GC (guanylate cyclase)-G is the most recently identified member of the receptor GC family. However, the regulation of its activity and protein expression in the mammalian olfactory system remains unclear. In the present study, we used a GC-G-specific antibody to validate that the GC-G protein is expressed in Grueneberg ganglion neurons, a newly recognized olfactory subsystem co-expressing other cGMP signalling components such as the cGMP-regulated PDE2A (phosphodiesterase 2A) and the cGMP-gated ion channel CNGA3 (cyclic nucleotide-gated cation channel α-3). Further molecular and biochemical analyses showed that heterologously expressed GC-G protein, specifically the C-terminal cyclase domain, was directly stimulated by bicarbonate in both in vivo cellular cGMP accumulation assays in human embryonic kidney-293T cells and in vitro GC assays with a purified recombinant protein containing the GC domain. In addition, overexpression of GC-G in NG108 neuronal cells resulted in a CO2-dependent increase in cellular cGMP level that could be blocked by treatment with acetazolamide, an inhibitor of carbonic anhydrases, which implies that the stimulatory effect of CO2 requires its conversion to bicarbonate. Together, our data demonstrate a novel CO2/bicarbonate-dependent activation mechanism for GC-G and suggest that GC-G may be involved in a wide variety of CO2/bicarbonate-regulated biological processes such as the chemosensory function in Grueneberg ganglion neurons.
GCs (guanylate cyclases), which produce the intracellular second messenger cGMP, mediate a broad spectrum of physiological processes, such as arterial blood pressure and volume homoeostasis, endochondral ossification, intestinal electrolyte and fluid transport, epithelial cell growth and differentiation, and phototransduction in the retina . However, knowledge of the complete range of cellular functions modulated by cGMP is expanding and perhaps not fully understood. In mammals, seven isoforms of GCs were designated GC-A to GC-G in order of their discovery . These GCs share an organized protein domain structure of at least four recognizable motifs: an extracellular ligand-binding domain, followed by a single transmembrane segment and a cytoplasmic region that can be further subdivided into a protein kinase-like domain and a C-terminal cyclase catalytic domain. GCs can be activated by extracellular peptide ligands , weakly by intracellular calcium-binding proteins [2,3] or via the small GTPase Rac and p21-activated kinase pathway .
The murine olfactory system is composed of at least four anatomically separate subsystems: the main olfactory epithelium, vomeronasal organ, septal organ and GG (Grueneberg ganglia); each subsystem is likely to have distinct functions [5,6]. GC-D was shown to be expressed in a small subset of rat main olfactory epithelial neurons. Previous genetic and electrophysiological studies demonstrated that GC-D-positive OSNs (olfactory sensory neurons) are stimulated by the peptides uroguanylin and guanylin [7,8]. In addition, GC-D-positive OSNs are responsible for CO2 detection, responses that require the activity of type II CA (carbonic anhydrase) to catalyse the conversion of CO2 to bicarbonate [9,10]. Advanced biochemical studies demonstrated that bicarbonate can directly activate the cGMP-generating activity of GC-D [10,11].
GC-G is the latest identified member of the GC family [12–14]. Use of a commercially available antibody revealed anti-GC-G immunoreactive staining localized in the GG neurons of the olfactory subsystem . However, the specificity of this antibody (i.e. its cross-reactivity with other receptor GCs) has not been comprehensively examined. Therefore, whether GC-G is indeed expressed in olfactory tissues and how GC-G is regulated in GG neurons remain unclear. In the present study, we produced and used an anti-GC-G-specific antibody to validate that GC-G protein is expressed in GG neurons and demonstrated that bicarbonate activates GC-G by directly acting on the cytosolic cyclase catalytic domain of GC-G.
All chemicals were of reagent grade from Sigma. Anti-FLAG M2 and anti-Myc 9E10 monoclonal antibodies were purchased from Sigma and Covance respectively. Anti-CA II antiserum was from Santa Cruz Biotechnology. The vector-based shRNA (short hairpin RNA) to knock down mouse GC-G (clone TRCN0000012167) or a non-targeting shRNA for GFP (green fluorescent protein; clone TRCN0000072178) were obtained from The RNAi Consortium .
