Islets of Langerhans contain γ-aminobutyrate (GABA) and may use it as an intercellular transmitter. In β-cells, GABA is stored in synaptic-like microvesicles and secreted through Ca2+-dependent exocytosis. Vesicular inhibitory amino acid transporter (VIAAT), which is responsible for the storage of GABA and glycine in neuronal synaptic vesicles, is believed to be responsible for the storage and secretion of GABA in β-cells. However, a recent study by Chessler et al. indicated that VIAAT is expressed in the mantle region of islets. In the present study, we investigated the precise localization of VIAAT in rat islets of Langerhans and clonal islet cells and found that it is present in α-cells, a minor population of F-cells and αTC6 cells, and clonal α-cells but not in β-cells, δ-cells, or MIN6 m9-cells (clonal β-cells). Combined biochemical, immunohistochemical, and electronmicroscopical evidence indicated that VIAAT is specifically localized with glucagon-containing secretory granules in α-cells. ATP-dependent uptake of radiolabeled GABA, which is energetically coupled with a vacuolar proton pump, was detected in digitonin-permeabilized αTC6 cells as well as in MIN6 m9 cells. These results demonstrate that functional neuronal VIAAT is present in glucagon-containing secretory granules in α-cells and suggest that the ATP-dependent GABA transporter in β-cells is at least immunologically distinct from VIAAT. Because glucagon-containing secretory granules also contain vesicular glutamate transporter and store l-glutamate, as demonstrated by Hayashi et al., the present results suggest more complex features of the GABAergic phenotype of islets than previously supposed.
γ-Aminobutyrate (GABA) is an inhibitory chemical transmitter in the central and peripheral nervous systems (1,2). Islets of Langerhans contain GABA, the enzymes that catalyze its biosynthesis (GAD) and metabolism (GABA transaminase), and GABA receptors at concentrations comparable with levels in the central nervous system (3–7). In particular, β-cells contain a much higher concentration of GABA than other types of islet cells. In β-cells, GABA is stored in synaptic-like microvesicles (SLMVs), endocrine counterparts of neuronal synaptic vesicles, but not in insulin granules (8). ATP-dependent vesicular GABA transport activity has been detected in isolated SLMVs (9). Furthermore, β-cells and MIN6 cells (clonal β-cells) secrete GABA through Ca2+-dependent and -independent pathways (10–14). Exogenous GABA has been shown to inhibit glucagon secretion by way of GABAA receptors on α-cells (15–17). These results indicate that β-cells are the major sites of GABA signal appearance in the islets and that GABA functions as a paracrine-like intercellular chemical transmitter that regulates islet function.
Vesicular inhibitory amino acid transporter (VIAAT) is responsible for the storage of GABA and glycine in synaptic vesicles through active transport at the expense of ATP hydrolysis by a vacuolar proton pump (18,19). The cDNA encoding VIAAT has been cloned, and its primary amino acid sequence has been determined (19). Because VIAAT is a potential probe for GABA signal appearance in neurons, it can be supposed that islet cells, especially β-cells, express VIAAT. Very recently, it was found that islet cells actually contain VIAAT (20). Unexpectedly, however, immunohistochemical evidence obtained with antibodies against either the COOH-terminal or NH2-terminal region of VIAAT indicated that VIAAT immunoreactivity is predominantly present in non−β-cells and probably α-cells, although the precise localization of VIAAT at the cellular or subcellular level has not yet been determined (20).
l-glutamate may also function as an intercellular messenger in the islets of Langerhans (21). We have shown that vesicular glutamate transporter (VGLUT), which is responsible for the vesicular storage of l-glutamate and thus a potential probe for l-glutamate signal appearance, is expressed in α-cells and is specifically associated with glucagon-containing secretory granules (21,22). In fact, α-cells cosecrete stoichiometric amounts of l-glutamate and glucagon upon stimulation by either β-aderenergic receptors or low-glucose conditions (22). Thus, it is of interest to determine whether α-cells express VIAAT; if so, the subcellular localization of VIAAT is of special interest with reference to the localization of VGLUT.
