Biology Reference
In-Depth Information
5
m
M
10
m
M
15
m
M
50
m
m
FIGURE 6.1
Representative confocal fluorescence images of entangled MT networks. Using our assembly
protocol, the entangled MT networks ranging from 5 to 50
m
M are isotropic and
homogeneous.
Adapted from
Yang et al. (2012)
. Reproduced by permission of The Royal Society of Chemistry.
black. Based on our typical pinhole diameter of
100 nm (equivalent to 1 Airy unit)
and objective lens, we estimate our
z
-resolution to be
m
1
m.
6.1.2.2
Dynamic network formation determined by confocal microscopy
Time-lapsed confocal fluorescence imaging allows the initial stages of network
formation to be observed. To achieve this, we use an upright Fluoview 1000 laser
scanning system (Olympus) with an environmental chamber to control the sample
temperature. In this case, tubulin proteins are kept ice-cold and loaded into the
capillary tubes just prior to imaging. There is no separate incubation step; rather,
the environmental chamber is maintained at 30-35
C, and polymerization is initi-
ated
in situ
and the resulting network growth observed. The lapsed time between the
placement of the sample tube on the microscope and the first recorded frame is typ-
ically 2-3 min. Since the sample volume is small, temperature equilibration is fast,
and we assume that the assembly of the network starts immediately upon placing the
tube inside the environmental control box. In detail, we use 559 nm laser excitation
through a 25
, NA 1.05 water-immersion objective, typically with scan size of
1024 pixels
2
, scan rate of 10
1024
125 nm/pixel.
We find a waiting time between each frame of 15-30 s, and total recording time of
m
s/pixel, and magnification of
10-20 min is sufficient to capture the dynamics of network formation.
6.1.2.3
Analysis of network mesh size
The average mesh size of the MT network can be determined through analysis of
two-dimensional confocal microscopy slices (
Yang et al., 2012
). First, the raw im-
ages are converted to binary images by thresholding to suppress background fluores-
cence and pixel noise while retaining the gross structural features of the network.
Threshold cutoff values are determined by visually comparing the intensities of
the brightest pixels of the background to those of the dimmest pixels on the MTs.
After thresholding, the distance between nearest-neighbor MT pixels within each
row and column is determined. The result is similar to the radial distribution of
