconds after irradiation from a 633-nm HeNe or 532-nm Nd:YAG laser, the cells labeled with WS-TDI remained clearly fluorescent after four minutes (Figure 3).

Figure 3. Cells loaded with Alexa 647/dextran (a), a styryl dye (b) or a water-soluble form of terrylene diimide (c) show good fluorescence at the start, but only the latter maintains its intensity for more than a few seconds. Reprinted with permission of the American Chemical Society.

Cellular biologists interested in the kinetics of molecular interactions use dual-color fluorescence cross-correlation spectroscopy to track the movements of two different fluorophores in the same region. However, to excite the fluorophores, two lasers, usually of different wavelengths, must be carefully aligned to the same confocal spot. And, although the technique can be performed with a single laser, crosstalk and unwanted Förster resonance energy transfer (FRET) between the fluorescence particles typically occurs.

To improve the technique, researchers at Riken in Wako and at Hokkaido University in Sapporo, both in Japan, have developed a fluorescent protein with a Stokes shift of ~180 nm. Because of this large shift, they dubbed the protein “Keima,” after the knightlike piece used in the Japanese game shogi. They report in the May issue of Nature Biotechnology that the variant, which is based on a mutated nonfluorescent protein taken from the stony coral Montipora spp., exhibits maximum absorption at 440 nm and maximum emission at 620 nm.

They tested the protein in a single-laser fluorescence cross-correlation spectroscopy setup, pairing it with cyan fluorescent protein and exciting both fluorophores with a 458-nm argon-ion laser. They found that the combination enabled simple but efficient use of the technique because Keima and CFP have nearly the same excitation spectra and easily separable emissions. Furthermore, no energy transfer occurred between the two proteins.

The scientists then used the technique to detect the calcium-dependent association between calmodulin and calmodulin-dependent kinase I (CaMKI). They created various pairings of the new fluorescent protein and CFP fused with either calmodulin or CaMKI, excited the fluorophores with the 458-nm laser and viewed the resulting fluorescence using a Zeiss confocal microscope and 40x, 1.2-NA objective. They found that, in general, the amplitude of the cross-correlation remained low when each sample lacked calcium but increased when calcium was added.

They also tested Keima’s ability to aid multicolor cell imaging. Using a Hamamatsu Photonics CCD camera and Keima, CFP and YFP simultaneously in transfected rat cardiac muscle cells, they observed the concentration of free calcium ions in the cytosol and the mitochondrial morphology within the highly motile cells. Similarly, they used two variants of Keima in conjunction with four other fluorophores to visualize the interactions among the plasma membrane, endoplasmic reticulum, Golgi apparatus, microtubules, mitochondria and nucleus of living cells (Figure 4).

Figure 4. A single 458-nm laser was used to illuminate cells with up to six different fluorophores, including two variants derived from the stony coral Montipora spp. Images courtesy of Atsushi Miyawaki of the Riken Brain Science Institute.

They note that there are chromatic effects related to the protein’s emissions that must be resolved, but suggest that the fluorophore and its variants should solve longstanding problems in multicolor imaging.

When FRET is desired rather than avoided, enhanced cyan and yellow fluorescent proteins have found widespread use as acceptor-donor pairs. They also are commonly employed as complementary fluorophores for dual-color imaging. However, they can be limited by their sensitivity to heat and by their tendency to aggregate and self-quench.

Enhancing the enhanced

Now researchers at the University of Amsterdam in the Netherlands have developed optimized variants of enhanced CFP and YFP. These “superfluorescent proteins” exhibit high brightness compared with their respective original forms. The investigators created the variants by selective mutation starting from Venus, another alternate form of YFP that folds more quickly and efficiently than enhanced YFP, and from Cerulean, a bright variant of enhanced CFP.

In the May 3 ASAP edition of Biochemistry, the scientists report that they tested the brightness of “super” CFP and YFP in samples of E. coli. Using a fluorescence microplate reader made by BioTek Instruments Inc. of Winooski, Vt., they found that E. coli expressing the variant SCFP3A were nine times brighter than bacteria expressing enhanced CFP and 1.2 times brighter than those expressing Cerulean. Furthermore, bacteria that expressed super YFP were 12 times brighter than those that expressed enhanced YFP and about two times brighter than those that expressed Venus.

To test the capabilities of the new variants, the investigators transfected cells with either their YFP or CFP variants and imaged the fluorescence with a CCD camera from Roper Scientific of Tucson, Ariz., connected to a Leica stereomicroscope. They noted that SCFP3A, which has a quantum yield of 0.56, fared less well in live mammalian cells than bacteria but was still 1.5 times brighter than enhanced CFP. Similarly, their YFP variant was 1.5 times brighter than enhanced YFP in mammalian cells.

Because enhanced CFP and enhanced YFP frequently are used together as acceptor-donor pairs in FRET, the researchers also tested the new variants to see whether they would be effective for such use. They used a Zeiss inverted microscope equipped with a frequency-domain fluorescence lifetime imaging detector containing a modulated image intensifier from Lambert Instruments of Leutingewolde, the Netherlands, to determine the fluorescence lifetime of the particles. They excited the super CFP with a helium-cadmium laser from Melles Griot of Carlsbad, Calif., and the super YFP with an argon-ion laser, also from Melles Griot.

Figure 5. Two variants of a “super” cyan fluorescent protein cannot be discriminated via their intensities alone (left, arrows). Because they have different fluorescence lifetimes, however, phase-lifetime images aid discrimination (right). Reprinted with permission of the American Chemical Society.

They found that the proteins paired very well for FRET-based studies because the high quantum yield and extended lifetime of the cyan variant — combined with the high extinction coefficient of the yellow variant — result in a marked difference in the CFP fluorescence lifetime in a FRET and non-FRET situation. They also found that pairing two versions of the cyan variant — both spectrally the same but one with a much lower lifetime than the other — permits dual-color imaging while staying in the cyan emission range (Figure 5).