Cell culture and transfection
HEK-293T [human embryonic kidney-293 cells expressing the large T-antigen of SV40 (simian virus 40)] and NG108 neuronal cells were maintained in DMEM (Dulbecco's minimal essential medium) supplemented with 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin and 100 μg/ml streptomycin at 37 °C in an atmosphere of 5% CO2. Cells were transfected using Lipofectamine™ 2000 (Invitrogen).
Construction of expression plasmids
The FLAG-tagged FL (full-length) GC-G expression plasmid (FLAG–GC-G) was constructed as described previously . The ΔECD+KLD or ΔCYC expression constructs code for a mutant protein lacking residues 44–455 and 556–833 or residues 834–1003 respectively. The CYC expression plasmid encodes only the C-terminal cyclase domain (residues 692–1100). All other expression plasmids (e.g. FLAG–GC-A to E) were constructed in a similar fashion in the pFLAG-CMV1 vector (Sigma).
Establishment of stable cell lines expressing GC-G
HEK-293T cells were co-transfected with the plasmid encoding FLAG–GC-G and the pcDNA4 vector (Invitrogen) containing the antibiotic zeocin resistance gene. Two days post-transfection, stable GC-G-expressing cells were selected with 400 μg/ml zeocin (Invitrogen). Zeocin-resistant cell clones were examined by anti-FLAG Western blotting to determine the expression of the FLAG-tagged GC-G, and the positive clones were amplified for further analysis.
Immunoprecipitation and Western blot analysis
Two days after transfection, cell lysates were clarified by centrifugation at 10000 g for 20 min at 4 °C. Samples underwent immunoprecipitation, then Western blot analysis was performed as previously described .
All animal procedures were performed according to the protocols approved by the Institutional Animal Care and Utilization Committee, Academia Sinica. Heads of 1-day-old mice (C57BL/6) were dissected, fixed in 4% paraformaldehyde overnight at 4 °C, and embedded in paraffin. Sections were cut at 5 μm. Anti-GC-G immunohistochemical analysis was performed as described previously .
Preparation of GST (glutathione transferase)-fusion proteins
GST-fusion proteins were constructed in the expression vector pGEX-4T (GE Healthcare). The PCR fragment encoding the GC-G cyclase domain (amino acids 849–1048) was ligated into the vector at the BamRI/XhoI sites to generate the GST–GC-G-CYC fusion protein. The fusion protein was expressed in Escherichia coli and purified from the soluble fraction of the bacterial lysate with glutathione–Sepharose beads according to the manufacturer's instructions (GE Healthcare).
GC activity assay
The concentration of cGMP was measured by the CatchPoint cGMP fluorescent assay kit (Molecular Devices). To demonstrate the biocarbonate activation, transfected cells were cultured in bicarbonate-free L-15 medium (GIBCO) in a CO2-deprived incubator for 24 h. Cells were then collected into new tubes and stimulated with bicarbonate at the indicated concentrations for 20 min at 37 °C together with 1 mM isobutylmethylxanthine [a PDE (phosphodiesterase) inhibitor]. Cell lysates were prepared, and cellular cGMP content was determined according to the manufacturer's instructions (Molecular Devices).
For the in vitro GC activity assay, purified recombinant GST–GC-G-CYC protein (1 μg) was mixed with 50 μl of 2× reaction buffer (50 mM Hepes, pH 7.4, 300 mM NaCl, 2 mM isobutylmethylxanthine, 4 mM ATP, 4 mM GTP, 1 mg/ml BSA, 10 mM MgCl2 and 20 mM sodium azide) and the indicated concentrations of bicarbonate to a final volume of 100 μl. After incubation for 20 min at 37 °C, the reaction was stopped by adding 250 μl of ice-cold 100% ethanol. After centrifugation (3000 g, 5 min, 4 °C), the supernatant was dried in a speed-vacuum concentrator, and cGMP content was measured using the cGMP assay kit.
