SEARCH

SEARCH BY CITATION

Keywords:

  • Ca2+/calmodulin-dependent protein kinase II;
  • immunoelectron microscope;
  • neurokinin 1 receptor;
  • pre-Bötzinger complex;
  • rat;
  • synapse

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgement
  7. References

The pre-Bötzinger complex (pre-BötC) in the ventrolateral medulla oblongata is a presumed kernel of respiratory rhythmogenesis. Ca2+-activated non-selective cationic current is an essential cellular mechanism for shaping inspiratory drive potentials. Ca2+/calmodulin-dependent protein kinase II (CaMKII), an ideal ‘interpreter’ of diverse Ca2+ signals, is highly expressed in neurons in mediating various physiological processes. Yet, less is known about CaMKII activity in the pre-BötC. Using neurokinin-1 receptor as a marker of the pre-BötC, we examined phospho (P)-CaMKII subcellular distribution, and found that P-CaMKII was extensively expressed in the region. P-CaMKII-ir neurons were usually oval, fusiform, or pyramidal in shape. P-CaMKII immunoreactivity was distributed within somas and dendrites, and specifically in association with the post-synaptic density. In dendrites, most synapses (93.1%) examined with P-CaMKII expression were of asymmetric type, occasionally with symmetric type (6.9%), whereas in somas, 38.1% were of symmetric type. P-CaMKII asymmetric synaptic identification implicates that CaMKII may sense and monitor Ca2+ activity, and phosphorylate post-synaptic proteins to modulate excitatory synaptic transmission, which may contribute to respiratory modulation and plasticity. In somas, CaMKII acts on both symmetric and asymmetric synapses, mediating excitatory and inhibitory synaptic transmission. P-CaMKII was also localized to the perisynaptic and extrasynaptic regions in the pre-BötC.

Abbreviations used
BSA

bovine serum albumin

CaMKII

Ca2+/calmodulin-dependent protein kinase II

I CAN

non-selective cationic current

ir

immunoreactive

mGluR1

group 1 metabotropic glutamate receptors

NGS

normal goat serum

NK1R

neurokinin 1 receptor

PBS

phosphate buffered saline

P

phospho

pre-BötC

pre-Bötzinger complex

PSD

post-synaptic density

TRP

transicent receptor potential

VGCC

voltage-gated Ca2+ channel

The pre-Bötzinger complex (pre-BötC) in the ventrolateral medulla is a functionally and anatomically defined site that is necessary and sufficient for inspiratory neural rhythms (Smith et al. 1991; Rekling and Feldman 1998; Feldman and Del Negro 2006). The pre-BötC neurons generate inspiratory drive potentials by synaptic input evoking post-synaptic currents that depend on intrinsic membrane properties (Rekling and Feldman 1998; Feldman and Del Negro 2006). Two types of inward currents, the persistent sodium current and the Ca2+-activated non-selective cationic current (ICAN), are important for respiratory rhythm generation (Thoby-Brisson and Ramirez 2001; Pena et al. 2004; Del Negro et al. 2005; Ptak et al. 2005; Koizumi and Smith 2008). A recently proposed group-pacemaker hypothesis assumes that ICAN activated by elevated cytosolic Ca2+ is an essential cellular mechanism for shaping the inspiratory drive potentials, which is fully evoked by ionotropic and metabotropic glutamatergic synaptic inputs (Pace et al. 2007; Mironov 2008; Pace and Del Negro 2008; Rubin et al. 2009; Del Negro et al. 2010). AMPA receptors initiate convergent signaling pathways that evoke post-synaptic ICAN underlying inspiratory drive potential (Mironov 2008; Pace and Del Negro 2008). Hence, respiratory network activity is a consequence of a complex orchestration that involves ion channels, receptors, neuropeptides, and biogenic amines (Doi and Ramirez 2008, 2010). The activated ion channels and receptors exert excitatory or inhibitory effects on the respiratory network via different second messenger systems. While there is increasing evidence of membrane's ion channels and receptors in contribution to respiratory rhythmogenesis and control, less attention is paid to the intracellular second messenger systems in the pre-BötC.

