Correlative fluorescence and electron microscopy

Journal of Neuroscience Methods 117 (2002) 81 /85 www.elsevier.com/locate/jneumeth

Correlative fluorescence and electron microscopy of biocytin-filled neurons with a preservation of the postsynaptic ultrastructure Youri Morozov, Ilgam Khalilov, Yehezkel Ben-Ari *, Alfonso Represa INMED/INSERM U29, 163 Route de Luminy, BP 13, 13009 Marseille, France Received 20 February 2002; received in revised form 19 March 2002; accepted 20 March 2002

Abstract Several techniques enable to inject intracellularly neurons with dyes and to use light and electron microscopy to correlate the physiological data with the morphological properties of the neuron. However, the ultrastructure of the neuron is usually obscured by the injected dye thus notably precluding the analysis of the postsynaptic specialisation and that of the other organelles. To overcome this problem, we have developed a technique based on fluorophore- and ultra small gold-conjugated streptavidins. We report, that this method facilitates the identification of intracellular organelles of the biocytin-filled neuron and of postsynaptic densities. This method is valid for the study of early postnatal neurons that are particularly refractory to this type of analysis. The procedure introduced here consists of the following steps: (1) injection of biocytin into the neuron by a patch-clamp pipette, (2) aldehyde fixation, (3) reaction with a fluorophore-conjugated streptavidin, (4) analysis with a fluorescence microscope, (5) formation of avidin /biotin complexes (ABC), (6) reaction with an ultra small gold-conjugated streptavidin, (7) silver enhancement of gold, (8) postfixation with osmium tetroxide and embedding in resin, (9) ultrathin sectioning and analysis with an electron microscope. Using this method, we show that in early postnatal hippocampal neurons, that have been injected with biocytine, it is possible to determine the morphology of the dendritic and axonal trees (including very thin details such as spines and filopodia) and to identify the localisation of the symmetric and asymmetric synapses on dendrites of the injected neuron. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Rat hippocampus; Synapses; Growth cones; Filopodia; Streptavidin; Ultra small gold; Fluorophore; ABC

1. Introduction The method of visualisation of biocytin-filled neurons through the formation of avidin /biotin horseradish peroxidase (HRP) complexes, followed by a peroxidase-3-3?-diaminobenzidine×/4HCl (DAB) reaction has become a classical procedure for light microscopy analysis. The HRP /DAB revelation of neurons was successfully used for electron microscopic (EM) analysis (Somogyi and Soltesz, 1986). However, the HRP /DAB method is usually used for EM analysis of the environment of the labelled neuron but not for the analysis of its

* Corresponding author. Tel.: /33-491-82-8100; fax: /33-491-828101. E-mail address: [email protected] (Y. Ben-Ari).

inner structures (Somogyi and Soltesz, 1986; Gulyas et al., 1999; Megias et al., 2001; Papp et al., 2001). Indeed, the end product of the HRP /DAB reaction is diffuse, highly electron dense and completely fills the inner space of the labelled cell, thus preventing the analysis of its cytoplasm and organelles. With this technique it is difficult to identify postsynaptic densities and cytoskeletal components of the HRP /DAB labelled cells. Furthermore, it is impossible to combine the HRP / DAB revelation with the labelling of intracellular proteins by using gold-conjugated antibodies, since the gold particles are difficult to observe in such a dark background. Recently, several other methods for correlative fluorescence and electron microscopy have been introduced. The intracellular injection of the mixture of Rhodamine

0165-0270/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 0 2 7 0 ( 0 2 ) 0 0 0 7 6 - 6

