135
Brain Research, 527 (1990) 135-139 Elsevier BRES 24262
Short Communications i
Epitopes of gap junctional proteins localized to neuronal subsurface cisterns T. Yamamoto 1, E.L. Hertzberg 2 and J.I. Nagy 1 1Department of Physiology, Facultyof Medicine, University of Manitoba, Winnipeg, Man. (Canada) and 2Department of Neuroscience, Albert Einstein College of Medicine, Bronx, New York, NY 10461 (U.S.A.) (Accepted 29 May 1990) Key words: Calcium; Connexin32; Endoplasmic reticulum; Gap junction; Motoneuron; Subsurface cistern
Several lines of evidence indicate the existence of channels that mediate the movement of calcium from the extracellular space directly into some intraeellular calcium storage compartment and from one intracellular membrane-bounded compartment to another. The possibility that such channels resemble intercellular communication pathways formed by gap junction proteins (connexins) was investigated in rat brain. Antibodies against a rat liver gap junction protein (connexin32) were found to recognize several distinct proteins on Western blots of brain homogenates. In motoneurons these antibodies immunohistochemicallylabelled portions of neuronal endoplasmic reticulum membranes that form subsurface dsterns (SSCs) adjacent to the plasma membrane. These results suggest that SSCs and connexin-likeproteins may be involved in the process of calcium mobilization in neurons. Recent studies of intracellular calcium sequestration and mobilization in a variety of cell types including neurons have provided evidence that calcium is stored within and released from one or more intracellular compartment; that movement of calcium between these compartment(s) is regulated by second messenger systems; and that some particular compartment is loaded with calcium directly from the extraceUular space 6, 13,17,18.20,23. Models designed to accommodate this evidence have invoked the existence of communication between intracellular calcium storage organelles and between these organelles and the extracellular space via channels similar to those mediating intercellular communication at gap junctions 1'11'12']7. However, there is no morphological evidence as yet for the formation of gap junctions between intracellular organelles, nor is there biochemical or physiological evidence for a functional association of the structural proteins (connexins) of gap junctions with cellular elements other than the plasma membrane. To investigate the possibility that such proteins or closely related proteins may be utilized to form intracellular channels through incorporation into cellular membrane compartments involved in calcium homeostasis, we conducted biochemical and immunohistochemical analyses of the proteins and cellular elements in rat brain recognized by anti-connexin antibodies. Three different membrane proteins of molecular mass
26, 32 and 43 kDa with various degrees of sequence homology are currently considered to form gap junctional channels between cells in a variety of mammalian tissues 2's'14'21'22. We used a polyclonal antibody against rat liver connexin32 and a monoclonal antibody (designated 92B) against a peptide corresponding to the amino acid sequence 224-234 in the carboxy-terminus region of connexin3222. The specificity of the polyclonal antibody was demonstrated by its recognition of connexin32 on Western blots of liver9'1° and brain homogenates (unpublished observations). In addition, our ultrastructural studies have shown that the polyclonal antibody immunolabels neuronal and glial gap junctions in brain 19'25 and that both the polyclonal and the monoclonal antibodies label gap junctions in liver28. The monoclonal antibody was prepared as previously described 1°. The specificity of this antibody for liver connexin32 was demonstrated by immunofluorescence in sections of liver and b y its recognition of connexin32 on Western blots 1°. Isolated rat liver gap junctions and brain homogenate were resolved by SDS-PAGE gels and processed for Western blots. Antibody binding was visualized using a rabbit anti-mouse Ig secondary antibody and then 125I-Protein A, followed by autoradiography. For light microscopy, rats were perfused with either 400 ml of 0.1 M sodium phosphate buffer, pH 7.4, containing 4% paraformaldehyde or with 200 ml of 0.1 M sodium phosphate buffer,
Correspondence: J.I. Nagy, Department of Physiology, Faculty of Medicine, University of Manitoba, 770 Bannatyne Avenue, Winnipeg, Manibota, Canada R3E 0W3. 0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
136 pH 7.4, containing 4% paraformaldehyde followed by 200 ml of 50 mM sodium borate buffer, pH 9.0, containing 4% paraformaldehyde. Robust immunohistochemical labelling was obtained with these as well as a variety of other fixation conditions. Sections cut on a freezing microtome were incubated with antibody 92B (1:7000) for 48 h at 4 °C. The sections were then processed by the peroxidase-antiperoxidase (PAP) method using a rabbit antimouse second antibody followed by mouse PAP, and photographed with differential interference contrast optics. In control sections, no labelling was seen after omission of antibody 92B or after adsorption of this antibody with the synthetic peptide against which it was produced. For electron microscopy, rats were perfused with 4% paraformaldehyde and 0.2% glutaraldehyde. Vibratome sections of the brainstem were processed for immunohistochemistry as described above, postfixed with osmium tetroxide and flat embedded in Epon. Ultrathin sections were counterstained with lead citrate. While we are as yet uncertain whether antibody 92B labels gap junctions in brain, it did recognize several potentially novel proteins revealed by Western blots and a distinct intracellular structure as seen by immunohistochemical analyses. On Western blots of proteins separated on 12.5% acrylamide gels routinely used for
a 1
analysis of gap junction proteins, antibody binding was observed, as expected, to connexin32 (apparent mol. wt. of 27,000 in this gel system), its 47,000 mol. wt. dimer and higher oligomers (Fig. la, lane 1). In brain homogenates (Fig. la; lane 2), binding was observed not only to a protein comigrating with connexin32 but to 35 and 70 kDa proteins as well. Similar analysis undertaken with gradient gels permitting resolution of higher molecular weight proteins (Fig. lb) demonstrated antibody binding to 150 and 175 kDa proteins in the brain sample in addition to the lower molecular weight proteins resolved on the 12.5% SDS polyacrylamide gels.