Ca2+/calmodulin-dependent protein kinase II (CaMKII) is an ubiquitous second messenger highly expressed in neurons, mediating many diverse physiological processes in response to increases in intracellular Ca2+ (Lisman et al. 2002; Colbran and Brown 2004). In the pre-BötC, cytosolic Ca2+ transients shaped by Ca2+ influx are closely coincident with ICAN activity. Fluorescent imaging of Ca2+ activities with calcium-sensitive dye and two-photon calcium imaging have been used in the isolated slice preparation to determine intrinsic rhythmicity (Frermann et al. 1999; Koshiya and Smith 1999; Mironov 2008; Morgado-Valle et al. 2008; Del Negro et al. 2011). CaMKII is an ideal ‘interpreter’ of diverse Ca2+ signals, and its distribution matches the localization of Ca2+ signals (De Koninck and Schulman 1998; Colbran and Brown 2004). Therefore, understanding CaMKII expression and its subcellular distribution is essential and important for understanding the intracellular Ca2+ signaling activity and neuronal functions. Such information, however, is not yet available for the pre-BötC.

The pre-BötC neurons express high levels of neurokinin-1 receptor (NK1R) and somatostatin (Gray et al. 1999; Wang et al. 2001; Stornetta et al. 2003). Saporin-mediated destruction of the pre-BötC NK1R-immunoreactive (ir) neurons causes sleep-disordered breathing (Mckay et al. 2005) and fatal respiratory pathology (Gray et al. 2001). Similarly, acute silencing of a subset of neurons expressing somatostatin immunoreactivity in the pre-BötC stops spontaneous breathing in awake adult rats (Tan et al. 2008). Using NK1R as a marker of the pre-BötC, we examined phospho (P)-CaMKII (the active form of CaMKII) expression in the pre-BötC. In particular, we focused on the synaptic relationship between P-CaMKII- and NK1R-ir neurons. This study provides cellular and structural evidence for CaMKII activity in the pre-BötC.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgement
  7. References

All experiments were performed on adult Sprague-Dawley rats (230–250 g, Animal Center of the Fourth Military Medical University). A total of 10 animals were used in this study. Protocols were approved by the Northwest China Committee of Experimental Animal Care, and their regulations were in accordance with NIH guidelines. All efforts were made to minimize animal suffering and the number of animals used.

Six rats were used for P-CaMKII or NK1R immunoperoxidase single labeling, and P-CaMKII/NK1R double labeling. Animals were deeply anesthetized with 1% sodium pentobarbital intraperitoneally (50 mg/kg), and perfused transcardially with 150 mL warm 0.9% saline followed by 500 mL ice cold 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for half an hour. The brainstems were removed and postfixed in the same fixative for 1 h at 4°C. They were then cryoprotected in 30% sucrose in 0.1 M phosphate buffer overnight at 4°C. Four sets of serial coronal sections were cut at 12 μm thickness on a cryostat (CM1900, Leica, Heidelberger, Germany), and mounted on gelatin-coated slides for immunohistochemical staining. All subsequent immunohistochemical procedures were done at 20°C.

Immunoperoxidase histochemistry

For single-labeled immunohistochemistry, slides were blocked for 2 h in phosphate buffered saline (PBS, pH 7.4) containing 5% bovine serum albumin (BSA), 5% normal goat serum (NGS), and 0.5% Triton X-100. Slides were then incubated overnight in primary antibody of rabbit anti- P-CaMKII (1 : 500, Cell Signaling Technology, Danvers, MA, USA) or NK1R (1 : 5000, Sigma, St. Louis, MO, USA) diluted in PBS containing 1% BSA, 1% NGS, and 0.5% Triton X-100. After rinsing in PBS, slides were then incubated with biotinylated anti-rabbit IgG (1 : 200, Jackson, West Grove, PA, USA) diluted in PBS containing 1% BSA and 1% NGS for 4 h. Slides were rinsed with PBS and then incubated with ABC solution (1 : 300, Sigma) for 4 h. The reaction was detected with glucose oxidase-3, 3′-diaminobenzidine method (Shu et al. 1988). Slides were then dehydrated and coverslipped. Sections were examined under a microscope (BX-51, Olympus, Tokyo, Japan). The landmarks of the brainstem were defined as described by Guyenet et al. (2002) and according to the atlas of Paxinos and Watson (1998). The Bregma levels of the intervening sections with P-CaMKII or NK1R immunoreactivity were then determined by their location relative to the landmarks (Fig. 1).