82

Y. Morozov et al. / Journal of Neuroscience Methods 117 (2002) 81 /85

and HRP with a subsequent HRP /DAB reaction, or the injection of Lucifer Yellow with a subsequent antiLucifer Yellow immunogold reaction have been described (Cabirol-Pol et al., 2000). Bergles et al. (2000) has used injections of biocytin with streptavidin conjugated with ultra small gold for the visualisation of oligodendrocyte precursors. We have recently reported that in both rodents and primate, during maturation, hippocampal neurons have a sequential formation of GABA and glutamate synapses and that dendritic synapses are formed before the perisomatic ones (Tyzio et al., 1999; Khazipov et al., 2001). To analyse this sequence and to determine its underlying mechanisms, it is essential to reconstruct, by electron microscopy, the injected neurons and to determine in a quantitative manner the formation of symmetric and asymmetric synapses. This type of study is, however, hampered by both the difficult fixation and preservation of the immature tissue and the absence of suitable methods that enable to visualise, with fine details, the morphology of the injected neuron. In the present report, we describe a method that enables to visualise the postsynaptic elements in immature neurons that have been injected with biocytine. Biocytin was first labelled by streptavidin conjugated with a fluorescent dye and in a second step it was labelled by streptavidin conjugated with ultra small gold. The staining with the fluorophore provided a bright staining of axonal and dendritic trees with fine details such as spine-like processes and filopodia. At the EM level, small electrondense particles provided a good staining that did not mask intracellular organelles and postsynaptic membranes.

2. Materials and methods Postnatal (P6) rats were deeply anaesthetised with sodium pentobarbital. After brain removal, coronal hippocampal slices (400 mm thick) were obtained by using a vibratome. The slices were incubated in an oxygenated artificial cerebrospinal fluid (ACSF) composed of 126 mM NaCl, 3.5 mM KCl, 2 mM CaCl2, 1.3 mM MgCl2, 25 mM NaHCO3, 1.2 mM NaHPO4, 10 mM of glucose, for at least 1 h. They were then fully submerged and superfused in the recording chamber with oxygenated ACSF at a rate of 2/3 ml/min. Pyramidal neurons and interneurons in the CA1 and CA3 zone were recorded under visual control with an Axioscope (Karl Zeiss, Germany), using the patchclamp technique in the whole cell-configuration and injected with 0.5% solution of biocytin. The slices were fixed with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) (PFA) with an addition of 2% glutaraldehyde overnight at /4 8C. The permeabilisation of the slices was conducted as follows: incuba-

tion in a mixture of 25% sucrose and 10% glycerol in 0.01 M phosphate buffer (pH 7.4) for 2 h for cryoprotection with subsequent freezing in liquid nitrogen and thawing in 0.02 M phosphate buffered saline pH 7.4 (PBS). After this the slices were immersed for 1 /2 h in steptavidin conjugated with Cy3TM (Jackson ImmunoResearch, West Grove, PA) or Alexa Fluor 488 (Molecular Probes, Leiden, The Netherlands) both diluted 1:200 in PBS with 5% normal goat serum (NGS). Then the slices were mounted on gelatine-coated slides with Gel mount (Biomeda, Foster City, CA) and analysed with an Olympus Fluoview FV 500 laser scanning microscope. The slices selected for EM were removed from the slides and re-sectioned with a freezing microtome in slices of 100 mm in thickness. For a subsequent connection of ultra small gold-conjugated streptavidin and for the multiplication of the sites of connection, avidin / biotin complexes (ABC) were formed on the molecules of the fluorophore-labelled streptavidin. Formation of ABC was fulfilled according to the manufacturer’s instructions (Vector laboratories, Burlingame, CA) for 5 /6 h at /4 8C. After this, the biotin in ABC was labelled with ultra small gold-conjugated streptavidin (Aurion, Wageningen, The Netherlands) diluted 1:20 in PBS with 5% NGS, 0.04% cold water fish skin gelatine and 0.05% sodium azide, overnight at /4 8C. The slices were rinsed with PBS for 1 h after every step of pseudoimmunoreaction. For the visualisation of ultra small gold, the slices were rinsed several times with distilled water and silver-enhanced with Aurion R-Gent SE-ME (Aurion, Wageningen, The Netherlands), for 10 /15 min according to the manufacturer’s instructions. The slices were studied and photographed with a regular light microscope Nikon Eclipse E800. The slices selected for subsequent preparation were post-fixed with 0.1% OsO4 in PBS for 30 min and rinsed in several changes of PBS and deionised water. En bloc staining with 1% aqueous uranyl acetate for 40 min was followed by the dehydration with methanol by a progressive lowering of temperature until /20 8C and embedding in Lowicryl HM 20 with UV-polymerisation. Ultrathin sections were cut with an ultramicrotome, mounted on formvar-coated grids, contrasted with 2% aqueous uranyl acetate and Reynolds lead citrate and observed with a Karl Zeiss EM 912 electron microscope. All reagents were purchased from Merck Eurolab (France), unless specified.

