Intraneuronal localization of Nogo-A in the rat
Abstract
Nogo-A is known to be a myelin-associated protein with strong inhibitory effect on neurite outgrowth and has been considered one of the major factors that hinder fiber regeneration in the central nervous system. Recent studies have demonstrated widespread occurrence of nogo-A mRNA and Nogo-A protein in neurons. Our concurrent immunohistochemical study substantiated the widespread distribution of neuronal Nogo-A. The present study was thus focused on its intraneuronal distribution in the central nervous system, using Western blotting, immunohistochemical, and immunogold electron microscopic techniques. Western blotting of the nucleus, cytoplasm, and membrane subcellular fractions of the cerebellum and spinal cord tissues demonstrated that all three fractions contained Nogo-A. Nogo-A immunoreactivity could be identified under confocal microscope in the nucleus, perikayon, and proximal dendrite and along the cell membrane. Under the electron microscope, the perikaryonal Nogo-A immunogold particles were mainly distributed at polyribosomes and rough endoplasmic reticulum, suggesting its relationship with translation process. The immunogold particles could also be found beneath or on the plasma membrane. In the nucleus, the Nogo-A immunogold particles were found to be localized at the chromatins of the nucleus, indicating its possible involvement in gene transcription. The presence of Nogo-A in the nucleus was further supported by transfection of COS-7L cells with nogo-A. This study provides the first immunocytochemical evidence for intraneuronal distribution of Nogo-A. Apparently, the significance of Nogo-A in the central nervous system is far more complex than what has been envisioned. J. Comp. Neurol. 458:1–10, 2003. © 2003 Wiley-Liss, Inc.
With the advances achieved during the past two decades, it is the consensus today that effective regeneration of the adult mammalian central nervous system (CNS) is possible (David and Aguayo, 1981; Schwab and Bartholdi, 1996). Many factors may influence the regeneration, some permissive, but many inhibitory (reviewed by Hagg et al., 1993; Jelsma and Aguayo, 1994; Schwab et al., 1993). Among the inhibitory factors, the myelin-associated proteins have been considered to be important (McKerracher et al., 1994; Mukhopadhyay et al., 1994; Schwab and Bartholdi, 1996). Schwab and his colleagues have isolated proteins NI 35 and NI 250, which strongly inhibit neurite outgrowth (Bandtlow et al., 1993; Caroni and Schwab, 1988a,b; Spillmann et al., 1998). Administration of the monoclonal antibody against NI 35/250, IN-1, has successfully induced fiber regeneration and sprouting in injured spinal cord with functional recovery (Bregman et al., 1995; Brosamle et al., 2000; Buffo et al., 2000; Merkler et al., 2001; Thallmair et al., 1998; von Meyenburg et al., 1998; Z'Graggen et al., 1998). It has also been reported by Schwab and colleagues that IN-1 can induce sprouting of the rubrospinal tract in adult rats with bilateral complete transection of the corticospinal tracts, resulting in a nearly complete recovery of fine movements of the arms and hands (Raineteau et al., 2001).
The IN-1-responsive protein NI 250 has been cloned by three different laboratories and denominated Nogo (Chen et al., 2000; Grandpré et al., 2000; Prinjha et al., 2000). Both the human and rat Nogos have three major alternative isoforms, namely, Nogo-A, -B, and -C. It has been reported that Nogo-A is distributed mainly in oligodendrocytes, whereas Nogo-B and -C in certain neurons and several nonneural tissues (Chen et al., 2000; Grandpré et al., 2000; Goldberg and Barres, 2000). Immunocytochemical and functional studies by Chen et al. (2000) and Grandpré et al. (2000) suggest that Nogo-A may be present on oligodendrocyte cell surfaces, as well as the endoplasmic reticulum and Golgi (Chen et al., 2000). The presence of Nogo-A at the membrane of the oligodendrocyte has recently been demonstrated in the optic nerve by immunogold electron microscopy (Huber et al., 2002).