image

Figure 1. Coronal sections showing immunoperoxidase labeling of neurokinin 1 receptor (NK1R) (a–d) or phospho-Ca2+/calmodulin-dependent protein kinase II (P-CaMKII) (e–h) at levels of −12.2 mm and −12.5 mm, relative to Bregma, in the ventral medulla oblongata in rats. Boxes outline the pre-Bötzinger complex (pre-BötC) regions (a, c, e, g). Higher magnifications illustrate NK1R- (arrows in b, d) or P-CaMKII (arrows in f, h) immunoreactive (ir) neurons in relevant boxes. Two populations of NK1R-ir neurons, large (b) and small fusiform neurons (f) are visualized in the pre-BötC. Dash circles in a, c, e, g show the region of the nucleus ambiguus at each level. Scale bars = 200 μm (a, c, e, g), 20 μm (b, d, f, h).

Download figure to PowerPoint

Immunofluorescent double labeling

Slides were blocked as described above, and incubated overnight with a cocktail of primary antibodies of rabbit anti-P-CAMKII and guinea pig anti-NK1R (1 : 800, Chemicon, Temecula, CA, USA). Secondary antibody was biotinylated anti-guinea pig IgG, diluted to 1 : 200 (Jackson). Diluting solutions of antibodies were the same as above. Slides were finally incubated with Texas Red conjugated streptavidin and Alexa 488-conjugated anti-rabbit IgG (Molecular Probes, Eugene, OR, USA), diluted to 1 : 500 in PBS containing 0.3% Triton X-100 for 4 h. After rinsing in PBS, slides were then coverslipped with anti-fading medium and examined with a confocal laser-scanning microscope (Fluoview 1000, Olympus) using laser beams of 543 nm and 488 nm with appropriate emission filters for Texas Red (590–610 nm) and Alexa 488 (510–525 nm), respectively. The observation was carried out with an ultraviolet-corrected objective lens (UPLAPO 40x). Digital images were captured by Fluoview application software (Olympus, Tokyo, Japan), arranged, and contrast-enhanced by the computer. Controls in which primary antibodies were replaced with normal serum or pre-immune serum showed no labeling above background.

Automated image analysis was performed with Image-Pro Plus software (Media Cybernetics, Inc., Rockville, MD, USA). NK1-ir neurons in red, P-CaMKII-ir neurons in green, and P-CaMKII/NK1R-ir neurons in yellow were counted, and the proportion of P-CaMKII/NK1R-ir neurons in total NK1R-ir neurons was then calculated in the pre-BötC. Areas of P-CAMKII-ir neurons were determined with ‘area measurement’ function and were presented as mean ± SD. Twenty areas containing the pre-BötC (10 sections) in each brainstem and a total of 120 areas in six animals were collected and analyzed.

Pre-embedding immunogold-silver cytochemistry

Four rats were deeply anesthetized with 1% sodium pentobarbital intraperitoneally (50 mg/kg) and perfused transcardially with 150 mL warm 0.9% saline, followed by 500 mL ice cold mixture of 4% paraformaldehyde, 0.05% glutaraldehyde, and 15% (v/v) saturated picric acid in 0.1 M phosphate buffer (pH 7.4) for 2 h. The brainstems were removed and postfixed by immersion in the same fixative for 3 h at 4°C. Serial coronal sections of 50 μm thickness were prepared with a vibratome (VT 1000S, Leica) for P-CAMKII/NK1R immunohistochemical double labeling. A region encompassing and slightly larger than the pre-BötC was prepared for vibratome sectioning, and approximately 18–20 sections were collected from each brainstem. This region might include a small part of the rostral ventral respiratory group or the BötC. Pre-BötC neurons could not be accurately identified until after immunohistochemical staining when NK1R-ir neurons were identified under the light microscope. The sections were placed in PBS containing 25% sucrose and 10% glycerol for 1 h for cryoprotection. After a freeze-thaw treatment, sections were immersed in PBS containing 5% BSA and 5% NGS for 4 h to block non-specific immunoreactivity. All immunohistochemical procedures were done at 20°C.

