20 ms, and ideally within 6–12 ms, were suggested to be
appropriate for determining diffusion coefficients of monomeric GFP [29, 37], and approaches to measure actual
bleached area have also been suggested [29].
8. To achieve sufficient temporal resolution during the acquisition
of fluorescence recovery, it is necessary to record the recovery
at intervals which are smaller than the half-time of the fastest
recovery component. Although as fast an acquisition as 1/10
of the half-time is recommended in some publications [28], in
our experience, for recording E-cad-GFP, 20 s intervals are
sufficient to detect a diffusional component with the half-time
of about 25 s. For calibration, the aim is to achieve sufficient
temporal resolution while maintaining maximal possible spatial
resolution.
9. The photobleaching during the acquisition of the fluorescence
recovery should be minimized as much as possible to obtain the
most accurate information about the dynamics [40]: deviation
by more than 10–15% is usually suggestive of excessive acquisitional bleaching [28].
10. It has been suggested that the duration of recovery acquisition
should be 7–10 times longer than the characteristic half-time
of the slowest detected component [28]. In our experience,
three times longer acquisition is sufficient for a reliable estimation of both half-times and maximum recovery [15]. Furthermore, theoretically the signal ought to completely recover with
a given time. For E-cad in the Drosophila this has been calculated to be a period of 2 h, through recovery of the “immobile”
fraction by slow E-cad degradation and the addition of newly
synthesized protein [41].
11. We use average intensity projections instead of maximum projections, as it considers not only protein concentration (intensity), but also distribution along the z-axis, i.e., junction width,
and is reflective of the total protein amount. For example, a
junction with the same E-cad density, but shorter in the z-axis
will result in the same intensity when a maximum projection is
used as a control and might skew the interpretation of the
results.
12. There are several plugins which allow one to perform registration of a time series, i.e., compensation for tissue movement
over time, such as “Register Virtual Stack Slices” in ImageJ.
However, even when using such scripts, we recommend manually tracking or checking to control for ROI position when
measuring intensities. For example, changes in the membrane
curvature are observed in epithelial cells, which might shift the
ROI position even if the cell position is correctly registered.
156
Joshua Greig and Natalia A. Bulgakova
appropriate for determining diffusion coefficients of monomeric GFP [29, 37], and approaches to measure actual
bleached area have also been suggested [29].
8. To achieve sufficient temporal resolution during the acquisition
of fluorescence recovery, it is necessary to record the recovery
at intervals which are smaller than the half-time of the fastest
recovery component. Although as fast an acquisition as 1/10
of the half-time is recommended in some publications [28], in
our experience, for recording E-cad-GFP, 20 s intervals are
sufficient to detect a diffusional component with the half-time
of about 25 s. For calibration, the aim is to achieve sufficient
temporal resolution while maintaining maximal possible spatial
resolution.
9. The photobleaching during the acquisition of the fluorescence
recovery should be minimized as much as possible to obtain the
most accurate information about the dynamics [40]: deviation
by more than 10–15% is usually suggestive of excessive acquisitional bleaching [28].
10. It has been suggested that the duration of recovery acquisition
should be 7–10 times longer than the characteristic half-time
of the slowest detected component [28]. In our experience,
three times longer acquisition is sufficient for a reliable estimation of both half-times and maximum recovery [15]. Furthermore, theoretically the signal ought to completely recover with
a given time. For E-cad in the Drosophila this has been calculated to be a period of 2 h, through recovery of the “immobile”
fraction by slow E-cad degradation and the addition of newly
synthesized protein [41].
11. We use average intensity projections instead of maximum projections, as it considers not only protein concentration (intensity), but also distribution along the z-axis, i.e., junction width,
and is reflective of the total protein amount. For example, a
junction with the same E-cad density, but shorter in the z-axis
will result in the same intensity when a maximum projection is
used as a control and might skew the interpretation of the
results.
12. There are several plugins which allow one to perform registration of a time series, i.e., compensation for tissue movement
over time, such as “Register Virtual Stack Slices” in ImageJ.
However, even when using such scripts, we recommend manually tracking or checking to control for ROI position when
measuring intensities. For example, changes in the membrane
curvature are observed in epithelial cells, which might shift the
ROI position even if the cell position is correctly registered.
156
Joshua Greig and Natalia A. Bulgakova