It has been briefly reported that nogo-A mRNA is also present in some neurons of the CNS in the adult animal and is particularly conspicuous at certain stages of development (Chen et al., 2000; Grandpré et al., 2000). The in situ hybridization study on nogo-A mRNA in rats by Josephson et al. (2001) has demonstrated that neurons of the adult brain are generally positive. Very recently, Schwab's group has published their study on the patterns of Nogo mRNA and protein expression in the developing and adult rat (Huber et al., 2002), describing, among other factors, the widespread expression of neuronal Nogo-A. Our concurrent Nogo-A immunohistochemical study in the rat is consistent with the general concept that Nogo-A is present in neurons of wide areas of the CNS, but with details supplementary to the previous studies. The present study, therefore, was mainly focused on the intraneuronal localization of Nogo-A, which had never been described before. The evidence from our light and electron microscopic immunocytochemistry as well as Western blotting demonstrated the occurrence of Nogo-A in the nucleus, perikaryon, and the membrane of the neuron. Its close relationship with chromatin- and mRNA rich-structures suggests possible involvement in neuronal transcriptional and translational processes.
MATERIALS AND METHODS
All efforts were made to minimize animal suffering and the number of animals used. Use of the animals was approved by the Animal Care and Use Committee of our university.
Preparation of Nogo-A polyclonal antibody
DNA constructs.
The coding sequence of KIAA0886 clone (Nagase et al., 1998) is the full-length sequence of human nogo-A. The plasmid was generously provided by the Kazusa DNA Research Institute (Chiba, Japan). The full-length coding sequence was inserted into pIRES2-EGFPN1 (Clontech, San Francisco, CA) to yield a mammalian expression vector, designated pIRES2-EGFPN1-hnogo-A. Nogo-A and -B have a common region of 172 amino acids at the N-terminus (Chen et al., 2000; Grandpré et al., 2000; Huber et al., 2002). Therefore, for raising specific antibody against Nogo-A, this common region was avoided by digesting KIAA0886 with EcoRI and NotI to yield the coding sequence of the C-terminus 572-amino acid residues (621–1,192). The fragment was inserted into pGEX-4T-3 (Amersham Pharmacia Biotech, Uppsala, Sweden) plasmid to yield a prokaryotic expression vector, designated pGEX-4T-3-Thnogo-A. The full-length sequences of nogo-B and nogo-C were prepared by polymerase chain reaction (PCR; Advantage-HF 2 PCR kit, Clontech) from human fetal brain Quick-Clone™ cDNA (Clontech) using human nogo-C-specific primer pairs 5′-gtggaattcagatggacggtcagaag-3′ and 5′-cagtcgactcagctttgcgcttcaat-3′ (EcoRI and SalI sites are underlined) and nogo-B primer pairs 5′-acggtaccagccatggaagacctggac-3′ and 5′-gagctcgagctttgcgcttcaatccag-3′ (kPnI and BamHI sites are underlined). The complete coding region of hnogo-A can also be amplified by the same primer pairs) and confirmed by sequencing. The following prokaryotic and mammalian expression vectors were constructed: pGEX-KG-hnogo-C, pEGFP N1-hnogo-C, pEGFP N1-hnogo-B, and pEGFP N1-hnogo-A.
Protein expression and antiserum raising.
Glutathione-S-transferase (GST) fusion constructs, pGEX-4T-3-Thnogo-A and pGEX-KG-hnogo-C, were expressed and purified from Escherichia coli as described by Guan and Dixon (1991) and Heape et al. (1999). The fusion protein GST-Thnogo-A was emulsified with Freund's adjuvant (Sigma, St. Louis, MO) and injected subcutaneously into New Zealand white rabbits at intervals of 2–4 weeks. The serum bleeds were evaluated by enzyme-linked immunosorbent assay (ELISA) and Western blotting. Purification of the IgGs was performed by thiophilic adsorption chromatography (Clontech). The IgGs were then passed through a column of GST-hnogo-C protein conjugated to CNBr-activated Sepharose 4B (Amersham Pharmacia Biotech, Arlington Heights, IL) to exclude the IgG fractions that recognize the epitopes of GST and the C terminus common to all three Nogo isofroms. The flowthrough was collected. The specificity of the purified anti-Nogo-A IgG was evaluated by Western blot.
Characterization of the antibody.
Two methods were used, Western blotting of transfected COS-7L cells and spinal cord, and immunocytochemistry of primary oligodendrocyte culture.
Western blotting.
