Single Molecule Spectroscopy of the Green Fluorescent Protein

Abstract

The Green Fluorescent Protein (GFP) and its mutants have received considerable attention as the only genetically encoded proteins that fluoresce without the addition of external cofactors. They currently find widespread application in molecular biology as highly specific labels expressable in eukaryotic as well as in prokaryotic cells. Mutants with different spectral properties are commercially available. Parallel to the introduction of the GFP as a label, optical single molecule detection techniques at room temperature became available. The unique properties of the GFP also raised interest in its application as a single molecule fluorophore. Two classes of single molecule experiments with the GFP can be distinguished. While in the first, the main focus is laid on using single molecule spectroscopy in order to investigate the photophysics of the GFP, in the second class it finds use as a marker for other proteins. A review of optical single molecule experiments with the GFP has recently been published [1]. Successful single molecule detection has been demonstrated using confocal microscopy, widefield imaging, and near field microscopy [2-4]. Already the very first experiments with single GFP molecules immobilized in aqueous gels, showed that it is a comparatively instable fluorophore that photobleaches after emitting approximately 105 photons [5]. Even worse for single molecule applications proved the fact that its fluorescence emission is interrupted by frequent excursions to a multitude of dark states. The dark state population leads to intermittencies in the fluorescence emission on time scales from ms to hours [2]. Temporal trajectories of single molecule fluorescence have long been used to deduce the photodynamics of transitions into and out of dark states. This analysis however necessitates sufficiently long trajectories with only one or two dark states present. Due the GFP's low photostability, it was so far impossible to derive meaningful photokinetic data of the GFP by using this technique. One has to conclude that GFP is a far worse single molecule fluorophore than it would be expected from extrapolating results of ensemble measurements. Instead of using trajectories of one individual molecule, data of many subsequent single molecule measurements can be averaged in order to obtain better statistics. If the samples are homogeneous, this will lead to the same results as the method described above. Fluorescence correlation spectroscopy (FCS) of fluorophores diluted in an appropriate solvent can be employed in this case. The molecules now diffuse through the focus of a confocal microscope. If the dilution is chosen so that only one molecule at a time is present in the focal volume, then fluorescence bursts of molecules diffusing through the focal volume are observed. From these bursts, autocorrelation functions of the temporal distribution of the fluorescence emission can be constructed. The autocorrelation curves are then used to derive information about the internal photodynamics of the sample molecules. It is obvious that the dynamic range accessible with FCS is limited by the diffusion time of the molecules of ∼ 1 ms for the longer times and by the detector response time of ∼ 100 ns for the short times. It turns out that this limitation can be advantageous in investigations of the GFP as it suppresses contributions from dark state dynamics outside this time window. window. FCS has successfully been used to study the influence of external parameters such as pH and buffer strength on the photodynamics of various mutants of the GFP [6, 7]. While the photostability of the GFP is of no concern in these experiments, also here a detailed analysis is often hampered by the fact that too many dark states are contributing to the photodynamics in most mutants. It has however been shown that simultaneous illumination of some variants with a weak second excitation beam can be used to almost completely depopulate a dark state. In all cases known to date this dark state contains a protonated chromophore, while the ground state of the variant is stabilized on the anionic form of the chromophore. Under two color illumination, these molecules exhibit a significantly more intense fluorescence emission. Two color FCS can be used for a detailed analysis of the photodynamics also of GFP molecules which in the accessible time window exhibit the photodynamics of a four level systems, i.e. containing two dark states [8,9]. The knowledge acquired from the FCS experiments and simultaneous two color excitation can be employed to boost the fluorescence signal in single molecule imaging experiments of some GFP mutants [10]. While this experimental technique can be used for certain systems, the application of the GFP in single molecule imaging applications remains difficult. This would only change with the appearance of mutants, which are more photostable and better stabilized on one emissive state. Until then, FCS seems to be the single molecule technique best suited for investigations with the GFP as a fluorophore as FCS does not suffer from the poor photostability of the GFP. A helpful comparison of the suitability of several common dyes and several GFP mutants for single molecule experiments has recently been published [11].