- Email: [email protected]
calmodulin-dependent protein kinase kinase β in the rat central nervous system
Neuroscience Research 39 (2001) 175– 188 www.elsevier.com/locate/neures
Immunohistochemical localization of Ca2 + /calmodulin-dependent protein kinase kinase b in the rat central nervous system Yasuhisa Nakamura a,*, Sachiko Okuno b, Takako Kitani b, Kazuyoshi Otake a, Fumi Sato a, Hitoshi Fujisawa b a
Section of Neuroanatomy, Graduate School of Medical and Dental Research, Tokyo Medical and Dental Uni6ersity, Yushima, Bunkyo-ku, Tokyo 113 -8519, Japan b Department of Biochemistry, Asahikawa Medical College, Midorigaoka, Asahikawa 078 -8510, Japan Received 17 August 2000; received in revised form 5 October 2000; accepted 6 October 2000
Abstract We examined regional and intracellular distribution of Ca2 + /calmodulin-dependent protein kinase kinase b (CaM-KK b), which activated Ca2 + /calmodulin-dependent protein kinase I and IV (CaM-K I and IV) immunohistochemically in the central nervous system of the rat by light and electron microscopy. Although most neurons in the brain and spinal cord exhibited the immunoreactivity, no labeled neurons were observed in the globus pallidus or entopeduncular nucleus, and only a small number of neurons showed weak immunoreactivity in the substantia nigra pars reticulata. In general, the immunoreactivity was observed both in the cytoplasm and cellular nucleus, although the immunoreactivity was not found in the cellular nucleus in some large neurons such as in the mesencephalic trigeminal nucleus, lateral vestibular nucleus or gigant cellular reticular formation. As to motoneurons of the cranial nerve nuclei and the anterior horn of the spinal cord, they revealed the immunoreactivity both in the cytoplasm and nucleus. The reaction product appeared as fine granules in the cytoplasm and nucleus under light microscopy. Electron microscopic observations confirmed that the reaction product was localized mainly on the Golgi apparatus or on the nuclear chromatin. Immunolabeling for antibody against CaM-KK b was discussed with the distribution of CaM-K I, IV and another CaM-KK, CaM-KK a, in the central nervous system. © 2001 Elsevier Science Ireland Ltd and the Japan Neuroscience Society. All rights reserved. Keywords: CaM-kinase I; CaM-kinase IV; CaM-kinase kinase a; CaM-kinase kinase b; Immunohistochemistry
1. Introduction There have been several investigations concerning Ca2 + /calmodulin-dependent protein kinase kinase
(CaM-KK), which activates CaM-K IV and CaM-K I by phosphorylation (Okuno and Fujisawa, 1993; Okuno et al., 1994; Sugita et al., 1994); these CaM-Ks play important roles as Ca2 + -responsive multifunc-
Abbre6iations: 4, trochlear nucleus; 5, motor trigeminal nucleus; 7, facial nucleus; AD, anterodorsal nucleus; AH, anterior horn; AON, anterior olfactory nucleus; AV, anteroventral nucleus; BL, basolateral amygdaloid nucleus; BM, basomedial amygdaloid nucleus; CA1, CA1 sector of Ammon’s horn; CA3, CA3 sector of Ammon’s horn; CA4, CA4 sector of Ammon’s horn; CblCx, cerebellar cortex; CbrCx, cerebral cortex; Ce, central amygdaloid nucleus; Cpu, caudate-putamen; DG, dentate gyrus; dTh, dorsal thalamus; EPl, external plexiform layer; Gl, glomerular layer; GP, globus pallidus; GrC, cerebellar granule cell layer; GRF, gigantocellular reticular formation; GrO, olfactory granule cell layer; Hip, hippocampal formation; Hyp, hypothalamus; IC, inferior colliculus; IGL, intermediate gray layer; IOd, dorsal accessory olive; IOm, medial accessory olive; IOp, principal olive; IP, cerebellar interpositus nucleus; Lat, cerebellar lateral nucleus; LC, locus coeruleus; LD, laterodorsal nucleus; lfp, longitudinal fasciculus of pons; LVN, lateral vestibular nucleus; MD, mediodorsal nucleus; Me, medial amygdaloid nucleus; Med, cerebellar medial nucleus; Mi, mitral cell layer; Mol, molecular layer; NRTP, pontine tegmental reticular nucleus; P, Purkinje cell layer; PBNm, medial parabrachial nucleus; Pi, piriform cortex; PN, pontine nuclei; py, pyramidal tract; RN, red nucleus; Rt, thalamic reticular nucleus; SC, superior colliculus; SGL, superficial gray layer; SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticulata; VA, ventral anterior nucleus; Vmes, mesencephalic trigeminal nucleus. * Corresponding author. Tel.: + 81-3-5803-5148; fax: + 81-3-5803-5151. E-mail address: [email protected] (Y. Nakamura). 0168-0102/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd and the Japan Neuroscience Society. All rights reserved. PII: S0168-0102(00)00209-1
176
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
tional protein kinases, in controlling a variety of neuronal functions in response to an increase in intracellular Ca2 + . Because CaM-KK was first investigated to phosphorylate CaM-K IV, it has been called CaM-K IV K (Okuno and Fujisawa, 1993). Independently, CaM-K I activator was also found, which phosphorylated CaM-K I and CaM-K IV (Lee and Edelman, 1994; Selbert et al., 1995). When crude extracts of rat cerebral cortex, brainstem, cerebellum and the bacteria transformed with an expression vector (pET) carrying CaM-K IV K cDNA were immunotitrated with antiserum to the cloned enzyme (CaM-K IV K), almost all the activity of the recombinant enzyme was immunoprecipitated (Okuno et al., 1996). During the study of immunotitration, it became apparent that only approximately 46, 56, and 25% of the activity of cerebral cortex, brainstem, and cerebellum, respectively, were immunoprecipitated by addition of an excess of the antiserum (Okuno et al., 1996). Thus, the cloned CaM-K IV K accounted for only one-fourth of total CaM-KK activity in the cerebellum where CaM-K IV exists most abundantly (Ohmstede et al., 1989; Jensen et al., 1991). This investigation has suggested that most of the activity is attributable to other CaM-KK. Then another isoform of CaM-K IV K was demonstrated in the rat brain. Thereafter, the first CaM-K IV K was named CaM-KK a (Okuno et al., 1996), and another one, which has been purified and characterized was named CaM-KK b (Okuno et al., 1997; Anderson et al., 1998). The distribution of CaMKK a was demonstrated by means of immunohistochemistry by using the antibody against CaM-KK a, which was located in the neuronal nuclei of most neurons with uniform density and also in the cytoplasm in some regions (Nakamura et al., 1996). In this study, newly developed antibody against CaM-KK b (Kitani et al., 1997) was used for immunohistochemistry to localize this enzyme in the rat central nervous system by light and electron microscopy.
paraformaldehyde solution or 3.5% paraformaldehyde and 0.1% glutaraldehyde mixture. After removal, the brains were cut serially at 40 mm into frontal or sagittal sections on a freezing microtome for light microscopy after soaking in 30% phosphate buffered sucrose for 2 days. For electron microscopy, brains were cut at 100 – 150 mm immediately after perfusion with a microslicer. Procedures for immunohistochemistry for light and electron microscopy were described previously (Nakamura et al., 1995, 1996). Briefly, free floating sections in 0.1 M phosphate buffer containing 0.1% Triton X-100 were incubated overnight in a refrigerator with purified antibody (1:2000 –3000). Then the sections were incubated in biotinylated anti-rabbit immunoglobulin and reacted with Vectastain ABC kit (Elite, Vector Labs, CA, USA). To detect peroxidase, 0.05% 3,3%-diaminobenzidine (DAB) was used with 0.003% H2O2, 0.005% nickel chloride and cobalt acetate in the phosphate buffer. For control, some sections were processed omitting primary antibody in the incubation, or were processed with immunoglobulin G (IgG) obtained from normal rabbit serum instead of the antibody. Brain structures were identified according to the rat stereotaxic atlas by Paxinos and Watson (1998). For light microscopy, the sections were mounted on gelatinized glass slides; some of them were counterstained with 0.5% Cresyl violet. For electron microscopy, target nuclei for the observation were trimmed from the immunostained sections and embedded in Epon after osmication, dehydration, and block-staining with 1% uranyl acetate in absolute ethanol. Ultrathin sections were examined under a Hitachi H-7100 electron microscope with or without uranyl acetate –lead citrate staining. In addition to the immunohistochemistry for CaMKK b, we stained frozen sections of rat brains with antibody against CaM-K I, which was made from a synthetic peptide, corresponding to the carboxyl-terminal 20 amino acids of CaM-K I. Crude serum obtained from immunized rabbits was purified with affinity chromatography.
2. Materials and methods Molecular cloning and sequencing of CaM-KK b was made (Kitani et al., 1997). Thereafter, rabbit antibody against CaM-KK b was prepared by immunization of a peptide corresponding to the carboxyl-terminal 24 amino acids of CaM-KK b purifying crude serum obtained from immunized rabbits by means of affinity chromatography, and the specificity of this antibody was reported (Kitani et al., 1997). Eleven rats (male, Wistar, Nippon Bio-Supp. Center, Tokyo, Japan, 7 weeks) were used in the immunohistochemical study. The rats, under deep pentobarbital anesthesia (50 mg/kg, i.p.), were killed by perfusion through the ascending aorta with phosphate buffered 4%
3. Results We examined the immunohistochemical reaction with an antibody against CaM-KK b (Fig. 1a). Although two kinds of perfusate, with glutaraldehyde or without glutaraldehyde, were used, the results apparently did not differ. Most neurons in the central nervous system showed immunoreactivity. Control sections, which were reacted without primary antibody did not show any immunostaining in the central nervous system (Fig. 1b). Control sections reacted with IgG showed similar feature as observed without primary antibody. During this study, immunolabeling was observed in the cytoplasm,
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
and in many neurons, it was also observed in cellular nucleus; however, the intensity of staining was not uniform. The immunohistochemical reaction for the antibody against CaM-KK b in the central nervous system is summarized in Table 1.
3.1. Telencephalon In the main olfactory bulb, most mitral cells and granule cells were stained moderately (Fig. 2). In the glomerular layer, fine fibers surrounding the glomeruli were stained densely (Fig. 2). Mitral cells and granule cells in the accessory olfactory bulb showed the same appearance as those in the main olfactory bulb. In the olfactory tubercle, pyramidal neurons were stained most densely, which is shown in Fig. 1a. In the cerebral neocortex, immunoreactivity of neurons was dense in layers II and III, and neurons in layers IV – VI showed less reactivity (Fig. 3). Stainability was rather uniform in every area of the cerebral cortex. Some large pyramidal neurons showed only cytoplasmic staining, although many smaller neurons showed immunostained cellular nuclei in immunolabeled cytoplasm. Labeled radial strands, which were the apical dendrites of pyramidal neurons were also found among neuronal somata (Fig. 3). Electron microscopy of these labeled dendrites showed reaction product adjacent to microtubules or near mitochondria. In the hippocampal formation, weak to moderate immunoreaction was observed in the pyramidal cells of the subiculum and Ammon’s horn, and granule cells of the dentate gyrus (Fig. 4). Well
177
stained processes extending from the pyramidal layer to the stratum radiatum were observed in the CA3 and CA4 sectors of the Ammon’s horn. Occasionally, a few densely stained probable interneurons were observed in the stratum radiatum. Among the limbic cortices, pyramidal neurons in the piriform cortex showed densest stainability, although in the other limbic cortices immunoreaction appeared as similar as in the neocortices. Comparing the immunostaining in various cortical regions, pyramidal neurons in the olfactory tubercle and piriform cortex showed the densest reactivity, and those in most neocortical regions showed not so dense as those in the above two regions. Pyramidal neurons in the hippocampal formation (CA1 –CA4) showed weaker immunostaining than in the other cortical regions. In the subcortical telencephalic nuclei, such as the septal nuclei, amygdaloid body and caudate-putamen, immunoreactivity was observed but not homogeneous within each subnucleus. In the septal nuclei, staining was rather weak except for the dorsal part of the lateral septal nucleus, where some neurons were well stained both in the somata and primary dendrites. In the amygdaloid body, the central nucleus was stained densest (Fig. 5), and the other nuclei such as the basolateral nucleus appeared less dense, and the anterior or cortical nuclei was stained weakest. Immunoreaction was observed both in the cytoplasm and nucleus in these amygdaloid neurons. In the caudate-putamen, both medium-sized and large neurons were stained moderately (Fig. 6). No immunoreaction in the cytoplasm or nucleus was observed in the globus pallidus (homologous to the primate external segment of the globus pallidus) or in the entopeduncular nucleus (the internal segment of the primate globus pallidus) (Fig. 6). In the claustrum, moderate labeling of neurons was observed. In the preoptic area, immunostaining was weak.