Cell or tissue preparation: COS-7L cell lines (Life Technologies, Gaithersberg, MD) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS). The COS-7L cells grown in 35-mm plates at 40–80% confluency were transiently transfected with 2 μg pIRES2-EGFP N1-hnogo-A, pEGFP N1- hnogo-C, pEGFP N1- hnogo-B, and pEGFP N1- hnogo-A using Superfect reagent according to the manufacturer's instructions (Qiagen, Hilden, Germany). Sixty hours later, the cells were trypsinized or scraped off and harvested. The proteins were extracted for 2 hours at 4°C after passing through a 27-gauge needle (20 times) in 200 μl of the CHAPS buffer containing protease inhibitors (Spillmann et al., 1998). The COS-7L cells transfected with mock pIRES2-EGFP served as control. The preparation of spinal cord extract will be described later in the Distribution of Nogo-A section.
Procedure.
Aliquots of cellular lysate or tissue samples were separated by electrophoresis on 10% sodium dodecyl sulfate-polyacrylamide minigels (SDS-PAGE) under reducing conditions. The proteins were electrotransferred to polyvinylidene difluoride membranes (PVDF; 0.2 μm, Boehringer Mannheim, Germany). The membranes were treated with 1% blocking solution (w/v) in Tris-buffered saline (pH 7.4) for 1 hour and incubated overnight at 4°C in 0.5% blocking solution with the primary antibody, anti-Nogo-A (1:4,000) or anti-green fluorescent protein (GFP) polyclonal antibody (Clontech, 1:1,000). After incubation with POD-labeled secondary antibodies, the signals were revealed by BM Chemiluminescence Western Blotting kit (Boehringer Mannheim). Rainbow colored protein Markers (Bio-Rad, Hercules, CA) were used to determine the protein size.
Primary oligodendrocyte culture and immunocytochemistry.
Oligodendrocyte cultures were prepared from mechanically dissociated cerebral cortex of neonatal rat pups as described (McCarthy and DeVellis, 1980; Lang et al., 1996). The cells were grown to confluency in DMEM containing 10% FCS in polylysine-coated tissue culture flasks. The oligodendrocytes were collected and replated onto polylysine/laminin-coated coverslips at a density of ∼1,000 cells/cm2 in defined DMEM or DMEM/F12 medium (both from Life Technologies). Half of the medium was renewed three times a week. The cells on the coverslips were rinsed in phosphate-buffered saline (PBS), pH 7.2, fixed in cold 4% paraformaldehyde for 10 minutes, rinsed in PBS, and permeabilized in ice-cold 100% methanol. After rinsing, the cells were incubated overnight at 4°C with primary anti-Nogo-A antibody (1:300) diluted in 3% goat serum. The cells were rinsed and then incubated for 1 hour with Alexa Fluor 488 goat anti-rabbit IgG (1:400; Molecular Probes, Eugene, OR). After rinsing, the coverslips were mounted with Vectashield (Vector, Burlingame, CA).
Distribution of Nogo-A
The distribution of Nogo-A was first studied with immunohistochemistry at the light microscopic level. After the distribution in the perikaryons and nuclei of neurons was found, the intraneuronal distribution of Nogo-A was first verified by counterstaining the Nogo-A-immunostained tissue sections with a nucleus marker, Hoechst 33342 (Molecular Probes). In addition, Western blotting of subcellular fractions, transfection of COS cells, and immunogold electron microscopy techniques were used to investigate this issue.
Immunohistochemistry.
The animals were anesthetized by intraperitoneal injection of sodium pentobarbital (45 mg/kg) and perfused transcardially with 0.9% saline followed by ice-cold 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The brain and the spinal cord were removed, postfixed in the same fixative for 2–4 hours, and cryoprotected in 30% sucrose at 4°C for 24 hours. Twenty-micrometer sections were cut on a cryostat. The sections were treated for 1 hour in PBS containing 3% bovine serum albumin and 0.3% Triton X-100 and incubated overnight at 4°C with the primary antibody against Nogo-A (1:300). For visualization of the antibody, the sections were incubated with Alexa Fluor 488 goat anti-rabbit IgG (1:400, Molecular Probes). The sections were examined under a versatile Olympus BX60 microscope equipped with a Leica F300 digital camera and an Olympus confocal microscope FV300. For adsorption control, the Nogo-A antibody was adsorbed with 10−5.5 M of the C-terminus 572-amino acid fragment (621–1,192), which had been used for producing antibody.