P-CAMKII was detected by the immunogold silver-staining method and NK1R by the immunoperoxidase method. The immunogold silver-staining and the immunoperoxidase method have well been used for double labeling under the electron microscope (Liu et al. 2003; Liu et al. 2004, Liu et al. 2005). Immunogold silver-enhanced particles are round or oval with distinct boundary and identical electron density, distributed singularly or in small group, which can be easily distinguished from immunoperoxidase reaction product that is flocculent, irregular, uneven, and with no clear boundary, distributed in puff. Sections were incubated overnight in a cocktail of the primary antibodies of rabbit anti-P-CAMKII and guinea pig anti-NK1R, diluted in concentrations described above in PBS containing 1% BSA and 1% NGS. They were then washed in PBS and incubated overnight in a mixture of secondary antibodies, anti-rabbit IgG conjugated to 1.4 nm gold particles (Nanoprobes, Stony Brook, NY, USA) at a 1 : 100 dilution and biotinylated anti-guinea pig IgG at a 1 : 200 dilution. After rinsing, sections were postfixed in 2% glutaraldehyde in PBS for 45 min. Silver enhancement was performed in the dark with HQ Silver Kit (Nanoprobes) for visualization of P-CAMKII immunoreactivity. Before and after the silver enhancement step, sections were rinsed several times with de-ionized water. They were then incubated in the ABC solution (Sigma) for 4 h and visualized by the glucose oxidase-3, 3′-diaminobenzidine method.

Immunolabeled sections were fixed with 0.5% osmium tetroxide in 0.1 M phosphate buffer for 1 h, dehydrated in graded ethanol series, then in propylene oxide, and finally flat-embedded in Epon 812 (SPI-CHEM, West Chester, PA, USA). After polymerization, sections were examined under the light microscope. Three to four sections containing both P-CAMKII and NK1R immunoreactivity in the pre-BötC were selected from each rat, trimmed under a stereomicroscope, and mounted onto blank resin stubs. Ultrathin sections were cut with an ultramicrotome (EM UC6, Leica) and mounted on mesh grids (6–8 sections/grid). They were then counter-stained with uranyl acetate and lead citrate, and observed under a JEM-1230 electron microscope (JEOL LTD, Tokyo, Japan). Electron micrographs were captured by Gatan digital camera and its application software (832 SC1000, Gatan, Warrendale, PA, USA). The pictures were arranged and contrast-enhanced by the computer.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgement
  7. References

The pre-BötC neurons defined by NK1R immunoreactivity are localized ventral to the nucleus ambiguus, caudal to the retrofacial nucleus (the nucleus ambiguus, pars compacta), and rostral to the anterior tip of the lateral reticular nucleus (Guyenet and Wang 2001; Wang et al. 2001; Guyenet et al. 2002). Identification of pre-BötC neurons in this study was based on this anatomical delineation, in agreement with our previous studies (Liu et al. 2003; Liu et al. 2004, Liu et al. 2005). Figure 1 illustrates pre-BötC NK1R-ir neurons at levels of −12.5 mm and −12.2 mm relative to Bregma. The Bregma level was assigned with reference to the landmark at the caudal end of the facial motor nucleus (−11.6 mm), according to the nomenclature of Paxinos and Watson (1998). NK1R immunoreactivity, present along the plasma membrane of somas and dendrites, highlighted some neurons in the pre-BötC (Fig. 1a, b, e, and f). Two populations of NK1R-ir neurons were observed in this region, including a subtype of large multipolar neurons (Fig. 1a and b) and a population of small fusiform ones (Fig. 1e and f). These were consistent with previous descriptions (Guyenet et al. 2002). The small fusiform NK1R-ir pre-BötC neurons are thought to be interneurons that drive respiratory rhythmogenesis (Guyenet et al. 2002). Using NK1R immunoreactivity as a marker of the pre-BötC, we detected P-CaMKII immunoreactivity in a series of coronal sections (Fig. 1c, d, g and h).

Light microscopic examination of P-CaMKII immunoreactivity in the pre-BötC

P-CaMKII-ir neurons in the pre-BötC were generally oval, fusiform, or pyramidal (Figs 1c, d, g, h, and 3). The area of P-CaMKII-ir neurons is displayed in Fig. 2. P-CaMKII-ir product was distributed within somas and contiguous primary processes (Figs 1d, h, and 3). Double-labeled neurons were clearly observed in the pre-BötC, as P-CaMKII-ir somas and processes had plasma membranes that were NK1R-ir (Fig. 3). Some small fusiform NK1R-ir neurons, the putative rhythmogenic neurons, were also found with P-CaMKII immunoreactivity (small arrows, Fig. 3b; arrows, Fig. 3d). Counts of double-labeled neurons were processed in 60 sections of six brainstems. Among 966 NK1R-ir neurons, 63.7% (615/966) were double-labeled with P-CaMKII in the pre-BötC. Among the double-labeled neurons, 19.3% (119/615) were small fusiform. P-CaMKII-ir neurons larger than 300 μm2 were devoid of NK1R immunoreactivity (star, Fig. 3d).

image

Figure 2. Area distribution of phospho-Ca2+/calmodulin-dependent protein kinase II (P-CaMKII)-ir neurons examined in the pre-BötC. Dark line shows a trend of the area.