b 2
3 MW
4 MW 175
150
7O -.-'85 .*-.27
47 27
i~
Fig. 1. Western blot analysis of proteins in isolated rat liver gap
junctions (lanes 1 and 3) and brain homogenate (lanes 2 and 4) with a site-specificmonoclonal antibody against connexin32(amino acids 224-234). Samples were processed for SDS-PAGE and resolved on either 12.5% (a) or 3-15% (b) polyacrylamide gels. Antibody binding is observed to the 27 kDa fiver gap junction protein (connexin32) and its oligomers in the isolated gap junction preparation and to 27, 35, 70, 150 and 175 kDa proteins in the brain homogenate.
'
~
'
~
' i~:
~
n*~ ¸
Fig. 2. Light microscope immunohistochemical localization of structures in motoneurons labelled by site-specific anti-connexin32 antibody 92B. a: facial motor nucleus in a section counterstained with thionin, b: higher magnification of two motoneurons in a non-counterstained section. Label is seen as puncta (arrows) near the surface of the neuronal soma and initial dendrites. Some light diffuse nuclear staining is also evident (arrowheads). (Scale bar, 50
/~m)
137 Light microscope immunohistochemistry revealed punctate labelling with antibody 92B in many neurons throughout the CNS. However, brainstem and spinal alpha motoneurons, such as those in the facial motor nucleus (Fig. 2a, counterstained section), exhibited the most conspicuous labelling with puncta about 1.0/~m in width and up to 5/~m in length. In most other brain
C
~¸ ~i~~ ~ I
•
J~
~
~iI
regions the puncta had diameters of about 0.5/~m or less. In motoneurons the puncta surrounded the somal periphery (Fig. 2a), indicating localization of label at or near the cell surface, and extended into the proximal dendrites where they are reduced in size (Fig. 2b). Seen en face the puncta were roughly circular with diameters of about 2-4/~m. Some very faint granular labelling was associated with the nucleus and sparsely distributed within the adjacent cytoplasm. At the electron microscope level, antibody 92B labelled various structures in motoneurons including free and membrane bound ribosomes, endoplasmic reticulum (ER) (Fig. 3a) and some multivesicular bodies (not shown) at discrete sites near the somal and dendritic plasmalemmal membranes (Fig. 3b) and subjacent cytoplasm. This pattern of labelling was also seen with the polycional anti-connexin32 antibody (not shown). A segment of the ER closely apposed to the plasma membrane was most densely labelled with both antibodies. Within this segment the superficial and deep E R membranes were themselves in close apposition and nearly obscured the lumen which was narrowed to a gap of 2-5 nm. The density of immunostaining at the cytoplasmic surface of the E R was clearly greater than that of the lumenal surface (Fig. 3c). A single mitochondrion was often seen closely applied to the immunolabelled region of ER. Some labelling was also present at the inner surface of the plasma membrane and at the surface of the adjacent mitochondrion. It remains to be determined whether the punctate labelling seen with antibody 92B by LM in other brain regions is also localized at regions of E R near the cell surface. The immunolabelled configuration of E R membranes we observe corresponds morphologically to previously described subsurface cisterns 3,7,15,24,26. Common characteristics of cisterns include their separation from the plasma membrane by an average distance of about 10 nm .(.Fig. 3. Electron microscope localization of structures in facial motoneurons labelled by site-specific anti-connexin32 monoclonal antibody 92B. a: dense labelling is restricted to regions of endoplasmic reticulum (large arrow) closely apposed to the plasma membrane of the motoneuron soma (MS) at sites contacted by large axon terminals (ax). Some label is also associated with surrounding ribosomes (small arrow) while ER and ribosomes deeper within the cell soma are unlabeHed(arrowhead). Unlabelled gaps along the ER cisternal membranes are occasionally observed and may be an artifact of tissue processing, b: medium and c: high magnificationof labelled ER regions. The width of the ER lumen between the white arrows in (c) is 2 rim. The bilayer leaflets of the immunolabelled ER membranes tended to separate and could be individually resolved. A mitochondrion is applied to the labelled ER in (b) and a second set of weakly labelled ER membranes is closely apposed to the densely stained subsurface cistern in (c) (arrowhead). TPM, plasma membrane of axon terminal; SPM, somal plasma membrane; OER, outer ER membrane; IER, inner ER membrane (Scale bars: a,b, 0.5/~m; c, 100 nm).
