Journal list menu
Possible ATP trafficking by ATP-shuttles in the olfactory cilia and glucose transfer across the olfactory mucosa
Abstract
Odor transduction in the cilia of olfactory sensory neurons involves several ATP-requiring enzymes. ATP is generated by glycolysis in the ciliary lumen, using glucose incorporated from surrounding mucus, and by oxidative phosphorylation in the dendrite. During prolonged stimulation, the cilia maintain ATP levels along their length, by unknown means. We used immunochemistry, RT-PCR, and immunoblotting to explore possible underlying mechanisms. We found the ATP-shuttles, adenylate and creatine kinases, capable of equilibrating ATP. We also investigated how glucose delivered by blood vessels in the olfactory mucosa reaches the mucus. We detected, in sustentacular and Bowman's gland cells, the crucial enzyme in glucose secretion glucose-6-phosphatase, implicating both cell types as putative glucose pathways. We propose a model accounting for both processes.
Abbreviations
ACIII, adenylate cyclase type III
AK, adenylate kinase
CK-BB, brain isoform of creatine kinase
Cr, creatine
G6Pase, glucose-6-phosphatase
GLUT, glucose transporter
OE, olfactory epithelium
OSN, olfactory sensory neuron
PCr, phosphocreatine
SCs, sustentacular cells
The olfactory mucosa consists of the olfactory epithelium (OE), at its surface, and the lamina propria, beneath. The epithelium is mostly populated by olfactory sensory neurons (OSNs) and surrounding sustentacular cells (SCs). These two cell types project, respectively, apical cilia and microvilli into a mucus layer that coats the epithelial surface. The lamina propria is irrigated by blood vessels and contains the Bowman's glands, whose ducts extend and open to the mucosal surface (Fig. 4). These glands and the SCs are thought to generate mucus, whose composition is largely undetermined [1, 2].
Odor transduction occurs in olfactory cilia, triggered when an odorant binds to an odor receptor. This receptor couples to a G-protein that activates adenylate cyclase ACIII that generates cyclic AMP. This second messenger opens cyclic nucleotide-gated channels, raising luminal Ca2+ and thereby opening Ca2+-dependent Cl− channels. Depolarizing currents through both channels generates the receptor potential, which is maintained during prolonged exposure to odorants. During such long stimuli, key olfactory enzymes, such as ACIII, kinases, and ATPases, continue to consume ATP in the cilia [2-4]. How the cilia maintain ATP levels along their entire length and fulfill these energy demands is not well understood.
The ATP used by the cilium has at least two origins: oxidative phosphorylation in the knob mitochondria and glycolysis within the cilium [5]. Inhibitors of oxidative phosphorylation and suppression of glycolysis, either by removing the extracellular glucose or with glycolytic inhibitors, significantly reduce the odor response in dissociated OSNs [5].
Olfactory sensory neurons possess a single dendrite having a dendritic knob at its apical end, whence the cilia emanate. These structures (> 30 μm long, < 0.2 μm wide) lack mitochondria and so do not generate ATP via oxidative phosphorylation. The knob can accommodate only 2–3 mitochondria [6], so, assuming that the ATP diffusion coefficient within a cilium is the same as in the cytoplasm (~ 0.9 × 10−6 cm2·s−1) [7], it is unlikely that the ATP generated in the knob could reach distant regions of the cilium, especially during extended chemotransduction activity, when its effective diffusion would drop drastically. An ATP molecule probably moves about 10 μm in 500 ms when cilia are at rest [7].