In this study, we found that functional VIAAT is expressed and localized with glucagon-containing secretory granules in α- and αTC6 cells but not in β-cells, δ-cells, or MIN6 cells. VIAAT and VGLUT are coexpressed and colocalized in α-cells. It is suggested that vesicular GABA transporter immunologically distinct from VIAAT is present in β-cells, and therefore the mode of action of GABA in the islets of Langerhans is more complex than previously supposed.
RESEARCH DESIGN AND METHODS
Animals and cell cultures.
Preparation of anti-VIAAT antibodies.
The DNA fragment encoding the seventh transmembrane-loop region of VIAAT (236 bp), which corresponds to D364-C437 (19), was amplified by PCR and then cloned into the EcoRI and XhoI sites of expression vector pGEX4T-2 (Amersham Pharmacia Biotech) to obtain a GST fusion-expressing plasmid. The plasmid was transformed to Escherichia coli BL21, and the resultant transformant was cultured and harvested after induction with 1 mmol/l isopropyl 1-thio-β-d-galactoside for 3 h. The E. coli cells were suspended in 50 mmol/l Tris-HCl (pH 8.0) 50 mmol/l NaCl, 1 mmol/l EDTA, and 1 mmol/l dithiothreitol and were sonicated. The lysate was centrifuged, and the resultant supernatant was applied to a 1 ml glutathione-Sepharose 4B column. The GST fusion proteins were eluted with 16 mmol/l reduced glutathione in buffer A. The GST fusion protein was injected twice at 2-week intervals into a rabbit with complete adjuvant and then with incomplete adjuvant.
Samples were denatured with SDS sample buffer containing 1% SDS and 10% β-mercaptoethanol, and then Western analysis was conducted as described (22).
The published procedure was used (22). In brief, a paraformaldehyde-fixed pancreas was sectioned, and the sections were mounted on slide glass. The sections were washed with PBS, incubated with the same buffer containing 0.1% Triton X-100 for 30 min, further washed with 2% goat serum in phosphate-buffered saline containing 0.5% BSA, and finally reacted with antibodies at 1 μg/ml or 1,000× diluted (VIAAT antibodies and others) in PBS containing 0.5% BSA for 1 h at room temperature. Samples were washed three times with PBS and then reacted with the second antibodies for 1 h at room temperature. The second antibodies used were Alexa Fluor 568-labeled anti-mouse IgG at 1 μg/ml or Alexa Fluor488-labeled anti-rabbit IgG at 2 μg/ml. These second antibodies were obtained from Molecular Probes. Finally, the immunoreactivity was examined under an Olympus FV300 confocal laser microscope. The antibodies used in the present study were as follows: either rabbit or mouse polyclonal antibodies against VGLUT2 were prepared as described previously (21). Mouse monoclonal antibodies against glucagon, insulin, and synaptophysin (SY38) were obtained from Sigma and Progen, respectively. Rat monoclonal antibodies against somatostatin were from Chemicon. Guinea pig polyclonal antisera against rat pancreatic polypeptide were from Linco Research.
The animals were anesthetized with ether and then perfused intracardially with saline, followed by 0.2% glutaraldehyde and 4% paraformaldehyde in 0.1 mol/l phosphate buffer (pH 7.4). Then each pancreas was cut into small pieces, which were washed with 0.1 mol/l cacodylate buffer (pH 7.4), stained with uranyl acetate for 2 h, dehydrated, and then embedded in London resin white for 2 days at −20°C. For double labeling immunoelectronmicroscopy, the London resin white–embedded immunogold method was used with a mixture of rabbit anti-VIAAT antiserum (100 diluted) and mouse monoclonal antibodies against glucagon in PBS containing gelatin, 2% goat serum, and 0.5% BSA at 37°C for 30 min, followed by a mixture of secondary antibodies conjugated with colloidal gold particles for 15 min at room temperature (22). After washing with 0.1 mol/l sodium cacodylate buffer (pH 7.4), the sections were postfixed in 5% glutaraldehyde in the same buffer and then observed under a Hitachi H-7100S electron microscope.