Download figure to PowerPoint

image

Figure 3. Immunofluorescent micrographs showing neurokinin 1 receptor (NK1R)/phospho-Ca2+/calmodulin-dependent protein kinase II (P-CaMKII) double-labeled neurons in the pre-BötC. P-CaMKII immunoreactivity is visualized with Alexa 488 (green) and NK1R with Texas Red (red). Colocalization of two different fluorophores is observed in somas (arrows in a–d), and processes (arrowheads in a, b, d). Small fusiform NK1R-ir neurons are also found double labeled with P-CaMKII (small arrows in b, arrows in d). Neurons singularly labeled with NK1R or P-CaMKII are detectable in the region (stars in a–d). Scale bars = 20 μm.

Download figure to PowerPoint

Electron microscopic examination of P-CaMKII in the pre-BötC

Under the electron microscope, immunogold silver-enhanced particles in single or small cluster, indicative of P-CaMKII immunoreactivity, were detected in somas (Fig. 4) and dendrites (Figs 4c, 5 and 6) in the pre-BötC. In particular, P-CaMKII-ir particles were found densely localized to the post-synaptic density (PSD) of asymmetric (arrows, Figs 4c and 5) and symmetric (arrows, Figs 4a, d, 6c and d) synapses. The types of synapses were characterized according to Peters et al. (1991) description. Basically, asymmetric synapses possess wider synaptic cleft and thicker PSD than those of symmetric synapses. Synaptic vesicles in asymmetric synapses are round, whereas they are a mixture of elongated and round ones in symmetric synapses. P-CaMKII synaptic associations were different between the soma and the dendrite in the pre-BötC. In somas, of 21 synapses detected with P-CaMKII immunoreactivity, 8 (38.1%) were of the symmetric type, which is exemplified in Fig. 4a and d. The rest were of the asymmetric type (arrow in Som, Fig. 4c). However, in dendrites, they were mostly localized to asymmetric synapses (arrows in Den, Figs 4c and 5). Of 87 synapses that were P-CaMKII-ir, 81 (93.1%) were of the asymmetric type. Only a few were of the symmetric type (arrows, Fig. 6c and d). Some synapses were devoid of P-CaMKII immunoreactivity (arrow in Den3, Fig. 5a; arrow with T2, Fig. 6a). In addition, P-CaMKII immunoreactivity was found to be expressed at the ‘perisynaptic’ region, referred to as a ~ 100–200 nm site from the PSD edge (double arrowheads, Fig. 6a and b). The ‘extrasynaptic’ region, the plasma membrane beyond the synaptic and perisynaptic regions, was also detectable with P-CaMKII immunoreactivity (unfilled arrows with arrowheads, Figs 4, 5d and e). P-CaMKII-ir particles were also associated with rough endoplasmic reticula and ribosomes. The mitochondrion and Golgi complex were not immunoreactive (Fig. 4). A few particles scattered singly and not associated with P-CaMKII-labeled profiles were regarded as non-specific background labeling (circles with ‘a’ beside them, Fig. 5c and e).

image

Figure 4. Electron micrographs showing neurokinin 1 receptor (NK1R)/phospho-Ca2+/calmodulin-dependent protein kinase II (P-CaMKII) double-labeled neurons in the pre-BötC. Immunoperoxidase reaction product, indicative of NK1R immunoreactivity, is mainly localized along the inner surface of the plasma membrane, the extrasynaptic region (small arrowheads in b, c) and in the soma as well (large arrowheads in b, d). Immunogold particles, indicative of P-CaMKII immunoreactivity, are densely distributed in somas (a–d). They are specifically associated with the post-synaptic density (PSD) of asymmetric (arrows in c) and symmetric (arrows in a, d) synapses. The particles are also localized at the extrasynaptic region (unfilled arrows with arrowheads in a–d). NK1R/P-CaMKII double-labeled neurons are identified in the region (b–d). Axons don't contain NK1R or P-CaMKII immunoreactivity (a, b). Som, soma; T, terminal; Den, dendrite; Axo, axon. Scale bars = 0.5 μm.