138 creating what here will be referred to as the cisternal cleft m and the close apposition of their E R membranes producing a structure that resembles a gap junction. In most neurons, the cisterns are located at the plasma m e m b r a n e opposite astrocytic processes and only occasionally opposite axon terminals. Cisterns in motoneurons seem to represent a special case since they are always opposite an axon terminal and they face a specific class of terminal (termed type C) which are c o m m o n on motoneuronal somas and are thought to be cholinergic 4'5. Thus, motoneurons may be ideal for studies of cistern function and the possible mediation of the action of C-terminal transmitters via these cisterns. It has been speculated much earlier that in analogy to the relationship between the sarcoplasmic reticulum and the transverse tubule in cardiac and skeletal muscle, subsurface cisterns may allow ionic communication between the cell exterior and the E R 7'15'16. A m o n g current views of how this may occur with respect to Ca 2+ are models 6'7'13'15A638'2°'23'27wherein the Ca 2+ level in some --
intracellular compartment, perhaps the cisternal cleft, regulates the rate of Ca 2÷ entry into the cell across the plasma membrane. A modification of such models can be reconciled with morphological similarities between subsurface cisternal membranes and gap junctions and with the existence of a cisternal protein which appears to be recognized by antibodies against a gap junction protein. For example, the cross-reactivity observed may indicate the presence of a channel-forming protein in an area of close m e m b r a n e apposition as found in gap junctions. 1 Berridge, M.J. and Taylor, C.W., Inositol trisphosphate and calcium signaling, Cold Spring Harbor Symp. Quant. Biol., 53 (1988) 927-933. 2 Beyer, E.C., Paul, D.L. and Goodenough, D.A., Connexin43: a protein from rat heart homologous to a gap junction protein from liver, J. Cell Biol., 105 (1987) 2621-2629. 3 Bunge, R.P., Bunge, M.B. and Peterson, E.R., An electron microscopic study of cultured rat spinal cord, J. Cell Biol., 24 (1965) 163-191. 4 Connaughton, M., Priestley, J.V., Sofroniew, M.V., Eckenstein, F. and Cuelio, A., Input to motoneurones in the hypoglossal nucleus of the rat: light and electron microscope immunocytochemistry for choline acetyltransferase, substance P, and enkephalins using monoclonai antibodies, Neuroscience, 17 (1986) 205-224. 5 Davidoff, M.S. and Irintchev, A.P., Acetylcholinesterase activity and type C synapses in the hypoglossal, facial and spinal-cord motor nuclei of rats: an electron-microscope study, Histochemistry, 84 (1986) 514-524. 6 Ghosh, T.K., Muilaney, J.M., Tarazi, El. and Gill, D.L., GTP-activated communication between distinct inositol 1,4, 5-trisphosphate-sensitive and -insensitive calcium pools, Nature, 340 (1989) 236-239. 7 Henkhart, M., Landis, D.M.D. and Reese, T.S., Similarity of junctions between plasma membranes and endoplasmic reticulum in muscle and neurons, J. Cell Biol., 70 (1976) 338-347. 8 Hertzberg, E.L., A detergent-independent procedure for the isolation of gap junctions from rat liver, J. Biol. Chem., 259 (1984) 9936-9943.
