Investigation of Transport Mechanisms and Regulation of Intracellular Zn2+ in Pancreatic α-Cells*

During insulin secretion, pancreatic α-cells are exposed to Zn2+ released from insulin-containing secretory granules. Although maintenance of Zn2+ homeostasis is critical for cell survival and glucagon secretion, very little is known about Zn2+-transporting pathways and the regulation of Zn2+ in α-cells. To examine the effect of Zn2+ on glucagon secretion and possible mechanisms controlling the intracellular Zn2+ level ([Zn2+]i), we employed a glucagon-producing cell line (α-TC6) and mouse islets where non-β-cells were identified using islets expressing green fluorescent protein exclusively in β-cells. In this study, we first confirmed that Zn2+ treatment resulted in the inhibition of glucagon secretion in α-TC6 cells and mouse islets in vitro. The inhibition of secretion was not likely via activation of KATP channels by Zn2+. We then determined that Zn2+ was transported into α-cells and was able to accumulate under both low and high glucose conditions, as well as upon depolarization of cells with KCl. The nonselective Ca2+ channel blocker Gd3+ partially inhibited Zn2+ influx in α-TC cells, whereas the L-type voltage-gated Ca2+ channel inhibitor nitrendipine failed to block Zn2+ accumulation. To investigate Zn2+ transport further, we profiled α-cells for Zn2+ transporter transcripts from the two families that work in opposite directions, SLC39 (ZIP, Zrt/Irt-like protein) and SLC30 (ZnT, Zn2+ transporter). We observed that Zip1, Zip10, and Zip14 were the most abundantly expressed Zips and ZnT4, ZnT5, and ZnT8 the dominant ZnTs. Because the redox state of cells is also a major regulator of [Zn2+]i, we examined the effects of oxidizing agents on Zn2+ mobilization within α-cells. 2,2′-Dithiodipyridine (-SH group oxidant), menadione (superoxide generator), and SIN-1 (3-morpholinosydnonimine) (peroxynitrite generator) all increased [Zn2+]i in α-cells. Together these results demonstrate that Zn2+ inhibits glucagon secretion, and it is transported into α-cells in part through Ca2+ channels. Zn2+ transporters and the redox state also modulate [Zn2+]i.

Blood glucose homeostasis is maintained by appropriate secretion of insulin and glucagon from pancreatic islet ␤and ␣-cells, respectively (1,2). It is quite clear that when ambient glucose concentrations rise, glucagon secretion is inhibited and insulin secretion is stimulated. The mechanisms that control the suppression of glucagon secretion under high glucose conditions have not been completely agreed upon, but it is likely that ␣-cells sense and respond to changes in blood glucose through direct and indirect mechanisms (3)(4)(5)(6). To date the most likely candidates are glucose itself, paracrine and endocrine factors, and neuronal modulation. Some of the paracrine/endocrine factors that could facilitate the suppression of glucagon secretion include insulin and other factors released upon ␤-cell exocytosis, including ␥-aminobutyric acid and Zn 2ϩ (3,7,8).
In pancreatic ␤-cells a fraction of the intracellular Zn 2ϩ pool is stored with insulin in intracellular vesicles as a complex of zinc-insulin (9). The concentration of Zn 2ϩ in these vesicles is about 20 mM (10,11). Following exocytosis of the intracellular vesicles, it is likely that the Zn 2ϩ released into the extracellular islet space would be transported back into the host cell or into neighboring cells. The effects of ␤-cell secretory products, Zn 2ϩ and insulin, on glucagon secretion are still controversial and may be species-specific. It was shown that Zn 2ϩ had inhibitory actions on glucagon secretion from rat islets (8,12,13). Recently it was found that switching off pancreatic artery infusions of Zn 2ϩ stimulated glucagon secretion in rats (14). Contrary to these experiments, a lack of Zn 2ϩ effect on glucagon secretion was observed in mouse islets (15).
In mammalian cells the uptake of Zn 2ϩ and its outward transport are regulated by two families of membrane proteins that work in opposite directions. The influx of Zn 2ϩ is facilitated by ZIP 3 (SLC39) proteins (16 -20) and by different types of Ca 2ϩ channels: dihydropyridine-sensitive Ca 2ϩ channels in heart cells (21) and the N-methyl-D-aspartic acid and ␣-amino-3-hydroxy-5-methyl-4-isoxazole propionate/kainate-activated Ca 2ϩ -permeable channels in neurons (22,23). The efflux of Zn 2ϩ or intracellular Zn 2ϩ sequestration is promoted by CDF/ ZnT (SLC30) family of transporters, which work in opposition to the ZIP transporters (16 -20). In addition, a Na ϩ /Zn 2ϩ exchange mechanism that is different from both ZnT1 and Na ϩ /Ca 2ϩ exchanger utilizes the transmembrane electrochemical Na ϩ gradient and mediates extrusion of Zn 2ϩ in some cells (24). A Zn 2ϩ /H ϩ antiport mechanism, associated with intracellular ZnTs, could be responsible for Zn 2ϩ accumulation in intracellular organelles (25). It is likely that ZnT proteins serve as secondary active transporters, using the gradient of other ions to drive the Zn 2ϩ transport (20).