HEK-293T cells were plated on a 35-mm glass-bottom dish. One day later, cells were transiently transfected with use of FuGENE™ 6 (Roche) and an expression plasmid encoding a fluorescent indicator cGMP (α-FlincG)  alone or together with another construct encoding GC-G. After 24 h, transfected cells were stimulated with 30 mM NaHCO3 in culture medium for 15 min. Imaging experiments involved spinning disk confocal microscopy (Ultra-View, PerkinElmer). Cells were imaged on an inverted microscope (IX71, Olympus) with a 60× oil-immersion objective and maintained at 37 °C during the experiments. Intracellular cGMP level reflected by the intensity of the fluorescent cGMP biosensor was recorded by ratiometric excitation at 410 and 480 nm and emission at 510 nm . Images and data were analysed using of MetaMorph Software (Molecular Devices).
Data are expressed as means±S.E.M. Differences between groups were analysed by unpaired Student's t test. P<0.05 was considered statistically significant.
Specificity of the anti-GC-G antibody and expression of GC-G in GG neurons
To validate the specificity of the anti-GC-G antibody we developed previously , we first heterologously expressed each isoform of receptor GCs (GC-A to G) in HEK-293T cells. Western blot analysis revealed that the anti-GC-G antibody specifically recognized the recombinant GC-G protein, but not other receptor GCs (Figure 1A). We then examined the expression of GC-G by immunohistochemistry in the frontal nasal sections containing the GG neurons of neonatal mice using this anti-GC-G-specific antibody. The GG neurons were positive for anti-GC-G immunoreactivity (Figures 1B and 1C). Pre-incubation of anti-GC-G antibody with its corresponding immunogen or omission of the primary antibody resulted in no staining (results not shown), thus demonstrating the specificity of anti-GC-G immunostaining in the GG neurons.
Specificity of the anti-GC-G antibody and expression of GC-G in GG neurons
Bicarbonate increases GC-G activity in cells
Because bicarbonate can directly stimulate the activity of GC-D in a subset of OSNs [9,10], we investigated whether the cyclase activity of GC-G expressed in GG is regulated by bicarbonate in cells. After transient transfection, the expression of the FLAG–GC-G was confirmed by Western blot analysis (Figure 1A) and shown to be targeted on the cell surface of HEK-293T cells by flow cytometry (Figure 2A). Because bicarbonate in the extracellular medium can be transported into cells within minutes and has been used to study its effect on the cyclase activity of sAC (soluble adenylate cyclase) expressed in HEK-293T cells , we adopted a similar system to test the bicarbonate effects on GC-G activity. The addition of NaHCO3 to the extracellular medium stimulated cellular cGMP accumulation in HEK-293T cells expressing GC-G, but not in vector-transfected cells (Figure 2B). This activation was dose-dependent for NaHCO3, with a significant stimulation at as low as 5 mM and saturation at approx. 30–50 mM. This observation agrees with previous studies of the effective bicarbonate concentration on sAC or GC-D [10,11,20] and is within the physiological ranges of intracellular bicarbonate levels for various cellular systems [20,21]. In addition, GC-G activity was stimulated equally well by NaHCO3 and KHCO3, but not NaCl (Figure 2C), which suggests that GC-G activation was specific to the bicarbonate ion and excluded both Na+ ions and altered ionic strength as regulators of GC-G activity.
GC-G expressed in HEK-293T cells is activated by bicarbonate
Consistent with this finding, a GC-G-specific sequence of RNAi (RNA interference)-mediated silencing effectively reduced the GC-G protein expression (>80% suppression of protein expression), which concomitantly diminished not only the basal but also bicarbonate-stimulated cGMP production in the GC-G stable cell line. Together, these results suggest that GC-G is indeed the main factor regulated by bicarbonate in a cell system (Figure 3). Furthermore, bicarbonate-stimulated intracellular cGMP production by GC-G could be dynamically monitored by a fluorescent indicator of cGMP (α-FlincG) (Figure 4) . These results suggest that GC-G cyclase activity can be activated by bicarbonate in a cellular context in the absence of any factor specific to GG neurons.