Immunoelectron microscopy.
A pre-embedding immunogold-silver staining technique was used. The animals were deeply anesthetized with 1% sodium pentobarbital intraperitoneally (45 mg/kg body weight) and perfused transcardially with 150 ml warm saline, followed by a 500-ml ice-cold mixture of 4% paraformaldehyde, 0.05% glutaraldehyde, and 15% (v/v) saturated picric acid in 0.1 M phosphate buffer (pH 7.4) for 2 hours. The spinal cord was removed and postfixed by immersion in the same fixative for 3 hours at 4°C. Tissue sections of 50 μm thickness were prepared on a vibratome (Leica VT 1000S) and placed into 0.01 M PBS (pH 7.4) containing 25% sucrose and 10% (v/v) glycerol for 1 hour for cryoprotection.
After freeze-thaw treatment, the sections were immersed in 0.01 M PBS containing 5% bovine serum albumin and 5% normal goat serum for 2 hours and then incubated within rabbit anti-Nogo-A antibody, diluted to 1:1,000 in 1% bovine serum albumin and 1% normal goat serum for 48 hours at 4°C. The sections were then washed in 0.01 M PBS and incubated with second antibody (1:100, goat anti-rabbit IgG conjugated to 1.4-nm gold particles; Nanoprobes, Stony Brook, NY) at room temperature overnight. After rinsing, the sections were postfixed with 2% glutaraldehyde in 0.01 M PBS for 45 minutes. Silver enhancement was performed in the dark with an HQ Silver Kit (Nanoprobes). Before and after the silver enhancement, the sections were rinsed with deionized water several times. The sections were fixed with 0.5% osmium tetroxide in 0.1 M phosphate buffer for 1 hour and then dehydrated with graded ethanol, replaced with propylene oxide, and finally flat-embedded in Epon 812. Areas of the ventral horn of the spinal cord containing Nogo-A-like immunoreactivity were selected, trimmed under a stereomicroscope, and mounted onto blank resin stubs. Ultrathin sections were prepared on an LKB Nova Ultratome (Bromma, Sweden), counterstained with uranyl acetate and lead citrate, and examined under a JEM-100SX electron microscope. The light and electron microscopic photos were prepared with Photoshop; nothing has been added to, or deleted from, the histology.
Western blotting of brain tissue extracts and subcellular fractions.
Tissue blocks of the olfactory bulb, cerebral cortex, hippocampus, thalamus, hypothalamus, cerebellar cortex, corpus callosum, and spinal cord from the normal adult Sprague-Dawley rats (180–200 g) were dissected out on ice immediately after decapitation of the rat, frozen in liquid nitrogen, and stored at −70°C. All subsequent preparation steps were carried out at 4°C. The tissues were homogenized in CHAPS extraction buffer (Spillmann et al., 1998), containing 60 mM CHAPS, 100 mM Tris-HCl (pH 8.0), and 10 mM EDTA (pH 8.0) with a complete protease inhibitor cocktail (Boehringer Mannheim) in a ratio of 1: 50 (w/v) and left for 2 hours.The supernatant was collected after centrifugation at 15,000g for 30 minutes.
Subcellular fractions of the rat spinal cord and cerebellum were prepared by differential centrifugation as described by Hack et al. (2000). The spinal cord or cerebellum tissue was immersed in 10–20 volumes of the homogenization buffer containing 0.32 M sucrose and 5 mM HEPES-NaOH (pH 7.4), 0.2 mM calcium chloride, and a mixture of protease inhibitors. The tissue was homogenized on ice with a Teflon pestle (20 strokes) followed by a glass pestle (40 strokes). The homogenate was centrifuged at 1,000g for 10 minutes to obtain the pellet of the nuclear fraction. The supernatant was centrifuged at 9,800g for 20 minutes to obtain a membrane pellet and a cytosolic supernatant. The pellet was suspended in 5 mM HEPES with 1.5 M sucrose to lyse the possible unbroken cells. The cytosolic components were then removed by centrifugation. The nuclear or membrane fraction was resuspended in the CHAPS extraction buffer described above to half of the volume of the original homegenization buffer and extracted for 2 hours at 4°C. All the fractions thus obtained represent crude fractions of the tissues. An equal volume of each fraction (5 μl) was subjected to SDS-PAGE. The Western blotting procedure has been described above.