Recently a genome-wide association study identified novel risk loci for type 2 diabetes mellitus (26 -28). These loci include a nonsynonymous polymorphism in the zinc transporter SLC30A8, which is expressed in insulin-producing ␤-cells (29). Based upon these recent linkage studies, the mechanisms of Zn 2ϩ transport as well as the modulation of intracellular Zn 2ϩ in islet cells takes on an important clinical significance. Despite the importance of Zn 2ϩ in islet of Langerhans function, a limited number of articles are dedicated to Zn 2ϩ transport (30 -33). It was demonstrated that Zn 2ϩ accumulation into insulinsecreting ␤-cells occurred by the following two pathways: through membrane Zn 2ϩ transporters under low glucose conditions, and through L-type voltage-gated Ca 2ϩ channels (VGCC) under high glucose (32). To date, Zn 2ϩ transport mechanisms have not been studied in other islet cell types, including ␣-cells.
Taking into account the Zn 2ϩ -rich environment surrounding ␣-cells, it is expected that they should have an effective mechanism for maintaining Zn 2ϩ homeostasis. Dysfunction of the transporters responsible for Zn 2ϩ export can lead to dramatic changes in cell viability, because an excess of Zn 2ϩ can induce oxidative damage, mitochondrial depolarization, and opening of mitochondrial permeability transition pores (34 -36). In addition, as stated above, Zn 2ϩ released from ␤-cells may be a required component of the "off switch" for glucagon secretion. Thus the inhibition of glucagon secretion may require Zn 2ϩ transport and the modulation of the cytoplasmic Zn 2ϩ level ([Zn 2ϩ ] i ). ␣-Cells are electrically active and equipped with different types of channels, including K ATP channels and four types of VGCC, which could be involved in regulating Zn 2ϩ as in ␤-cells (3,37). In addition, ␣-cells may also modulate [Zn 2ϩ ] i through ZnTs and Zips. Finally, it is well known that free Zn 2ϩ in cells is modulated by redox state and the level of reactive oxygen species in cells (38,39), which in turn are regulated by glucose and fatty acid metabolism and ␤-oxidation. Thus the redox state of the ␣-cell may also regulate ␣-cell [Zn 2ϩ ] i .
In this study we showed an inhibitory effect of Zn 2ϩ on glucagon secretion in a glucagon-producing ␣-cell line (␣-TC). Using the Zn 2ϩ -selective dye FluoZin-3 AM (32, 40) we studied, for the first time, Zn 2ϩ transport in intact mouse pancreatic islet ␣-cells, as well as in the ␣-TC cell line, and showed the possibility of Zn 2ϩ transport through both Ca 2ϩ channels and Zn 2ϩ influx transporter(s). We demonstrated the expression of ZIP and ZnT family of Zn 2ϩ transporters in ␣-TC cells, as well as in isolated mouse islets. In addition, we investigated the regulation of [Zn 2ϩ ] i by oxidizing agents in ␣-cells. Our results suggest that cytoplasmic Zn 2ϩ homeostasis in ␣-cells is maintained by a concerted action of Zn 2ϩ transporters, Ca 2ϩ channels, and the redox state of thiol groups.
Dispersed Islet Cells from Mouse Insulin Promoter-Green Fluorescent Protein (MIP-GFP) Mice-To identify non-␤-cells we used MIP-GFP mice (CD1 background), in which GFP is specifically expressed in the islet ␤-cells. MIP-GFP mice were a gift from Dr. M. Hara, University of Chicago, Chicago, IL (41). Islets were isolated as described previously (32). To obtain dispersed cells, isolated islets were incubated for 10 min in Ca 2ϩ -free phosphate-buffered solution supplemented with 2 mM EGTA, 3 mM glucose, 100 units/ml penicillin, and 100 g/ml streptomycin. Islets then were centrifuged and incubated with dispase II, followed by addition of RPMI 1640 medium with 11.1 mM glucose, 10% bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 10 mM HEPES, pH 7.4. The suspension was centrifuged, and the pellet was resuspended in the same medium. The cells were plated on 22-mm glass coverslips coated with poly-L-lysine and maintained for 1-3 days at 37°C and 5%CO 2 , 95% air. We selected GFP-negative (non-green) cells for experiments.
Glucagon Secretion Assay-␣-TC cells were plated onto 24-well plates and cultured for 48 h in Dulbecco's modified Eagle's medium with L-glutamine supplemented with serum, streptomycin, and penicillin at 37°C. Culture medium was aspirated, and cells were preincubated for 15 min with buffer containing (mM) 130 NaCl, 5 KCl, 2 CaCl 2 , 1 MgCl 2 , 5 NaHCO 3 , 10 HEPES, pH 7.4, supplemented with 20 mM glucose. Buffer was then aspirated, and cells were incubated in 0.5 ml of buffer with 1 or 20 mM glucose with or without different concentra-tions of Zn 2ϩ for 1 h at 37°C. After incubation, 450 l of medium was collected and centrifuged at 2000 rpm, and supernatant was assayed for glucagon using glucagon radioimmunoassay kit (Linco Research) according to the manufacturer's instructions. The DNA amount in samples was determined spectrophotometrically for normalization. The procedure for mouse (FVB background) islet glucagon secretion assay was similar.
Fluorescent Measurements-During fluorescent measurements we used the incubation and perfusion buffer containing (mM) 130 NaCl, 5 KCl, 2 CaCl 2 , 1 MgCl 2 , 5 NaHCO 3 , 10 HEPES, pH 7.4, or the same medium without Ca 2ϩ . In selected experiments, to depolarize the cells, 30 or 50 mM NaCl was replaced with 30 or 50 mM KCl. Fluorescent experiments were carried out using an Olympus BX51W1 fluorescent microscope fitted with a 20ϫ/0.95 water immersion objective and cooled CCD camera equipped with a magnification changer (U-TVCAC, Olympus). For excitation, a xenon lampbased DeltaRam high speed monochromator from Photon Technology International (PTI, Lawrenceville, NJ) was used. For control of the monochromator and video camera, as well as for analysis of fluorescent imaging, ImageMaster 3 software (PTI) was used as we have described previously (32). The cells were transferred to an open chamber on the microscope stage for imaging and perfused at 1 ml/min. All time-dependent experiments were performed at 36 -37°C using TC-324B Heater Controller (Warner Instruments, Hamden, CT). In static experiments a Delta T Culture Dish Controller (Bioptechs, Butler, PA) was used for heating. One to 30 randomly chosen cells were analyzed per coverslip.