Total RNAs extracted from isolated islets and αTC6 cells (1 μg) were transcribed into cDNA in a final volume of 20 μl reaction buffer comprising 0.2 mmol/l each dNTP, 10 mmol/l dithiothreitol, 25 pmol random octamers, and 200 units Molony murine leukemia virus reverse transcriptase (Amersham). After a 1-h incubation at 42°C, the reaction was terminated by heating at 90°C for 5 min. For PCR amplification, the 10-fold diluted synthesized cDNA solution was added to the reaction buffer comprising 0.6 mmol/l total dNTP (150 μmol/l each dNTP), 25 pmol primers, and 1.5 units Ampli Taq-Gold DNA polymerase (Perkin Elmer). Thirty-five temperature cycles were conducted as follows: denaturation at 94°C for 30 s, annealing at 54°C for 30 s, and extension at 72°C for 1 min. The amplification products were analyzed by polyacrylamide gel electrophoresis. The sequences of the oligonucleotides used as primers were based on the published sequences of a rat VIAAT-specific sense primer, 5′-GAGACATTCATTATCAGCGCGG-3′ (bases 328–349) and antisense primer, 5′-CCATCTTCGTTCTCCTCGTACA-3′ (bases 595–616) (19), or a mouse VIAAT specific sense primer, 5′-CTGCCCTCTCTCGAAGGCAAC-3′ (bases 1,008–1,028) and antisense primer, 5′-GGATAGGACAGCAGCGCCTTG-3′ (25). DNA sequencing was performed by the chain-termination method.
ATP-dependent uptake of radiolabeled GABA.
The previously published (21) method was used with slight modification. Cultured cells (4.0 × 105 cells/35-mm dish) were washed with 1 ml buffer comprising 20 mmol/l MOPS-Tris (pH. 7.0), 0.3 mol/l sucrose, 0.1 mol/l KCl, and 2 mmol/l Mg acetate. Then, the cells were permeabilized with digitonin at 10 μmol/l for 10 min at 37°C. The medium was then replaced with fresh buffer containing ATP-Tris (pH. 7.0) at 2 mmol/l. In some experiments, reagents at the indicated concentrations were also included in the assay medium (see the legend to Fig. 3). Then, GABA uptake was started by the addition of radiolabeled GABA (2.5 μCi at 0.1 mmol/l; specific activity, 60 Ci/mmol; New England Nuclear) at 37°C. After a 15-min incubation, uptake was terminated by washing the cells twice with 1 ml ice-cold 20 mmol/l MOPS-Tris (pH. 7.0) containing 0.3 mol/l sucrose. The radioactivity in the culture medium and cell lysate was then counted with a liquid scintillation counter. In some experiments, an extract of cell lysate was applied on a CAPCELL PAK C18 column (2.0 × 250 mm; Siseido), and then the content of GABA was determined by high-performance liquid chromatography and amperometric detection, as previously described (12). Degradation of GABA was negligible (<3%) under the assay conditions used.
VIAAT is localized in glucagon-containing secretory granules.
As the first step of this study, we raised antibodies against VIAAT and investigated the expression of VIAAT in islets of Langerhans. The antibodies recognized the seventh transmembrane-loop region of VIAAT (73 amino acids), which is different from the regions recognized by the antibodies used by Chessler et al. (20). The antibodies recognized two kinds of polypeptides in synaptic vesicles with apparent molecular masses of 53 and 57 kDa and a single polypeptide of islet cells with an apparent molecular mass of 53 kDa (Fig. 1A). The immunoreactivity disappeared when the antibodies were pretreated with an antigenic polypeptide. These results indicated the VIAAT is present in islet cells, confirming the previous observation (20).