Download figure to PowerPoint

image

Figure 5. Electron micrographs showing phospho-Ca2+/calmodulin-dependent protein kinase II (P-CaMKII) immunoreactivity at asymmetric synapses in dendrites. P-CaMKII-ir particles are densely localized to the post-synaptic density (PSD) of asymmetric synapses (arrows in a–e). Some are associated with the extrasynaptic region (unfilled arrows with arrowheads in d, e). Some synapses lack P-CaMKII immunoreactivity (arrow in Den3 in a). NK1R-ir product is observed in dendrites (arrowheads in a–c). The product can be seen at the perisynaptic region (arrowheads in a, b, large arrowheads in c). Immunogold particles in circles with ‘a’ next to them (c, e) indicate non-specific background labeling. T, terminal; Den, dendrite; Axo, axon; a, artifact. Scale bars = 0.2 μm.

Download figure to PowerPoint

image

Figure 6. Electron micrographs showing phospho-Ca2+/calmodulin-dependent protein kinase II (P-CaMKII) immunoreactivity at synaptic and perisynaptic regions in dendrites. Large arrows show synapses. Small arrows point to P-CaMKII-ir particles in association with the post-synaptic density (PSD) in asymmetric (a, b) and symmetric (c, d) synapses. Some particles are observed at the perisynaptic region (double arrowheads in a, b). Arrowheads (b, d) show NK1R immunoreactivity. T, terminal; Den, dendrite; Axo, axon. Scale bars in a, b = 0.5 μm, in c, d = 0.2 μm.

Download figure to PowerPoint

Immunoperoxidase reaction product indicative of NK1R immunoreactivity was localized mainly at extrasynaptic regions in the pre-BötC (small arrowheads, Fig. 4b and c), confirming our previous studies (Liu et al. 2002, 2003, 2005). NK1R immunoreactivity was also detectable at perisynaptic regions (arrowheads, Fig. 5a and b; large arrowheads, Fig. 5c). At times, NK1R-ir product was found in the cytoplasm (large arrowheads, Fig. 4b and d; arrowheads, Fig. 6b and d). Colocalization of P-CaMKII with NK1R was identified in the pre-BötC (Figs 4b–d, 5a–c, 6b and d), consistent with our light microscopic observations. Axons were devoid of either P-CaMKII or NK1R immunoreactivity (Figs 4a, b, 5c, e, and 6a).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgement
  7. References

This study provides morphological evidence of P-CaMKII expression in the pre-BötC. P-CaMKII immunoreactivity is distributed in somas and dendrites, and is specifically associated with synapses. P-CaMKII synaptic associations are different between somas and dendrites. In somas, 38.1% of synapses examined with P-CaMKII expression are of the symmetric type, whereas in dendrites, more than 90% of them are asymmetric. The identification of P-CaMKII at asymmetric synapses in dendrites implies that pre-BötC CaMKII likely senses and monitors Ca2+ activity and phosphorylates post-synaptic proteins, including glutamatergic receptors, to modulate excitatory post-synaptic activity and contribute to respiratory modulation and plasticity. In somas, CaMKII acts on both symmetric and asymmetric synapses that may modulate neuronal excitatory and inhibitory balance. P-CaMKII is also expressed at the ‘perisynaptic’ and ‘extrasynaptic’ regions in pre-BötC neurons.