3.2. Diencephalon
Fig. 1. Sagittal sections of the rat brain stained immunohistochemically with (a) and without (b) the antibody against CaM-KK b. Strong reactivity (a) and no reactivity (b) are shown from the anterior olfactory nucleus to the lower brainstem. Counterstain was not made for any of the light microscopic sections shown in Figs. 1–17. Scale bar= 5 mm.
In the dorsal thalamus neurons in the anterodorsal (AD), laterodorsal (LD), ventroanterior-ventrolateral complex (VA –VL) and midline nuclei were labeled moderately, and those in other nuclei such as the intralaminar or ventral posterior nuclei, were weakly labeled (Fig. 7). In the hypothalamus, labeling of neurons in the suprachiasmatic, paraventricular and mammillary nuclei was moderate, and in other nuclei such as the anterior or ventromedial nuclei, it was weak. In the habenular nucleus, neurons in the medial nucleus showed moderate labeling, but in the lateral nucleus very weak. Neurons in the thalamic reticular nucleus (Fig. 7) and subthalamic nucleus were moderately labeled, and neurons in the zona incerta were stained weakly.
178
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Table 1 The immunohistochemical reaction for the antibody against CaMKK b in the central nervous systema Regions of the CNS examined
Relative intensity of immunostaining
Olfactory system Main olfactory bulb Mitral cells Granule cells Glomerular layer
++ ++ +++
Accessory olfactory bulb Mitral cells Granule cells Anterior olfactory nucleus
++ ++ ++
Olfactory tubercle Pyramidal cells Islands of Calleja
++++ ++
Neocortex Frontal cortex Parietal cortex Temporal cortex Occipital cortex
(I–IV) (I–IV) (I–IV) (I–IV)
Limbic cortex Piriform cortex Cingulate cortex
+++, +++, +++, +++,
(V–VI) (V–VI) (V–VI) (V–VI)
++ ++ ++ ++
Entorhinal cortex Subicular cortex Dentate gyrus Ammon’s horn
++++ (II, IV) +++, (I, III–VI) ++ (II, IV) +++, (I, III–VI) ++ (II, IV) +++, (I, III–VI) ++ +++ ++ ++ (CA1–2) ++, (CA3–4) +
Septum Lateral septal nucleus Medial septal nucleus Nuclei of diagonal band Bed nucleus of the stria terminalis
+/++ + + +
Preoptic area Medial preoptic nucleus Lateral preoptic nucleus
+ −/+
Amygdala Anterior amygdaloid nucleus Cortical amygdaloid nucleus Medial amygdaloid nucleus Lateral amygdaloid nucleus Basolateral amygdaloid nucleus Basomedial amygdaloid nucleus Central amygdaloid nucleus
+ + + + ++ + +++
Basal ganglia Caudate-putamen Nucleus accumbens Globus pallidus Ventral pallidum Entopeduncular nucleus Subthalamic nucleus Claustrum Substantia nigra pars reticulata pars compacta Ventral tegmental area
++ ++ − − − ++ ++ −/+ ++ ++
Retrosplenial cortex Perirhinal cortex
Table 1 (Continued) Regions of the CNS examined
Relative intensity of immunostaining
Thalamic reticular nucleus Zona incerta
++ +
Hypothalamus Anterior hypothalamic area Lateral hypothalamic area Suprachiasmatic nucleus Supraoptic nucleus Paraventricular nucleus Arcuate nucleus Ventromedial nucleus Dorsomedial nucleus Premammillary nucleus Medial mammilary nucleus Lateral mammillary nucleus Posterior hypothalamic area
+ + ++ + ++ + + + ++ ++ ++ +
Habenula Medial habenular nucleus Lateral habenular nucleus
++ −/+
Dorsal thalamus Anterodorsal nucleus ++ Anteroventral nucleus + Anteromedial nucleus + Mediodorsal nucleus + Ventral anterior and ventral lateral ++ complex Ventroposteiror nuclei + Ventral medial nucleus + Laterodorsal nucleus ++ Midline nuclei ++ Intralaminar nuclei + Parafascicular nucleus + Posterior thalamic nuclei + Lateral geniculate nucleus + Medial geniculate nucleus + Lower brainstem Red nucleus Interstitial nucleus of Cajal Superior colliculus Superficial gray layer Deep gray layers
+++ + ++ +
Inferior colliculus External nucleus ++ Central nucleus ++ Pedunculopontine tegmental nucleus+ Medial parabrachial nucleus + Lateral parabrachial nucleus ++ Interpeduncular nucleus +++ Locus coeruleus ++++ Trigeminal nuclei Mesencephalic nucleus Motor nucleus Principal sensory nucleus Spinal nucleus Superior olivary complex Trapezoid nucleus Pontine reticular formation Pontine tegmental reticular nucleus Pontine nuclei Central gray Raphe nuclei
++ ++ ++ +/++ ++ ++ + +++ +++ + ++
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
179
Table 1 (Continued) Regions of the CNS examined
Relative intensity of immunostaining
Medullary retiuclar formation (GRF etc.) Lateral reticular nucleus Vestibular nuclei Cochlear nuclei Inferior olivary complex Prepositus hypoglossi nucleus External cuneate nucleus Dorsal column nuclei Nucleus of the solitary tract Motor nuclei of the cranial nerves
+++ +++ ++ +++ ++ ++ ++ ++ + ++
Cerebellum Cerebellar nuclei Purkinje cells Granule cells Molecular layer
++ + + +++
Spinal cord Layer I of the posterior horn Other layers of the posterior horn Reticular nucleus Inermediate zone Anterior horn
++ + + + ++
a ++++, +++, ++ and + indicate densest, dense, moderate and weak immunostainability, respectively.
Fig. 3. The frontal sensory-motor cortex (layers III– VI) shows that some large neurons (arrowheads) are devoid of staining in their nuclei. Apical dendrites are well stained as vertical arrays. Scale bar =100 mm.
3.3. Mesencephalon In the mesencephalon, superior colliculus neurons showed weak to moderate labeling (Fig. 8). Neurons in the superficial gray layer were stained denser than those in the other layers. Most large neurons in the intermediate gray layer showed no immunolabeling in their cellular nuclei. In the inferior colliculus, neurons both in the central and external nuclei were stained moderately. Neurons in the central gray matter were weakly labeled. In the tegmentum, cytoplasm of the magnocellular red nucleus neurons were densely stained, and nuclei of some neurons showed denser staining than the cytoplasm, while nuclei of some neurons were devoid of immunostaining (Fig. 9). Although abundant substantia nigra pars compacta neurons and ventral tegmental area neurons were stained moderately, only a small number of neurons in the substantia nigra pars reticulata was weakly stained (Fig. 9). Axons in the crus cerebri showed dense immunolabeling. At the pontomesencephalic junction, the pedunculopontine tegmental nucleus neurons were weakly stained both in the cytoplasm and nucleus. Neurons in the mesencephalic reticular formation were also weakly labeled. Description of the motoneruons will be summarized below.
3.4. Pons and cerebellum
Fig. 2. A coronal section of the olfactory bulb shows fine fibers labeled densely surrounding the olfactory glomeruli (Gl). Scale bar = 100 mm.
In the pons, neurons in the pontine nuclei were stained densely (Fig. 10). Pontine tegmental reticular nucleus neurons were also densely stained (Fig. 10); in these nuclei both the cytoplasm and nucleus of neurons showed immunoreactivity. In addition to neuronal labeling, corticofugal axons in the pons were labeled
180
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Fig. 4. The dorsal hippocampal formation shows immunolabeling in the dentate gyrus (DG) and Ammon’s horn (CA1– CA4). Arrowheads point to neuronal labeling in the stratum radiatum. Scale bar = 500 mm.
densely. Neurons in the medial and lateral parabrachial nuclei were stained weakly and moderately, respectively (Fig. 11). In the locus coeruleus, almost all neurons were very densely stained (Fig. 11); the stainability was heavier in the nucleus than the cytoplasm. As to the trigeminal nuclei, large neurons of the mesencephalic trigeminal nucleus showed moderate staining in the cytoplasm, although cellular nuclei were devoid of labeling (Fig. 11). In the main sensory nucleus, neurons were also moderately labeled. Most neurons in the superior olivary nuclear complex and the nucleus of the trapezoid body were labeled moderately both in the cytoplasm and nucleus. Pontine reticular formation neurons were weakly labeled. The cerebellar cortex appeared very dense under low magnification (Fig. 1a). In the Purkinje cell layer, Purkinje cells were observed to be stained weakly both in the cytoplasm and nucleus (Fig. 12). In the granular layer, weakly labeled granule cells appeared like small rings, which suggested that only a thin layer of cytoplasm was labeled, and any labeling of cellular nuclei was not observed (Fig. 12). It was hard to identify immunolabeled larger neurons such as Golgi cells in this layer. Although the molecular layer was stained densely, no labeled cell bodies were discernible, but only fine strands of fibers, presumably parallel fibers, ran vertically and horizontally. Electron microscopy proved that these immunolabeled fibers were parallel fibers and their axon terminals (Fig. 13). Packs of fibers of small diameter, which was a characteristic feature of the parallel fiber, contained densely stained microtubules, and some labeled axon terminals included densely stained synaptic vesicles. These axon terminals were characterized by round synaptic vesicles in them and asymmetric synaptic contacts. In the cerebellar medial, interpositus and lateral nuclei, moderately labeled neurons were seen — most cellular nuclei were not labeled (Fig. 14).
3.5. Medulla oblongata and spinal cord In the medulla oblongata, immunoreactivity of neurons was observed in many nuclei such as the inferior olivary complex, dorsal column nuclei and medullary reticular formation. In the lateral vestibular nucleus, cellular nuclei of most large neurons were not stained
Fig. 5. Part of the amygdaloid body shows different stainability among subnuclei (BL, BM, Ce and Me). Scale bar = 500 mm.
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
181
showed immunoreactivity (Fig. 17a). The cytoplasm of all neurons was stained densely and almost all cellular nuclei were stained much denser. Neurons of the motor trigeminal (Fig. 17b) and abducens nuclei showed similar stainability as observed in the oculomotor and trochlear nuclei. In the facial (Fig. 17c), ambigual and hypoglossal nuclei, the appearance of immunoreactivity was similar as other motor cranial nerve nuclei. Motoneurons in the anterior horn of the cervical spinal cord also showed the same stainability of both the cytoplasm and nucleus as other cranial motoneurons (Fig. 17d).
3.7. Electron microscopic obser6ations We observed neurons of the cerebral cortex, caudateputamen, cerebellar cortex, locus coeruleus and motoneurons of the trigeminal and facial nerve nuclei electron microscopically. At a lower magnification, reaction product was observed in the cytoplasm and the nucleus as electron dense fine grains, which were less than 0.3 mm in diameter. In the nucleus, reaction product was located mostly on dense aggregations of nuclear chromatin structures as if immunoreaction was carried to restricted areas (Fig. 18). In the cytoplasm,
Fig. 6. Densely stained caudate-putamen (CPu) contrasts with the globus pallidus (GP) which does not show any immunoreaction. Scale bar= 500 mm.