Cotransfection of COS-7L cells with nogo-A and the constitutively active form of Notch1.
pEGFP N1-hnogo-A and pDsRed1-N1-NICD plasmids were prepared. The enhanced green fluorescent protein (EGFP) was used as a protein marker. The DsRed1-N1-NICD served as a nuclear marker. The coding sequence of the cytoplasmic portion of the rat Notch1 receptor (amino acids 1,744–2,531) was amplified by PCR from a rat brain cDNA library (Life Technologies) and inserted into a mammalian expression vector, pDsRed1-N1 (Clontech), to create a constitutively active form of Notch1 tagged with red fluorescence protein. The constructs were confirmed by sequencing.
COS-7L cells were seeded on 25-mm coverslips and transfected with a total of 2 μg plasmid DNA consisting of 1 μg pEGFP N1-hnogo-A and 1 μg pDsRed1-N1-NICD using Superfect reagent (Qiagen). Twenty-four hours after transfection, the cells were washed in PBS and fixed for 20 minutes in 4% paraformaldehyde. The cells were then washed in PBS for 30 minutes and mounted on glass slides with Vectashield (Vector).
RESULTS
Characterization of the Nogo-A antibody
Western blot of the lysate of the mock plasmid transfected COS-7L cells showed negative staining with the Nogo-A antibody (lane 1 of Fig. 1A). That of the COS-7L cells with full-length hnogo-A transfection displayed several bands, with the topmost band at 220 kDa, which matched the size of Nogo-A (lane 2 of Fig. 1A). Western blot of the spinal cord showed a very strong band at 220 kDa and a faint band at the lower end (lane 3 of Fig. 1A).
A: Western blots for Nogo-A. Lane 1, COS-7L cells transfected with mock vector as control. Lane 2, COS-7L cells transfected with hnogo-A. Lane 3, spinal cord. The Nogo-A antibody specifically recognized a 220-kDa band, representing human and rat Nogo-A. The additional minor lower molecular weight immunoreactive bands in lane 2 may be proteolytic fragments of full-length Nogo-A protein revealed by trypsin (Spillmann et al., 1998). B: Western blots for Nogo isoforms. Lane 1, COS-7L cells transfected with mock vector as control. Lane 2, COS-7L cells transfected with enhanced green fluorescent protein (EGFP)-hnogo-A. Lane 3, COS-7L cells transfected with EGFP-hnogo-B. Lane 4, COS-7L cells transfected with EGFP-hnogo-C. Lane 5, spinal cord. The antibody specifically recognizes a 220-kDa band in the spinal cord and Nogo-A-transfected, but not Nogo-B and -C transfected, COS-7L cells.
Western blotting of COS-7L cells transfected with EGFP-fusioned hnogo-A, hnogo-B, or hnogo-C and stained with antibody against EGFP demonstrated that all three transfected nogos were expressed (data not shown). The Nogo-A antibody specifically stained hNogo-A, but not Nogo-B or -C (Fig. 1B).
The cultured primary oligodendrocytes and their processes were strongly Nogo-A immunoreactive (Fig. 2 Aa).
Confocal microscopy. A: Oligodendrocytes. a: Cultured oligodendrocytes. The cell bodies and their processes are clearly stained. Their nuclei are very weak in Nogo-A immunoreactivity (arrows). b,c: Strongly Nogo-A-immunoreactive spinal cord oligodendrocytes. Their processes wrap around myelinated fibers. Note the granular appearance of the cytoplasmic (arrowheads) and weak nuclear immunoreactivity (arrows). B: Spinal cord motor neurons. Note the nuclear immunoreactivity with spared nucleoli (arrows) and the Nissl-like cytoplasmic staining (arrowheads), extending into the proximal dendrites. Double arrowheads shows peripheral Nogo-A immunoreactivity, presumably in relation to the cell membrane. C: Cerebellar cortex. The Purkinje cells are strongly Nogo-A-immunoreactive. Note the fine granular cytoplasmic staining, the nuclear staining (arrows), and the presumable membrane staining (arrowheads) of the Purkinje cells. The granular neurons are weakly immunoreactive. For contrast, the astrocytes were double-immunostained with anti-glial fibrillary acidic protein (GFAP; red). D: COS-7L cells cotransfected with pEGFP N1-hnogo-A and pDsRed1-N1-NICD. Arrows in a–c point to the colocalized cells. a: DIC microphotograph of the COS cells. b: GFP-marked Nogo-A (green). Many of the nuclei are strongly GFP-fluorescent. Weaker fluorescence can be seen in the cytoplasm. c: Merging of the GFP-positive and DsRed-Notch1-ICD-marked nuclei (red), verifying the presence of Nogo-A in the nuclei. The difference in green/red ratio determines the different shades in the merged color. Scale bars = 25 μm in A–C; 50 μm in D.