To investigate the cellular distribution of VIAAT, indirect immunofluorescence microscopy was performed (Fig. 1B). VIAAT immunoreactivity is confined to the mantle region of islets and is colocalized with glucagon but not with insulin or somatostatin, indicating the presence of VIAAT in α-cells but not in β- or δ-cells. It should be noted that part of the immunoreactivity of pancreatic polypeptide is colocalized with VIAAT, indicating that a population of F-cells also expresses VIAAT (Fig. 1B, arrows).
Very recently, we showed that VGLUT2 and VGLUT1, which are responsible for the storage of l-glutamate in neuronal synaptic vesicles, are expressed in α-cells, but not in β- or δ-cells, and are specifically localized in glucagon-containing secretory granules (22). Consistently, double-labeling experiments involving antibodies against VGLUT2 and VIAAT clearly showed that the two immunoreactivities are codistributed (Fig. 1C), indicating that α-cells express both VIAAT and VGLUT.
Double-labeling immunoelectronmicroscopy may reveal the subcellular localization of VIAAT in α-cells. As shown in Figs. 1D and E, immunogold particles for VIAAT (20-nm particles) were specifically associated with glucagon-containing secretory granules, which can be labeled by 10-nm gold particles. Only a background level of labeling was observed in other subcellular regions, including the nucleus and mitochondria. Only a background level of labeling was also observed in exocrine acinar cells (Fig. 1D), β-cells (Fig. 1E), and δ-cells (data not shown). Quantitatively, the labeling densities for VIAAT in glucagon-containing secretory granules and other areas including the cytoplasm and nuclei of α-cells were 4.8 ± 0.5 and 0.03 ± 0.01 immunogold particles/μm2 for four independent experiments. Labeling densities of less than 0.06 ± 0.02 immunogold particles/μm2 for four independent experiments were also observed in secretory granules and in the cytoplasm of β- and δ-cells. Neither control serum nor antiserum preabsorbed with an antigenic peptide for VIAAT gave any specific labeling (data not shown). Taken together, these results demonstrated that VIAAT is specifically associated with glucagon-containing secretory granules in islet α-cells.
Expression and localization of VIAAT in αTC6 cells.
αTC6 cells, clonal α-cells, retain the characteristics of α-cells; the-cells store l-glutamate and glucagon in their secretory granules and secrete them upon low glucose stimulation (22,26). VGLUT2 is expressed and localized in the glucagon-containing secretory granules for the vesicular storage of l-glutamate (21,22). It would thus be interesting to determine the expression and localization of VIAAT in the cells because analysis of the function of VIAAT in α-cells would be much simpler.
RT-PCR analysis indicated expression of the VIAAT gene in αTC6 cells (Fig. 2A). Western blotting indicated that antibodies to VIAAT recognized a polypeptide of 53 kDa (Fig. 2B). Like β-cells, αTC6 cells possess SLMVs, secretory vesicle-like organelles distinct from glucagon-containing secretory granules (26). Immunohistochemistry demonstrated that VIAAT immunoreactivity is confined to particle-like structures, which are codistributed with glucagon but not with synaptophysin, a marker of SLMVs (Fig. 2C). Furthermore, αTC6 cells were extensively homogenized, and their organelles were separated by sucrose density gradient centrifugation. As shown in Fig. 2D, upon centrifugation VIAAT migrated together with glucagon to the bottom of the centrifuge tube (fra. 2–4), while synaptophysin was recovered in the upper portion of the centrifuge tube (fra. 5–9). VGLUT2 was similarly distributed with VIAAT on such centrifugation, as expected. Taken together, these results demonstrate that VIAAT is associated with glucagon-containing secretory granules but not with SLMVs in αTC6 cells.
VIAAT in αTC6 cells is functional.