(Fig. 14). Neurons in the cochlear nuclei, the cytoplasm and nucleus were densely stained. Immunoreactivity of the nucleus of the solitary tract was weak. In the spinal trigeminal nucleus, many medium sized or small neurons revealed weak or moderate cytoplasmic and nuclear staining, although nuclei of some large neurons were not stained. Neurons in the inferior olivary nuclear complex, both the cytoplasm and nucleus were stained moderately (Fig. 15). Dorsal column nuclei neurons were also labeled moderately both in the cytoplasm and nucleus. In the medullary reticular formation, neurons were stained densely (Fig. 16). Although many neurons in the gigantocellular reticular formation did not show nuclear staining, medium-sized or small neurons in the parvocellular, lateral, and paramedian reticular nucleus showed immunoreactivity both in the cytoplasm and nucleus. In the upper cervical spinal cord, most neurons in the dorsal horn, intermediate zone and ventral horn were immunolabeled both in the cytoplasm and nucleus. Among them, layer I neurons in the posterior horn and ventral neurons were stained rather densely than the others.
3.6. Motoneurons in the cranial ner6e nuclei and the spinal cord Motoneurons in the oculomotor and trochlear nuclei
Fig. 7. The rostral part of the dorsal thalamus does not show uniform labeling among subnuclei (AD, AV, LD, MD and VA). Scale bar = 500 mm.
182
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Fig. 10. The basal part of the pontine region shows immunolabeling of neurons in the pontine nuclei (PN) and pontine tegmental reticular nucleus (NRTP). Longitudinally oriented fibers (lfp) are labeled moderately. Scale bar =200 mm.
CaM-KK b was synthesized in the endoplasmic reticulum and transported to and concentrated in the Golgi apparatus. Fig. 8. Part of the superior colliculus shows weak immunolabeling in the superficial (SGL) and intermediate gray layers (IGL). Some labeled neurons are devoid of nuclear staining (arrowheads). Scale bar= 100 mm.
immunoreactive materials were located mainly on vacuoles or vesicles of the Golgi apparatus, Nissl substance or adjacent to mitochondria. It was conceivable that
3.8. CaM-K I immunohistochemistry As has been mentioned by Picciotto et al. (1995), immunolabeling of CaM-K I was localized in the cytoplasm of neurons in the rat central nervous system; no immunolabeled cellular nucleus was found. Although
Fig. 9. The mesencephalic tegmentum shows weak labeling in the substantia nigra pars compacta (SNc) and almost no labeling in the substantia nigra pars reticulata (SNr) although a part of its neuropile is stained densely. Many magnocellular red nucleus (RN) neurons are devoid of nuclear staining (arrowheads). Scale bar =500 mm.
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
immunolabeled neurons were localized in the central nervous system as in the previous study (Picciotto et al., 1995), immunoreaction of neuronal cell bodies was moderately immunolabeled in the territories of the globus pallidus, entopeduncular nucleus and substantia nigra pars reticulata, in rather thick stained neuropile, where labeled neurons had not been observed (Picciotto et al., 1995). A table (Table 2) and a figure (Fig. 19) summarize these results comparing the stainability for CaM-KKs and CaM-Ks in some representative nuclei. This figure
183
shows different staining of the cytoplasm and cellular nucleus with antibodies against CaM-KK b (Fig. 19a and b), CaM-K I (Fig. 19c) and CaM-K IV (Fig. 19d) at higher magnification. An antibody against CaM-KK b stained the cytoplasm alone (Fig. 19a) or both the cytoplasm and nucleus (Fig. 19b), however, an antibody against CaM-K I stained exclusively the cytoplasm (Fig. 19c), and antibodies against CaM-K IV (Fig. 19d) and CaM-KK a stained chiefly the cellular nucleus. Because nuclear staining for CaM-KK a (Nakamura et al., 1996) was virtually the same as that for CaM-K IV, this finding was not shown here. Fig. 19d was taken from a histological section, which was prepared before (Nakamura et al., 1995). 4. Discussion
Fig. 11. The dorsal part of the pontine region shows dense labeling of the locus coeruleus neurons (LC) both in the cytoplasm and nucleus, and only cytoplasmic labeling in the mesencephalic trigeminal nucleus neurons (Vmes). Scale bar =100 mm.
In the present paper, we described the distribution of CaM-KK b-like immunolabeling in the rat central nervous system. This enzyme is known to activate CaM-K IV (Kitani et al., 1997; Okuno et al., 1997) and also CaM-K I (Okuno et al., 1997) by phosphorylation. Another isoform of CaM-KK b had already been cloned and named as CaM-KK a (Okuno et al., 1996), whose distribution in the central nervous system in the rat has been reported (Nakamura et al., 1996). The distribution of CaM-KK b was, however, quite different from that of CaM-KK a, which was localized mainly in the cellular nuclei of almost all neurons (Nakamura et al., 1996). In addition to the activation of CaM-K IV, both CaM-KK a and CaM-KK b also phosphorylate CaM-K I, which is another Ca2 + / calmodulin-dependent multifunctional protein kinase, and has been reported to be activated upon phosphorylation by CaM-K I K (Lee and Edelman, 1994; Selbert et al., 1995). However, exact location of CaM-K I K in the brain or in cell organelles has not been reported. We, therefore, will not discuss the phosphorylation of CaM-Ks by CaM-K I K at present.
4.1. Relationship of CaM-KK i with CaM-K I and CaM-K IV
Fig. 12. A part of the cerebellar cortex shows weak labeling of the Purkinje cells (P). Because of dense labeling of parallel fibers, the molecular layer (white bar with Mol) appears very dark. In the granular layer (GrC) the granule cell cytoplasm shows a ring-like appearance. Scale bar =50 mm.
Because CaM-KKs phosphorylate CaM-K I and CaM-K IV to activate their function, it is reasonable to believe that their locations are in the same brain areas and in the same cellular organelles. CaM-K I is distributed in the cytoplasm throughout the rat central nervous system from the telencephalon, diencephalon, mesencephalon, cerebellum, lower brainstem to the spinal cord (Picciotto et al., 1995). Although neurons in the substantia nigra was not reported to be labeled immunohistochemically, we found that neurons in both pars compacta and reticulata of the substantia nigra had been labeled moderately. In addition, neurons in the globus pallidus and entopeduncular nucleus had
184
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Fig. 13. An electron micrograph shows the molecular layer of the cerebellum. Some labeled parallel fibers contained densely stained microtubules. Arrows point to labeled parallel fiber terminals in which round synaptic vesicles are observed. Scale bar = 0.5 mm.
also been labeled, although results of immunoreaction in these nuclei were not described in their report (Picciotto et al., 1995). In the present study CaM-KK b-like immunoreactivity was found to be located in the cytoplasm and nucleus of most neuron groups in the brain and spinal cord, except in the globus pallidus, entopeduncular nucleus and substantia nigra pars reticulata. CaM-KK b, therefore, can activate CaM-K I in all neurons in the central nervous system, except in the above three basal ganglia. The distribution of CaM-K IV was not uniform in the central nervous system (Nakamura et al., 1995). Dense immunostaining was observed in the cerebral cortex, hippocampal formation, caudate-putamen, dorsal thalamus, cerebellar granular layer, and the dorsal horn of the spinal cord. Immunolabeled neurons, however, were not observed in the globus pallidus, entopeduncular nucleus, substantia nigra (both pars compacta and pars reticulata), ventral tegmental area, medial habenular nucleus, interpeduncular nucleus, mesencephalic trigeminal nucleus, principal sensory trigeminal nucleus, dorsal column nuclei, motor cranial nerve nuclei, or the anterior horn of the cervical spinal cord. Contrary to CaM-K I, CaM-K IV was localized mainly in the cellular nucleus, and also in the cytoplasm in some neurons such as cerebral pyramidal, rubral, lateral vestibular and cerebellar granule neurons. Nuclear CaM-K IV, therefore, can be phosphorylated by CaMKK a and CaM-KK b, and cytoplasmic CaM-K IV by CaM-KK b when it is co-localized in neurons. As to the location of CaM-KK b in neurons, sometimes it appeared both in the cytoplasm and nucleus. In some nuclei, such as the cerebral cortex or lateral vestibular nucleus, CaM-KK b was localized in the cytoplasm only in some large neurons, but was local-
ized both in the cytoplasm and cellular nucleus in other neurons. These observations suggest that CaM-KK b can be translocated from the cytoplasm to the nucleus easily, the mechanism yet unknown.
4.2. Polyclonal and monoclonal antibodies against CaM-KK i Recently, an immunohistochemical study dealing with the localization of CaM-KK b in the rat brain was published (Sakagami et al., 2000). The authors raised a novel monoclonal antibody against CaM-KK b and used it for their study. The result of their report was quite different from our present study. They described immunolabeling as restricted to the cytoplasm throughout the brain. In the cerebral cortical region, the difference between the stainability with the two antibodies was not so manifest. In the basal ganglia immunoreac-
Fig. 14. A coronal section of the cerebellar nuclear region shows weak to moderate labeling of the cerebellar medial (Med), interpositus (IP) and lateral (Lat), and lateral vestibular (LVN) nuclear neurons. Scale bar =500 mm.
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Fig. 15. Neurons of the inferior olivary nuclear complex (IOd, IOm and IOp) show moderate labeling. Descending fibers in the pyramis (py) are also labeled moderately. Scale bar = 100 mm.
Fig. 16. Some large neurons in the gigantocellular reticular formation (GRF) are devoid of nuclear staining (arrowheads). Scale bar =100 mm.
tivity of neurons in the globus pallidus, entopeduncular nucleus and substantia nigra pars reticulata was shown with the monoclonal antibody, although we did not observe it with the polyclonal antibody. Purkinje cells in the cerebellar cortex, and neurons in the cerebellar nuclei were reportedly devoid of labeling with the monoclonal antibody. One of the most conspicuous findings with the monoclonal antibody was that immunoreaction for motoneurons of the cranial nerve nuclei and the anterior horn of the spinal cord was not detected.
185
With the polyclonal antibody, we revealed cytoplasmic and nuclear immunolabeling of motoneruons. Considering these observations, we could not conclude that the monoclonal and polyclonal antibodies showed the same antigen –antibody reaction in the brain and spinal cord. Here we discuss the immunohistochemical different expressions between monoclonal and polyclonal antibodies. The monoclonal antibody was produced by using amino acid residues 520 –587 near the carboxyl terminal (Sakagami et al., 2000), and polyclonal antibody by using amino acid residues 564 –587 (Kitani et al., 1997). Because both antibodies showed the same Western Blot pattern of doublet near 67 kDa, coinciding with that of CaM-KK b (Kitani et al., 1997), these two antibodies are thought to react specifically with CaM-KK b. When cDNA for CaM-KK b was expressed in Escherichia coli, the crude extract showed the ability to phosphorylate the CaM-K IV peptide, and showed relatively high Ca2 + /calmodulin-independent activity (Kitani et al., 1997). When cDNA for CaM-KK a was expressed in E. coli, the crude extract also showed Ca2 + /calmodulin-independent activity, but the ratio of the Ca2 + /calmodulin-independent activity to the total activity was much higher when the b isoform was expressed than when the a isoform was expressed (Kitani et al., 1997). These results suggested that the b isoform of CaM-KK was susceptible to proteolytic attack. Because truncation of the carboxyl terminal regions of CaM-K II (Hagiwara et al., 1991; Yamagata et al., 1991) and CaM-K IV (Cruzalegui and Means, 1993) by proteolysis or mutagenesis produced the Ca2 + /calmodulin-independent forms of the enzymes, CaMKK b might also be influenced by proteolysis like these CaM-Ks. Previous studies of tissue distribution of CaM-KK b in a Western Blot analysis demonstrated a doublet (two immunoreactive bands) (Kitani et al., 1997; Anderson et al., 1998), which suggested that degradation into fragments occurred on the CaM-KK b molecule (Kitani et al., 1997). These phenomena might cause varying results of antigen –antibody reaction in the immunohistochemical study by using CaM-KK b as an antibody. It has been said that fixation procedures involved some causes of unstable staining in the immunohistochemistry — with paraformaldehyde fixation, there is a possibility that the epitope is hidden within a cell structure or is destroyed during fixation (Harlow and Lane, 1988). Sakagami et al. (2000) indicated that the different stainability by monoclonal and polyclonal antibodies might occur by using different fixatives; 2% paraformaldehyde and picric acid, or 4% paraformaldehyde, respectively. However, the difference of immunostaining between using monoclonal and polyclonal antibodies seems too great to be explained by the difference between the two kinds of fixatives. It
186
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
will be necessary to carry out the immunohistochemical staining procedures by using both antibodies under a same condition.