Dynamic Measurements of Zn 2ϩ , Ca 2ϩ , Plasma Membrane (⌬⌿ P ), and Mitochondrial Membrane Potential (⌬⌿ m )-For Zn 2ϩ measurements, coverslips with ␣-TC cells, or dispersed islet cells, were loaded with 2 M FluoZin-3 AM for 50 min in incubation buffer in the presence of 2 mM glucose at 37°C and 5% CO 2 , 95% air. Then coverslips with ␣-TC cells or islet cells were subsequently incubated for 10 min in the same buffer without dye. FluoZin-3 AM was excited at 480 nm, and emission was measured with a 525-nm band pass filter using 505 nm beam splitter. Changes in intracellular Ca 2ϩ concentrations were assessed using Fura-2 AM. Coverslips with cells were loaded with 2.5 M Fura-2 AM for 45 min in incubation buffer under the same conditions as with FluoZin-3 AM. Cells were excited by dual excitation at 340/ 380 nm, and emission was detected by a 510-nm band pass filter using 415-nm beam splitter. Cell ⌬⌿ P was measured by negatively charged oxonol dye DiBAC 4 (3). Coverslips with ␣-TC cells or islet cells were loaded with 100 nM dye for 20 min in incubation buffer. The perfusion solution contained the same concentration of oxonol dye. The excitation and emission wavelengths were similar to FluoZin-3 AM. Mitochondrial ⌬⌿ m was measured by positively charged dye Rh123 with excitation and emission wavelengths similar to FluoZin-3 AM. Cells were loaded with 5 M Rh123 for 45 min. To minimize possible photodamaging effects, the shutter on the monochromator was closed for 3 s between each acquisition.

Measurements of Cells Simultaneously Stained with Fluo-Zin-3 AM, MitoTracker Red CMXRos, and Hoechst 33342-
␣-TC cells were loaded with 2 M FluoZin-3 AM for 30 min in incubation buffer in the presence of 2 mM glucose at 37°C and 5% CO 2 , 95% air. Then 1 M of mitochondria-staining dye MitoTracker Red and 2 M of DNA-specific dye Hoechst 33342 were added, and cells were incubated for 20 min. Cells were washed in incubation buffer, transferred to the chamber on the microscope stage, and maintained at 36 -37°C. Excitation wavelengths for MitoTracker Red and Hoechst 33342 were 540 and 350 nm, respectively, and emission was measured using 660-nm band pass filter and 550-nm beam splitter for Mito-Tracker Red and 465-nm band pass filter and 400-nm beam splitter for Hoechst 33342.
Confocal Fluorescent Measurements-Confocal imaging was performed using a Zeiss LSM510 laser scanning microscope. Coverslips with ␣-TC cells co-loaded with FluoZin-3 AM (2 M) and MitoTracker Red (500 nM) or FluoZin-3 AM and endoplasmic reticulum (ER)-staining dye FM 4-64 (2 M) were transferred into the chamber on the stage of the microscope fitted with a 40 ϫ 0.75 water immersion objective. For Fluo-Zin-3 AM and MitoTracker Red, as well as for FluoZin-3 AM and FM 4-64 excitations, the 488-and 514-nm argon laser lines were used, and emissions were acquired using FITC and TRITC set of filters, respectively. The images were analyzed using Zeiss LSM Image software.
Measurements of NAD(P)H Fluorescence-Changes in NAD(P)H redox state of ␣-TC were observed at perfusion of cells with incubation buffer in the presence of 1 mM glucose at 37°C. The excitation wavelength for NAD(P)H autofluorescence was 360 nm, and emission was measured using 465-nm band pass filter and 400-nm beam splitter. To assess the maximal signal (maximal NAD(P)H reduction) in cells, the mitochondrial electron transport chain was inhibited by azide.
Assessment of ␣-TC Cells Apoptosis and Necrosis-For apoptosis and necrosis measurements, cells plated on glass coverslips were incubated at 37°C during 1 h with different concentrations of Zn 2ϩ in incubation buffer containing 1 mM glucose. Then coverslips were simultaneously incubated with annexin V-FITC (0.4 g/ml) and propidium iodide (PI, 1 M) for 20 min at room temperature in darkness in incubation buffer with 1 mM glucose. Coverslips were washed in the same buffer, transferred to an open chamber on the microscope stage, and maintained at 36 -37°C. The fluorescence of annexin V-FITC was excited at 480 nm and emission measured with 525-nm band pass filter using a 505-nm beam splitter. The PI fluorescence was excited at 540 nm and emission measured with a 660-nm band pass filter using a 550-nm beam splitter. Percent of apoptosis and necrosis was calculated as the ratio of annexin V or PI-positive cells to all cells in the field of view. For these experiments ϳ400 -500 cells in the field of view were analyzed per coverslip.
