Polarization-dependent fluorescence correlation spectroscopy for studying structural properties of proteins in living cell

Rotational diffusion measurement is predicted as an important method in cell biology because the rotational properties directly reflect molecular interactions and environment in the cell. To prove this concept, polarization-dependent fluorescence correlation spectroscopy (pol-FCS) measurements of purified fluorescent proteins were conducted in viscous solution. With the comparison between the translational and rotational diffusion coefficients obtained from pol-FCS measurements, the hydrodynamic radius of an enhanced green fluorescent protein (EGFP) was estimated as a control measurement. The orientation of oligomer EGFP in living cells was also estimated by pol-FCS and compared with Monte Carlo simulations. The results of this pol-FCS experiment indicate that this method allows an estimation of the molecular orientation using the characteristics of rotational diffusion. Further, it can be applied to analyze the degree of molecular orientation and multimerization or detection of tiny aggregation of aggregate-prone proteins.

Supplementary Note 1, Optical calibration method. To calibrate an optical system for the pol-FCS system, rhodamine 6G (Rh6G), which is widely used as a standard fluorescent molecule for FCS measurements, was measured. The autocorrelation function (ACF) and the CCFs of the Rh6G under X-NN optical condition are shown in Supplementary Fig. 1. In the case of ACF, it was observed a large fraction of after-pulsing in the microsecond region of the correlation functions ( Supplementary Fig. 1b: red arrow) and a decrease of the correlation functions in the nanosecond region ( Supplementary Fig. 1b: green arrow). However, those were not in the CCFs because the after-pulsing, detector dead-time, and specific noise independently occurred in each APD. This result indicates that these fractions caused by the detector noise can be removed from the correlation functions in the range from nanoseconds to microseconds because the detector dead-time and after-pulsing fraction of ACF were canceled via CCFs calculation. Nonlinear curve fitting to CCF without the rotational diffusion component was performed because this experiment aimed to determine the character of the optical system from the shape of the CCF affected by translational diffusion. In this fitting, s was 5 to obtain the reliable value of the parameter.
However, when s was fixed, the standard residuals were tiny. Therefore, s value was believed to be a suitable value for these experiments.
Supplementary Note 2, Estimation method of triplet time. Triplet time was obtained using Eq. S1 and was found to be 1.24 s, which was measured using ConfoCor 3 (Supplementary Fig. 3).
In terms of this quantitative analysis, translational and rotational diffusion times clearly decreased as the concentration of glycerol was increased. Furthermore, the fraction of rotational diffusion was changed. The difference between the results of experiments without glycerol and those using glycerol is significant (among the series with glycerol were observed. It means that some changes are related to the glycerol presence, and the results are independent on the solution viscosity.
Supplementary Note 3, Numerical simulation of the fraction of rotational diffusion in pol-FCS using the Monte Carlo method.
Fractions of rotational diffusion were numerically simulated for EGFP oligomers using the Monte Carlo method. Here, we assumed for simplicity that the transition dipole directions of absorption and emission of EGFPs were the same, and the fluorescence lifetime is much faster than the rotational diffusion time. Furthermore, we assumed that the fluorescence quantum yield of the EGFPs and the quantum efficiency of the photon detector are 1.0 The probability that fluorophore absorbs a photon can be expressed using the normalized transition moment of absorption and emission of the n-th EGFP in the EGFP oligomer 1,2 : , where is the unit vector with the polarization direction of the excitation laser (the polarizer direction). Similarly, the probability that a photon emitted by an excited EGFP can be detected is as follows: , where is the unit vector with the polarization direction of the detected fluorescent light (the analyzer direction). In that case, the detected fluorescence intensity using the detector is given by (S4) where is the intensity of the excitation laser at the EGFPs position.
and are the rotation angles, and those were randomly defined in the range of 0 to 2π. An EGFP oligomer rotates maintaining the orientation among all the EGFPs included by the EGFP oligomer.
The auto-correlation function of the fluorescence intensity signal is given by: where , . (S10) In the numerical simulations, initial orientations of each EGFP included in the EGFP oligomer were randomly defined. Next, the fluorescence intensity signal was generated using Eqs. S5 and S6. Finally, the fraction of rotational diffusion was obtained using Eq. S8.