3. Results and discussion Pyramidal neurons or interneurons were recorded with the patch clamp techniques from postnatal slices (P6) and injected with biocytine. We obtained a very bright staining of the injected neurons with the


Fig. 1. Micrographs of a biocytin-filled basket interneuron in the CA1 zone of the postnatal hippocampus (P6) revealed with streptavidin / Alexa Fluor 488. (A) Picture of the axonal and dendritic trees with low magnification of confocal microscope. (B) Detail of the framed area in A, depicting a dendrite with a growth cone, filopodia and spine-like processes. Scale bar: A, 100 mm; B, 10 mm. Abbreviations: a, axon; c, growth cone; d, dendrite; f, filopodia; s, spine-like processes.

streptavidin /fluorophore technique. Neuronal cell bodies, dendritic and axonal trees were readily visualised (Fig. 1). Visual background was absent. Judging by the distribution of the visualised dendrites and axons through the thickness of the slices, as it was seen with optical slicing (for some cells more than 100 mm, data not shown), there was a suitable penetration of the revealing reagents. We have also often observed growth cones of the dendrites (Fig. 1B) and of the axons (data not shown), suggesting that the procedure is effective to reveal the dendrites and the axons until their natural termination. We could identify fine details such as spinelike processes and filopodia (Fig. 1B) confirming the high sensitivity of the method. Fig. 2 illustrates a CA1 pyramidal neuron of the postnatal (P6) rat hippocampus revealed first by fluorescence (Fig. 2A), then by a regular light microscope (Fig. 2B) and finally by EM (Fig. 2C). The fluorescence study provided the morphology of the dendritic and axonal trees of the biocytin-filled neuron (Fig. 2A). Regular light microscopy allowed us to identify and to orient the labelled neuron (before and after embedding it in the resin) and thus to facilitate ultrathin sectioning (Fig. 2B). EM analysis allowed easy identification of the labelled neuron with electrondense particles in the dendrites and the cell body (Fig. 2C). With EM the labelling of the dendrites was reconstructed for a long

83

distance with a negligible silver background. The preservation of the tissue enabled the identification of the postsynaptic densities and allowed us to distinguish between symmetric and asymmetric synaptic contacts (Fig. 3). The neurons reconstructed with the method based on streptavidin /fluorophores are quite similar to those obtained with the classical HRP /DAB method (Gaiarsa et al., 2001; Tyzio et al., 1999). Our protocol proves to be particularly suitable in combination with confocal microscopy notably with a Z -direction scanning for optical slicing and a three-dimensional analysis. Moreover, the protocol gives an opportunity to make a rapid analysis of the distribution of the axonal and dendritic trees with confocal microscopy prior to a more detailed analysis of the best samples with EM. The routine procedure of the preparation of biological subjects for EM includes post-fixation with 1% solution of OsO4 and embedding in epoxy resin. We have found that such a strong concentration of osmium tetroxide creates significant difficulties in the identification of the biocytin-filled cell for two reasons: a partial oxidation and decoloration of silver, on one hand, and the obscuration of the tissue, on the other hand. The use of a solution of OsO4 with a concentration reduced to 0.1% with subsequent embedding in a relatively hydrophilic Lowicryl HM20 let us obtain a stable visualisation of the labelled neuron with a good preservation of the membrane structure. The embedding in the acrylatebased resin is especially important for the preservation of the immature tissue. This makes the method valid for a study of the early post-natal neurons whose membranes are particularly labile with the hydrophobic epoxy resin. Several excellent studies by Freund, Buzsaki and coworkers have illustrated the use of the HRP /DAB method for a quantitative EM analysis of excitatory and inhibitory synaptic inputs (Gulyas et al., 1999; Megias et al., 2001; Papp et al., 2001). This method, however, requires a postembedding immunogold labelling of GABA (Hodgson et al., 1985; Somogyi and Hodgson, 1985). This is the consequence of the darkness of the intracellular end-products of the HRP /DAB reaction that precludes the identification of the postsynaptic membrane and its symmetric or asymmetric features. We propose the present method as a suitable alternative in studies in which the ultrastructure and postsynaptic density can be identified and quantified without the use of anti-GABA or other antibodies. In conclusion, this method is particularly adequate for the morphological characterisation of maturing neurons and allows the identification of excitatory and inhibitory synapses without using complicated postembedding immunolabelling.