ZIP and ZnT mRNA Expression Analysis in ␣-TC6 Cells-Total RNA was isolated from ␣-TC cells, mouse islets (FVB background), and whole mouse brain using TRIzol Reagent (Invitrogen) according to the manufacturer's instruction. The extracted total RNA was treated with rDNase I (Ambion, Houston, TX). One g of the isolated RNA was reverse-transcribed using Moloney murine leukemia virus reverse transcriptase according to the manufacturer's instructions (Invitrogen). The resulting cDNA was used for amplification in quantitative real time PCR (qPCR). qPCR was performed as described previously (32). Primers were designed using Primer Express version 2.0 software (Applied Biosystems, Foster City, CA). Primer sequences are indicated in supplemental Table 1. 10 ng of ␣-TC6 cDNA per well was used as the template for amplification. The real time PCR protocol employed was as follows: heat activation of polymerase at 95°C for 3 min, followed by 40 cycles of 95°C for 10 s, 65°C for 15 s, and 72°C for 20 s. Readings were carried out on an ABI Prism 7900HT Sequence Detection System (Applied Biosystems) and compared against a standard curve created from mouse genomic DNA by serial dilutions. Data were normalized to mouse ␤-actin mRNA.
Statistics-All experiments were repeated three or more times in ␣-TC cells or dispersed islet cells, and typical results are presented. The raw data were processed using PSI-PLOT. Student's t test was used for determination of statistical significance. p values less than 0.05 were considered statistically significant.
Similar changes were observed in isolated mouse islets (Fig. 1B). This result is in agreement with the role of glucose as one of the physiological inhibitors of glucagon secretion. As seen in Fig. 1 in ␣-TC cells, the inhibitory effect of 10 M Zn 2ϩ on glucagon secretion is greater than that of 20 M Zn 2ϩ . Experiments using an expanded concentration range showed that a further increase in Zn 2ϩ concentration led to a further decline of the inhibition (supplemental Fig. 1A). To investigate the possible effect of Zn 2ϩ treatment on apoptosis or necrosis, we incubated ␣-TC cells for 1 h with buffer supplemented with 1 mM glucose in the presence of various Zn 2ϩ concentrations and measured cell viability (supplemental Fig. 1B). The analysis of fluorescence intensity of annexin V and PI showed that compared with control, treatment with 10 M Zn 2ϩ had no effect on the necrotic processes, whereas at 20, 50, and 100 M Zn 2ϩ necrosis was significantly increased (supplemental Fig. 1B). None of the Zn 2ϩ concentrations used had any significant effect on the degree of apoptosis (not shown). Based on these results, it is likely that the mechanism underlying the ␣-TC secretory response to high concentrations of Zn 2ϩ involves necrosis. We hypothesize that the increased glucagon secretion (compared with 10 M Zn 2ϩ ) in the presence of 20, 50, or 100 M Zn 2ϩ could be explained by an interplay between a direct inhibitory effect of Zn 2ϩ and injury/partial destruction of cells resulting in the release of intracellular glucagon.
Fluorescent Visualization of Zn 2ϩ Accumulation in ␣-TC Cells-Our previous results showed the partial co-localization of FluoZin-3 and a mitochondrial marker in ␤-cells (32). In this study ␣-TC cells stained simultaneously with FluoZin-3 AM and the mitochondrial marker dye MitoTracker Red ( Fig. 2A,  upper panel) or FluoZin-3 AM and the ER indicator FM 4-64 ( Fig. 2A, bottom panel) were imaged using confocal microscopy. Experiments showed nonhomogeneous fluorescence of FluoZin-3 in the cell interior with insignificant nuclear fluorescence and prominent location of mitochondria or ER in the perinuclear region. Partial co-localization of FluoZin-3 and MitoTracker Red or FluoZin-3 and FM 4-64 was observed. Punctate Zn 2ϩ staining and significant co-localization suggest that the distinct amount of extractable Zn 2ϩ is present not only in the cytoplasm but also in other cellular compartments like mitochondria and ER.
In addition, we examined the Zn 2ϩ distribution in ␣-TC cells after treatment with Zn 2ϩ (Fig. 2). ␣-TC cells were simultaneously stained with FluoZin-3 AM, MitoTracker Red, and DNA-specific dye Hoechst 33354 and incubated at low (1 mM, Fig. 2B) or high glucose (20 mM, Fig. 2C) conditions. As seen in Fig. 2, B and C, there was relatively weak fluorescence in the cell interior before Zn 2ϩ challenge, with nuclear fluorescence below that of the cytoplasm. After Zn 2ϩ loading for 15 min, FluoZin-3 fluorescence was elevated more or less evenly both in the cytoplasm and other cellular compartments (Fig. 2, B and C), whereas Mito-Tracker or Hoechst fluorescence did not change. It is noteworthy that Zn 2ϩ can transport into the mitochondria (42)(43)(44)(45)(46), and Zn 2ϩ -transporting systems are present in the endoplasmic reticulum (47) and the Golgi apparatus (48). Despite the relatively large pores in the nuclear membrane, which can allow Zn 2ϩ to enter into the nucleus, the increase in fluorescence observed at the nuclear region does not necessarily indicate penetration of Zn 2ϩ into nucleus. Most likely, it reflects the accumulation of Zn 2ϩ in the space surrounding the nucleus. These observations demonstrated that at both low and high glucose conditions ␣-cells were able to accumulate Zn 2ϩ .