ATP-shuttles, such as creatine kinase (CK) and adenylate kinase (AK), commonly equilibrate ATP along finger-like organelles like cilia and flagella [8-10]. Raff and Blums [11] performed a theoretical analysis of ATP levels in ciliary structures where ATP is originated from a point source and is hydrolyzed along them. They found that, in order to maintain ATP within physiological values in structures longer than 20 μm, there is an absolute need for ATP regenerating systems like the creatine phosphate shuttle. As stated in [12], these kinases function as ‘phosphoryl wires’ for transferring a phosphoryl group between ATP-generating and ATP-consuming sites. For rapidly equilibrating chemical reactions, a flux wave can travel faster than the diffusing reactants [12]. In the case of the CK shuttle, a phosphate group is transferred by CK from ATP to creatine (Cr) forming phosphocreatine (PCr), which is then used by CK to phosphorylate ADP, generating ATP [13]. Notably, PCr diffuses ~ 40% faster than ATP [14]. It is conceivable that ATP-shuttles could make ATP continually available for the transduction machinery all along the cilium. The enzymes creatine and adenylate kinase (AK) can perform other cellular functions besides their shuttle activity [10, 15].
We previously found that glucose, delivered by the blood vessels of the lamina propria [16], is present at millimolar levels in the mucus [5]; the pathways through which glucose reaches the mucus have not been established. In general, glucose levels within cell bodies are extremely low, because once entering a cell, it is rapidly turned into glucose-6-phosphate by the glycolytic enzyme glucokinase. In order for a cell to secrete glucose through glucose transporters, G6P must be dephosphorylated back to glucose by glucose-6-phosphatase (G6Pase). There are three known isoforms of this enzyme [17]: the classical G6Pase that is highly expressed in the liver, G6Pase2, which is found almost exclusively in pancreatic β-cells [18], and the ubiquitous G6Pase3. It has not been determined which cell types of the olfactory mucosa might be expressing G6Pase.
Here, we explored the presence of ATP-shuttles that might function as a possible mechanism to equilibrate ATP levels along the cilium. We also investigated possible pathways transferring glucose from the blood vessels to the mucus, where it is taken up by the cilia for ATP generation.
Methods
Animals
Sprague Dawley male rats of 18–40 days old were obtained from the animal facility of the Pontificia Universidad Católica de Chile. The animals were handled and sacrificed with isoflurane inhalation followed by decapitation, according to the guidelines of the Animal Ethics Committee of the Universidad de Chile and Fondo Nacional de Ciencia y Tecnología de Chile (FONDECYT).
Dissociation of the olfactory mucosa
Olfactory mucosae were obtained using a simplification of the method described in [19]; we used mechanical dissociation instead of enzymatic treatment. The complete olfactory mucosa was removed and placed in divalent cation-free Ringer's solution containing (mm): 140 NaCl, 5 KCl, 10 Glucose, 10 HEPES, pH 7.4 at 37 °C. The tissue was carefully disrupted with fine tweezers, cut into small pieces (~ 1 mm2), and gently passed several times through a 1 mL pipette tip. The suspension was transferred to a Falcon tube and centrifuged at 300 r.p.m. for 15 min; the supernatant was discarded and the cells in the bottom were immediately used for the immunofluorescence.
Immunofluorescence
The immunocytochemistry was done according to Mura et al. [20]. In the fixation step, we added 2 μm Taxol for preserving cilia. The primary antibodies were mouse anti-CK-BB 1 : 100 (Abcam, Cambridge, MA, USA; ab125114), mouse anti-AK4 1 : 100 (Santa Cruz, Paso Robles, CA, USA; sc-271161), and rabbit anti-G6Pase3 1 : 250 (Bioss Antibodies, Woburn, MA, USA; bs-13253R-Biotin); the anti-CK-BB and anti-AK4 antibodies were also used for immunohistochemistry and western blots. The antibodies were diluted in blocking solution (5% BSA in PBS) and incubated overnight at 4 °C under gentle shaking. The same protocol was used for the negative controls, omitting the incubation with the primary antibodies. The secondary antibodies were goat anti-mouse Alexa fluor 488 1 : 200 (Abcam; ab150113), goat anti-rabbit Alexa 488 1 : 250 (Invitrogen, Carlsbad, CA, USA; A27034), Streptavidin Alexa fluor 488 1 : 500 (Thermo Fisher, Waltham, MA, USA; S323545), goat anti-rabbit Alexa fluor 546 conjugate 1 : 500 (Invitrogen; A-11035), and donkey anti-goat Alexa fluor 350 1 : 200 (Invitrogen; A-21081) prepared in phosphate buffer saline (PBS).