RNAi-mediated knockdown of GC-G reduced bicarbonate-stimulated cGMP production in a GC-G stable cell line
Fluorescence imaging showing that bicarbonate stimulates GC-G activity
The GC domain is essential for bicarbonate stimulation
To investigate which domain is functionally involved in bicarbonate activation, we engineered three additional deletion constructs expressing (i) the mutant GC-G protein without the extracellular and kinase-like domains (ΔECD+KLD); (ii) the mutant protein lacking the functional cyclase domain (ΔCYC); or (iii) the protein fragment containing only the C-terminal cyclase domain (CYC) (Figure 5A). Western blot analysis confirmed that all of these mutant proteins were expressed in HEK-293T cells (Figure 5B). Furthermore, the addition of bicarbonate-stimulated cellular cGMP production in HEK-293T cells expressed the ΔECD+KLD protein to the same degree as the FL GC-G protein (Figure 5C), which suggests that the extracellular and kinase-like domains of GC-G are not important for bicarbonate activation. However, deletion of the functional cyclase domain in the ΔCYC protein completely abolished the bicarbonate-stimulated cGMP increase. Likewise, the CYC protein containing only the C-terminal cyclase catalytic domain was responsive to bicarbonate activation. Together, these data demonstrate that the cyclase domain of GC-G is critical for bicarbonate sensing and activation, at least when overexpressed in HEK-293T cells.
The GC catalytic domain of GC-G is required for bicarbonate activation
Recombinant protein containing the GC-G GC domain can be directly stimulated by bicarbonate
To validate whether bicarbonate directly acts on the cyclase domain of GC-G and to eliminate further the possibility of yet-unknown cellular factors involved in activation, we produced and purified the recombinant GST-fusion protein containing the GC-G cyclase domain (residues 849–1078) (Figure 6A) to examine the effects of bicarbonate. The addition of bicarbonate significantly enhanced the cyclase activity of the purified protein in a dose-dependent manner (Figure 6B), which is similar to the bicarbonate effects on GC-G expressed in the cells. The median effective concentration (EC50) of bicarbonate activation was ~15 mM, which is within the range of the physiological concentrations of bicarbonate . Most importantly, the bicarbonate stimulation was not caused by alteration in pH value of the reaction solution after the addition of bicarbonate, because neither the recombinant GC-G-CYC protein alone nor bicarbonate-stimulated cyclase activity of purified GC-G-CYC protein was sensitive to pH changes ranging from 7.2 to 7.8 (Figure 6C). Together, our results suggest that bicarbonate can directly activate the cyclase domain of GC-G without any additional factor from cells.
Bicarbonate directly activates the recombinant cyclase domain of GC-G
CO2 stimulates the cyclase activity of GC-G in neuronal cells
Because bicarbonate can be catalytically converted from CO2 by CAs under physiological conditions, we next tested whether GC-G could be regulated by CO2 in a cellular context. We chose NG108 neuronal cells for this study, because these cells contain low basal cGMP level and express both mRNA and protein levels of type II CA (Figures 7A and 7B). Transfection of the GC-G expression plasmid resulted in a 3-fold increase of cellular cGMP level, as compared with vector-transfected neuronal cells cultured without CO2 (Figure 7C). However, exposure to CO2 further enhanced the cGMP levels in these GC-G-expressing cells (Figure 7C). Interestingly, this CO2-induced cGMP increase was completely blocked by treatment with AZ (acetazolamide), an inhibitor of CAs.