Antibody preadsorption and omission of primary antibody controls showed negative oligodendrocytic and neuronal staining, including the membrane, cytoplasm, and nucleus. (Fig. 3).
Adsorption control of the Nogo-A antibody. A: Ventral horn of the spinal cord stained with the normal antibody. B: Stained with the preadsorbed antibody. C: A DIC microphotograph taken at the same location as in B. Scale bar = 100 μm.
General distribution of Nogo-A
Western blotting.
Western blots of the olfactory bulb, cerebral cortex, hippocampus, thalamus, hypothalamus, cerebellar cortex, corpus callosum, and spinal cord all displayed a 220-kDa band, with occasional faint bands in different tissues (Fig. 4).
Western blots of different CNS tissues. In all the tissues studied, there is a strongly stained 220-kDa band.
Immunohistochemistry.
The oligodendrocytes in the white matter of the central nervous system were strongly immunoreactive for Nogo-A. In the cross sections of the spinal cord their processes could be seen to wrap around the myelinated fibers. Confocal scanning demonstrated a granular appearance of the cytoplasm and weak immunoreactivity in the nucleus (Fig. 2 Ab,c).
Most, but not all, of the neurons were found to be Nogo-A-immunoreactive. The density of the Nogo-A immunoreactvity varied markedly among different regions. The olfactory brain was generally strongly immunoreactive, particularly the piriform cortex (Fig. 5A). The somatic motor neurons of the cranial nerves (Fig. 5B) and spinal cord ventral horn were strongly immunoreactive (Fig. 2B). The anterodorsal nucleus was the most strongly immunoreactive nucleus in the thalamus, whereas the rest of the thalamus displayed very weak, if any, Nogo-A immunoreactivity (Fig. 5C). In the cerebellum, the Purkinje cells stood out prominently. Their basal dendrites could be identified in thicker sections, less so in confocal optic sectioned pictures (see Fig. 7B). Some of the cells in the molecular layer were fairly well stained. Neurons in the granular layer stained very weakly (Fig. 2C).
Examples of Nogo-A-immunoreactive areas. A: Piriform cortex. B: Hypoglossal nucleus as an example for cranial somatic motor nuclei. C: Anterior dorsal nucleus of the thalamus. Scale bar = 500 μm.
Cerebellar cortex, confocal microscopy. Double labeling of Nogo-A immunoreactivity and Hoechst nuclear staining. A: Hoechst staining. The granular layer is strongly immunoreactive. The nuclei of the Purkinje cells are much more weakly stained, as is the case for the majority of neurons. Arrows show double-labeled nuclei of the Purkinje cells. B: Nogo-A immunoreactivity. The Purkinje cells with immunonegative nucleoli can be clearly identified. Scale bar = 200 μm.
Intraneuronal distribution of Nogo-A
Western blotting of cell fractions.
Western blots of the membrane and cytosolic and nuclear fractions of the spinal cord all displayed a strongly stained Nogo-A band. The cytosolic fraction appeared to be the densest. Different fractions of the cerebellar cortex also showed Nogo-A bands. In the cerebellum, the membrane fraction showed weakest staining. The Nogo-A band in the nuclear fraction was denser than the cytosolic fraction (Fig. 6).
Western blots of subcelluar fractions. A: Spinal cord. B: Cerebellum. Me, membrane fraction; Cyto, cytosol fraction; Nu, nucleus fraction. All three fractions from both the spinal cord and cerebellum display a Nogo-A band. In the cerebellum, the membrane fraction shows the weakest staining.
Cotransfected COS-7L cells.