For immunohistochemistry, rats were anesthetized with isoflurane and sacrificed by decapitation. The turbinates were fixed overnight in 4% PFA in PBS at 4 °C. The fixed tissue was cryoprotected in 30% sucrose in PBS for 16 h at 4 °C and embedded and frozen in Tissue-tek OCT (Sakura Finetek, Torrance, CA, USA). Fourteen micrometer thick coronal cryosections on microslides (Superfrost PWS; VWR, 48311-703) were permeabilized with 0.3% Triton X-100, 5% BSA overnight at 4 °C, washed four times with PBS, incubated with fluorescently labeled secondary antibodies in PBS for 2 h at room temperature, and then mounted in fluorsave (Calbiochem, San Diego, CA, USA). Primary antibodies were mouse anti-CK-BB 1 : 100, mouse anti-AK4 1 : 100, and rabbit anti-ACIII 1 : 200. Secondary antibodies were goat anti-mouse Alexafluor 488 1 : 200 (Abcam A-150113) and Alexafluor 546 (Abcam A-11035)[21]. No staining was observed in cryosections incubated with secondary antibodies only.
Ciliary-enriched and deciliated epithelium membranes
We followed the protocol of Kuhlmann et al. [22], using olfactory epithelia (OE) of rats instead of mice. For obtaining deciliated epithelia membranes, we homogenized the material remaining after detaching the cilia from the tissue in sucrose buffer (0.3 m sucrose, 1 mm EDTA, 15 mm Tris-Cl pH 7.4). This homogenate was centrifuged at 1000 g for 10 min at 4 °C, the supernatant was centrifuged at 7700 g for 10 min at 4 °C, and the new supernatant was centrifuged at 100 000 g for 60 min at 4 °C.
Microsomal fraction from cerebellum
This fraction was obtained after homogenizing the tissue in sucrose buffer. The centrifugation steps were 1000 g for 10 min at 4 °C; 10 000 g for 30 min at 4 °C; and 100 000 g for 60 min at 4 °C.
Western blot
Western blots were performed as in [5]. Isolation of a membrane fraction enriched in olfactory cilia was prepared according to [22]. On top of the aforementioned immunofluorescence antibodies, we used rabbit adenylate cyclase type III 1 : 200 (Santa Cruz; sc-588) and mouse actin 1 : 5000 (Sigma Aldrich, St. Louis, MO, USA; 3853). Actin was used as a loading control and ACIII as a cilia marker.
Reverse transcription-PCR of dissociated olfactory sensory neurons and control tissues
A suspension of dissociated cells was obtained as described above and transferred to a Petri dish containing RNase-free Ringer's solution (mm): 140 NaCl, 5 KCl, 1 CaCl2, 1 MgCl2, 10 HEPES, 10 glucose, 1 Na-pyruvate, pH 7.4. A drop of the suspension was placed on a coverslip and the individual OSNs were identified by visual inspection under an inverted microscope (Nikon Eclipse TE2000-S, Nikon,Taoyuan City, Taiwán) equipped with a 40× objective. The OSNs were collected by means of sterile borosilicate micropipettes (~ 4 μm tip) built in a micropipette puller (HEKA Elektronik, Lambrecht (Pfalz), Germany; PIP-6). Individual OSNs were gently drawn into the pipette by negative pressure and transferred to a DNase/RNase-free PCR tube, breaking the pipette tip against the bottom. Approximately 150 OSNs were collected in each tube with several pipettes. The tube contained 16 μL of concentrated PCR buffer; after adding the cells, the final concentration was (mm) Tris-HCl, 500 KCl, 6, MgCl2, 2 DTT (Invitrogen), 8 U RNase OUT, and 1 U DNase (Thermo Fisher). The procedure continued with a standard retrotranscription protocol, where the enzyme was the reverse transcriptase M-MLV (Invitrogen). Primers were designed using Primer-BLAST with the DNA sequence encoding for rat AK1 (Access number NM_024349.3), AK4 (Access number