For live cell imaging, use an inverted epifluorescence microscope with an incubation system, such as a stage-top incubator, to maintain the temperature of the cells at 37°C (see Note 5).

Use a high magnification and high numerical aperture objective lens, such as PlanApo N TIRF (60×, NA  =  1.45) oil immersion objective, to identify single spots in the cytoplasm and suitable to ensure the collection of faint fluorescence signals from single EGFP molecules.

Illuminate the specimen with blue laser power of a few μW/μm² at the sample surface. Generally, the intensity distribution of the laser beam is approximated by a Gaussian function, which will result in different efficiencies of excitation by inhomogeneous illumination. Therefore, the beam line should be expanded by a beam expander, and the intensity data should be acquired within a limited area of irradiation around the center to avoid too large a variation in the irradiation intensity (for example, within 10% variation).

Place a dichroic mirror (Q505LP) and emission filter (FF01-520/35-25) combination optimal for collecting EGFP fluorescence emission in the microscope (see Note 6). An excitation filter is not required, because the introduction of an excitation filter in the laser line may cause uneven illumination owing to the interference fringe.

To observe the quantized photobleaching of single EGFP molecules, acquire images at 20 frames per second, or faster, using a cooled electron-multiplying charge-coupled device (EM-CCD) camera (a cooled EM-CCD camera is best suited for quantification of the fluorescence intensity of single EGFP molecules). The electron-multiplying gain register can increase the signal more than several hundredfold before any readout noise is added, therefore keeping the noise level low. Furthermore, a cooled EM-CCD camera incorporating back illuminated type CCD chips is recommended for its high quantum efficiency (see Note 7).

Load purified EGFP at the concentration of ng/ml level in a saline solution such as PBS onto a pre-cleaned glass coverslip to be absorbed nonspecifically (see Note 8). Excess free EGFP is then washed with the saline solution prior to observation.

Acquire series of images of single EGFP molecules to measure changes in fluorescent intensity before and after quantized photobleaching. We typically monitor the intensity for at least 10 s at video rate under conditions where single EGFP molecules photobleach in a few seconds on average.

Measure the fluorescence intensities with select regions within the image series using image analysis software (see Note 9). The region of interest (ROI) for specifying a fluorescence molecule is made by the circle with the smallest diameter to cover the Airy disk. The ROI for calculating the background fluorescence is made by a slightly larger diameter circle. The margin between the smaller ROI and the larger ROI, which forms a concentric ring that surrounds the Airy disk, is used to calculate the background fluorescence intensity.

With the transfected CEFs cells on the microscope, acquire single image snapshots under the same optics and imaging conditions (optics setup, camera frame rate) as in the quantification of the intensities of single EGFP molecules (see Note 10).

To analyze the fluorescence intensities of mRNA molecules in the transfected CEFs, draw ROIs the same size as those used in the quantification of the intensities of single EGFP molecules. To extract the position of bright spots in the cytoplasm of CEFs, dynamic thresholding is an effective approach and is described below.

Place in the laser path a field stop that corresponds to a tracking area (15 μm in diameter) at the specimen plane, so that EGFP molecules outside of the tracking area are not illuminated.

Capture images of 10.6  ×  10.6 μm rectangle field in the tracking area by an EM-CCD camera every 8 ms with intense laser illumination (a few μW/μm² at the sample surface) because β-actin mRNA molecules at the leading edge of CEFs are present at a high density and they move rapidly. A large number of fluorescent spots are photobleached by the intense laser and the remaining spots decrease in density such that they can be distinguished as discrete spots. Track the remaining fluorescent spots and newly appearing fluorescent spots from outside of the illumination area.

Background smoothing of noisy images: Several spatial or time averaging methods are commonly used. Spatial averaging methods, which are based on a spatial low-pass filter including a mean filter, a median filter and a Gaussian filter, remove noise as well as details on the targets to give blurred images. The intensity of target spots is sometimes lost by spatial averaging and a faint spot cannot be distinguished. On the other hand, time averaging, also called moving average or rolling average, averages a series of a certain number of time data to smooth short-term fluctuations such as background noises. Time averaging is likely to have little effect on the intensities of the target spots, because relatively constant values of pixels against time are less affected by the time averaging process. However, time averaging intrinsically involves inevitable effects, especially on the observation of diffusing molecules that are permanently or temporarily confined within some compartments. The effect of time averaging upon changing the exposure time and frame rate has been reported (16). The effect of time averaging is that the position of a molecule appears to be more centralized to the compartments especially when the averaging time is comparable to, or longer than, the time for the target molecules to reach the compartment boundary. In order to reduce such an effect, we process the images using a four-frame exponential moving average with a 2/3 smoothing factor and end this process with spatial averaging of neighboring 3  ×  3 pixels. All the averaging processes are performed with ImageJ.