84


Fig. 2. Pyramidal neuron in the CA1 zone of the postnatal hippocampus (P6) labelled for correlative microscopy. (A) Confocal micrograph of the neuron revealed with streptavidin-Cy3TM. (B) Light micrograph of the cell body and the apical dendrite of the neuron shown in A after embedding in Lowicril HM 20. (C) Electron micrograph of the cell body and a fragment of the apical dendrite of the neuron shown in A and B. Scale bar: A, B, 50 mm; C, 5 mm. Abbreviation: b, cell body; d, apical dendrite; s, silver particles.

Fig. 3. Interneuron in the CA3 zone of the postnatal hippocampus (P6). A: Light micrograph of the interneuron after embedding in Lowicril HM 20. Symmetric (B) and asymmetric (C) synapses on the dendrites of the interneuron shown in A. Scale bar: A, 50 mm; B, C, 0.5 mm. Abbreviations: as, bouton of an asymmetric synapse; ss, bouton of a symmetric synapse; so, stratum orients; sr, stratum radiatum. Arrows ‘b’ and ‘c’, approximate localisation of the synapses, shown in B and C, respectively.


Acknowledgements We are grateful to Professor P. Somogyi and Dr I. Guillemain for critical reading and discussion of the manuscript, Dr L. Aniksztejn, M. Demarque, G. Medyna and E. Morozova for technical assistance. Youri Morozov is presently a recipient of an INMED fellowship.

References Bergles DE, Roberts JDB, Somogyi P, Jahr CE. Glutamatergic synapses on oligodendrocyte precursor cells in the hippocampus. Nature 2000;405:187 /91. Cabirol-Pol MJ, Mizrahi A, Simmers J, Meyrand P. Combining laser scanning confocal microscopy and electron microscopy to determine sites of synaptic contact between two identified neurons. J Neurosci Methods 2000;97:175 /81. Gaiarsa JL, Khalilov I, Gozlan H, Ben-Ari Y. Morphology of CA3 non-pyramidal cells in the developing rat hippocampus. Dev Brain Res 2001;127:157 /64. Gulyas AI, Megias M, Emri Z, Freund TF. Total number and ratio of excitatory and inhibitory synapses converging onto single inter-

85

neurons of different types in the CA1 area of the rat hippocampus. J Neurosci 1999;19:10082 /97. Hodgson AJ, Penke B, Erdei A, Chubb IV, Somogyi P. Antisera to gaminobutyric acid. I. Production and characterization using a new model system. J Histochem Cytochem 1985;33:229 /39. Khazipov R, Esclapez M, Caillard O, Bernard C, Khalilov I, Tyzio R, Hirsch J, Dzhala V, Berger B, Ben-Ari Y. Early development of neuronal activity in the primate hippocampus in utero. J Neurosci 2001;21:9770 /81. Megias M, Emri Z, Freund TF, Gulyas AI. Total number and distribution of inhibitory and excitatory synapses on hippocampal CA1 pyramidal cells. Neuroscience 2001;102:527 /40. Papp E, Leinekugel X, Henze DA, Lee J, Buzsaki G. The apical shaft of CA1 pyramidal cells is under gabaergic interneuronal control. Neuroscience 2001;102:715 /21. Somogyi P, Hodgson AJ. Antisera to g-aminobutyric acid. III. Demonstration of GABA in Golgi-impregnated neurons and in conventional electron microscopic sections of cat striate cortex. J Histochem Cytochem 1985;33:249 /57. Somogyi P, Soltesz I. Immunogold demonstration of GABA in synaptic terminals of intracellularly recorded, horseradish peroxidase-filled basket cells and clutch cells in the cat’s visual cortex. Neuroscience 1986;19:1051 /65. Tyzio R, Represa A, Jorquera I, Ben-Ari Y, Gozlan H, Aniksztejn L. The establishment of GABAergic and glutamatergic synapses on CA1 pyramidal neurons is sequential and correlates with the development of the apical dendrite. J Neurosci 1999;19:10372 /82.