4.3. Relationship of CaM-KK i to neurons in the cerebellar cortex Because cerebellar granule cells contained abundant
CaM-K IV in the cellular nucleus, this enzyme was named CaM-K Gr at first (Ohmstede et al., 1989; Jensen et al., 1991). An enzyme, which phosphorylates to activate CaM-K IV, and which has been named CaMKK a (Okuno et al., 1994), however, was not found in the granule cell (Nakamura et al., 1996). If CaM-KK b had been observed in the cellular nucleus in the present study, it might phosphorylate CaM-K IV in the nucleus,
Fig. 17. Motoneurons of some cranial nerve nuclei (a–c) and cervical spinal cord (d) show both cytoplasmic and nuclear immunostaining. Arrowheads point to representative labeled neurons. Scale bar = 50 mm.
Fig. 18. This electron micrograph shows a large motoneuron of the facial nucleus. In the nucleus (Nu) reaction products (arrows) are located on dense chromatin structures, and in the cytoplasm on the Golgi apparatus (Go). Electron staining with uranyl acetate or lead citrate for thin sections was not made. AT, axon terminal. Scale bar = 2 mm.
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
187
Table 2 Comparison of immunostainability for CaM-kinases and CaM-kinase kinases in some representative nucleia CaM-KK b *1 Cerebral cortex (Pyramidal neuron)
CaM-KK a *2
CaM-K I *3
CaM-K IV *4
c (+n)
n
c
n
c+n (–) (–)
n n n
c c c
n (–) (–)
Cerebellar cortex Purkinje cell Granule cell
c c
n (–)
c c
(–) n
Motoneuron
c+n
n
c
(–)
Basal ganglia Striatum Globus pallidus Substantia nigra Pars reticulata
a c, Cytoplasmic immunostaining; n, nuclear immunostaining; c(+n), some neurons show nuclear staining, (–), no immunostaining; *1, present study; *2, Nakamura et al. (1996); *3, Picciotto et al. (1995); *4, Nakamura et al. (1995).
but CaM-KK b existed only in the cytoplasm. Furtherobservations will be required to solve the problem of what kind of enzyme can activate CaM-K IV in the cerebellar granule cell. CaM-KK b might activate CaM-K I in the cytoplasm of the cerebellar granule cell, and it might also activate CaM-K I in the Purkinje cell cytoplasm, because Purkinje cells contain CaM-K I, although CaM-K IV was not found in them.
4.4. Relationship of CaM-KK i to neurons in the basal ganglia In the caudate-putamen, CaM-KK b was observed both in the cytoplasm and nucleus, and it might activate CaM-K I in the cytoplasm and CaM-K IV in the nucleus, respectively. On the other hand, CaM-KK b was not localized immunohistochemically in the globus pallidus, entopeduncular nucleus or substantia nigra pars reticulata. This observation coincided well with ours showing that CaM-K IV was not found in these nuclei. Lack of immunolabeling may be explained in two ways; first, the amount of the production of CaMK IV and CaM-KK b in these nuclei is very poor, because these nuclei of the basal ganglia may not need such phosphorylation processes. Second, these enzymes are metabolized too fast to be detected morphologically in these nuclei. In the present study, CaM-K I was localized immunohistochemically in the cytoplasm of these three basal ganglia neurons. Thus, the problem is raised about how CaM-K I could be phosphorylated in these basal ganglia. It should be solved by future studies.
4.5. Relationship of CaM-KK i to motoneurons Most motoneurons showed both cytoplasmic and nuclear immunoreaction of CaM-KK b. Among large
neurons in the central nervous system, motoneurons, some cerebral large pyramidal neurons and magnocellular rubral neurons were stained both in the cytoplasm and nucleus. Other large neurons in many nuclei such as the mesencephalic trigeminal nucleus, lateral vestibular nucleus and gigantocellular reticular formation did not show nuclear labeling. Immunohistochemistry with the antibody against CaM-KK a also revealed immunolabeling of both the cytoplasm and nucleus in the rat motoneurons, although cytoplasmic staining was weak (Nakamura et al., 1996). On the other hand, CaM-K IV has not been found in the motoneuron of
Fig. 19. Immunostaining of some brain areas with antibodies against CaM-KK b (KK b), CaM-K I (KK I) and CaM-K IV (KK IV) are shown at higher magnification. (a) and (b) show giant pyramidal neurons in the layer V of the frontal cortex and motoneurons of the hypoglossal nucleus, respectively stained for CaM-KK b. (c) shows pyramidal neurons in the layer V of the parietal cortex stained for CaM-K I, and (d) shows cellular nuclei of neurons in the layer V of the frontal cortex stained for CaM-K IV. Scale bar = 50 mm.
188
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
the cranial nerve nuclei or in the anterior horn of the spinal cord (Nakamura et al., 1995), although CaM-K I has been shown in the motoneuron cytoplasm (Picciotto et al., 1995). CaM-KK b along with CaM-KK a, therefore, can activate CaM-K I in the motoneuron cytoplasm, because CaM-KK b can activate both CaMK I and CaM-K IV (Okuno et al., 1997). At present, no CaM-Ks have been found in the motoneuron nucleus. If CaM-KK b is translocated from the cytoplasm to the nucleus, what is the role of this enzyme there. Recently, CaM-KK has been proved to activate protein kinase B by phosphorylation (Yano et al., 1998). Even if CaM-Ks are not present in the nucleus, other substances such as protein kinase B might become a substrate for CaM-KK b. Further studies search for unknown substrates for these kinase kinases in the motoneuron nucleus are required.
Acknowledgements The authors thank Mie Taguchi for her technical assistance. This work was partly supported by Grantsin-Aid nos. 08680810 and 11680735 for Scientific Research (C) (Yasuhisa Nakamura) and 10102002 for Specially Promoted Research (Hitoshi Fujisawa) from the Ministry of Education, Science, Sports and Culture.
References Anderson, K.A., Means, R.L., Huang, Q.-H., Kemp, B.E., Goldstein, E.G., Selbert, M.A., Edelman, A.M., Fremeau, R.T., Means, A.R., 1998. Components of a calmodulin-dependent protein kinase cascade. Molecular cloning, functional characterization and cellular localization of Ca2 + /calmodulin-dependent protein kinase kinase b. J. Biol. Chem. 273, 31880–31889. Cruzalegui, F.H., Means, A.R., 1993. Biochemical characterization of the multifunctional Ca2 + /calmodulin-dependent protein kinase type IV expressed in insect cells. J. Biol. Chem. 268, 26171– 26178. Hagiwara, T., Ohsako, S., Yamauchi, T., 1991. Studies on the regulatory domain of Ca2 + /calmodulin-dependent protein kinase II by expression of mutated cDNA in Escherichia coli. J. Biol. Chem. 266, 16401– 16408. Harlow, E., Lane, D., 1988. Antibodies. A Laboratory Manual. Cold Spring Harbor Laboratory, pp. 361–365. Jensen, K.F., Ohmstede, C.-A., Fisher, R.S., Sahyoun, N., 1991. Nuclear and axonal localization of Ca2 + /calmodulin-dependent protein kinase type Gr in rat cerebellar cortex. Proc. Natl. Acad.
Sci. USA 88, 2850– 2853. Kitani, T., Okuno, S., Fujisawa, H., 1997. Molecular cloning of Ca2 + /calmodulin-dependent protein kinase kinase b. J. Biochem. 122, 243– 250. Lee, J.C., Edelman, A.M., 1994. A protein activator of Ca2 + / calmodulin-dependent protein kinase Ia. J. Biol. Chem. 269, 2158– 2164. Nakamura, Y., Okuno, S., Sato, F., Fujisawa, H., 1995. An immunohistochemical study of Ca2 + /calmodulin-dependent protein kinase IV in the rat central nervous system: light and electron microscopic observations. Neuroscience 68, 181– 194. Nakamura, Y., Okuno, S., Kitani, Y., Otake, K., Sato, F., Fujisawa, H., 1996. Distribution of Ca2 + /calmodulin-dependent protein kinase kinase a in the rat central nervous system: an immunohistochemical study. Neurosci. Lett. 204, 61 – 64. Ohmstede, C.A., Jensen, K.F., Sahyoun, N.E., 1989. Ca2 + /calmodulin-dependent protein kinase enriched in cerebellar granule cellsidentification of a novel neuronal calmodulin-dependent protein kinase. J. Biol. Chem. 264, 5866– 5875. Okuno, S., Fujisawa, H., 1993. Requirement of brain extract for the activity of brain calmodulin-dependent protein kinase IV expressed in Escherichia coli. J. Biochem. 114, 167– 170. Okuno, S., Kitani, T., Fujisawa, H., 1994. Purification and characterization of Ca2 + /calmodulin-dependent protein kinase IV kinase from rat brain. J. Biochem. 116, 923– 930. Okuno, S., Kitani, T., Fujisawa, H., 1996. Evidence for the existence of Ca2 + /calmodulin-dependent protein kinase IV kinase isoforms in rat brain. J. Biochem. 119, 1176– 1181. Okuno, S., Kitani, T., Fujisawa, H., 1997. Purification and characterization of Ca2 + /calmodulin-dependent protein kinase kinase b from rat cerebellum. J. Biochem. 121, 155– 160. Paxinos, G., Watson, C., 1998. The Rat Brain in Stereotaxic Coordinates, fourth ed. Academic Press, San Diego. Picciotto, M.R., Zoli, M., Bertuzzi, G., Nairn, A.C., 1995. Immunochemical localization of calcium/calmodulin-dependent protein kinase I. Synapse 20, 75 – 84. Sakagami, H., Umemiya, M., Kondo, H., 2000. Distinct immunohistochemical localization of two isoforms of Ca2 + /calmodulin-dependent protein kinase kinases in the adult rat brain. Eur. J. Neurosci. 12, 89 – 99. Selbert, M.A., Anderson, K.A., Huang, Q.-H., Goldstein, E.G., Means, A.R., Edelman, A.M., 1995. Phosphorylation and activation of Ca2 + /calmodulin-dependent protein kinase IV by Ca2 + / calmodulin-dependent protein kinase Ia kinase: phosphorylation of threonine 196 is essential for activation. J. Biol. Chem. 270, 17616– 17621. Sugita, R., Mochizuki, H., Ito, T., Yokokura, H., Kobayashi, R., Hidaka, H., 1994. Ca2 + /calmodulin-dependent protein kinase kinase cascade. Biochem. Biopyhys. Res. Commun. 203, 694–701. Yamagata, Y., Czernik, A.J., Greengard, P., 1991. Active catalytic fragment of Ca2 + /calmodulin-dependent protein kinase II. Purification, characterization, and structural analysis. J. Biol. Chem. 266, 15391– 15397. Yano, S., Tokumitsu, H., Soderling, T.R., 1998. Calcium promotes cell survival through CaM-K kinase activation of the protein-kinase-B pathway. Nature 396, 584– 587.