VIAAT is responsible for vesicular storage of GABA using an electrochemical gradient established by a vacuolar proton pump (18,19). To determine whether VIAAT in αTC6 cells is functional, we measured the uptake of radiolabeled GABA by digitonin-permeabilized cells in the presence and absence of ATP. Upon digitonin treatment, the cells become permeable to various exogenous compounds such as ATP and radiolabeled GABA (21), so ATP-dependent GABA transport can be detected. It was found that the αTC6 cells took up radiolabeled GABA in an ATP-dependent and -independent manner (Fig. 3A). On the other hand, ATP-dependent uptake was not observed in PC12 or COS7 cells (Fig. 3A). The ATP-dependent GABA uptake was sensitive to SF6847, a proton conductor, and bafilomycin A1, a vacuolar proton pump inhibitor, while ATP-independent uptake was insensitive to these treatments. Omission of Mg ions in the assay medium had a similar effect. Properties of the ATP-dependent uptake are consistent with those of VIAAT in neuronal synaptic vesicles (18) and recombinant VIAAT expressed in PC12 cells (19).
To obtain further evidence for the participation of VIAAT in the ATP-dependent GABA uptake process, release of the accumulated GABA from the αTC6 cells was measured (Fig. 3B). The radiolabeled GABA that was ATP-dependently accumulated was maintained in the cells when ATP was included in the assay medium. The radioactivity was gradually released upon omission of ATP. Addition of either SF6847 or bafilomycin A1 caused rapid loss of the radioactivity even in the presence of ATP. In contrast, neither bafilomycin A1 nor SF6847 affected the ATP-independent association of radiolabeled GABA. Taken together, these results constitute evidence for occurrence of a functional VIAAT in αTC6 cells.
ATP-dependent GABA transporter in MIN6 m9 cells is immunologically distinct from VIAAT.
Finally, ATP-dependent GABA uptake was measured in digitonin-permeabilized MIN6 m9 cells (Fig. 4A). The MIN6 m9 cells took up GABA upon the addition of ATP, which is sensitive to bafilomycin A1 and proton conductors, like αTC6 cells, confirming the presence of the ATP-dependent GABA transporter reported previously (9,10). The degree of the ATP-dependent uptake is comparable with that in αTC6 cells. As expected, the ATP-dependent uptake of radiolabeled GABA in MIN6 m9 as well as αTC6 cells was sensitive to cold GABA but not to cold l-glutamate (Fig. 4B). The ATP-dependent uptake of radiolabeled GABA in MIN6 m9 cells seemed to be more sensitive to cold GABA than that in αTC6 cells. More importantly, immunohistochemical studies indicated that MIN6 m9 cells do not exhibit any immunoreactivity for VIAAT (Fig. 4C). These results suggest that the ATP-dependent GABA transporter in MIN6 m9 cells is at least immunologically distinct from VIAAT in α-cells.
VIAAT is responsible for the vesicular storage of GABA in neurons (18,19). In the first step of the present study, we confirmed the observation of Chessler et al. (20) that VIAAT is present in the islet mantle but not β-cells. We further studied the localization of VIAAT in more detail and found that it is present in glucagon-containing secretory granules in α-cells.
The presence of VIAAT in islets was determined in immunological and molecular biological studies. Antibodies raised against the amino-terminus and COOH-terminal regions of neuronal VIAAT immunostained the mantle region of islets but did not recognize any organelles in β-cells (20). Antibodies against the conserved region of VIAAT immunostained α-cells, a subpopulation of F-cells and αTC6 cells, in several species but failed to immunostain β-cells (Fig. 1) or MIN6 m9 cells (Fig. 4). Isolated islets and αTC6 cells contain VIAAT mRNA or VIAAT polypeptide (Figs. 1 and 2). On the other hand, MIN6 m9 cells contain neither VIAAT mRNA nor VIAAT polypeptide (data not shown), although the cells exhibit functional ATP-dependent GABA uptake activity (Fig. 3) and can secrete GABA upon stimulation (10,11,22). The properties of ATP-dependent GABA uptake by αTC6 cells and MIN6 cells, e.g., inhibitor sensitivity and driving force, are similar (Fig. 3) (9,10,18,19). These results suggest that the VIAAT is similar to but distinct from the ATP-dependent GABA transporter in SLMVs of β-cells. It is highly probable that at least two distinct vesicular GABA transporter moieties are expressed in islets—neuronal VIAAT is present in α-cells, and an immunologically distinct ATP-dependent GABA transporter is present in SLMVs of β-cells. It is noteworthy that recombinant VIAAT expressed in PC12 cells (19) exhibits similar sensitivity to cold GABA as that in αTC6 cells (Fig. 4B), while the effect of cold GABA on the radiolabeled GABA uptake in MIN6 m9 cells seems to be more sensitive (Fig. 4B) (10). These results are consistent with the idea that VIAAT and a distinct vesicular GABA transporter are present in α- and β-cells, respectively.