Strong green fluorescence of the GFP-Nogo-A could be identified in the nuclei of the COS-7L cells. Weaker fluorescence was also present in the cytoplasm. The presence of Nogo-A in the nucleus was verified by its colocalization with the nuclear marker DsRed-NICD. (Fig. 2D).
Immunocytochemistry at the light microscopic level
Consistent with Western blotting, Nogo-A immunoreactivity could be found in the perikaryon, nucleus, and cell membrane. Although neuronal Nogo-A was found to be widely expressed, the nuclear immunoreactivity was far less so. It occurred consistently in somatic motor nuclei and the Purkinje cells of the cerebellum. However, it did not appear in all neurons. Furthermore, the relative density of cytoplasmic versus nuclear Nogo-A immunoreactivity varied. In some neurons the cytoplasm was the densest immunoreactive part, whereas in others the nuclei were the stronger counterpart. There were also neurons with primarily cytoplasmic or nuclear Nogo-A immunoreactivity. The Nogo-A immunoreactivity of the Purkinje cells appeared granular in both the cytoplasm and the nucleus (Fig. 2C). In the spinal motor neurons, the cytoplasmic Nogo-A immunoreactivity often occurred in coarse patches, bearing a resemblance to the Nissl bodies, and extended into the proximal dendrites (Fig. 2B). The nucleoli of the nuclei were negative for Nogo-A immunoreactivity. (Figs. 2B,C, 7B). Counterstaining the sections with Hoechst dye further verified the nuclear location of Nogo-A immunoreactivity (Fig. 7). With the help of the confocal microscope, a thin immunoreactive layer could often be identified along the periphery of both the Purkinje cells and the spinal motor neurons, presumably in relation to the cell membranes (Fig. 2B,C).
Immunogold electron microscopy.
The spinal cord motor neurons were studied. Most of the immunogold particles were found in the nucleus, soma, and dendrites. In the nucleus, the gold particles were distributed in close relationship to the chromatins, sparing the nucleolus (Fig. 8A). In the perikaryon, they were mainly localized at polyribosomes and rough endoplasmic reticula (Fig. 8B,C). Not infrequently the gold particles could be found beneath the membrane. A few could actually be identified on the membrane itself (Fig. 8C). A substantial amount of gold particles could also be ascertained in the proximal dendrites, mostly on polyribosomes and along or at the plasma membrane (Fig. 9).
Nogo-A immunogold electron microscopy of spinal cord motor neuron. A: Nogo-A in the nucleus. The gold particles are distributed at chromatins. The nucleolus is free from any gold particle. B,C: Nogo-A in perikaryon. The gold particles are mainly located at polyribosomes (small arrows) and rough endoplasmic reticula (large arrows). Many are found beneath or on the membrane (arrowheads in C). Scale bars = 0.5 μm.
Nogo-A immunoelectron microscopy of the proximal dendrite (Den) of a spinal cord motor neuron. Also shown is a part of a soma (Som). The immunogold particles are mainly distributed on polyribosomes (arrows). Many are located beneath or on the membrane (arrowheads). Scale bar = 1 μm.
DISCUSSION
Special care was taken to prove the specificity of the Nogo-A antibody. The 572-amino acid antigen against which the antiserum body was raised was the C-terminus of Nogo-A, avoiding the common N-terminus for both Nogo-A and -B. The antibody fractions that recognize the 188-amino acid C-terminus common for all three isoforms of Nogo were absorbed by a Nogo-C-conjugated affinity purification column. The specificity of the resultant antibody was proved by Western blots, immunostaining of cultured oligodendrocytes, and antigen adsorption control. For immunoelectron microscopy, the immunogold grains were specifically distributed at the target regions with little background, as shown in Figures 8 and 9. In Figure 8A, the nucleolus is clear of any gold grains, whereas the rest of the nucleus was heavily labeled. In Figures 8C and 9, the structures next to the labeled soma and dendrites show only individual scattered background gold grains. Specific labeling of particular organelles further supports the specificity of the immunogold reactivity.