Time-dependent Kinetics of Zn 2ϩ Transport in ␣-TC and Dispersed Islet MIP-GFP-negative Cells at Low and High Glucose Conditions and under KCl Depolarization-To examine the dynamics of Zn 2ϩ influx, a perfusion system was employed using 5-10 M Zn 2ϩ . Fig. 3A shows the kinetics of FluoZin-3 fluorescence was almost linear (at least during the 25 min), reflecting the linearity of increase in [Zn 2ϩ ] i . Treatment with the artificial Zn 2ϩ ionophore pyrithione sharply increased the rate of Zn 2ϩ entry (Fig. 3A). The subsequent application of the cell-permeable highly selective Zn 2ϩ chelator TPEN, which has much higher affinity for Zn 2ϩ than for Ca 2ϩ or Mg 2ϩ (49,50), resulted in the decline of fluorescence (decrease of[Zn 2ϩ ] i ) to initial levels.
In insulin-secreting ␤-cells, plasma membrane potential is tightly regulated by K ATP channel activity, and an increased glucose concentration results in closure of the K ATP channel and depolarization of cells. ␣-Cells are also electrically active, but unlike ␤-cells an increase in glucose concentrations leads to a more hyperpolarized state of ␣-cell plasma membrane (51)(52)(53). To check the effect of glucose on the rate of Zn 2ϩ accumulation, we perfused ␣-TC cells with low followed by high glucose concentrations (Fig. 3B). As seen in Fig. 3B, perfusion with Zn 2ϩ at 1 mM glucose resulted in an increase in FluoZin-3 fluorescence. No significant change in the rate of Zn 2ϩ accumulation was observed when cells were challenged with 20 mM glucose after low glucose perfusion. Thus in ␣-cells the change in glucose concentration from low to high did not have any significant effect on the rate of Zn 2ϩ accumulation. Switching from high to low glucose (Fig. 3C) also had little effect on the rate of Zn 2ϩ influx.
Depolarization of ␣-TC cells with KCl resulted in an increase of the Zn 2ϩ accumulation rate. Fig. 3D and pie chart i demonstrated the effect of KCl on the rate of Zn 2ϩ influx following low glucose treatment. In ␣-TC cells (75 cells in four independent experiments), the increase in the rate of Zn 2ϩ accumulation because of perfusion with KCl was observed in 57 cells (76%). In 14 cells (18.7%) KCl did not change the rate of Zn 2ϩ accumulation, and in 4 cells (5.3%) treatment led to a decrease. As seen in Fig. 3B, treatment with KCl following high glucose also resulted in increased Zn 2ϩ influx in ␣-TC cells. In 50 investigated cells (four independent experiments), we observed an increased rate of Zn 2ϩ accumulation in 32 cells (64%), no effect in 14 cells (28%), and a decrease in 4 cells (8%). Fig. 3B, pie chart ii, illustrates these data.
To determine whether the conditions used above resulted in changes of ⌬⌿ P in ␣-cell, we used the potential-sensitive anionic dye DiBAC 4 (3). The initial level of DiBAC 4 (3) fluorescence was monitored at low glucose conditions, and stimulation with high glucose led to some decrease in DiBAC 4 (3) fluorescence (hyperpolarization) (Fig. 3E). Further perfusion with 30 mM KCl resulted in a fluorescence increase, reflecting depolarization of these cells (Fig. 3E). It is of note that fast increases in intracellular Zn 2ϩ because of treatment with Zn 2ϩ /pyrithione also caused depolarization of both the plasma (Fig. 3F) and mitochondrial membrane (Fig. 3G). Perfusion with FCCP caused more complete depolarization of mitochondria (Fig. 3G).
Additionally, we performed a similar study in primary ␣-cells to look at the Zn 2ϩ transporting characteristics. It is known that pancreatic islets contain four types of cells (␣-, ␤-, ␦-, and PPcells), and the dominant are ␤-, ␣-, and ␦-cells with following distribution in rodents: ␤-cells, 75-80%; ␣-cells, 10 -18%, and ␦-cells, 5-10% (54 -58). We used dispersed islet cells isolated from MIP-GFP mice, where ␤-cells are GFP-positive. In our study the smallest GFP-negative dispersed islet cells in each field of view were chosen for fluorescent measurements, and the responses of these cells were evaluated. We took into account that two types (␣and ␦-cells) of GFP-negative cells are possible and that the mean diameter of ␣-cells (10.6 m) is smaller than ␦-cells (11.8 m) (59, 60). Fig. 4 shows representative experiments reflecting the rate of Zn 2ϩ accumulation in primary ␣-cells both at low and high glucose conditions, as well as at treatment with KCl. A depolarizing concentration of KCl changed the rate of Zn 2ϩ accumulation in most cells. In 17 investigated GFP-negative cells (five independent experiments) treated with KCl following low glucose incubation, we observed an increased rate of Zn 2ϩ accumulation in 11 cells (64.7%), no effect in 5 cells (29.4%), and decreased Zn 2ϩ influx in 1 cell (5.9%) (Fig. 4A). In 12 investigated GFP-negative cells (three independent experiments) treated with KCl following high glucose, increased rate of Zn 2ϩ accumulation was observed in 10 cells (83.3%), no effect in 1 cell (8.3%), and decreased Zn 2ϩ influx in 1 cell (8.3%) (Fig. 4B). Pie charts in Fig. 4 indicate the percentage of cells responding to KCl. Similar to our observations in ␣-TC cells, switching from high to low glucose did not change the rate of Zn 2ϩ accumulation (Fig. 4C). Because the mean diameter of ␣and ␦-cells is distinguished only by a 10% difference, one is never completely sure that all traces presented in Fig. 4 belong to ␣-cells. However, taking into account that the average abundance of ␣-cells (10 -18%) is significantly higher than that of ␦-cells (5-10%), we can assume that the majority of traces presented in Fig. 4 reflect the ␣-cell responses.