NM_017135.3), G6Pase2 (Access number: XM_017592318.1), and G6Pase3 (Access number NM_176077.3). The primers for G6Pase described in [23] were used. RT-PCR reactions were carried out with My Taq Red Mix 2x, according to the manufacturer instructions. The primers were as follows: for CK-BB (30 cycles), 5′–3′ (ACCTCAACCCAGACAACCTG) forward and (CTGGACAGGGCTTCTACTGC) reverse (Tm = 55 °C); for AK1 (30 cycles), 5′–3′ (CGGCTTCTTGATTGACGGCT) forward (Tm = 61 °C) and (CACAATGCCGCGTTTGTCAT) reverse (Tm = 60 °C); for AK4 (35 cycles), 5′–3′ (CTCAAGTGCTGGGGGTTGAT) forward and (CTGCGTCCTTGTACCGTCTT) reverse (Tm = 60 °C); for G6Pase (30 cycles), 5′–3′ (ACTGGTTCAACCTCGTCTT) forward and (CGAAAGATAGCGAGAGTAGA) reverse (Tm = 55 °C); for G6Pase2 (30 cycles), 5′–3′ (GCTCAGGGTCACAAGCTCAT) forward and (CAGAGCACAGAGCAAGTGGA) reverse (Tm = 60 °C); for G6Pase3 (35 cycles), 5′–3′ (GGTGTTATCCTTGGCTGGCT) forward and (TCCTAGGTGTACTCGCCGAA) reverse (Tm = 60 °C).
The mRNAs from cerebellum, muscle, liver, and lung were obtained by a standard TRIZOL method. The products were analyzed by electrophoresis on a 2% agarose gel.
Results
To investigate mechanisms that maintain steady ATP levels in olfactory cilia, we sought to determine expression of the ATP-carriers creatine and AKs that could fulfill this role. Additionally, we studied pathways that might be involved in transferring glucose from the lamina propria of the olfactory mucosa to the mucus embedding the olfactory cilia. We used immunocytochemistry, immunohistochemistry, immunoblotting, tissue RT-PCR, and RT-PCR of dissociated OSNs.
The ATP-shuttle creatine kinase is expressed in the olfactory cilia
We explored the expression of CK-BB in the olfactory cilia by means of immunocytochemistry. Figure 1A–E reveal the presence of the enzyme in the cilia in a representative isolated OSN. Figure 1A provides a phase contrast view of the cell, Fig. 1B shows the fluorescence after labeling the cell with anti-CK-BB antibody, and Fig. 1C is the overlay of images A and B; the inset in Fig. 1C is a magnified view of the apical part of the OSN. The phase contrast image of a control OSN not treated with the primary antibody and its fluorescent image are shown in Fig. 1D,E. These observations show that CK-BB is present in the olfactory cilia, consistent with the notion that it could participate as an ATP-shuttle in these organelles. Labeling is also observed in the rest of the cell, possibly reflecting the involvement of the kinase in other cellular functions.
For further evidence for the ciliary presence of CK-BB, we conducted immunohistochemistry of the OE. The ciliary layer immunoreacted to the antibody against CK-BB and to an antibody against ACIII (Fig. 1F–H). Controls without the primary antibodies are shown (Fig. 1I–K). As a control, we also tested both antibodies on deciliated epithelium and, as expected, no ciliary labelings was observed; nevertheless, the somatic CK-BB labeling was still present (Fig. 1L–N). Therefore, this experiment confirms the presence of CK-BB in the olfactory cilia.
As an additional complementary experiment to examine whether the enzyme is present in the cilia, we searched for it in immunoblots of a ciliary-enriched membrane preparation. A cerebellar microsomal membrane preparation was used as positive control and deciliated epithelium to control for cilia enrichment, using ACIII as cilia marker (Fig. 1O; quantification shown bellow; N = 4). The labeling of the ciliary fraction by the anti-CK-BB antibody supports the evidence provided by immunocytochemistry and immunohistochemistry results.