The presence of nogo-A mRNA in neurons was briefly reported by Chen et al. (2000) and Grandpré et al. (2000). Josephson et al. (2001) reported the expression of nogo-A mRNA in neurons of wide areas of fetal and adult human and rat brain. The authors found that neurons in the adult rat brain were generally positive. Very prominent nogo-A mRNA and nogo-ABC mRNA signals were obtained from neurons of the hippocampus, piriform cortex, red nucleus, and oculomotor nucleus. Recently, Huber et al. (2002), using both in situ hybridization and immunohistochemical techniques, described the neuronal distribution of nogo-A mRNA and Nogo-A protein in the developing and adult rat. They have mainly studied the forebrain areas, cerebellum, spinal cord, retina, and dorsal root ganglion. Our Nogo-A immunohistchemical study is consistent with their conclusion that Nogo-A is widely expressed in neurons of the CNS and also adds more details. The preferential expression of Nogo-A in the olfactory cortex, somatic motor neurons, and cerebellar Purkinje cells suggests that Nogo-A may play different roles in different functional systems.
Chen et al. (2000) suggested that Nogo-A might be present on the cell surface and the endoplasmic reticulum and Golgi of oligodendrocytes. Molecular structural analysis by Grandpré et al. (2000) suggests that Nogo-A associates primarily with the endoplasmic reticulum membrane and, to a lesser extent, with the plasma membrane. In the present study, evidence from Western blotting and light and electron microscopic studies demonstrated the presence of Nogo-A in neuronal perikaryon, on the plasma membrane, and in the nucleus. Our confocal microscopic examination demonstrated Nogo-A immunoreactivity at the periphery of the neuron, suggesting a membrane relationship. However, it is beyond the resolution of confocal microscope to pinpoint the immunoreactivity on the membrane itself. Our electron microscopic study clearly demonstrated that Nogo-A immunogold particles did occur near the membrane of the somas and dendrites, some actually located at the membrane. The occurrence of Nogo-A at the cell membrane is expected from its molecular structure. The Nogo-66 domain has been reported to be the extracellular segment joining two transmembrane domains (Grandpré et al., 2000). It is possible that other transmembrane domains exist as well (Huber and Schwab, 2000; Simonen et al., 2000). In the neuronal cytoplasm, Nogo-A immunogold particles were identified at mRNA-rich polyribosomes and rough endoplasmic reticula of neuronal perikarya, suggesting that Nogo-A may be involved in translation processes.
The finding of nuclear Nogo-A immunoreactivity was unexpected. We first found it in our immunohistochemical study, and various verification approaches were then taken. The nuclear localization of Nogo-A immunoreactivity was confirmed by counterstaining with Hoechst 33342, a DNA dye. Western blotting of the nuclear fractions of the spinal cord and cerebellar tissues lent further support. The CNS tissues contain both neurons and glia cells. Thus, positive Western blotting of the nucleus fraction of brain tissue cannot by itself specifically prove the occurrence of Nogo-A in neuronal nuclei. Nevertheless, little immunostaining of oligodendrocytic nuclei would mean that the Nogo-A band in the nuclear fractions mainly represented neuronal Nogo-A. The subcellular fraction method we used is the routine molecular biological method, with which a minor perikaryon contamination of the nuclear fraction could not be ruled out. In our Western blot of the nuclear fraction of the cerebellum, it could be clearly seen that the nuclear fraction had a stronger band of the molecular size of Nogo-A than the cytosolic fraction, which could not be explained by perikaryon contamination.
Our purpose was to prove the presence of Nogo-A in neuronal nuclei, and our Western blotting data served the purpose. Analysis with PSORT software (http://psort.ims.u-tokyo.ac.jp) revealed two potential nuclear localization signals near the N- and C-termini of Nogo-A. Our COS-7L transfection experiments proved that, indeed, Nogo-A could be translocated into the cell nuclei. Our immunogold electron microscopic study further localized the Nogo-A to chromatins, indicating that it may be involved in regulation of gene transcription. It is thus likely that its presence in the nucleus is dynamically influenced by the status of neurons for particular purposes.
The widespread neuronal distribution of Nogo-A, particularly the multiple intracellular topologies in the CNS, as demonstrated in the present study, reveal a new spectrum for Nogo-A function. Apparently the significance of Nogo-A in the CNS is far more complex than what has been envisioned. Further evidence is needed before more solid speculations on its functions can be substantiated.
Acknowledgements
The authors are grateful for the generous donation of the human nogo-A cDNA clone from the Kazusa DNA Research Institute (Chiba, Japan). We appreciate the technical assistance of Dr. Bin Lang and Ms. M.L. Zhang.