Effect of K ATP Channel and Selective and Nonselective Ca 2ϩ Channels Inhibitors on Zn 2ϩ Accumulation-Previously it was shown that a K ATP channel antagonist increased and L-VGCC inhibitors decreased Zn 2ϩ influx in ␤-cells under high glucose-and KCl-depolarized conditions (32). In contrast to ␤-cells the K ATP channel inhibitor tolbutamide had no measurable effect on the rate of Zn 2ϩ influx, despite tolbutamide-induced depolarization of ␣-cells both at low and high glucose (not shown). We also did not observe any visible inhibitory effect of nitrendipine, a dihydropyridine L-VGCC inhibitor, on the rate of Zn 2ϩ influx in ␣-TC cells either at low or high glucose (not shown). To verify whether other types of Ca 2ϩ channels could be involved in Zn 2ϩ transport across the plasma membrane, we treated cells with the nonselective Ca 2ϩ channel blocker. Gd 3ϩ partially blocked Zn 2ϩ entry under both low (Fig. 5A) and high glucose (Fig. 5B) conditions (by 25 Ϯ 3%, n ϭ 3, p Ͻ 0.05 and 24.6 Ϯ 6.4% n ϭ 4, p Ͻ 0.05, respectively). Nitrendipine did not significantly attenuate Zn 2ϩ influx in KCl-depolarized ␣-TC cells (16.5 Ϯ 7.2%, n ϭ 3, p ϭ 0.15; Fig.  5C). However, an inhibitory effect of Gd 3ϩ on Zn 2ϩ accumulation in KCl-treated cells was observed (23 Ϯ 2.5%, n ϭ 4; p Ͻ 0.05) (Fig.  5D). These findings indicated that Zn 2ϩ transport in ␣-cells under physiological conditions is not associated with L-VGCC but rather with other types of Ca 2ϩ channels. Based on experiments with several organic Ca 2ϩ channel antagonists, it is suggested that the effects of Gd 3ϩ on Zn 2ϩ channels in cortical neurons may be direct and inhibitory in nature (61). Gd 3ϩ has been shown to inhibit Zn 2ϩ transport and Zn 2ϩ -induced depolarization in kidney cells, and the inhibition does not depend on the activation of Cl Ϫ or Na ϩ channels (62). Thus, in this study, it cannot be ruled out that Gd 3ϩ directly affects the Zn 2ϩ transport pathway(s). It is of note that in our study ␣-TC cells can exhibit spontaneous oscillations of intracellular Ca 2ϩ ([Ca 2ϩ ] i ) both at low and high glucose. Treatment with Gd 3ϩ and nitrendipine caused the inhibition of spontaneous Ca 2ϩ oscillations (not shown), which indicated the presence of L-VGCC and other types of VGCCs.
Expression of ZIP and ZnT Family of Zinc Transporters in ␣-TC6 Cells, Mouse Islets, and Mouse Brain-Our observations illustrated that Gd 3ϩ partially inhibited Zn 2ϩ influx both at low and high glucose (Fig. 5, A and B), suggesting other possible Zn 2ϩ transport mechanisms are active, including plasma membrane Zn 2ϩ transporters. To begin to address this possibility, we used qPCR to evaluate the expression of zinc influx (Zip) and efflux (ZnT) transporter genes in the ␣-TC cells. We examined the expression level of the K ATP channel Kir6.2 subunit (KCNJ11), which is present in ␣-TC cells (3,37), for comparison/reference. Zip1, Zip3, Zip10, and Zip14, as well as ZnT1, ZnT4, ZnT5, ZnT6, ZnT7, and ZnT8 transcripts are expressed in ␣-TC cells (Fig. 6A). Of all the transporter genes expressed, Zip1 and Zip14 were the most abundant influx transporters, and ZnT4, ZnT5, and ZnT8 were the dominant efflux transporters (Fig. 6A). We also investigated the level of Zip and ZnT transporter genes in mouse islets (Fig. 6B). These data indicated the presence of several transporter genes where the predominant efflux transporters were ZnT5 and ZnT8 (Fig. 6B). Our observation of the expression of ZnT8 gene in mouse islets was in line with recent studies, which showed the localization of ZnT8 in ␤-cell secretory vesicle membranes (29,63); however, our results suggest that this gene is also expressed in the ␣-cell. Furthermore, our immunohistochemical data showed expression of ZnT8 protein in both insulin and glucagon positive dispersed mouse islet cells (Fig. 6D). Mouse brain was used as a positive control for expression of both Zip and ZnT transporter genes (Fig. 6C).
Effect of Zn 2ϩ on K ATP Currents in ␣-Cells-To determine whether Zn 2ϩ was inhibiting glucagon secretion via an effect on the K ATP channel, we examined whether extracellularly applied Zn 2ϩ had any effect on K ATP channel current density in mouse islet ␣-cells. Because capacitance of cells is proportional to the size of cells, ␣-cells were identified by their small size (capacitance Ͻ4 pF) and the presence of Na ϩ current when depolarized from a holding potential of Ϫ70 mV (6). K ATP current measured in these cells was very small under the recording conditions applied, and 10 M Zn 2ϩ was found to have no significant effect on the current density (Fig. 7). As was mentioned above, in these experiments we anticipate that the majority of cells recorded would be ␣-cells.
Mobilization of Intracellular Zn 2ϩ Pool and Zn 2ϩ Accumulation After Treatment with ϪSH Group Modifiers-It is known that excess Zn 2ϩ in cells is buffered by binding to cysteine, histidine, or glutamate residues of metallothioneins, one of the major intracellular Zn 2ϩ stores in most of cells (64). To examine the intracellular protein-bound Zn 2ϩ pool, we treated cells with a variety of compounds that affect Zn 2ϩ binding by causing oxidative/nitrosative stress or by chemical modification of ϪSH groups in the proteins.