The previous results imply that the CB-KK mRNA should be found in the OSNs. Accordingly, using RT-PCR of dissociated OSNs, we detected the mRNA of the brain CK-BB isoform of creatine kinase (Fig. 1P, lane 2). This experiment was conducted in isolated OSNs collected from dissociated olfactory epithelia to make sure that this mRNA proceeded exclusively from this cell type. We used cerebellum and lung as positive and negative controls, respectively (lanes 3 and 4). This finding is in agreement with the expression of CK-BB in OSNs.
The evidence from these four experiments supports the notion that the CK-BB ATP-shuttle is implicated in fulfilling the energy requirements of the olfactory cilia.
The ATP-shuttle adenylate kinase is expressed in the olfactory cilia
We used the same approach to determine whether AK is expressed in the cilia. The presence of the enzyme in the cilia was revealed by immunocytochemistry in dissociated neurons (Fig. 2A–E). As with the CK-BB, we conducted immunohistochemistry of the OE, confirming the localization of this ATP-shuttle in the cilia (Fig. 2F–K). There is also the presence of the enzyme in other parts of the cell, possibly indicating its participation in other processes. An immunoblot of a ciliary-enriched membrane fraction further supports the idea that AK is expressed in the cilia (Fig. 2L; N = 4). RT-PCR of dissociated OSNs indicated the presence of mRNA for the isoforms AK1 and AK4 (Fig. 2M, line 2, both gels); positive controls for AK1 and AK4 were muscle and cerebellum, respectively (lanes 3); the negative controls were liver and lung, respectively (lanes 4).
Together, these results suggest that the AK ATP-shuttle participates in equilibrating ATP levels within the olfactory cilia.
Presence and distribution of glucose-6-phosphatase in the olfactory mucosa
To test for the presence of G6Pase in the olfactory mucosa, we first examined the expression of the protein by immunohistochemistry of the olfactory mucosa (Fig. 3A–H). Several Bowman's glands in the lamina propria and SCs in the OE are labeled by an anti-GTPase3 antibody (Fig. 3B,D,E); the OE corresponds to the diagonal band located at the top left part of the images. The Bowman's glands are also revealed by an antibody against the gland marker SOX9 (Fig. 3C–E). The overlays of the images in Figure 3B,C, and Figure 3A–C are presented in Fig. 3D and E, respectively. The fluorescence image of controls not treated with the primary antibodies and the respective phase contrast image are shown in Fig. 3F–H.
We also examined whether G6Pase3 localized to the SCs using immunocytochemistry of dissociated SCs. Figure 3I depicts a phase contrast image of a group of SCs, Fig. 3J displays the corresponding immunofluorescence image with the anti-G6Pase3 antibody, and Fig. 3K shows the overlay of the two images. Labeling of the enzyme is observed in the cell bodies and, importantly, in the microvilli. Figure 3L,M are the phase contrast and fluorescence images of the control without the primary antibody of a different SCs group. Combined, these results indicate that G6Pase3 is found both in Bowman's gland cells and in the apical region of SCs.
We then searched for the mRNAs of the three different isoforms of the enzyme by means of RT-PCR in extracts of olfactory mucosa and liver (positive control for both isoforms [24]). In the olfactory mucosa, we detected G6Pase 3 (Fig. 3N, lane 4), but neither G6Pase (lane 2) nor G6Pase 2 (data not shown). The respective positive controls are also shown (lanes 1 and 5). This result indicates the presence of the enzyme mRNA in the mucosa, as predicted from the immunohistochemistry and cytochemistry.
These observations are consistent with the view that the SCs and the Bowman's gland cells are involved in secreting glucose to the mucus.
Discussion
This study presents two main contributions: (a) the discovery of two ATP-shuttle enzymes in the olfactory cilia most probably implicated in equilibrating ATP levels along the entire cilium to fulfill the requirements of chemotransduction; and (b) the detection of G6Pase3 in the olfactory mucosa, specifically in the cells believed to secrete glucose to the mucus, namely, the sustentacular and Bowman's gland cells.