P-CaMKII in pre-BötC neurons

CaMKII, a serine/threonine kinase, is important for Ca2+-sensitive neuronal activity. Local changes in Ca2+ concentration drive calmodulin binding and CaMKII activation that is controlled further by autophosphorylation at Thr286 to generate an autonomously active form of the kinase (Hudmon and Schulman 2002; Lisman et al. 2002; Schulman 2004). Thus, an antibody against P-CaMKII (Thr286) was used in this study to detect CaMKII activity. We focused on CaMKII activity in the pre-BötC for four main reasons: first, pre-BötC neurons that generate inspiratory drive potentials are critically dependent on ICAN activated by Ca2+ influx (Pace et al. 2007; Mironov 2008; Pace and Del Negro 2008; Rubin et al. 2009), and Ca2+ influx will activate CaMKII. Second, channels mediating ICAN are thought to belong to the activation of the transicent receptor potential (TRP) family, TRPM4 and/or TRPM5, which are activated by intracellular Ca2+ and PIP2 (Crowder et al. 2007; Mironov 2008). Third, ICAN activity is tightly coupled to the activity of ionotropic and metabotropic glutamatergic receptors, including NMDA, AMPA, and group 1 metabotropic glutamate receptors (mGluR1), which are targets of CaMKII (Soderling et al. 2001; Lisman et al. 2002; Bayer et al. 2006; Kristensen et al. 2011). Fourth, optical imaging studies that use cell permeant forms of Ca2 + -sensitive dyes (Frermann et al. 1999; Koshiya and Smith 1999; Mironov 2008; Morgado-Valle et al. 2008) or intracellular dialysis of Oregon Green BAPTA dyes (Del Negro et al. 2011) display Ca2+ transients during inspiration in active pre-BötC neurons. Indeed, we found that most NK1R-ir neurons, including some small fusiform neurons that are putative rhythmogenic neurons, expressed P-CaMKII, indicating a robust Ca2+ activity in the pre-BötC. Sources of Ca2+ influx can be through voltage-gated Ca2+ channels (VGCC), NMDA receptor-mediated Ca2+ entry, and IP3-mediated intracellular Ca2+ release (Pace et al. 2007). AMPA receptors-mediated VGCC is thought to be critical in evoking ICAN activity, whereas NMDA receptors are not essential for respiratory rhythmogenesis (Funk et al. 1997; Morgado-Valle and Feldman 2007; Pace et al. 2007; Pace and Del Negro 2008). A recent study has reported that N-type and P/Q-type calcium channels are required for stable breathing and sighing, and an absence of P/Q-type calcium channels in mice with a genetic ablation of Ca(v)2.1 increases breathing disturbances that lead to early mortality (Koch et al. 2013). L-type VGCC is a CaMKII substrate (Hudmon et al. 2005; Jenkins et al. 2010). The tethered CaMKII is persistently activated in a transition that is sharply dependent on the frequency of Ca2+ oscillations, thereby serving as both a Ca2+ spike frequency detector to monitor Ca2+ channel activity and a resident kinase effector to regulate Ca2+ channel activity (De Koninck and Schulman 1998; Hudmon et al. 2005). High voltage-activated VGCC is the dominant source of somatic Ca2+ in rhythmic respiratory neurons, and the calcium oscillations provide further evidence for rhythmic activation of calcium-dependent conductances or second messenger systems in contributing to rhythmogenesis and modulation (Frermann et al. 1999). Hence, CaMKII is a good second messenger candidate that was, indeed, extensively expressed at synaptic, perisynaptic, and extrasynaptic regions in pre-BötC neurons.

P-CaMKII synaptic activity in the pre-BötC

CaMKII is a remarkably abundant protein at asymmetric synapses and constitutes up to 8% of total proteins in the PSD (Kennedy et al. 1983; Peng et al. 2004; Cheng et al. 2006). The PSD is an electron-dense protein network that contains glutamate receptors and various membrane proteins anchored to cytoskeletal scaffolding molecules (Okabe 2007; Sheng and Hoogenraad 2007). Most rapid excitatory synaptic transmission occurs through NMDA and AMPA receptors within the PSD (Newpher and Ehlers 2008). Synaptically targeted CaMKII is maximally activated by Ca2+ entry and is in a prime position to phosphorylate synaptic targets including AMPA and NMDA receptors (Strack and Colbran 1998; Soderling et al. 2001; Mayadevi et al. 2002; Tsui et al. 2005). NMDA receptor is one of the first CaMKII docking proteins to be identified. It places CaMKII in a strategic location to respond to focal Ca2+ influx through the NMDA receptor (Bayer et al. 2001; Merrill et al. 2005). CaMKII activity induces synaptic potentiation by increasing AMPA receptor trafficking, enhancing AMPA channel conductance, and clustering AMPA receptors, thereby affecting the magnitude of synaptic transmission that contributes to the enhancement of synaptic strength and synaptic plasticity, including learning and memory (Benke et al. 1998; Derkach et al. 1999; Colbran and Brown 2004; Merrill et al. 2005).