The redox status of the cells is known to play a vital role in the regulation of different cellular processes. We investigated the effect of oxidative modification of ϪSH groups on the change of the intracellular Zn 2ϩ profile. A small amount (10 M) of ϪSH group oxidant DTDP led to a significant increase of the Fluo-Zin-3 fluorescence level, indicating elevation of free intracellular Zn 2ϩ in ␣-TC cells (Fig. 8A) or in GFP-negative dispersed mouse islet cells (supplemental Fig.  2A). Other oxidizing compounds menadione (generator of superoxide) (Fig. 8B) or SIN-1 (generator of peroxynitrite by simultaneous production of nitric oxide and superoxide) (Fig. 8C) also elevated FluoZin-3 fluorescence, further supporting mobilization of intracellular Zn 2ϩ by redox state. Products of lipid peroxidation can trigger multiple signaling cascades. HNE, a reactive aldehydic product of lipid peroxidation, which itself increases reactive oxygen species production, can form adducts with proteins containing histidine, cysteine, or lysine residues (65,66). As seen in Fig. 8D, treatment of ␣-cells with HNE also led to mobilization of intracellular Zn 2ϩ . The cells treated with the above oxidizing reagents were loaded with cell-impermeable dye PI, and no fluorescence increase was detected (not shown), which indicated the maintenance of cell membrane integrity during the time of experiments. NAD(P)H autofluorescence was used to monitor the changes in cellular redox state following oxidant treatment. We found that perfusion of ␣-TC cells with 20 M DTDP led to a drop in NAD(P)H fluorescence, indicating the oxidation of pyridine nucleotides (not shown). NEM, a sulfhydryl alkylating agent that covalently modifies cysteine residues in proteins, also significantly increased intracellular Zn 2ϩ mobilization in both ␣-TC cells (Fig. 8E) and GFPnegative dispersed mouse islet cells (supplemental Fig. 2B). Similarly, iodoacetamide, another alkylating agent for cysteine and histidine residues in proteins, also mobilized intracellular Zn 2ϩ (Fig. 8F).
The comparison of Fig. 8, B and C with E and F, shows the difference in the maximal values of FluoZin-3 fluorescence after treatment with various compounds. This diversity reflects the ability of compounds used to extract Zn 2ϩ from intracellular Zn 2ϩ stores, as well as the duration of treatment.
Our data suggest that these reagents extracted Zn 2ϩ from Zn 2ϩ -containing metallothioneins or other possible Zn 2ϩ stores. Because this extractable Zn 2ϩ reflects a pool of total Zn 2ϩ present in the cells, these results demonstrated a significant amount of bound Zn 2ϩ in this cell type. Thus, bound Zn 2ϩ reflects another significant pool that is highly modulated by redox state and potentially oxidative stress. To further examine this, we investigated whether the removal of oxidants or alkylating agents as well as a shift in redox state to more reduced conditions resulted in a decrease in [Zn 2ϩ ] i . As seen in Fig. 8G, the DTDP-stimulated increase in [Zn 2ϩ ] i was reversible. The fast increase in [Zn 2ϩ ] i was observed (Fig. 8G) if cells were perfused with 20 M DTDP. The removal of the ϪSH group oxidant from the perfusion medium led to restoration of the initial FluoZin-3 fluorescence during ϳ10 min (Fig. 8G). Similar effects were observed during the treatment of ␣-TC cells with or SIN-1 (not shown). Removal of the ϪSH group modifier (NEM, 15 M) resulted in a decrease in the rate of [Zn 2ϩ ] i accu-mulation but not to the restoration of the initial fluorescence level (not shown), possibly indicating the higher affinity of NEM to its target. We also used the thiol-reducing agent DTE to study the effect of ϪSH group reduction on [Zn 2ϩ ] i . Experiments showed that perfusion of ␣-TC (Fig. 8H) with DTE (100 M) following exposure to DTDP (20 M) led to the dramatic decrease (during 2-3 min) in FluoZin-3 fluorescence indicating the decreased level of free Zn 2ϩ in these cells.

DISCUSSION
In pancreatic ␤-cells Zn 2ϩ is co-secreted with insulin so it is anticipated that islet cells are exposed to high levels of Zn 2ϩ especially under high glucose conditions. ⌱t is possible that ␣-cells possess an effective Zn 2ϩ influx system that in part regulates glucagon secretion as well as other important cellular processes. ␣-Cells also likely have an effective efflux system to protect from excess Zn 2ϩ influx, because high levels of Zn 2ϩ lead to cell damage and death (34,36). Thus Zn 2ϩ likely influences ␣-cell function at many different levels requiring an effective means to maintain Zn 2ϩ homeostasis. The [Zn 2ϩ ] i in cells may be determined by a number of mechanisms as follows: Zn 2ϩ accumulation through the Ca 2ϩ channels, ZIP plasma membrane import transporter(s), Zn 2ϩ efflux catalyzed by ZnT export transporter(s), and redox state of cells. Our data suggest all four modes of regulation are possible in the ␣-cell. A defect in any one of these mechanisms may result in changes of [Zn 2ϩ ] i leading to ␣-cell dysfunction.