Another significant finding in the present study is that VIAAT is localized with glucagon-containing secretory granules. The localization of glucagon and VIAAT is correlated well, indicating that essentially all the α-cells express VIAAT. To our knowledge, this is the first report of the presence of VIAAT in secretory vesicles other than synaptic vesicles. It should be stressed that glucagon-containing secrerory granules also contain VGLUTs and accumulate l-glutamate (21,22). The presence of VGLUTs in glucagon-containing secretory granules indicates that l-glutamate and glucagon are costored and coreleased from α-cells under low-glucose conditions (22). In fact, low-glucose treatment triggers the secretion of stoichiometric amounts of l-glutamate and glucagon from isolated islets as well as αTC6 cells (22,26). Since VGLUTs are present in almost all glucagon-containing secretory granules (21,22), the present results imply that VIAAT and VGLUTs are colocalized in glucagon-containing secretory granules, which must be a peculiar property of glucagon-containing secretory granules.
One can expect that glucagon-containing secretory granules also contain GABA as well as l-glutamate and that α-cells can cosecrete them under low glucose conditions. In fact, islet α-cells contain an isoform of GAD and produce GABA for vesicular storage and secretion (20). Since VIAAT also recognizes glycine as a substrate (18,19), it is also possible that α-cells accumulate glycine in secretory granules and secrete it with glucagon. An attempt to confirm cosecretion of endogenous GABA with glucagon from αTC6 cells, however, was unsuccessful because αTC6 cells do not contain a detectable amount of GABA due to a lack of GAD (S.U., R.M., S. Yatsushiro, M.O., Y.M., unpublished observations). Secretion of GABA or glycine from islet α-cells awaits further study.
The physiological significance of VIAAT in glucagon-containing secretory granules in α-cells is thus currently not clear. However, it has been shown that l-glutamate secreted by α-cells stimulates ionotropic glutamate receptors and triggers the secretion of GABA through enhanced exocytosis of GABA-containing SLMVs from β-cells (22). GABA in turn may bind to GABAA receptors on α-cells, causing inhibition of glucagon secretion (15–17). Thus, the possible cosecretion of GABA and l-glutamate from α-cells, as deduced in the present study, may participate in the regulatory mechanism for glucagon secretion: GABA from α-cells may function as an autocrine transmitter. It is also noteworthy that l-glutamate may function as an autocrine transmitter and inhibit the secretion of glucagon by way of metabotropic glutamate receptor type 8 on α-cells (27). l-glutamate- and GABA-mediated signaling may be involved in the regulation of glucagon secretion.
In conclusion, we obtained evidence that α-cells contain both VIAAT and VGLUTs in secretory granules. Our results suggest that VIAAT in α-cells is immunologically distinct from the ATP-dependent GABA transporter in SLMVs of β-cells. The present results suggest that GABA signaling in islets of Langerhans is more complex than previously supposed. Further studies, especially identification of the vesicular GABA transporter moiety in β-cells, are important for revealing all GABAergic properties of islets.
This study was supported by grants-in-aid for Scientific Research from the Ministry of Education, Science, and Sports; the Society for Research on Umami Taste; the Takeda Foundation; and the Kazato Research Foundation. M.H. was supported by a Research Fellowship from the Japan Society for the Promotion of Science for Young Scientists.
We are grateful to Dr. S. Yatsushiro for help in the initial stages of this study and Dr. K. Hamaguchi and Prof. S. Seino for providing the αTC6 cells and MIN6 m9 cells, respectively.