CO2 activates the cyclase activity of GC-G expressed in neuronal cells
In the present study, we developed a GC-G-specific antibody to validate the unique expression of GC-G in GG neurons and revealed bicarbonate as a novel activator of GC-G activity in HEK-293T cells. However, the structure and molecular basis for the interaction between bicarbonate and the GC-G cyclase domain remains to be further elucidated. Interestingly, CO2 increased the cellular cGMP accumulation in the GC-G-transfected neuronal cells, which was blocked by treatment with AZ. These data suggest that CO2 must be converted into bicarbonate for the stimulatory effect on GC-G. Our results agree with recent findings for GC-D showing that GC-D-positive OSNs co-expressing type II CA can sense a near-atmospheric concentration of CO2, possibly through the CA II-mediating conversion of CO2 into bicarbonate for GC-D activation [9–11]. Furthermore, at least two downstream cGMP signalling proteins, the cGMP-regulated PDE2A and the cGMP-gated ion channel CNGA3 (cyclic nucleotide-gated cation channel α-3), are expressed in both GC-D-positive OSNs and GC-G-positive GG neurons [15,22,23]. Although CA II mRNA was not detected in GG neurons , its protein expression has yet to be comprehensively examined by immunohistochemistry, and other forms of CAs (16 different isoenzymes ) may be expressed in GG. Therefore, GG cells may respond to CO2 in a GC-G-dependent manner. However, further studies are required to clarify this notion.
Recent studies demonstrated that neuronal cells in the GG, a newly discovered olfactory subsystem [25–29], mediate a chemosensory function in detecting an alarm pheromone  and respond to cool temperatures [23,31,32]. Although little is known about the identity of the alarm pheromone and the signalling cascade involved in the chemosensory function, the cGMPgated ion channel CNGA3 was shown to play a critical role in mediating the responses of the GG neurons to cool temperature . Thus, further investigations into whether GC-G-mediated cGMP signalling is also functionally coupled with CNGA3 and is involved in chemical or thermal sensation in GG is of interest.
In addition to its expression in GG neurons, GC-G protein was shown to be highly expressed in mouse sperm . An increase in cellular bicarbonate is involved in the regulation of sperm activities, including motility, chemotaxis, capacitation, and the acrosomal reaction . GC-G might participate in the aforementioned sperm processes in a bicarbonate-dependent manner. However, this suggestion warrants further investigation. Furthermore, because GC-G mRNA or protein can be detected in several other tissues by RT–PCR (reverse transcription–PCR) or immunohistochemistry [14,34,35], GC-G might be involved in other bicarbonate-dependent physiological or pathological processes. For example, elevated intracellular bicarbonate concentrations under certain metabolic processes  or with tissue ischaemia-reperfusion injury  may represent putative modulating signals for GC-G activation. In agreement with this suggestion, our recent gene-targeting study demonstrated that GC-G may be activated and acts as an early injury signal that promotes apoptotic and inflammatory responses in ischaemia/reperfusion-induced acute renal injury .
In summary, our results reveal a novel biochemical regulatory mechanism for an orphan GC receptor, GC-G. Given that GC-G is co-expressed with the cGMP-gated ion channel CNGA3, GC-G-positive GG neurons may transduce stimuli through an excitatory cGMP-mediated signalling mechanism. Investigations are required to further define the molecular regulation of GC-G and unravel the precise physiological function of the GC-G-mediated cGMP signal pathway in GG neurons and in other organs.
cyclic nucleotide-gated cation channel α-3
green fluorescent protein
human embryonic kidney-293 cells expressing the large T-antigen of SV40 (simian virus 40)
olfactory sensory neuron
soluble adenylate cyclase
short hairpin RNA
Ying-Chi Chao constructed all expression plasmids used in this study and determined the effects of CO2 and bicarbonate on GC-G activity. Chien-Jui Cheng, Hsiu-Ting Hsieh and Chih-Ching Lin performed the immunohistochemical analysis, cell imaging of cGMP indicator and production of recombinant GST fusion protein respectively. Chien-Chang Chen and Ruey-Bing Yang conceived and supervised the overall project, designed biochemical and molecular biology experiments, and co-wrote and edited the manuscript prior to submission.
This work was supported by the National Science Council [grant numbers NSC-97-2320-B-001-009-MY3 (to R.-B.Y.) and NSC 97-2320-B-038-018-MY3 (to C.-J.C)]. RNAi reagents were obtained from the National RNAi Core Facility located at the Institute of Molecular Biology/Genomic Research Center, Academia Sinica, supported by the National Research Program for Genomic Medicine, National Science Council [grant number NSC 97-3112-B-001-016].
These authors contributed equally to this work.