Dependence of fluorescent protein brightness on protein concentration in solution and enhancement of it

Fluorescent proteins have been widely used in biology because of their compatibility and varied applications in living specimens. Fluorescent proteins are often undesirably sensitive to intracellular conditions such as pH and ion concentration, generating considerable issues at times. However, harnessing these intrinsic sensitivities can help develop functional probes. In this study, we found that the fluorescence of yellow fluorescent protein (YFP) depends on the protein concentration in the solution and that this dependence can be enhanced by adding a glycine residue in to the YFP; we applied this finding to construct an intracellular protein-crowding sensor. A Förster resonance energy transfer (FRET) pair, involving a cyan fluorescent protein (CFP) insensitive to protein concentration and a glycine-inserted YFP, works as a genetically encoded probe to evaluate intracellular crowding. By measuring the fluorescence of the present FRET probe, we were able to detect dynamic changes in protein crowding in living cells.

For CFP-longliker-YFP1G construction, we designed the sense primers containing the sequence encoding the long-linker (GGSGGT×6) and C-terminal sequence of CFP containing the BspEI site and reverse primers containing the BspEI site. The primers were annealed and ligated into a pAL7 vector encoding GimRET between the BspEI/BspEI sites and then transformed into Rosseta2 (DE3). Fluorescence spectra of GimRET in the absence (black) or presence of 10 mg/mL tubulin (red) or microtubules (blue).
(d) Fluorescence spectra of GimRET in the absence (black) or presence of 9.7 mg/mL actin monomer (G-actin, red) or actin filament (F-actin, blue). (e) Fluorescence spectra of GimRET in 0 mg/mL (black) and 15 mg/mL (red) DNA solutions. (f) Fluorescence spectra of GimRET at 0 mM (black), 100 mM (red), 200 mM (blue), and 300 mM (green) KH 2 PO 4 . All spectra were obtained at a 440 nm excitation wavelength, and represent the average of four trials.
Lysozyme chloride was dissolved into 100 mM HEPES (pH 7.4), and then the solution was dialyzed to remove the chloride ion. Actin was purified from rabbit skeletal muscle (1). To make actin filaments, we added 200 mM KCl and 10 µM phalloidin into the buffer, 2 mM HEPES (pH 7.8), 0.2 mM ATP, 0.1 mM CaCl 2 , and 1 mM 2mercaptoethanol. After 24 h, the actin filaments were collected by centrifuging to remove the chloride ions. Tubulin was purified from bovine brain (2). To make microtubule, we increased the temperature of the solution to 37 °C for 15 min, and then added 1 mM taxol into the buffer, 100 mM PIPES (pH 6.9), 1 mM MgCl 2 , 1 mM EGTA, 1 mM GTP. To obtain the E. coli lysate, the cell pellet was mixed with 100 mM HEPES-NaOH (pH 7.4) until the weight of the wet pellet and buffer was the same, and sonicated for 30 s 20 times, resulting in E. coli solution (lysate:buffer = 1:1). The resultant protein concentration was 160 mg/mL. The E. coli solution was diluted in 100 mM HEPES-NaOH (pH 7.4) at a lysate:buffer ratio of 1:4 and 1:8. GimRET was diluted to 0.1 mg/mL in each E. coli lysate solution. 1 Spudich, J. A., & Watt, S. J Biol Chem. 246, 4866-4871 (1971) 2 Murofushi, H., et al. J Cell Biol. 103, 1911-1919(1986  (a-c) Dependence of GimRET ratio and diffusion coefficient on various solution conditions. GimRET was diluted to 1 μg/mL in 150 mg/mL, 200 mg/mL, or 250 mg/mL BSA (a), 7.5%, 10%, or 15% (v/v) PEG6000 (b), and 20%, 30%, or 40% (v/v) sucrose (c). BSA, PEG6000, and sucrose were dissolved in 100 mM HEPES-NaOH (pH 7.4). Red, intensity ratio of CFP and YFP1G; blue, corresponding diffusion coefficients obtained by FRAP. Plots represent the average of 12 trials. Error bars, standard deviation. (d-f) Correlations of the diffusion coefficient obtained by FRAP and the intensity ratio of GimRET in BSA (d), PEG6000 (e), and sucrose (f). Data values are from a, b, and c, respectively.
Error bars, standard deviation. Single and double asterisks correspond to P value is respectively < 0.05, and <0.01 in two sample t-test.    to P value is respectively < 0.01, and <0.001 in two sample t-test.
Cells expressing GimRET, CFP-YFP or stained with SNARF were incubated in 20 mM PBS (pH 7.4) for 10 min before observation, and the cells were observed under the fluorescent microscope. After the observation, we changed the medium from PBS at pH 7.4 to PBS at pH 8.5 or pH 6.5 on the microscope to change the extracellular pH.
After 30 min, we observed the same cells. Because intracellular pH was stable, we observed the cells 30 min after change of the medium. The observation was performed by using 60× objective lens. The fluorescent images were averaged 5 times. The data analysis was performed using homemade software.  asterisks correspond to P value is respectively < 0.01, and <0.001 in two sample t-test.
The procedure of the osmotic shock assay for E. coli by glycerol was performed following the previous report (4). The dehydrated osmotic pressure was calculated using the Norrish equation ( Colour bar indicates the ratio from 1.0 (black) to 3.0 (white).

Figure S16. Comparison of ratiometric images of SNARF-1 and GimRET in Neuro 2A cells.
Images of the total fluorescent intensity (left panels) and the intensity ratio (right panels) of Neuro 2A cell labelled with SNARF-1 (a) or expressing GimRET (b). The ratio value for SNARF-1 is the ratio of 640-700 nm and 500-560 nm, and that of GimRET is the ratio of 460-500 nm and 520-560 nm. Colour bars indicate the ratio from 1.0 (black) to 3.0 (white).  trials. The solid lines fits to a single exponential decay curve, f(t) = a·exp(-t/τ) or a·exp(t/τ). Error bars, standard deviations. Because photobleaching of the YFP1G channel originated from both photobleaching of YFP and CFP, the decay time of the YFP1G fluorescence was smaller than that of CFP. Differences between the decay times of CFP and YFP1G increased the intensity ratio. In our experimental setup, the time constant of the intensity ratio increase was 4680 h, which is negligibly longer than the observation period.  and YFP1G (green) in GimRET and the intensity ratio (red). All traces represent the average of four individual trials.
Error bars, standard deviations.
The procedure of the Measurement of folding and maturation of fluorescent protein in E. coli was performed following the previous report (7). E. coli transformed by the plasmid of each fluorescent protein were cultured for 12 h in LB medium. E. coli was collected by centrifuging and dissolved in M9 medium. From just after the induction by 1 µM IPTG, the fluorescent intensity of the E. coli were measured using Multilabel Reader 2030 ARVO X3 (Perkin Elmer, MA) at 37°C for 6 h. The excitation wavelength was set to 430 nm for CFP, and 510 nm for YFP1G and YFP.
The emission wavelength was set to 475-485 nm for CFP, and 535-545 nm for YFP1G and YFP.