allowed the imaging and full reconstruction of a retinal circular segment with a
diameter of 0.22 mm and approximate thickness of 0.03 mm [16]. Ongoing
exploration of the acquired retinal volume uncovers much greater complexity of the
retinal synaptic network than previously recognized, and has demonstrated the
existence of a number of new connection motifs and functions, as well as new
contact architectures that challenge the classical ultrastructural definition of a
chemical synapse [23].
A similar method based on conventional electron microscopy tissue preparation
is serial multiplex immunogold labeling [20]. In this case, the tissue is also prepared
using standard electron microscopy methods, embedded in plastic, and sectioned
into long series of ultrathin sections. Small subsets of sections are labelled with
antibodies to different neuropeptides using the immunogold technique where the
secondary antibodies are conjugated to gold particles that can be conveniently
observed under the electron microscope. As the case of small molecules like GABA
and other neurotransmitters, the antigenicity of small neuropeptides withstands the
tissue processing involved in EM sample preparation [24, 25]. The use of electron
microscopy to image the immunolabel avoids potential problems of registration of
light level and EM datasets. This approach was applied to serial sections spanning
the entire body of the larva of the marine annelid Platynereis dumerilii. Eleven
antibodies against neuropeptides were used and several molecularly identified
neurons were fully reconstructed and their synaptic partners were identified,
revealing a candidate chemotactic circuit in the larval head.
Unfortunately, the majority of antibodies cannot be used on conventionally
prepared EM tissue. Many of the chemicals and treatments traditionally used to
preserve good ultrastructure are detrimental to most antigens. The most notorious
culprit is osmium tetroxide, which is used to protect the tissue from extraction
during dehydration, as it binds to lipids and proteins. However, osmium also disrupts the tertiary structure of most proteins [26] preventing the binding of antibodies. Epoxy resins, which are typically the resin of choice for EM are also
associated with decreased antigenicity and decreased access of antibodies, as they
are known to covalently bind to the tissue components [27, 28]. Acrylic resins are
preferred for immunolabeling studies [29], but they generally have lower stability in
the electron beam and, especially, when osmium is omitted, can result in poor
ultrastructural preservation. The cumulative effects of the treatments at different
steps of tissue processing often result in an inverse relationship between the
preservation of ultrastructure and antigenicity. As seen in Fig. 6.1, postfixation with
osmium followed by Epon embedding provides the best ultrastructure and the worst
immunoreactivity for the synaptic proteins synapsin and VGluT1, while omitting
osmium and embedding in LRWhite gives the opposite results. Introducing a low
concentration of osmium to the LRWhite embedding protocol only partially
improves tissue integrity while reducing immunoreactivity.
One solution to this problem is offered by low-temperature methods for tissue
processing. Sample dehydration, a necessary step preceding resin embedding, is much
less damaging to tissue integrity at very low temperatures (<−70 °C), where lipid
extraction by organic solvents is minimized [30]. Under these conditions, osmium is no
6 Conjugate Immunofluorescence—SEM Array Tomography …
151
diameter of 0.22 mm and approximate thickness of 0.03 mm [16]. Ongoing
exploration of the acquired retinal volume uncovers much greater complexity of the
retinal synaptic network than previously recognized, and has demonstrated the
existence of a number of new connection motifs and functions, as well as new
contact architectures that challenge the classical ultrastructural definition of a
chemical synapse [23].
A similar method based on conventional electron microscopy tissue preparation
is serial multiplex immunogold labeling [20]. In this case, the tissue is also prepared
using standard electron microscopy methods, embedded in plastic, and sectioned
into long series of ultrathin sections. Small subsets of sections are labelled with
antibodies to different neuropeptides using the immunogold technique where the
secondary antibodies are conjugated to gold particles that can be conveniently
observed under the electron microscope. As the case of small molecules like GABA
and other neurotransmitters, the antigenicity of small neuropeptides withstands the
tissue processing involved in EM sample preparation [24, 25]. The use of electron
microscopy to image the immunolabel avoids potential problems of registration of
light level and EM datasets. This approach was applied to serial sections spanning
the entire body of the larva of the marine annelid Platynereis dumerilii. Eleven
antibodies against neuropeptides were used and several molecularly identified
neurons were fully reconstructed and their synaptic partners were identified,
revealing a candidate chemotactic circuit in the larval head.
Unfortunately, the majority of antibodies cannot be used on conventionally
prepared EM tissue. Many of the chemicals and treatments traditionally used to
preserve good ultrastructure are detrimental to most antigens. The most notorious
culprit is osmium tetroxide, which is used to protect the tissue from extraction
during dehydration, as it binds to lipids and proteins. However, osmium also disrupts the tertiary structure of most proteins [26] preventing the binding of antibodies. Epoxy resins, which are typically the resin of choice for EM are also
associated with decreased antigenicity and decreased access of antibodies, as they
are known to covalently bind to the tissue components [27, 28]. Acrylic resins are
preferred for immunolabeling studies [29], but they generally have lower stability in
the electron beam and, especially, when osmium is omitted, can result in poor
ultrastructural preservation. The cumulative effects of the treatments at different
steps of tissue processing often result in an inverse relationship between the
preservation of ultrastructure and antigenicity. As seen in Fig. 6.1, postfixation with
osmium followed by Epon embedding provides the best ultrastructure and the worst
immunoreactivity for the synaptic proteins synapsin and VGluT1, while omitting
osmium and embedding in LRWhite gives the opposite results. Introducing a low
concentration of osmium to the LRWhite embedding protocol only partially
improves tissue integrity while reducing immunoreactivity.
One solution to this problem is offered by low-temperature methods for tissue
processing. Sample dehydration, a necessary step preceding resin embedding, is much
less damaging to tissue integrity at very low temperatures (<−70 °C), where lipid
extraction by organic solvents is minimized [30]. Under these conditions, osmium is no
6 Conjugate Immunofluorescence—SEM Array Tomography …
151