Immunohistochemical localization of Ca2 + /calmodulin-dependent protein kinase kinase b in the rat central nervous system Yasuhisa Nakamura a,*, Sachiko Okuno b, Takako Kitani b, Kazuyoshi Otake a, Fumi Sato a, Hitoshi Fujisawa b a
Section of Neuroanatomy, Graduate School of Medical and Dental Research, Tokyo Medical and Dental Uni6ersity, Yushima, Bunkyo-ku, Tokyo 113 -8519, Japan b Department of Biochemistry, Asahikawa Medical College, Midorigaoka, Asahikawa 078 -8510, Japan Received 17 August 2000; received in revised form 5 October 2000; accepted 6 October 2000
Abstract We examined regional and intracellular distribution of Ca2 + /calmodulin-dependent protein kinase kinase b (CaM-KK b), which activated Ca2 + /calmodulin-dependent protein kinase I and IV (CaM-K I and IV) immunohistochemically in the central nervous system of the rat by light and electron microscopy. Although most neurons in the brain and spinal cord exhibited the immunoreactivity, no labeled neurons were observed in the globus pallidus or entopeduncular nucleus, and only a small number of neurons showed weak immunoreactivity in the substantia nigra pars reticulata. In general, the immunoreactivity was observed both in the cytoplasm and cellular nucleus, although the immunoreactivity was not found in the cellular nucleus in some large neurons such as in the mesencephalic trigeminal nucleus, lateral vestibular nucleus or gigant cellular reticular formation. As to motoneurons of the cranial nerve nuclei and the anterior horn of the spinal cord, they revealed the immunoreactivity both in the cytoplasm and nucleus. The reaction product appeared as fine granules in the cytoplasm and nucleus under light microscopy. Electron microscopic observations confirmed that the reaction product was localized mainly on the Golgi apparatus or on the nuclear chromatin. Immunolabeling for antibody against CaM-KK b was discussed with the distribution of CaM-K I, IV and another CaM-KK, CaM-KK a, in the central nervous system. © 2001 Elsevier Science Ireland Ltd and the Japan Neuroscience Society. All rights reserved. Keywords: CaM-kinase I; CaM-kinase IV; CaM-kinase kinase a; CaM-kinase kinase b; Immunohistochemistry
1. Introduction There have been several investigations concerning Ca2 + /calmodulin-dependent protein kinase kinase
(CaM-KK), which activates CaM-K IV and CaM-K I by phosphorylation (Okuno and Fujisawa, 1993; Okuno et al., 1994; Sugita et al., 1994); these CaM-Ks play important roles as Ca2 + -responsive multifunc-
Abbre6iations: 4, trochlear nucleus; 5, motor trigeminal nucleus; 7, facial nucleus; AD, anterodorsal nucleus; AH, anterior horn; AON, anterior olfactory nucleus; AV, anteroventral nucleus; BL, basolateral amygdaloid nucleus; BM, basomedial amygdaloid nucleus; CA1, CA1 sector of Ammon’s horn; CA3, CA3 sector of Ammon’s horn; CA4, CA4 sector of Ammon’s horn; CblCx, cerebellar cortex; CbrCx, cerebral cortex; Ce, central amygdaloid nucleus; Cpu, caudate-putamen; DG, dentate gyrus; dTh, dorsal thalamus; EPl, external plexiform layer; Gl, glomerular layer; GP, globus pallidus; GrC, cerebellar granule cell layer; GRF, gigantocellular reticular formation; GrO, olfactory granule cell layer; Hip, hippocampal formation; Hyp, hypothalamus; IC, inferior colliculus; IGL, intermediate gray layer; IOd, dorsal accessory olive; IOm, medial accessory olive; IOp, principal olive; IP, cerebellar interpositus nucleus; Lat, cerebellar lateral nucleus; LC, locus coeruleus; LD, laterodorsal nucleus; lfp, longitudinal fasciculus of pons; LVN, lateral vestibular nucleus; MD, mediodorsal nucleus; Me, medial amygdaloid nucleus; Med, cerebellar medial nucleus; Mi, mitral cell layer; Mol, molecular layer; NRTP, pontine tegmental reticular nucleus; P, Purkinje cell layer; PBNm, medial parabrachial nucleus; Pi, piriform cortex; PN, pontine nuclei; py, pyramidal tract; RN, red nucleus; Rt, thalamic reticular nucleus; SC, superior colliculus; SGL, superficial gray layer; SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticulata; VA, ventral anterior nucleus; Vmes, mesencephalic trigeminal nucleus. * Corresponding author. Tel.: + 81-3-5803-5148; fax: + 81-3-5803-5151. E-mail address: [email protected] (Y. Nakamura). 0168-0102/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd and the Japan Neuroscience Society. All rights reserved. PII: S0168-0102(00)00209-1
176
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
tional protein kinases, in controlling a variety of neuronal functions in response to an increase in intracellular Ca2 + . Because CaM-KK was first investigated to phosphorylate CaM-K IV, it has been called CaM-K IV K (Okuno and Fujisawa, 1993). Independently, CaM-K I activator was also found, which phosphorylated CaM-K I and CaM-K IV (Lee and Edelman, 1994; Selbert et al., 1995). When crude extracts of rat cerebral cortex, brainstem, cerebellum and the bacteria transformed with an expression vector (pET) carrying CaM-K IV K cDNA were immunotitrated with antiserum to the cloned enzyme (CaM-K IV K), almost all the activity of the recombinant enzyme was immunoprecipitated (Okuno et al., 1996). During the study of immunotitration, it became apparent that only approximately 46, 56, and 25% of the activity of cerebral cortex, brainstem, and cerebellum, respectively, were immunoprecipitated by addition of an excess of the antiserum (Okuno et al., 1996). Thus, the cloned CaM-K IV K accounted for only one-fourth of total CaM-KK activity in the cerebellum where CaM-K IV exists most abundantly (Ohmstede et al., 1989; Jensen et al., 1991). This investigation has suggested that most of the activity is attributable to other CaM-KK. Then another isoform of CaM-K IV K was demonstrated in the rat brain. Thereafter, the first CaM-K IV K was named CaM-KK a (Okuno et al., 1996), and another one, which has been purified and characterized was named CaM-KK b (Okuno et al., 1997; Anderson et al., 1998). The distribution of CaMKK a was demonstrated by means of immunohistochemistry by using the antibody against CaM-KK a, which was located in the neuronal nuclei of most neurons with uniform density and also in the cytoplasm in some regions (Nakamura et al., 1996). In this study, newly developed antibody against CaM-KK b (Kitani et al., 1997) was used for immunohistochemistry to localize this enzyme in the rat central nervous system by light and electron microscopy.
paraformaldehyde solution or 3.5% paraformaldehyde and 0.1% glutaraldehyde mixture. After removal, the brains were cut serially at 40 mm into frontal or sagittal sections on a freezing microtome for light microscopy after soaking in 30% phosphate buffered sucrose for 2 days. For electron microscopy, brains were cut at 100 – 150 mm immediately after perfusion with a microslicer. Procedures for immunohistochemistry for light and electron microscopy were described previously (Nakamura et al., 1995, 1996). Briefly, free floating sections in 0.1 M phosphate buffer containing 0.1% Triton X-100 were incubated overnight in a refrigerator with purified antibody (1:2000 –3000). Then the sections were incubated in biotinylated anti-rabbit immunoglobulin and reacted with Vectastain ABC kit (Elite, Vector Labs, CA, USA). To detect peroxidase, 0.05% 3,3%-diaminobenzidine (DAB) was used with 0.003% H2O2, 0.005% nickel chloride and cobalt acetate in the phosphate buffer. For control, some sections were processed omitting primary antibody in the incubation, or were processed with immunoglobulin G (IgG) obtained from normal rabbit serum instead of the antibody. Brain structures were identified according to the rat stereotaxic atlas by Paxinos and Watson (1998). For light microscopy, the sections were mounted on gelatinized glass slides; some of them were counterstained with 0.5% Cresyl violet. For electron microscopy, target nuclei for the observation were trimmed from the immunostained sections and embedded in Epon after osmication, dehydration, and block-staining with 1% uranyl acetate in absolute ethanol. Ultrathin sections were examined under a Hitachi H-7100 electron microscope with or without uranyl acetate –lead citrate staining. In addition to the immunohistochemistry for CaMKK b, we stained frozen sections of rat brains with antibody against CaM-K I, which was made from a synthetic peptide, corresponding to the carboxyl-terminal 20 amino acids of CaM-K I. Crude serum obtained from immunized rabbits was purified with affinity chromatography.
2. Materials and methods Molecular cloning and sequencing of CaM-KK b was made (Kitani et al., 1997). Thereafter, rabbit antibody against CaM-KK b was prepared by immunization of a peptide corresponding to the carboxyl-terminal 24 amino acids of CaM-KK b purifying crude serum obtained from immunized rabbits by means of affinity chromatography, and the specificity of this antibody was reported (Kitani et al., 1997). Eleven rats (male, Wistar, Nippon Bio-Supp. Center, Tokyo, Japan, 7 weeks) were used in the immunohistochemical study. The rats, under deep pentobarbital anesthesia (50 mg/kg, i.p.), were killed by perfusion through the ascending aorta with phosphate buffered 4%
3. Results We examined the immunohistochemical reaction with an antibody against CaM-KK b (Fig. 1a). Although two kinds of perfusate, with glutaraldehyde or without glutaraldehyde, were used, the results apparently did not differ. Most neurons in the central nervous system showed immunoreactivity. Control sections, which were reacted without primary antibody did not show any immunostaining in the central nervous system (Fig. 1b). Control sections reacted with IgG showed similar feature as observed without primary antibody. During this study, immunolabeling was observed in the cytoplasm,
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
and in many neurons, it was also observed in cellular nucleus; however, the intensity of staining was not uniform. The immunohistochemical reaction for the antibody against CaM-KK b in the central nervous system is summarized in Table 1.
3.1. Telencephalon In the main olfactory bulb, most mitral cells and granule cells were stained moderately (Fig. 2). In the glomerular layer, fine fibers surrounding the glomeruli were stained densely (Fig. 2). Mitral cells and granule cells in the accessory olfactory bulb showed the same appearance as those in the main olfactory bulb. In the olfactory tubercle, pyramidal neurons were stained most densely, which is shown in Fig. 1a. In the cerebral neocortex, immunoreactivity of neurons was dense in layers II and III, and neurons in layers IV – VI showed less reactivity (Fig. 3). Stainability was rather uniform in every area of the cerebral cortex. Some large pyramidal neurons showed only cytoplasmic staining, although many smaller neurons showed immunostained cellular nuclei in immunolabeled cytoplasm. Labeled radial strands, which were the apical dendrites of pyramidal neurons were also found among neuronal somata (Fig. 3). Electron microscopy of these labeled dendrites showed reaction product adjacent to microtubules or near mitochondria. In the hippocampal formation, weak to moderate immunoreaction was observed in the pyramidal cells of the subiculum and Ammon’s horn, and granule cells of the dentate gyrus (Fig. 4). Well
177
stained processes extending from the pyramidal layer to the stratum radiatum were observed in the CA3 and CA4 sectors of the Ammon’s horn. Occasionally, a few densely stained probable interneurons were observed in the stratum radiatum. Among the limbic cortices, pyramidal neurons in the piriform cortex showed densest stainability, although in the other limbic cortices immunoreaction appeared as similar as in the neocortices. Comparing the immunostaining in various cortical regions, pyramidal neurons in the olfactory tubercle and piriform cortex showed the densest reactivity, and those in most neocortical regions showed not so dense as those in the above two regions. Pyramidal neurons in the hippocampal formation (CA1 –CA4) showed weaker immunostaining than in the other cortical regions. In the subcortical telencephalic nuclei, such as the septal nuclei, amygdaloid body and caudate-putamen, immunoreactivity was observed but not homogeneous within each subnucleus. In the septal nuclei, staining was rather weak except for the dorsal part of the lateral septal nucleus, where some neurons were well stained both in the somata and primary dendrites. In the amygdaloid body, the central nucleus was stained densest (Fig. 5), and the other nuclei such as the basolateral nucleus appeared less dense, and the anterior or cortical nuclei was stained weakest. Immunoreaction was observed both in the cytoplasm and nucleus in these amygdaloid neurons. In the caudate-putamen, both medium-sized and large neurons were stained moderately (Fig. 6). No immunoreaction in the cytoplasm or nucleus was observed in the globus pallidus (homologous to the primate external segment of the globus pallidus) or in the entopeduncular nucleus (the internal segment of the primate globus pallidus) (Fig. 6). In the claustrum, moderate labeling of neurons was observed. In the preoptic area, immunostaining was weak.