The effects of glucose and Zn 2ϩ on glucagon secretion remain uncertain at the present time. In intact rat and mouse islets increased glucose decreases glucagon release (15,67,68). Contrary to this, glucose stimulates glucagon release in isolated rat pancreatic ␣-cells (13). In glucagonoma cell lines InR1-G9 (69) and ␣-TC1-9 (15), an increase in the glucose concentration led to a decrease in glucagon secretion. Our results in this study showed that glucose caused the inhibition of glucagon secretion (Fig. 1) in both an ␣-cell line and isolated islets. However, Zn 2ϩ reduced glucagon secretion both at low and high glucose conditions suggesting its effects are not entirely glucosedependent (Fig. 1). Our experiments also demonstrated that Zn 2ϩ accumulated in ␣-cells both at low and high glucose conditions (Fig. 2, B and C, Fig. 3, B and C, and Fig. 4, A and B). Unlike ␤-cells (32), treatment with tolbutamide or nitrendipine did not have any visible effect on the rate of Zn 2ϩ accumulation under the conditions studied. The nonselective Ca 2ϩ channel blocker Gd 3ϩ moderately attenuated Zn 2ϩ accumulation both at low and high glucose conditions (Fig. 5, A and B), as well as during depolarization of cells with KCl (Fig. 5D). These findings indicate that Zn 2ϩ is transported into ␣-cells in part through dihydropyridine-insensitive Ca 2ϩ channels. Four types of VGCC (L-, T-, N-, and R-type) and tetrodotoxin-sensitive voltage-gated Na ϩ -channels have been observed in ␣-cells and thus could be involved in Zn 2ϩ transport (3,37). Further studies are required to examine if other ion channels are involved.
As stated, the mechanism of the inhibition of glucagon release by Zn 2ϩ is unclear. It was suggested that Zn 2ϩ inhibits glucagon secretion in rat ␣-cells because of activation of the K ATP channel resulting in hyperpolarization of cells (12). However, the activation of the K ATP channel by Zn 2ϩ was not found in mouse ␣-cells (70). In these studies both groups used different protocols for measuring K ATP channel modulation. Our results are in agreement with previous findings in mouse; we see no effect of Zn 2ϩ on ␣-cell K ATP current, suggesting that under the conditions studied this mechanism is not involved in Zn 2ϩ modulation of glucagon secretion. The inhibition of secretion therefore appears to require transport as part of the mechanism (Fig. 7).
In addition to ion channels, another possible pathway of Zn 2ϩ transport is through plasma membrane Zn 2ϩ transporter(s). Although there are some data on Zn 2ϩ influx transporters in the pancreas, information about Zip transporters in ␣and ␤-cells is limited. Zip1 transcripts were previously found in human (71), but not mouse, pancreas (72). The expression of Zip5 mRNA in the whole mouse pancreas was also observed (73,74). Our previous work using quantitative real time PCR demonstrates a number of Zip transcripts in mouse insulinoma ␤-cells (MIN6), mouse islets, and mouse brain (32). In this work we have identified in glucagon-producing ␣-TC cells the expression of a number of genes of the ZIP and ZnT families (Fig. 6A). The most abundantly expressed Zip transporter genes in ␣-TC cells were Zip1, Zip10, and Zip14 (Fig. 6A). The expression profile of Zip transporters in mouse islets (Fig. 6B) and mouse brain (Fig. 6C) mainly confirms our previous results (32).
Unlike Zip transporters, the Zn 2ϩ efflux transporters (ZnT) in pancreatic ␤-cells are better characterized. Among the ZnT family of transporters, the ZnT8 was recently detected in insulin-secreting INS-1 cells and human pancreatic islets and was co-localized with insulin in these cells, suggesting they are in vesicular membranes (29,63). Our results identified the expression of ZnT1, ZnT4, ZnT5, ZnT6, ZnT7, and ZnT8 transporter genes in ␣-TC cells (Fig. 6A) and mouse islets (Fig. 6B), with ZnT4, ZnT5, and ZnT8 showing the highest level of expression in ␣-TC cells and ZnT5 and ZnT8 in islets. The high level of ZnT5 in ␣-TC cells (Fig. 6A) and islets (Fig. 6B) is in agreement with a previous report, in which ZnT5 was found to be abundantly expressed in pancreas (75). Our data demonstrated that ZnT8 was not exclusively a ␤-cell Zn 2ϩ transporter. It is noteworthy that others have shown that ZnT8 was also expressed in adipose tissue with comparable abundance as ZnT1, ZnT5, and ZnT7 (76). The role and localization of ZnT8 in ␣-TC cells are not clear. As stated above, in ␤-cells ZnT8 is thought to be localized to the insulin secretory granule, and overexpression studies in ␤-cells suggests it functions to modulate [Zn 2ϩ ] i and regulate (enhance) insulin secretion (29,63). It is important to FIGURE 7. Effect of Zn 2؉ on K ATP current in mouse islet ␣-cells. Extracellularly applied Zn 2ϩ (10 M) had no effect on K ATP current density. Results represent the average Ϯ S.E. of 9 cells.
binding and a decrease in [Zn 2ϩ ] i , could decrease the damage caused by excess of Zn 2ϩ and may have a protective role. Our experiments with the thiol-reducing agent DTE (Fig. 8H), which led to a more reduced intracellular environment and decreased [Zn 2ϩ ] i , support such a conclusion.
In summary the data presented demonstrate that Zn 2ϩ accumulates in glucagon-producing ␣-cells under low and high glucose conditions through both dihydropyridine-insensitive Ca 2ϩ channels and other Zn 2ϩ -transporting mechanisms.
[Zn 2ϩ ] i in ␣-cells is therefore likely regulated by Ca 2ϩ channels and Zn 2ϩ transporters as well as through the redox sensitivity of thiol groups.