We found densely distributed P-CaMKII immunoreactivity that was specifically associated with the PSD in pre-BötC neurons. A majority was localized to asymmetric synapses in dentrides, suggesting phosphorylation of excitatory post-synaptic proteins, including AMPA and NMDA receptors. According to the group-pacemaker model, pre-BötC glutamatergic synaptic transmission elevates post-synaptic Ca2+, thereby activating inward current ICAN and generating inspiratory bursts (Rekling and Feldman 1998; Mironov 2008; Rubin et al. 2009; Del Negro et al. 2010). AMPA receptor-mediated Ca2+ influx is essential to trigger inspiratory drive potential generation (Pace and Del Negro 2008). P-CaMKII synaptic tethering likely senses Ca2+ spike activity and phosphorylates glutamatergic receptors, thereby monitoring excitatory synaptic transmission that may contribute to inspiratory drive within the pre-BötC network.

P-CaMKII immunoreactivity was also localized to symmetric inhibitory synapses in the pre-BötC. This is not surprising, as GABA receptors have been described as substrates of CaMKII (Houston et al. 2007; Guetg et al. 2010; Marsden et al. 2010). CaMKIIα autophosphorylated at Thr286 is localized to GABAA receptors, triggering receptors’ insertion and enhancing inhibitory transmission in dissociated hippocampal and cortical cultures (Marsden et al. 2010). NMDA receptor-dependent internalization of GABABB receptors also requires CaMKII activation (Guetg et al. 2010). If CaMKII-mediated synaptic inhibition is present in pre-BötC neurons, the effect would be more prominent in somas than dendrites, as nearly 40% of synapses detected with P-CaMKII immunoreactivity were localized to symmetric synapses in somas, whereas only 7% were in dendrites. The discrepancy implies different roles of P-CaMKII between the soma and the dendrite, mediating excitatory and inhibitory balance in the former and excitatory-dominated effect in the latter.

P-CaMKII activity in perisynaptic and extrasynaptic regions in the pre-BötC

The perisynaptic region is compositionally or functionally distinct from the extrasynaptic region and contains mGluR1, NMDA, AMPA receptors, as well as the scaffolding protein homer (Newpher and Ehlers 2008; Groc et al. 2009; Opazo and Choquet 2011). It may have functional significance in controlling glutamate receptor exchange or reservoir for receptor trafficking into or out of the PSD (Newpher and Ehlers 2008; Opazo and Choquet 2011). Perisynaptic NMDA receptors can be activated by synaptically released glutamate but only with strong synaptic stimulation, whereas extrasynaptic ones may be activated by glutamate originating from sources other than the synaptically released glutamate (Groc et al. 2009). Consistently, P-CaMKII was expressed at perisynaptic and extrasynaptic regions in the pre-BötC, which may be coincident with Ca2+ activity and correspond to activities of glutamate receptors.

P-CaMKII functional implications

In in vitro slice studies, dendritic Ca2+ transients are thought to be critical for determining inspiratory drive in the pre-BötC, whereas somatic Ca2+ transients do not make significant contributions (Morgado-Valle et al. 2008). Two-photon imaging studies reveal that dendritic Ca2+ activity are followed by somatic Ca2+ transients to promote the generation of inspiratory bursts (Mironov 2008; Del Negro et al. 2011). CaMKII displays a propagating translocation, starting locally and spreading to the entire dendritic field of neurons in hippocampal cultures (Rose et al. 2009). The distribution of P-CaMKII at synapses, perisynaptic, and extrasynaptic regions indicates distinct functional significance of CaMKII in the pre-BötC. At least, P-CaMKII-mediated synaptic activities are different between the dendrite and the soma. Based on the proposed mechanism of pre-BötC rhythmicity that glutamate causes post-synaptic depolarization via AMPA receptors and acts at mGluR1, which collectively increase intracellular Ca2+ to evoke ICAN, P-CaMKII can sense and monitor Ca2+ activity on the one hand, and modulate glutamate receptors’ activities on the other hand, as it potentially participates in respiratory rhythmogenesis, control, and plasticity.

NK1R-ir neurons in the ventral respiratory group are functionally heterogeneous (Wang et al. 2002). The present morphological evidence cannot answer how NK1R/P-CaMKII-expressed neurons in the pre-BötC are actually contributed to respiratory activity. However, a recent study in in vitro brainstem slice provides functional evidence that CaMKII-mediated facilitation of glutamatergic synaptic transmission strengthens synchronous bursting activity within pre-BötC network (Mironov 2013), which corroborates our findings.

Acknowledgement

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgement
  7. References

This study was supported by Natural Science Foundation of China, Grant numbers: 31171102, 31100791, and 30870814. There are no potential financial or any other conflicts of interest regarding this manuscript.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgement
  7. References