3.2. Diencephalon
Fig. 1. Sagittal sections of the rat brain stained immunohistochemically with (a) and without (b) the antibody against CaM-KK b. Strong reactivity (a) and no reactivity (b) are shown from the anterior olfactory nucleus to the lower brainstem. Counterstain was not made for any of the light microscopic sections shown in Figs. 1–17. Scale bar= 5 mm.
In the dorsal thalamus neurons in the anterodorsal (AD), laterodorsal (LD), ventroanterior-ventrolateral complex (VA –VL) and midline nuclei were labeled moderately, and those in other nuclei such as the intralaminar or ventral posterior nuclei, were weakly labeled (Fig. 7). In the hypothalamus, labeling of neurons in the suprachiasmatic, paraventricular and mammillary nuclei was moderate, and in other nuclei such as the anterior or ventromedial nuclei, it was weak. In the habenular nucleus, neurons in the medial nucleus showed moderate labeling, but in the lateral nucleus very weak. Neurons in the thalamic reticular nucleus (Fig. 7) and subthalamic nucleus were moderately labeled, and neurons in the zona incerta were stained weakly.
178
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Table 1 The immunohistochemical reaction for the antibody against CaMKK b in the central nervous systema Regions of the CNS examined
Relative intensity of immunostaining
Olfactory system Main olfactory bulb Mitral cells Granule cells Glomerular layer
++ ++ +++
Accessory olfactory bulb Mitral cells Granule cells Anterior olfactory nucleus
++ ++ ++
Olfactory tubercle Pyramidal cells Islands of Calleja
++++ ++
Neocortex Frontal cortex Parietal cortex Temporal cortex Occipital cortex
(I–IV) (I–IV) (I–IV) (I–IV)
Limbic cortex Piriform cortex Cingulate cortex
+++, +++, +++, +++,
(V–VI) (V–VI) (V–VI) (V–VI)
++ ++ ++ ++
Entorhinal cortex Subicular cortex Dentate gyrus Ammon’s horn
++++ (II, IV) +++, (I, III–VI) ++ (II, IV) +++, (I, III–VI) ++ (II, IV) +++, (I, III–VI) ++ +++ ++ ++ (CA1–2) ++, (CA3–4) +
Septum Lateral septal nucleus Medial septal nucleus Nuclei of diagonal band Bed nucleus of the stria terminalis
+/++ + + +
Preoptic area Medial preoptic nucleus Lateral preoptic nucleus
+ −/+
Amygdala Anterior amygdaloid nucleus Cortical amygdaloid nucleus Medial amygdaloid nucleus Lateral amygdaloid nucleus Basolateral amygdaloid nucleus Basomedial amygdaloid nucleus Central amygdaloid nucleus
+ + + + ++ + +++
Basal ganglia Caudate-putamen Nucleus accumbens Globus pallidus Ventral pallidum Entopeduncular nucleus Subthalamic nucleus Claustrum Substantia nigra pars reticulata pars compacta Ventral tegmental area
++ ++ − − − ++ ++ −/+ ++ ++
Retrosplenial cortex Perirhinal cortex
Table 1 (Continued) Regions of the CNS examined
Relative intensity of immunostaining
Thalamic reticular nucleus Zona incerta
++ +
Hypothalamus Anterior hypothalamic area Lateral hypothalamic area Suprachiasmatic nucleus Supraoptic nucleus Paraventricular nucleus Arcuate nucleus Ventromedial nucleus Dorsomedial nucleus Premammillary nucleus Medial mammilary nucleus Lateral mammillary nucleus Posterior hypothalamic area
+ + ++ + ++ + + + ++ ++ ++ +
Habenula Medial habenular nucleus Lateral habenular nucleus
++ −/+
Dorsal thalamus Anterodorsal nucleus ++ Anteroventral nucleus + Anteromedial nucleus + Mediodorsal nucleus + Ventral anterior and ventral lateral ++ complex Ventroposteiror nuclei + Ventral medial nucleus + Laterodorsal nucleus ++ Midline nuclei ++ Intralaminar nuclei + Parafascicular nucleus + Posterior thalamic nuclei + Lateral geniculate nucleus + Medial geniculate nucleus + Lower brainstem Red nucleus Interstitial nucleus of Cajal Superior colliculus Superficial gray layer Deep gray layers
+++ + ++ +
Inferior colliculus External nucleus ++ Central nucleus ++ Pedunculopontine tegmental nucleus+ Medial parabrachial nucleus + Lateral parabrachial nucleus ++ Interpeduncular nucleus +++ Locus coeruleus ++++ Trigeminal nuclei Mesencephalic nucleus Motor nucleus Principal sensory nucleus Spinal nucleus Superior olivary complex Trapezoid nucleus Pontine reticular formation Pontine tegmental reticular nucleus Pontine nuclei Central gray Raphe nuclei
++ ++ ++ +/++ ++ ++ + +++ +++ + ++
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
179
Table 1 (Continued) Regions of the CNS examined
Relative intensity of immunostaining
Medullary retiuclar formation (GRF etc.) Lateral reticular nucleus Vestibular nuclei Cochlear nuclei Inferior olivary complex Prepositus hypoglossi nucleus External cuneate nucleus Dorsal column nuclei Nucleus of the solitary tract Motor nuclei of the cranial nerves
+++ +++ ++ +++ ++ ++ ++ ++ + ++
Cerebellum Cerebellar nuclei Purkinje cells Granule cells Molecular layer
++ + + +++
Spinal cord Layer I of the posterior horn Other layers of the posterior horn Reticular nucleus Inermediate zone Anterior horn
++ + + + ++
a ++++, +++, ++ and + indicate densest, dense, moderate and weak immunostainability, respectively.
Fig. 3. The frontal sensory-motor cortex (layers III– VI) shows that some large neurons (arrowheads) are devoid of staining in their nuclei. Apical dendrites are well stained as vertical arrays. Scale bar =100 mm.
3.3. Mesencephalon In the mesencephalon, superior colliculus neurons showed weak to moderate labeling (Fig. 8). Neurons in the superficial gray layer were stained denser than those in the other layers. Most large neurons in the intermediate gray layer showed no immunolabeling in their cellular nuclei. In the inferior colliculus, neurons both in the central and external nuclei were stained moderately. Neurons in the central gray matter were weakly labeled. In the tegmentum, cytoplasm of the magnocellular red nucleus neurons were densely stained, and nuclei of some neurons showed denser staining than the cytoplasm, while nuclei of some neurons were devoid of immunostaining (Fig. 9). Although abundant substantia nigra pars compacta neurons and ventral tegmental area neurons were stained moderately, only a small number of neurons in the substantia nigra pars reticulata was weakly stained (Fig. 9). Axons in the crus cerebri showed dense immunolabeling. At the pontomesencephalic junction, the pedunculopontine tegmental nucleus neurons were weakly stained both in the cytoplasm and nucleus. Neurons in the mesencephalic reticular formation were also weakly labeled. Description of the motoneruons will be summarized below.
3.4. Pons and cerebellum
Fig. 2. A coronal section of the olfactory bulb shows fine fibers labeled densely surrounding the olfactory glomeruli (Gl). Scale bar = 100 mm.
In the pons, neurons in the pontine nuclei were stained densely (Fig. 10). Pontine tegmental reticular nucleus neurons were also densely stained (Fig. 10); in these nuclei both the cytoplasm and nucleus of neurons showed immunoreactivity. In addition to neuronal labeling, corticofugal axons in the pons were labeled
180
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Fig. 4. The dorsal hippocampal formation shows immunolabeling in the dentate gyrus (DG) and Ammon’s horn (CA1– CA4). Arrowheads point to neuronal labeling in the stratum radiatum. Scale bar = 500 mm.
densely. Neurons in the medial and lateral parabrachial nuclei were stained weakly and moderately, respectively (Fig. 11). In the locus coeruleus, almost all neurons were very densely stained (Fig. 11); the stainability was heavier in the nucleus than the cytoplasm. As to the trigeminal nuclei, large neurons of the mesencephalic trigeminal nucleus showed moderate staining in the cytoplasm, although cellular nuclei were devoid of labeling (Fig. 11). In the main sensory nucleus, neurons were also moderately labeled. Most neurons in the superior olivary nuclear complex and the nucleus of the trapezoid body were labeled moderately both in the cytoplasm and nucleus. Pontine reticular formation neurons were weakly labeled. The cerebellar cortex appeared very dense under low magnification (Fig. 1a). In the Purkinje cell layer, Purkinje cells were observed to be stained weakly both in the cytoplasm and nucleus (Fig. 12). In the granular layer, weakly labeled granule cells appeared like small rings, which suggested that only a thin layer of cytoplasm was labeled, and any labeling of cellular nuclei was not observed (Fig. 12). It was hard to identify immunolabeled larger neurons such as Golgi cells in this layer. Although the molecular layer was stained densely, no labeled cell bodies were discernible, but only fine strands of fibers, presumably parallel fibers, ran vertically and horizontally. Electron microscopy proved that these immunolabeled fibers were parallel fibers and their axon terminals (Fig. 13). Packs of fibers of small diameter, which was a characteristic feature of the parallel fiber, contained densely stained microtubules, and some labeled axon terminals included densely stained synaptic vesicles. These axon terminals were characterized by round synaptic vesicles in them and asymmetric synaptic contacts. In the cerebellar medial, interpositus and lateral nuclei, moderately labeled neurons were seen — most cellular nuclei were not labeled (Fig. 14).
3.5. Medulla oblongata and spinal cord In the medulla oblongata, immunoreactivity of neurons was observed in many nuclei such as the inferior olivary complex, dorsal column nuclei and medullary reticular formation. In the lateral vestibular nucleus, cellular nuclei of most large neurons were not stained
Fig. 5. Part of the amygdaloid body shows different stainability among subnuclei (BL, BM, Ce and Me). Scale bar = 500 mm.
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
181
showed immunoreactivity (Fig. 17a). The cytoplasm of all neurons was stained densely and almost all cellular nuclei were stained much denser. Neurons of the motor trigeminal (Fig. 17b) and abducens nuclei showed similar stainability as observed in the oculomotor and trochlear nuclei. In the facial (Fig. 17c), ambigual and hypoglossal nuclei, the appearance of immunoreactivity was similar as other motor cranial nerve nuclei. Motoneurons in the anterior horn of the cervical spinal cord also showed the same stainability of both the cytoplasm and nucleus as other cranial motoneurons (Fig. 17d).
3.7. Electron microscopic obser6ations We observed neurons of the cerebral cortex, caudateputamen, cerebellar cortex, locus coeruleus and motoneurons of the trigeminal and facial nerve nuclei electron microscopically. At a lower magnification, reaction product was observed in the cytoplasm and the nucleus as electron dense fine grains, which were less than 0.3 mm in diameter. In the nucleus, reaction product was located mostly on dense aggregations of nuclear chromatin structures as if immunoreaction was carried to restricted areas (Fig. 18). In the cytoplasm,
Fig. 6. Densely stained caudate-putamen (CPu) contrasts with the globus pallidus (GP) which does not show any immunoreaction. Scale bar= 500 mm.