However, it appears that the region of closest apposition of immunoreactive membranes is between cisternal membranes rather than between cisternal and plasma membranes as anticipated in some models of calcium translocation from the extracellular space 12"17. This might suggest that channels form across the E R membranes themselves in specialized regions of certain classes of cisterns. When open, these channels may allow calcium to flow directly from the cisternal cleft into the cytoplasm or other underlying structures. Alternatively, the immunoreactivity observed in appositional plasma membrane might be indicative of the presence of other channelforming proteins which could interact directly with homologues in cisternal membranes providing for a more complex arrangement of channels from the extracellular space to the cisternal lumen. In either case, our results would be consistent with previous suggestions of intracellular communication mediated by channels analogous to those at gap junctions. In order to determine the nature of the subsurface cisternal protein that may form such channels, it will be necessary to isolate the various proteins in brain recognized by anti-connexin antibodies and establish their relationship to the known connexins. We thank L. Poison, A. Ochaiski, M. Sawchuk and R. Corpina for technical assistance. Supported by NIH (GM30667), MRC of Canada, and the Manitoba Health Research Council. J.I.N. is a MRC Scientist and E.L.H. is a recipient of a NIH Research Career Development Award (HD00713) and a Irma T. Hirschl Career Scientist Award. 9 Hertzberg, E.L. and Skibbens, R.V., A protein homologous to the 27,000 dalton liver gap junction protein is present in a wide variety of species and tissues, Cell, 39 (1984) 61-69. 10 Hertzberg, E.L., Disher, R.M., Tiller, A.A., Zhou, Y. and Cook, R.G., Topology of the Mr 27,000 liver gap junction protein. Cytoplasmic localization of amino- and carboxyl termini and a hydrophilic domain which is protease-hypersensitive, J. Biol. Chem., 263 (1988) 19105-19111. 11 Irvin, R.F. and Moor, R.M., Inositol (1,3,4,5) tetrakisphosphate-induced activation of sea urchin eggs requires the presence of inositol trisphosphate, Biochem. Biophys. Res. Commun., 146 (1987) 284-290. 12 Irvine, R.F., Moor, R.M., Pollock, W.K., Smith, P.M. and Wregget, K.A., Inositol phosphates: proliferation, metabolism and function, Phil. Trans. R. Soe. Lond. B, 320 (1988) 281-298. 13 Irvine, R.F., How do inositol 1,4,5-trisphosphate and inositol 1,3,4,5-tetrakisphosphate regulate intracellular Ca2+?, Biochem. Soc. Trans., 17 (1989) 6-8. 14 Kumar, N.M. and Gilula, N.B., Cloning and characterization of human and rat liver cDNAs coding for a gap junction protein, J. Cell Biol., 103 (1986) 767-776. 15 Le Beux Y.J., Subsurface cisterns and iameltar bodies: particular forms of the endoplasmic retieulum in the neurons, Z. Zellforsch., 133 (1972) 327-352. 16 McBurney, R.N. and Neering, I.R., Neuronal calcium homeostasis, Trends Neurosci., 10 (1987) 164-169. 17 Merritt, J.E. and Rink, T.J., Regulation of cytosolic free calcium in Fura-2-1oaded rat parotid acinar cells, J. Biol. Chem., 262 (1987) 17362-17369.
139 18 Mullaney, J.M., Chueh, S.-H., Ghosh, T.K. and Gill, D.L., Intracellular calcium uptake activated by GTP. Evidence for a possible guanine nucleotide-induced transmembrane conveyance of intraceUular calcium, J. Biol. Chem., 262 (1987) 13865-13872. 19 Nagy, J.I., Yamamoto, T., Shiosakaa, S., Dewar, K.M., Whittaker, M.E. and Hertzberg, E.L., Immunohistochemical localization of gap junction protein in rat CNS: a preliminary account. In E.L. Hertzberg and R.G. Johnson (Eds.), Modern Cell Biology, Vol. 7, Liss, New York, 1988, pp. 375-389. 20 Nahorski, S.R., Inositol polyphosphates and neuronal calcium homeostasis, Trends Neurosci., 11 (1988) 444-448. 21 Nicholson, B.J. and Zhang, J.-T., Multiple protein components in a single gap junction: cloning of a second hepatic gap junction (M~21,000). In E.L. Hertzberg and R.G. Johnson (Eds.), Modern Cell Biology, Vol. 7, Liss, New York, 1988, pp. 207-218. 22 Paul, D.L., Molecular cloning of cDNA for rat liver gap junction protein, J. Cell Biol., 103 (1986) 123-134. 23 Putney, Jr., J.W., A model for receptor-regulated calcium entry,
Cell Calcium, 7 (1986) 1-12. 24 Rosenbluth, J., Subsurface cisterns and their relationship to the neuronal plasma membrane, J. Cell BioL, 13 (1962) 405-421. 25 Shiosaka, S., Yamamoto, T., Hertzberg, E.L. and Nagy, J.I., Gap junction protein in rat hippocampus: correlative light and electron microscope immunohistochemical localization, J. Comp. Neurol., 281 (1989) 282-297. 26 Siegesmund, K.A., The fine structure of subsurface cisterns, Anat. Rec., 162 (1968) 187-196. 27 Takemura, H., Highes, A.R., Thastrup, O. and Putney, Jr., J.W., Activation of calcium entry by the tumor promoter thapsigargin in parotid acinar cells, J. Biol. Chem., 264 (1989) 12266-12271. 28 Yamamoto, T., Hertzberg, E.L. and Nagy, J.I., Antibodies against rat liver connexin32 recognize subsurface cisterns in motoneurons: immunohistochemical evidence for similarities between gap junctional and subcisternal proteins, Soc. Neurosci. Abstr., 1989, Vol. 15, Part 1,680.