We previously showed that both oxidative phosphorylation and glycolysis provide most of the ATP used by olfactory cilia to fuel chemotransduction, and we suggested that some mechanism might also be involved in equilibrating ATP levels along the cilia [5]. Pursuing this idea, we searched for a mechanism that could equilibrate ATP along the entire length of the cilium, such as ATP-shuttles, that could function as temporal ATP buffers [13]. In general, phosphagens are expressed in cells that, like the OSNs, exhibit variable rates of ATP consumption [25]. The CK-BB isoform that we detected in the olfactory cilia is commonly associated with processes requiring significant ATP turnover.
The ATP-shuttle systems have been shown to be important for equilibrating ATP along finger-like organelles such as sperm flagella [26] and Paramecium cilia [27]. Villar et al. [5] proposed that these ATP-shuttles might equilibrate ATP along the cilia in periods of high demand by the transduction machinery.
In a proteomics study in mice, Kuhlman et al. [22] found AK1 in a ciliary-enriched membrane fraction. By RT-PCR of dissociated OSNs, we found the mRNAs for isoforms AK1 and AK4 of this enzyme; the immunoblot of ciliary-enriched membranes also detected isoform AK4. If the different three-dimensional structures of the AK isoforms are assumed to be similar, it is possible that the antibody to AK4 also recognizes AK1 in our immunocytochemistry. The fact that Kuhlmann et al. [22] did not find AK4 might be due to a lower expression level of this isoform compared to AK1. Ibarra-Soria et al. [28] detected isoforms AK4 and AK1 in the whole mucosa, consistent with our results.
Nunez-Parra et al. [16] observed a stereotyped, polarized distribution of glucose transporters in cells of the olfactory mucosa. They detected GLUT1 in the blood vessels in the Bowman's gland cells and the basolateral membrane of the SCs, and GLUT3 in the microvilli and the cilia. Considering the differences in affinity for both transporters [29], they suggested that glucose is transferred directionally from the lamina propria to the epithelial surface. In their model, glucose delivered from the blood vessels of the lamina propria would be incorporated by the SCs and released to the mucus, where it could be used by the cilia as an energy source for sustained chemotransduction. Villar et al. [5] confirmed this speculation, documenting the presence of millimolar glucose in the mucus that the cilia can uptake via GLUT3 and use as an effective energy supply.
A necessary step preceding the secretion of glucose is the metabolization of G6P to glucose by G6Pase. If the SCs participate in glucose secretion, one would expect to find G6Pase in the vicinity of their apical membrane. Accordingly, isoform 3 of this enzyme was clearly expressed in the microvillar region. We also found well-defined G6Pase labeling in the Bowman's gland cells. Therefore, our evidence supports the idea proposed by Nunez-Parra et al. [16] of directional traffic of glucose, and it extends the model of Villar et al. [5] by suggesting possible pathways through which glucose might reach the mucus to be incorporated by the cilia.
Based on this and previous work [5, 16], we show a scheme that proposes how ATP levels are sustained and equilibrated in chemotransducing cilia (Fig. 4). In this model, glucose flows from the blood vessels to the mucus through two pathways: the SCs and the Bowman's gland cells, both involving G6Pase3. The model also proposes a role for the CK and AK as ATP-shuttles to maintain appropriate ATP levels for sustaining chemotransduction along the cilium during long odorant exposures.
Acknowledgements
This work was funded by a FONDECYT grant, No. 1140520 (JB) and FONDEQUIP grant EQM140131 (JB, CV). We thank Julio Alcayaga, Pabla Aguirre, Daniela Astudillo, Ricardo Delgado, Pablo Lazcano, Casilda Mura, Cristina Olmos, Gina Sánchez, and Lorena Saragoni for help in the experiments, and Peter O'Day for advice. We specially thank the Editor and the two reviewers for their help to improve the presentation of the manuscript. The authors acknowledge the services provided by UC CINBIOT Animal Facility funded by PIA CONICYT ECM-07.
Author contributions
JB provided funding; JB and CV conceived the idea, designed and supervised the experiments, wrote the manuscript and designed the figures. CA designed and performed experiments, designed the figures. KB performed experiments.
