Illuminated by understandings of brightness

Gary Boas

Förster resonance energy transfer (FRET) can offer tremendous insight into protein interactions inside cells. However, it typically does not provide quantitative information about the protein complex, telling the researcher only that interactions have occurred. Fluorescence correlation spectroscopy offers a possible alternative, but quantitative interpretation with the technique is difficult.

In the Feb. 27 issue of PNAS, investigators with the University of Minnesota, Minneapolis, reported a method based on fluorescence fluctuation spectroscopy that enabled them to measure quantitatively the stoichiometry of protein complexes in living cells. Fluorescence fluctuation spectroscopy is based on fluctuations in fluorescence intensity that result from proteins entering or leaving an observation volume of less than 1 fl.

Researchers have reported a method based on fluorescence fluctuation spectroscopy that provides quantitative measurements of the stoichiometry of protein complexes in living cells. The dual-channel setup shown here splits the fluorescence intensity according to color into different channels and records the intensity ratio, which is used to determine the protein coexpression ratio. A second measurement of the same sample,performed with the single-channel setup, helps to determine the brightness.

The researchers had argued previously that brightness analysis of such fluctuations holds promise for quantification of protein interactions. In the present study, they demonstrated that they could determine the stoichiometry of protein complexes by performing such analysis.

Two proteins

Studies using GFP as a marker have confirmed that the stoichiometry of protein complexes is contained within the brightness. Because the scientists monitored only a single fluorescent color, however, the brightness analysis addressed homocomplex formation only — interactions of two of the same protein. Quantitative analysis of heterocomplexes requires two fluorescent colors. The University of Minnesota researchers developed and reported a method that allows them to identify the stoichiometry of two different proteins — labeled with cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP), respectively — within a protein complex.

The investigators have been working for several years to develop spectroscopy techniques that can quantify protein interactions in living cells. Currently available techniques, such as FRET, either cannot quantify the interactions or acquire quantitative information only indirectly. The brightness analysis reported in the PNAS paper is much more robust and offers a means of determining the composition of homo- and heterocomplexes under stoichiometric binding conditions. The technique works also for protein complexes that show no FRET signal.

To explain the principle underlying the technique, researcher Joachim D. Müller offered an analogy: If we think of a fluorescent protein as a lightbulb, then brightness serves as a measure of how many lightbulbs are in a protein complex. As an example, consider a tetrameric protein complex A3B where three proteins A labeled with CFP and one protein B labeled with YFP come together. The trick to determining the composition of the protein complex is to measure its brightness while selectively turning the CFP and YFP lightbulbs on and off by changing the laser wavelength.

In this table, brightness analysis is illustrated for a tetrameric protein complex A3B, where protein A is labeled with CFP, and protein B carries YFP. Both CFP and YFP are coexcited at 905 nm; as a result, all four fluorescent proteins light up, producing a relative brightness four times that of a single protein. If the excitation wavelength is switched to 820 nm, only CFP is excited, producing a relative brightness of 3. Only CFP is excited when the wavelength is switched to 960 nm, leading to a relative brightness of only 1.

“By coexciting CFP and YFP, the four lightbulbs of the tetramer light up,” he explained. “If you change the excitation wavelength to where only CFP excites, only three lightbulbs light up.” The changes in the measured brightness while the CFP and YFP lightbulbs are turned on and off encode the stoichiometry of the protein complex.

The experimental setup was based on a modified microscope made by Carl Zeiss of Thornwood, N.Y., outfitted with a 63×, 1.4-NA plan apochromat oil-immersion objective. The beam of a mode-locked Ti:sapphire laser made by Spectra-Physics of Mountain View, Calif., pumped by an intracavity-doubled Nd:YVO₄ laser provided excitation.

The researchers located and positioned cells using the epifluorescence mode of the microscope. After selecting a cell, they turned on the Ti:sapphire laser for fluorescence fluctuation spectroscopy measurements. First, they determined the fluorescence intensity ratio of the emitted fluorescence for excitation at 905 nm. They achieved this by splitting the fluorescence into two detection channels.

The fluorescence intensities were then recorded by two avalanche photodiodes (APDs), made by PerkinElmer Optoelectronics of Vaudreuil, Quebec, Canada. The intensity ratio calculated from these measurements contains information about the coexpression ratio of the proteins labeled with CFP and YFP. This is a crucial measurement that is required for the correct interpretation of the subsequent brightness measurement. For this second measurement the scientists redirected the fluorescence emission to another microscope port, in which they recorded the fluorescence intensity with another avalanche photodiode for brightness analysis.

They faced some challenges during the study. Under normal conditions, the brightness could be affected by the presence of FRET. In the current study, though, they were able to mitigate the possible impact of FRET between two different-color fluorophores. “We found conditions where energy transfer does not change the brightness,” Müller said. “Basically, whatever you lose in brightness in protein A you gain in protein B.”

To test the technique, the investigators applied it to looking at nuclear receptor interactions and specifically at their binding to a coactivator. “To our surprise, we found that three nuclear receptors bind to the coactivator,” Müller said. “If you look at the structure of the nuclear receptor coactivator, it actually contains three binding sites, but it was previously assumed that only two are used.” He noted, though, that because they are a physics lab testing a new technique, the scientists looked at only truncated proteins. The findings must be followed up with a study of full-length proteins.

Indeed, the researchers are beginning to look at such proteins. “We are starting to address the biological part of the question,” Müller added. They also are working to refine the technique itself. Currently, determining the stoichiometry of a complex requires more than one measurement. The team hopes to resolve the composition of a complex directly using a more powerful analysis tool; this will require further research, Müller said.