Brain Research, 527 (1990) 135-139 Elsevier BRES 24262
Short Communications i
Epitopes of gap junctional proteins localized to neuronal subsurface cisterns T. Yamamoto 1, E.L. Hertzberg 2 and J.I. Nagy 1 1Department of Physiology, Facultyof Medicine, University of Manitoba, Winnipeg, Man. (Canada) and 2Department of Neuroscience, Albert Einstein College of Medicine, Bronx, New York, NY 10461 (U.S.A.) (Accepted 29 May 1990) Key words: Calcium; Connexin32; Endoplasmic reticulum; Gap junction; Motoneuron; Subsurface cistern
Several lines of evidence indicate the existence of channels that mediate the movement of calcium from the extracellular space directly into some intraeellular calcium storage compartment and from one intracellular membrane-bounded compartment to another. The possibility that such channels resemble intercellular communication pathways formed by gap junction proteins (connexins) was investigated in rat brain. Antibodies against a rat liver gap junction protein (connexin32) were found to recognize several distinct proteins on Western blots of brain homogenates. In motoneurons these antibodies immunohistochemicallylabelled portions of neuronal endoplasmic reticulum membranes that form subsurface dsterns (SSCs) adjacent to the plasma membrane. These results suggest that SSCs and connexin-likeproteins may be involved in the process of calcium mobilization in neurons. Recent studies of intracellular calcium sequestration and mobilization in a variety of cell types including neurons have provided evidence that calcium is stored within and released from one or more intracellular compartment; that movement of calcium between these compartment(s) is regulated by second messenger systems; and that some particular compartment is loaded with calcium directly from the extraceUular space 6, 13,17,18.20,23. Models designed to accommodate this evidence have invoked the existence of communication between intracellular calcium storage organelles and between these organelles and the extracellular space via channels similar to those mediating intercellular communication at gap junctions 1'11'12']7. However, there is no morphological evidence as yet for the formation of gap junctions between intracellular organelles, nor is there biochemical or physiological evidence for a functional association of the structural proteins (connexins) of gap junctions with cellular elements other than the plasma membrane. To investigate the possibility that such proteins or closely related proteins may be utilized to form intracellular channels through incorporation into cellular membrane compartments involved in calcium homeostasis, we conducted biochemical and immunohistochemical analyses of the proteins and cellular elements in rat brain recognized by anti-connexin antibodies. Three different membrane proteins of molecular mass
26, 32 and 43 kDa with various degrees of sequence homology are currently considered to form gap junctional channels between cells in a variety of mammalian tissues 2's'14'21'22. We used a polyclonal antibody against rat liver connexin32 and a monoclonal antibody (designated 92B) against a peptide corresponding to the amino acid sequence 224-234 in the carboxy-terminus region of connexin3222. The specificity of the polyclonal antibody was demonstrated by its recognition of connexin32 on Western blots of liver9'1° and brain homogenates (unpublished observations). In addition, our ultrastructural studies have shown that the polyclonal antibody immunolabels neuronal and glial gap junctions in brain 19'25 and that both the polyclonal and the monoclonal antibodies label gap junctions in liver28. The monoclonal antibody was prepared as previously described 1°. The specificity of this antibody for liver connexin32 was demonstrated by immunofluorescence in sections of liver and b y its recognition of connexin32 on Western blots 1°. Isolated rat liver gap junctions and brain homogenate were resolved by SDS-PAGE gels and processed for Western blots. Antibody binding was visualized using a rabbit anti-mouse Ig secondary antibody and then 125I-Protein A, followed by autoradiography. For light microscopy, rats were perfused with either 400 ml of 0.1 M sodium phosphate buffer, pH 7.4, containing 4% paraformaldehyde or with 200 ml of 0.1 M sodium phosphate buffer,
Correspondence: J.I. Nagy, Department of Physiology, Faculty of Medicine, University of Manitoba, 770 Bannatyne Avenue, Winnipeg, Manibota, Canada R3E 0W3. 0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
136 pH 7.4, containing 4% paraformaldehyde followed by 200 ml of 50 mM sodium borate buffer, pH 9.0, containing 4% paraformaldehyde. Robust immunohistochemical labelling was obtained with these as well as a variety of other fixation conditions. Sections cut on a freezing microtome were incubated with antibody 92B (1:7000) for 48 h at 4 °C. The sections were then processed by the peroxidase-antiperoxidase (PAP) method using a rabbit antimouse second antibody followed by mouse PAP, and photographed with differential interference contrast optics. In control sections, no labelling was seen after omission of antibody 92B or after adsorption of this antibody with the synthetic peptide against which it was produced. For electron microscopy, rats were perfused with 4% paraformaldehyde and 0.2% glutaraldehyde. Vibratome sections of the brainstem were processed for immunohistochemistry as described above, postfixed with osmium tetroxide and flat embedded in Epon. Ultrathin sections were counterstained with lead citrate. While we are as yet uncertain whether antibody 92B labels gap junctions in brain, it did recognize several potentially novel proteins revealed by Western blots and a distinct intracellular structure as seen by immunohistochemical analyses. On Western blots of proteins separated on 12.5% acrylamide gels routinely used for
a 1
analysis of gap junction proteins, antibody binding was observed, as expected, to connexin32 (apparent mol. wt. of 27,000 in this gel system), its 47,000 mol. wt. dimer and higher oligomers (Fig. la, lane 1). In brain homogenates (Fig. la; lane 2), binding was observed not only to a protein comigrating with connexin32 but to 35 and 70 kDa proteins as well. Similar analysis undertaken with gradient gels permitting resolution of higher molecular weight proteins (Fig. lb) demonstrated antibody binding to 150 and 175 kDa proteins in the brain sample in addition to the lower molecular weight proteins resolved on the 12.5% SDS polyacrylamide gels.