(Fig. 14). Neurons in the cochlear nuclei, the cytoplasm and nucleus were densely stained. Immunoreactivity of the nucleus of the solitary tract was weak. In the spinal trigeminal nucleus, many medium sized or small neurons revealed weak or moderate cytoplasmic and nuclear staining, although nuclei of some large neurons were not stained. Neurons in the inferior olivary nuclear complex, both the cytoplasm and nucleus were stained moderately (Fig. 15). Dorsal column nuclei neurons were also labeled moderately both in the cytoplasm and nucleus. In the medullary reticular formation, neurons were stained densely (Fig. 16). Although many neurons in the gigantocellular reticular formation did not show nuclear staining, medium-sized or small neurons in the parvocellular, lateral, and paramedian reticular nucleus showed immunoreactivity both in the cytoplasm and nucleus. In the upper cervical spinal cord, most neurons in the dorsal horn, intermediate zone and ventral horn were immunolabeled both in the cytoplasm and nucleus. Among them, layer I neurons in the posterior horn and ventral neurons were stained rather densely than the others.
3.6. Motoneurons in the cranial ner6e nuclei and the spinal cord Motoneurons in the oculomotor and trochlear nuclei
Fig. 7. The rostral part of the dorsal thalamus does not show uniform labeling among subnuclei (AD, AV, LD, MD and VA). Scale bar = 500 mm.
182
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Fig. 10. The basal part of the pontine region shows immunolabeling of neurons in the pontine nuclei (PN) and pontine tegmental reticular nucleus (NRTP). Longitudinally oriented fibers (lfp) are labeled moderately. Scale bar =200 mm.
CaM-KK b was synthesized in the endoplasmic reticulum and transported to and concentrated in the Golgi apparatus. Fig. 8. Part of the superior colliculus shows weak immunolabeling in the superficial (SGL) and intermediate gray layers (IGL). Some labeled neurons are devoid of nuclear staining (arrowheads). Scale bar= 100 mm.
immunoreactive materials were located mainly on vacuoles or vesicles of the Golgi apparatus, Nissl substance or adjacent to mitochondria. It was conceivable that
3.8. CaM-K I immunohistochemistry As has been mentioned by Picciotto et al. (1995), immunolabeling of CaM-K I was localized in the cytoplasm of neurons in the rat central nervous system; no immunolabeled cellular nucleus was found. Although
Fig. 9. The mesencephalic tegmentum shows weak labeling in the substantia nigra pars compacta (SNc) and almost no labeling in the substantia nigra pars reticulata (SNr) although a part of its neuropile is stained densely. Many magnocellular red nucleus (RN) neurons are devoid of nuclear staining (arrowheads). Scale bar =500 mm.
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
immunolabeled neurons were localized in the central nervous system as in the previous study (Picciotto et al., 1995), immunoreaction of neuronal cell bodies was moderately immunolabeled in the territories of the globus pallidus, entopeduncular nucleus and substantia nigra pars reticulata, in rather thick stained neuropile, where labeled neurons had not been observed (Picciotto et al., 1995). A table (Table 2) and a figure (Fig. 19) summarize these results comparing the stainability for CaM-KKs and CaM-Ks in some representative nuclei. This figure
183
shows different staining of the cytoplasm and cellular nucleus with antibodies against CaM-KK b (Fig. 19a and b), CaM-K I (Fig. 19c) and CaM-K IV (Fig. 19d) at higher magnification. An antibody against CaM-KK b stained the cytoplasm alone (Fig. 19a) or both the cytoplasm and nucleus (Fig. 19b), however, an antibody against CaM-K I stained exclusively the cytoplasm (Fig. 19c), and antibodies against CaM-K IV (Fig. 19d) and CaM-KK a stained chiefly the cellular nucleus. Because nuclear staining for CaM-KK a (Nakamura et al., 1996) was virtually the same as that for CaM-K IV, this finding was not shown here. Fig. 19d was taken from a histological section, which was prepared before (Nakamura et al., 1995). 4. Discussion
Fig. 11. The dorsal part of the pontine region shows dense labeling of the locus coeruleus neurons (LC) both in the cytoplasm and nucleus, and only cytoplasmic labeling in the mesencephalic trigeminal nucleus neurons (Vmes). Scale bar =100 mm.
In the present paper, we described the distribution of CaM-KK b-like immunolabeling in the rat central nervous system. This enzyme is known to activate CaM-K IV (Kitani et al., 1997; Okuno et al., 1997) and also CaM-K I (Okuno et al., 1997) by phosphorylation. Another isoform of CaM-KK b had already been cloned and named as CaM-KK a (Okuno et al., 1996), whose distribution in the central nervous system in the rat has been reported (Nakamura et al., 1996). The distribution of CaM-KK b was, however, quite different from that of CaM-KK a, which was localized mainly in the cellular nuclei of almost all neurons (Nakamura et al., 1996). In addition to the activation of CaM-K IV, both CaM-KK a and CaM-KK b also phosphorylate CaM-K I, which is another Ca2 + / calmodulin-dependent multifunctional protein kinase, and has been reported to be activated upon phosphorylation by CaM-K I K (Lee and Edelman, 1994; Selbert et al., 1995). However, exact location of CaM-K I K in the brain or in cell organelles has not been reported. We, therefore, will not discuss the phosphorylation of CaM-Ks by CaM-K I K at present.
4.1. Relationship of CaM-KK i with CaM-K I and CaM-K IV
Fig. 12. A part of the cerebellar cortex shows weak labeling of the Purkinje cells (P). Because of dense labeling of parallel fibers, the molecular layer (white bar with Mol) appears very dark. In the granular layer (GrC) the granule cell cytoplasm shows a ring-like appearance. Scale bar =50 mm.
Because CaM-KKs phosphorylate CaM-K I and CaM-K IV to activate their function, it is reasonable to believe that their locations are in the same brain areas and in the same cellular organelles. CaM-K I is distributed in the cytoplasm throughout the rat central nervous system from the telencephalon, diencephalon, mesencephalon, cerebellum, lower brainstem to the spinal cord (Picciotto et al., 1995). Although neurons in the substantia nigra was not reported to be labeled immunohistochemically, we found that neurons in both pars compacta and reticulata of the substantia nigra had been labeled moderately. In addition, neurons in the globus pallidus and entopeduncular nucleus had
184
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Fig. 13. An electron micrograph shows the molecular layer of the cerebellum. Some labeled parallel fibers contained densely stained microtubules. Arrows point to labeled parallel fiber terminals in which round synaptic vesicles are observed. Scale bar = 0.5 mm.
also been labeled, although results of immunoreaction in these nuclei were not described in their report (Picciotto et al., 1995). In the present study CaM-KK b-like immunoreactivity was found to be located in the cytoplasm and nucleus of most neuron groups in the brain and spinal cord, except in the globus pallidus, entopeduncular nucleus and substantia nigra pars reticulata. CaM-KK b, therefore, can activate CaM-K I in all neurons in the central nervous system, except in the above three basal ganglia. The distribution of CaM-K IV was not uniform in the central nervous system (Nakamura et al., 1995). Dense immunostaining was observed in the cerebral cortex, hippocampal formation, caudate-putamen, dorsal thalamus, cerebellar granular layer, and the dorsal horn of the spinal cord. Immunolabeled neurons, however, were not observed in the globus pallidus, entopeduncular nucleus, substantia nigra (both pars compacta and pars reticulata), ventral tegmental area, medial habenular nucleus, interpeduncular nucleus, mesencephalic trigeminal nucleus, principal sensory trigeminal nucleus, dorsal column nuclei, motor cranial nerve nuclei, or the anterior horn of the cervical spinal cord. Contrary to CaM-K I, CaM-K IV was localized mainly in the cellular nucleus, and also in the cytoplasm in some neurons such as cerebral pyramidal, rubral, lateral vestibular and cerebellar granule neurons. Nuclear CaM-K IV, therefore, can be phosphorylated by CaMKK a and CaM-KK b, and cytoplasmic CaM-K IV by CaM-KK b when it is co-localized in neurons. As to the location of CaM-KK b in neurons, sometimes it appeared both in the cytoplasm and nucleus. In some nuclei, such as the cerebral cortex or lateral vestibular nucleus, CaM-KK b was localized in the cytoplasm only in some large neurons, but was local-
ized both in the cytoplasm and cellular nucleus in other neurons. These observations suggest that CaM-KK b can be translocated from the cytoplasm to the nucleus easily, the mechanism yet unknown.
4.2. Polyclonal and monoclonal antibodies against CaM-KK i Recently, an immunohistochemical study dealing with the localization of CaM-KK b in the rat brain was published (Sakagami et al., 2000). The authors raised a novel monoclonal antibody against CaM-KK b and used it for their study. The result of their report was quite different from our present study. They described immunolabeling as restricted to the cytoplasm throughout the brain. In the cerebral cortical region, the difference between the stainability with the two antibodies was not so manifest. In the basal ganglia immunoreac-
Fig. 14. A coronal section of the cerebellar nuclear region shows weak to moderate labeling of the cerebellar medial (Med), interpositus (IP) and lateral (Lat), and lateral vestibular (LVN) nuclear neurons. Scale bar =500 mm.
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
Fig. 15. Neurons of the inferior olivary nuclear complex (IOd, IOm and IOp) show moderate labeling. Descending fibers in the pyramis (py) are also labeled moderately. Scale bar = 100 mm.
Fig. 16. Some large neurons in the gigantocellular reticular formation (GRF) are devoid of nuclear staining (arrowheads). Scale bar =100 mm.
tivity of neurons in the globus pallidus, entopeduncular nucleus and substantia nigra pars reticulata was shown with the monoclonal antibody, although we did not observe it with the polyclonal antibody. Purkinje cells in the cerebellar cortex, and neurons in the cerebellar nuclei were reportedly devoid of labeling with the monoclonal antibody. One of the most conspicuous findings with the monoclonal antibody was that immunoreaction for motoneurons of the cranial nerve nuclei and the anterior horn of the spinal cord was not detected.
185
With the polyclonal antibody, we revealed cytoplasmic and nuclear immunolabeling of motoneruons. Considering these observations, we could not conclude that the monoclonal and polyclonal antibodies showed the same antigen –antibody reaction in the brain and spinal cord. Here we discuss the immunohistochemical different expressions between monoclonal and polyclonal antibodies. The monoclonal antibody was produced by using amino acid residues 520 –587 near the carboxyl terminal (Sakagami et al., 2000), and polyclonal antibody by using amino acid residues 564 –587 (Kitani et al., 1997). Because both antibodies showed the same Western Blot pattern of doublet near 67 kDa, coinciding with that of CaM-KK b (Kitani et al., 1997), these two antibodies are thought to react specifically with CaM-KK b. When cDNA for CaM-KK b was expressed in Escherichia coli, the crude extract showed the ability to phosphorylate the CaM-K IV peptide, and showed relatively high Ca2 + /calmodulin-independent activity (Kitani et al., 1997). When cDNA for CaM-KK a was expressed in E. coli, the crude extract also showed Ca2 + /calmodulin-independent activity, but the ratio of the Ca2 + /calmodulin-independent activity to the total activity was much higher when the b isoform was expressed than when the a isoform was expressed (Kitani et al., 1997). These results suggested that the b isoform of CaM-KK was susceptible to proteolytic attack. Because truncation of the carboxyl terminal regions of CaM-K II (Hagiwara et al., 1991; Yamagata et al., 1991) and CaM-K IV (Cruzalegui and Means, 1993) by proteolysis or mutagenesis produced the Ca2 + /calmodulin-independent forms of the enzymes, CaMKK b might also be influenced by proteolysis like these CaM-Ks. Previous studies of tissue distribution of CaM-KK b in a Western Blot analysis demonstrated a doublet (two immunoreactive bands) (Kitani et al., 1997; Anderson et al., 1998), which suggested that degradation into fragments occurred on the CaM-KK b molecule (Kitani et al., 1997). These phenomena might cause varying results of antigen –antibody reaction in the immunohistochemical study by using CaM-KK b as an antibody. It has been said that fixation procedures involved some causes of unstable staining in the immunohistochemistry — with paraformaldehyde fixation, there is a possibility that the epitope is hidden within a cell structure or is destroyed during fixation (Harlow and Lane, 1988). Sakagami et al. (2000) indicated that the different stainability by monoclonal and polyclonal antibodies might occur by using different fixatives; 2% paraformaldehyde and picric acid, or 4% paraformaldehyde, respectively. However, the difference of immunostaining between using monoclonal and polyclonal antibodies seems too great to be explained by the difference between the two kinds of fixatives. It
186
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
will be necessary to carry out the immunohistochemical staining procedures by using both antibodies under a same condition.
