Biomedical Engineering Reference
In-Depth Information
Olivier et al. developed algorithms to automatically trace the entire cell differentiation process through
the successive cleavage stages in live zebra fish embryos. The experiments followed the embryos up to
the 10th cell division cycle, where there are ~1000 cells within an embryo. The detailed cell lineage tree
and location atlas should be invaluable for understanding how internal and external factors combine to
influence each cell, thereby orchestrating the complex steps of embryogenesis.
7.4 technical notes for SHG imaging of Microtubules
Since the SHG intensity from microtubules is weak, a number of practical details should be considered
for improving the signal-to-noise ratio of the observed signal. To optimize detection, photon counting
should be used instead of simple integration of the photocurrent of the photomultiplier tubes. Photon
counting will exclude the dark current of the photomultiplier tube when calculating pixel intensity, and
therefore accumulates less noise for long image acquisition. For example, an SHG image of microtubules
in acute mouse brain slice requires an acquisition time of ~20 s (Figure 7.6). In addition, narrowband
emission filters should exactly match the narrowband SHG emission; so, contaminating signals such
as broadband autofluorescence are minimized (Figure 7.5b). Similarly, a compromise should be made
between using short excitation wavelengths, which generate higher SHG intensity from microtubules
(Figure 7.5a), versus long wavelengths, which reduce photodamage. Theoretically, SHG is a scattering
process that involves a virtual state; so, photodamage is nonexistent. However, in practice, photodamage
can result from exciting other endogenous photon-absorbing molecules. The mechanisms for photodam-
age are often not known and should be calibrated for the specimen under study (Sacconi et al., 2006).
Fluorescence can overwhelm the weak SHG signal from microtubules; so, it is essential to check for
the signatures of SHG in every experiment. A few of the following factors will be sufficient to verify that
SHG is responsible for the measured intensity: a squared dependence on excitation intensity, dependence
on excitation polarization, tuning the excitation wavelength so that the emission wavelength shifts out-
side the narrowband emission filter range, and angular dependence of the SHG emission. So far, it has
been demonstrated by numerous studies that the SHG signals in mitotic spindles and in neurons origi-
nate from microtubules, but for other specimens, the identity of the SHG source needs to be verified.
The SHG emission has angular dependence. For microtubule bundles, the forward emission can be
~5-100 times stronger than the backward emission. Efficient collection of the forward emission requires
using a high-numerical aperture objective lens in the forward pathway that matches the field of view of
the objective lens used to focus the excitation beam. Moreover, collecting in the forward pathway from
thin samples such as cultured neurons is easier and should be attempted first. For imaging acute brain
tissue, there is a compromise between thick slices, where the forward emission would be more scattered,
versus thin slices, where the microtubule integrity may be disrupted. SHG signals from microtubules
are abolished from fixed or dying neurons.
The SHG intensity is also dependent on the polarization of the excitation laser source relative to the
molecular orientation of the tubulins. The optimal condition for SHG occurs when the laser polariza-
tion is parallel to the long axis of the microtubules (Figure 7.8). This alignment is possible over the entire
field of view when it is anatomically known that all microtubules orient similarly, for example, in the
apical dendrites in area CA1 of acute hippocampal slices (Figure 7.6h). Alternately, using circular polar-
ization or linear polarization with a constantly rotating sample will ensure that all microtubules within
the field of view are equally excited.
Finally, from the SHG signal of microtubules, a couple of studies have measured a hyperpolariz-
ability angle that matches well with the physical dimensions of tubulin (Psilodimitrakopoulos et al.,
2009, Odin et al., 2009), implicating tubulin as the basic unit that scatters and participates in SHG. For
a given spatial concentration of tubulins, it is possible to calculate the expected intensity and angular
distribution of the SHG signal (Moreaux et al., 2000, Williams et al., 2005). For a bundle of microtu-
bules, numerical simulation using typical values for axon and dendrite diameters and intermicrotubule
spacing had been carried out (Figure 7.9). The simulation results showed that in contrast to membrane