b 2
3 MW
4 MW 175
150
7O -.-'85 .*-.27
47 27
i~
Fig. 1. Western blot analysis of proteins in isolated rat liver gap
junctions (lanes 1 and 3) and brain homogenate (lanes 2 and 4) with a site-specificmonoclonal antibody against connexin32(amino acids 224-234). Samples were processed for SDS-PAGE and resolved on either 12.5% (a) or 3-15% (b) polyacrylamide gels. Antibody binding is observed to the 27 kDa fiver gap junction protein (connexin32) and its oligomers in the isolated gap junction preparation and to 27, 35, 70, 150 and 175 kDa proteins in the brain homogenate.
'
~
'
~
' i~:
~
n*~ ¸
Fig. 2. Light microscope immunohistochemical localization of structures in motoneurons labelled by site-specific anti-connexin32 antibody 92B. a: facial motor nucleus in a section counterstained with thionin, b: higher magnification of two motoneurons in a non-counterstained section. Label is seen as puncta (arrows) near the surface of the neuronal soma and initial dendrites. Some light diffuse nuclear staining is also evident (arrowheads). (Scale bar, 50
/~m)
137 Light microscope immunohistochemistry revealed punctate labelling with antibody 92B in many neurons throughout the CNS. However, brainstem and spinal alpha motoneurons, such as those in the facial motor nucleus (Fig. 2a, counterstained section), exhibited the most conspicuous labelling with puncta about 1.0/~m in width and up to 5/~m in length. In most other brain
C
~¸ ~i~~ ~ I
•
J~
~
~iI
regions the puncta had diameters of about 0.5/~m or less. In motoneurons the puncta surrounded the somal periphery (Fig. 2a), indicating localization of label at or near the cell surface, and extended into the proximal dendrites where they are reduced in size (Fig. 2b). Seen en face the puncta were roughly circular with diameters of about 2-4/~m. Some very faint granular labelling was associated with the nucleus and sparsely distributed within the adjacent cytoplasm. At the electron microscope level, antibody 92B labelled various structures in motoneurons including free and membrane bound ribosomes, endoplasmic reticulum (ER) (Fig. 3a) and some multivesicular bodies (not shown) at discrete sites near the somal and dendritic plasmalemmal membranes (Fig. 3b) and subjacent cytoplasm. This pattern of labelling was also seen with the polycional anti-connexin32 antibody (not shown). A segment of the ER closely apposed to the plasma membrane was most densely labelled with both antibodies. Within this segment the superficial and deep E R membranes were themselves in close apposition and nearly obscured the lumen which was narrowed to a gap of 2-5 nm. The density of immunostaining at the cytoplasmic surface of the E R was clearly greater than that of the lumenal surface (Fig. 3c). A single mitochondrion was often seen closely applied to the immunolabelled region of ER. Some labelling was also present at the inner surface of the plasma membrane and at the surface of the adjacent mitochondrion. It remains to be determined whether the punctate labelling seen with antibody 92B by LM in other brain regions is also localized at regions of E R near the cell surface. The immunolabelled configuration of E R membranes we observe corresponds morphologically to previously described subsurface cisterns 3,7,15,24,26. Common characteristics of cisterns include their separation from the plasma membrane by an average distance of about 10 nm .(.Fig. 3. Electron microscope localization of structures in facial motoneurons labelled by site-specific anti-connexin32 monoclonal antibody 92B. a: dense labelling is restricted to regions of endoplasmic reticulum (large arrow) closely apposed to the plasma membrane of the motoneuron soma (MS) at sites contacted by large axon terminals (ax). Some label is also associated with surrounding ribosomes (small arrow) while ER and ribosomes deeper within the cell soma are unlabeHed(arrowhead). Unlabelled gaps along the ER cisternal membranes are occasionally observed and may be an artifact of tissue processing, b: medium and c: high magnificationof labelled ER regions. The width of the ER lumen between the white arrows in (c) is 2 rim. The bilayer leaflets of the immunolabelled ER membranes tended to separate and could be individually resolved. A mitochondrion is applied to the labelled ER in (b) and a second set of weakly labelled ER membranes is closely apposed to the densely stained subsurface cistern in (c) (arrowhead). TPM, plasma membrane of axon terminal; SPM, somal plasma membrane; OER, outer ER membrane; IER, inner ER membrane (Scale bars: a,b, 0.5/~m; c, 100 nm).