4.3. Relationship of CaM-KK i to neurons in the cerebellar cortex Because cerebellar granule cells contained abundant
CaM-K IV in the cellular nucleus, this enzyme was named CaM-K Gr at first (Ohmstede et al., 1989; Jensen et al., 1991). An enzyme, which phosphorylates to activate CaM-K IV, and which has been named CaMKK a (Okuno et al., 1994), however, was not found in the granule cell (Nakamura et al., 1996). If CaM-KK b had been observed in the cellular nucleus in the present study, it might phosphorylate CaM-K IV in the nucleus,
Fig. 17. Motoneurons of some cranial nerve nuclei (a–c) and cervical spinal cord (d) show both cytoplasmic and nuclear immunostaining. Arrowheads point to representative labeled neurons. Scale bar = 50 mm.
Fig. 18. This electron micrograph shows a large motoneuron of the facial nucleus. In the nucleus (Nu) reaction products (arrows) are located on dense chromatin structures, and in the cytoplasm on the Golgi apparatus (Go). Electron staining with uranyl acetate or lead citrate for thin sections was not made. AT, axon terminal. Scale bar = 2 mm.
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
187
Table 2 Comparison of immunostainability for CaM-kinases and CaM-kinase kinases in some representative nucleia CaM-KK b *1 Cerebral cortex (Pyramidal neuron)
CaM-KK a *2
CaM-K I *3
CaM-K IV *4
c (+n)
n
c
n
c+n (–) (–)
n n n
c c c
n (–) (–)
Cerebellar cortex Purkinje cell Granule cell
c c
n (–)
c c
(–) n
Motoneuron
c+n
n
c
(–)
Basal ganglia Striatum Globus pallidus Substantia nigra Pars reticulata
a c, Cytoplasmic immunostaining; n, nuclear immunostaining; c(+n), some neurons show nuclear staining, (–), no immunostaining; *1, present study; *2, Nakamura et al. (1996); *3, Picciotto et al. (1995); *4, Nakamura et al. (1995).
but CaM-KK b existed only in the cytoplasm. Furtherobservations will be required to solve the problem of what kind of enzyme can activate CaM-K IV in the cerebellar granule cell. CaM-KK b might activate CaM-K I in the cytoplasm of the cerebellar granule cell, and it might also activate CaM-K I in the Purkinje cell cytoplasm, because Purkinje cells contain CaM-K I, although CaM-K IV was not found in them.
4.4. Relationship of CaM-KK i to neurons in the basal ganglia In the caudate-putamen, CaM-KK b was observed both in the cytoplasm and nucleus, and it might activate CaM-K I in the cytoplasm and CaM-K IV in the nucleus, respectively. On the other hand, CaM-KK b was not localized immunohistochemically in the globus pallidus, entopeduncular nucleus or substantia nigra pars reticulata. This observation coincided well with ours showing that CaM-K IV was not found in these nuclei. Lack of immunolabeling may be explained in two ways; first, the amount of the production of CaMK IV and CaM-KK b in these nuclei is very poor, because these nuclei of the basal ganglia may not need such phosphorylation processes. Second, these enzymes are metabolized too fast to be detected morphologically in these nuclei. In the present study, CaM-K I was localized immunohistochemically in the cytoplasm of these three basal ganglia neurons. Thus, the problem is raised about how CaM-K I could be phosphorylated in these basal ganglia. It should be solved by future studies.
4.5. Relationship of CaM-KK i to motoneurons Most motoneurons showed both cytoplasmic and nuclear immunoreaction of CaM-KK b. Among large
neurons in the central nervous system, motoneurons, some cerebral large pyramidal neurons and magnocellular rubral neurons were stained both in the cytoplasm and nucleus. Other large neurons in many nuclei such as the mesencephalic trigeminal nucleus, lateral vestibular nucleus and gigantocellular reticular formation did not show nuclear labeling. Immunohistochemistry with the antibody against CaM-KK a also revealed immunolabeling of both the cytoplasm and nucleus in the rat motoneurons, although cytoplasmic staining was weak (Nakamura et al., 1996). On the other hand, CaM-K IV has not been found in the motoneuron of
Fig. 19. Immunostaining of some brain areas with antibodies against CaM-KK b (KK b), CaM-K I (KK I) and CaM-K IV (KK IV) are shown at higher magnification. (a) and (b) show giant pyramidal neurons in the layer V of the frontal cortex and motoneurons of the hypoglossal nucleus, respectively stained for CaM-KK b. (c) shows pyramidal neurons in the layer V of the parietal cortex stained for CaM-K I, and (d) shows cellular nuclei of neurons in the layer V of the frontal cortex stained for CaM-K IV. Scale bar = 50 mm.
188
Y. Nakamura et al. / Neuroscience Research 39 (2001) 175–188
the cranial nerve nuclei or in the anterior horn of the spinal cord (Nakamura et al., 1995), although CaM-K I has been shown in the motoneuron cytoplasm (Picciotto et al., 1995). CaM-KK b along with CaM-KK a, therefore, can activate CaM-K I in the motoneuron cytoplasm, because CaM-KK b can activate both CaMK I and CaM-K IV (Okuno et al., 1997). At present, no CaM-Ks have been found in the motoneuron nucleus. If CaM-KK b is translocated from the cytoplasm to the nucleus, what is the role of this enzyme there. Recently, CaM-KK has been proved to activate protein kinase B by phosphorylation (Yano et al., 1998). Even if CaM-Ks are not present in the nucleus, other substances such as protein kinase B might become a substrate for CaM-KK b. Further studies search for unknown substrates for these kinase kinases in the motoneuron nucleus are required.
Acknowledgements The authors thank Mie Taguchi for her technical assistance. This work was partly supported by Grantsin-Aid nos. 08680810 and 11680735 for Scientific Research (C) (Yasuhisa Nakamura) and 10102002 for Specially Promoted Research (Hitoshi Fujisawa) from the Ministry of Education, Science, Sports and Culture.
References Anderson, K.A., Means, R.L., Huang, Q.-H., Kemp, B.E., Goldstein, E.G., Selbert, M.A., Edelman, A.M., Fremeau, R.T., Means, A.R., 1998. Components of a calmodulin-dependent protein kinase cascade. Molecular cloning, functional characterization and cellular localization of Ca2 + /calmodulin-dependent protein kinase kinase b. J. Biol. Chem. 273, 31880–31889. Cruzalegui, F.H., Means, A.R., 1993. Biochemical characterization of the multifunctional Ca2 + /calmodulin-dependent protein kinase type IV expressed in insect cells. J. Biol. Chem. 268, 26171– 26178. Hagiwara, T., Ohsako, S., Yamauchi, T., 1991. Studies on the regulatory domain of Ca2 + /calmodulin-dependent protein kinase II by expression of mutated cDNA in Escherichia coli. J. Biol. Chem. 266, 16401– 16408. Harlow, E., Lane, D., 1988. Antibodies. A Laboratory Manual. Cold Spring Harbor Laboratory, pp. 361–365. Jensen, K.F., Ohmstede, C.-A., Fisher, R.S., Sahyoun, N., 1991. Nuclear and axonal localization of Ca2 + /calmodulin-dependent protein kinase type Gr in rat cerebellar cortex. Proc. Natl. Acad.
Sci. USA 88, 2850– 2853. Kitani, T., Okuno, S., Fujisawa, H., 1997. Molecular cloning of Ca2 + /calmodulin-dependent protein kinase kinase b. J. Biochem. 122, 243– 250. Lee, J.C., Edelman, A.M., 1994. A protein activator of Ca2 + / calmodulin-dependent protein kinase Ia. J. Biol. Chem. 269, 2158– 2164. Nakamura, Y., Okuno, S., Sato, F., Fujisawa, H., 1995. An immunohistochemical study of Ca2 + /calmodulin-dependent protein kinase IV in the rat central nervous system: light and electron microscopic observations. Neuroscience 68, 181– 194. Nakamura, Y., Okuno, S., Kitani, Y., Otake, K., Sato, F., Fujisawa, H., 1996. Distribution of Ca2 + /calmodulin-dependent protein kinase kinase a in the rat central nervous system: an immunohistochemical study. Neurosci. Lett. 204, 61 – 64. Ohmstede, C.A., Jensen, K.F., Sahyoun, N.E., 1989. Ca2 + /calmodulin-dependent protein kinase enriched in cerebellar granule cellsidentification of a novel neuronal calmodulin-dependent protein kinase. J. Biol. Chem. 264, 5866– 5875. Okuno, S., Fujisawa, H., 1993. Requirement of brain extract for the activity of brain calmodulin-dependent protein kinase IV expressed in Escherichia coli. J. Biochem. 114, 167– 170. Okuno, S., Kitani, T., Fujisawa, H., 1994. Purification and characterization of Ca2 + /calmodulin-dependent protein kinase IV kinase from rat brain. J. Biochem. 116, 923– 930. Okuno, S., Kitani, T., Fujisawa, H., 1996. Evidence for the existence of Ca2 + /calmodulin-dependent protein kinase IV kinase isoforms in rat brain. J. Biochem. 119, 1176– 1181. Okuno, S., Kitani, T., Fujisawa, H., 1997. Purification and characterization of Ca2 + /calmodulin-dependent protein kinase kinase b from rat cerebellum. J. Biochem. 121, 155– 160. Paxinos, G., Watson, C., 1998. The Rat Brain in Stereotaxic Coordinates, fourth ed. Academic Press, San Diego. Picciotto, M.R., Zoli, M., Bertuzzi, G., Nairn, A.C., 1995. Immunochemical localization of calcium/calmodulin-dependent protein kinase I. Synapse 20, 75 – 84. Sakagami, H., Umemiya, M., Kondo, H., 2000. Distinct immunohistochemical localization of two isoforms of Ca2 + /calmodulin-dependent protein kinase kinases in the adult rat brain. Eur. J. Neurosci. 12, 89 – 99. Selbert, M.A., Anderson, K.A., Huang, Q.-H., Goldstein, E.G., Means, A.R., Edelman, A.M., 1995. Phosphorylation and activation of Ca2 + /calmodulin-dependent protein kinase IV by Ca2 + / calmodulin-dependent protein kinase Ia kinase: phosphorylation of threonine 196 is essential for activation. J. Biol. Chem. 270, 17616– 17621. Sugita, R., Mochizuki, H., Ito, T., Yokokura, H., Kobayashi, R., Hidaka, H., 1994. Ca2 + /calmodulin-dependent protein kinase kinase cascade. Biochem. Biopyhys. Res. Commun. 203, 694–701. Yamagata, Y., Czernik, A.J., Greengard, P., 1991. Active catalytic fragment of Ca2 + /calmodulin-dependent protein kinase II. Purification, characterization, and structural analysis. J. Biol. Chem. 266, 15391– 15397. Yano, S., Tokumitsu, H., Soderling, T.R., 1998. Calcium promotes cell survival through CaM-K kinase activation of the protein-kinase-B pathway. Nature 396, 584– 587.