borohydride [44, 45]. In our experience, the majority of antibodies used in AT do
not appear to be sensitive to glutaraldehyde, however, if such an effect is observed,
glutaraldehyde concentration used for fixation can be decreased down to 0.1%.
The fixed tissue is then cut into 100–200 lm slices using a Vibratome, and the
region of interest is dissected out. To prevent the formation of ice crystals during
low temperature processing, the tissue is cryoprotected with increasing concentrations of glycerol. It is then quick-frozen in a dry ice/ethanol bath and dehydration is
performed through freeze-substitution using methanol at −90 °C. Under these
milder conditions of dehydration, the stabilization of lipids and other tissue elements can be achieved with uranyl acetate, which is added to the methanol during
the dehydration steps. Unlike osmium, uranyl acetate does not have major impact
on the tertiary structure of proteins [46–48]. After the dehydration step, the sample
is slowly warmed up to −45 °C and infiltrated with the resin Lowicryl HM20.
Finally, the resin is polymerized using UV light at 0 °C. The polymerized block
with tissue can be stored for many years before use.
To prepare serial sections, standard electron microscopy procedures are followed. The block is trimmed to a trapezoid block face shape and then sectioned
with an ultramicrotome using a diamond knife. Typically 70 nm sections are collected, but depending on the needs, thickness can vary between 50 and 200 nm.
Thinner sections provide the advantage of better resolution at the EM level, but
have the drawbacks of lower contrast and increased effort for large volume
reconstructions. Glue is applied to the sides of the block pyramid to ensure that the
sections stick to each other and form a ribbon, which is then picked on a
coverslip. Coverslips are coated with carbon, which greatly increases the adherence
of the sections and also provides a conductive substrate required for the EM
imaging. Conventional indirect immunostaining is used, usually with three different
primary antibodies (Table 6.1) from different host species applied simultaneously,
followed by the corresponding secondary antibodies conjugated to fluorophores
with distinct excitation/emission spectra. While primary antibodies directly conjugated to fluorophores can also be used, the use of secondary antibodies provides
considerable signal amplification, resulting in a much stronger and more robust
immunofluorescence label. The coverslips are mounted on glass slides or
custom-made chambers using non-hardening mounting medium that also contains
the nuclear stain DAPI.
The immunolabeled samples are imaged using an automated epifluorescence
microscope. No additional advantage is offered by using a confocal microscope,
because there is no out of focus signal coming from the ultrathin sections used for
AT. The final resolution of the immunofluorescent images is at the theoretical limit
(*200 nm lateral resolution; z-resolution determined by section thickness) and can
be further improved to *100 nm lateral resolution using deconvolution [11].
Alternatively, to achieve 100 nm lateral resolution, the samples can be imaged with
structured illumination microscopy [38, 39]. The high resolution of
immunofluorescent images facilitates the registration of the light and electron
microscope images. Image acquisition begins with mapping of the ribbon of serial
sections. This is done at a low magnification, typically with a 10 Â objective, and a
156
K. D. Micheva and K. D. Phend
not appear to be sensitive to glutaraldehyde, however, if such an effect is observed,
glutaraldehyde concentration used for fixation can be decreased down to 0.1%.
The fixed tissue is then cut into 100–200 lm slices using a Vibratome, and the
region of interest is dissected out. To prevent the formation of ice crystals during
low temperature processing, the tissue is cryoprotected with increasing concentrations of glycerol. It is then quick-frozen in a dry ice/ethanol bath and dehydration is
performed through freeze-substitution using methanol at −90 °C. Under these
milder conditions of dehydration, the stabilization of lipids and other tissue elements can be achieved with uranyl acetate, which is added to the methanol during
the dehydration steps. Unlike osmium, uranyl acetate does not have major impact
on the tertiary structure of proteins [46–48]. After the dehydration step, the sample
is slowly warmed up to −45 °C and infiltrated with the resin Lowicryl HM20.
Finally, the resin is polymerized using UV light at 0 °C. The polymerized block
with tissue can be stored for many years before use.
To prepare serial sections, standard electron microscopy procedures are followed. The block is trimmed to a trapezoid block face shape and then sectioned
with an ultramicrotome using a diamond knife. Typically 70 nm sections are collected, but depending on the needs, thickness can vary between 50 and 200 nm.
Thinner sections provide the advantage of better resolution at the EM level, but
have the drawbacks of lower contrast and increased effort for large volume
reconstructions. Glue is applied to the sides of the block pyramid to ensure that the
sections stick to each other and form a ribbon, which is then picked on a
coverslip. Coverslips are coated with carbon, which greatly increases the adherence
of the sections and also provides a conductive substrate required for the EM
imaging. Conventional indirect immunostaining is used, usually with three different
primary antibodies (Table 6.1) from different host species applied simultaneously,
followed by the corresponding secondary antibodies conjugated to fluorophores
with distinct excitation/emission spectra. While primary antibodies directly conjugated to fluorophores can also be used, the use of secondary antibodies provides
considerable signal amplification, resulting in a much stronger and more robust
immunofluorescence label. The coverslips are mounted on glass slides or
custom-made chambers using non-hardening mounting medium that also contains
the nuclear stain DAPI.
The immunolabeled samples are imaged using an automated epifluorescence
microscope. No additional advantage is offered by using a confocal microscope,
because there is no out of focus signal coming from the ultrathin sections used for
AT. The final resolution of the immunofluorescent images is at the theoretical limit
(*200 nm lateral resolution; z-resolution determined by section thickness) and can
be further improved to *100 nm lateral resolution using deconvolution [11].
Alternatively, to achieve 100 nm lateral resolution, the samples can be imaged with
structured illumination microscopy [38, 39]. The high resolution of
immunofluorescent images facilitates the registration of the light and electron
microscope images. Image acquisition begins with mapping of the ribbon of serial
sections. This is done at a low magnification, typically with a 10 Â objective, and a
156
K. D. Micheva and K. D. Phend