138 creating what here will be referred to as the cisternal cleft m and the close apposition of their E R membranes producing a structure that resembles a gap junction. In most neurons, the cisterns are located at the plasma m e m b r a n e opposite astrocytic processes and only occasionally opposite axon terminals. Cisterns in motoneurons seem to represent a special case since they are always opposite an axon terminal and they face a specific class of terminal (termed type C) which are c o m m o n on motoneuronal somas and are thought to be cholinergic 4'5. Thus, motoneurons may be ideal for studies of cistern function and the possible mediation of the action of C-terminal transmitters via these cisterns. It has been speculated much earlier that in analogy to the relationship between the sarcoplasmic reticulum and the transverse tubule in cardiac and skeletal muscle, subsurface cisterns may allow ionic communication between the cell exterior and the E R 7'15'16. A m o n g current views of how this may occur with respect to Ca 2+ are models 6'7'13'15A638'2°'23'27wherein the Ca 2+ level in some --
intracellular compartment, perhaps the cisternal cleft, regulates the rate of Ca 2÷ entry into the cell across the plasma membrane. A modification of such models can be reconciled with morphological similarities between subsurface cisternal membranes and gap junctions and with the existence of a cisternal protein which appears to be recognized by antibodies against a gap junction protein. For example, the cross-reactivity observed may indicate the presence of a channel-forming protein in an area of close m e m b r a n e apposition as found in gap junctions. 1 Berridge, M.J. and Taylor, C.W., Inositol trisphosphate and calcium signaling, Cold Spring Harbor Symp. Quant. Biol., 53 (1988) 927-933. 2 Beyer, E.C., Paul, D.L. and Goodenough, D.A., Connexin43: a protein from rat heart homologous to a gap junction protein from liver, J. Cell Biol., 105 (1987) 2621-2629. 3 Bunge, R.P., Bunge, M.B. and Peterson, E.R., An electron microscopic study of cultured rat spinal cord, J. Cell Biol., 24 (1965) 163-191. 4 Connaughton, M., Priestley, J.V., Sofroniew, M.V., Eckenstein, F. and Cuelio, A., Input to motoneurones in the hypoglossal nucleus of the rat: light and electron microscope immunocytochemistry for choline acetyltransferase, substance P, and enkephalins using monoclonai antibodies, Neuroscience, 17 (1986) 205-224. 5 Davidoff, M.S. and Irintchev, A.P., Acetylcholinesterase activity and type C synapses in the hypoglossal, facial and spinal-cord motor nuclei of rats: an electron-microscope study, Histochemistry, 84 (1986) 514-524. 6 Ghosh, T.K., Muilaney, J.M., Tarazi, El. and Gill, D.L., GTP-activated communication between distinct inositol 1,4, 5-trisphosphate-sensitive and -insensitive calcium pools, Nature, 340 (1989) 236-239. 7 Henkhart, M., Landis, D.M.D. and Reese, T.S., Similarity of junctions between plasma membranes and endoplasmic reticulum in muscle and neurons, J. Cell Biol., 70 (1976) 338-347. 8 Hertzberg, E.L., A detergent-independent procedure for the isolation of gap junctions from rat liver, J. Biol. Chem., 259 (1984) 9936-9943.
However, it appears that the region of closest apposition of immunoreactive membranes is between cisternal membranes rather than between cisternal and plasma membranes as anticipated in some models of calcium translocation from the extracellular space 12"17. This might suggest that channels form across the E R membranes themselves in specialized regions of certain classes of cisterns. When open, these channels may allow calcium to flow directly from the cisternal cleft into the cytoplasm or other underlying structures. Alternatively, the immunoreactivity observed in appositional plasma membrane might be indicative of the presence of other channelforming proteins which could interact directly with homologues in cisternal membranes providing for a more complex arrangement of channels from the extracellular space to the cisternal lumen. In either case, our results would be consistent with previous suggestions of intracellular communication mediated by channels analogous to those at gap junctions. In order to determine the nature of the subsurface cisternal protein that may form such channels, it will be necessary to isolate the various proteins in brain recognized by anti-connexin antibodies and establish their relationship to the known connexins. We thank L. Poison, A. Ochaiski, M. Sawchuk and R. Corpina for technical assistance. Supported by NIH (GM30667), MRC of Canada, and the Manitoba Health Research Council. J.I.N. is a MRC Scientist and E.L.H. is a recipient of a NIH Research Career Development Award (HD00713) and a Irma T. Hirschl Career Scientist Award. 9 Hertzberg, E.L. and Skibbens, R.V., A protein homologous to the 27,000 dalton liver gap junction protein is present in a wide variety of species and tissues, Cell, 39 (1984) 61-69. 10 Hertzberg, E.L., Disher, R.M., Tiller, A.A., Zhou, Y. and Cook, R.G., Topology of the Mr 27,000 liver gap junction protein. Cytoplasmic localization of amino- and carboxyl termini and a hydrophilic domain which is protease-hypersensitive, J. Biol. Chem., 263 (1988) 19105-19111. 11 Irvin, R.F. and Moor, R.M., Inositol (1,3,4,5) tetrakisphosphate-induced activation of sea urchin eggs requires the presence of inositol trisphosphate, Biochem. Biophys. Res. Commun., 146 (1987) 284-290. 12 Irvine, R.F., Moor, R.M., Pollock, W.K., Smith, P.M. and Wregget, K.A., Inositol phosphates: proliferation, metabolism and function, Phil. Trans. R. Soe. Lond. B, 320 (1988) 281-298. 13 Irvine, R.F., How do inositol 1,4,5-trisphosphate and inositol 1,3,4,5-tetrakisphosphate regulate intracellular Ca2+?, Biochem. Soc. Trans., 17 (1989) 6-8. 14 Kumar, N.M. and Gilula, N.B., Cloning and characterization of human and rat liver cDNAs coding for a gap junction protein, J. Cell Biol., 103 (1986) 767-776. 15 Le Beux Y.J., Subsurface cisterns and iameltar bodies: particular forms of the endoplasmic retieulum in the neurons, Z. Zellforsch., 133 (1972) 327-352. 16 McBurney, R.N. and Neering, I.R., Neuronal calcium homeostasis, Trends Neurosci., 10 (1987) 164-169. 17 Merritt, J.E. and Rink, T.J., Regulation of cytosolic free calcium in Fura-2-1oaded rat parotid acinar cells, J. Biol. Chem., 262 (1987) 17362-17369.
139 18 Mullaney, J.M., Chueh, S.-H., Ghosh, T.K. and Gill, D.L., Intracellular calcium uptake activated by GTP. Evidence for a possible guanine nucleotide-induced transmembrane conveyance of intraceUular calcium, J. Biol. Chem., 262 (1987) 13865-13872. 19 Nagy, J.I., Yamamoto, T., Shiosakaa, S., Dewar, K.M., Whittaker, M.E. and Hertzberg, E.L., Immunohistochemical localization of gap junction protein in rat CNS: a preliminary account. In E.L. Hertzberg and R.G. Johnson (Eds.), Modern Cell Biology, Vol. 7, Liss, New York, 1988, pp. 375-389. 20 Nahorski, S.R., Inositol polyphosphates and neuronal calcium homeostasis, Trends Neurosci., 11 (1988) 444-448. 21 Nicholson, B.J. and Zhang, J.-T., Multiple protein components in a single gap junction: cloning of a second hepatic gap junction (M~21,000). In E.L. Hertzberg and R.G. Johnson (Eds.), Modern Cell Biology, Vol. 7, Liss, New York, 1988, pp. 207-218. 22 Paul, D.L., Molecular cloning of cDNA for rat liver gap junction protein, J. Cell Biol., 103 (1986) 123-134. 23 Putney, Jr., J.W., A model for receptor-regulated calcium entry,
Cell Calcium, 7 (1986) 1-12. 24 Rosenbluth, J., Subsurface cisterns and their relationship to the neuronal plasma membrane, J. Cell BioL, 13 (1962) 405-421. 25 Shiosaka, S., Yamamoto, T., Hertzberg, E.L. and Nagy, J.I., Gap junction protein in rat hippocampus: correlative light and electron microscope immunohistochemical localization, J. Comp. Neurol., 281 (1989) 282-297. 26 Siegesmund, K.A., The fine structure of subsurface cisterns, Anat. Rec., 162 (1968) 187-196. 27 Takemura, H., Highes, A.R., Thastrup, O. and Putney, Jr., J.W., Activation of calcium entry by the tumor promoter thapsigargin in parotid acinar cells, J. Biol. Chem., 264 (1989) 12266-12271. 28 Yamamoto, T., Hertzberg, E.L. and Nagy, J.I., Antibodies against rat liver connexin32 recognize subsurface cisterns in motoneurons: immunohistochemical evidence for similarities between gap junctional and subcisternal proteins, Soc. Neurosci. Abstr., 1989, Vol. 15, Part 1,680.
