the insulin secretory granule is the major site of katp channels of

The Insulin Secretory Granule Is the Major Site of KATPChannels of the Endocrine PancreasXuehui Geng,

1Lehong Li,

1Simon Watkins,

1Paul D. Robbins,

2and Peter Drain

1

With ATP sites on Kir6.2 that inhibit activity and ADPsites on SUR1 that antagonize the inhibition, ATP-sensitive potassium channels (KATP channels) are de-signed as exquisite sensors of adenine nucleotide levelsthat signal changes in glucose metabolism. If pancreaticKATP channels localize to the insulin secretory granule,they would be well positioned to transduce changes inglucose metabolism into changes in granule transportand exocytosis. Tests for pancreatic KATP channels lo-calized to insulin secretory granules led to the followingobservations: fluorescent sulfonylureas that bind thepancreatic KATP channel specifically label intracellularpunctate structures in cells of the endocrine pancreas.The fluorescent glibenclamides colocalize with Ins-C-GFP, a live-cell fluorescent reporter of insulin granules.Expression of either SUR1-GFP or Kir6.2-GFP fusionproteins, but not expression of GFP alone, directs GFPfluorescence to insulin secretory granules. An SUR1antibody specifically labels insulin granules identifiedby anti-insulin. Two different Kir6.2 antibodies specifi-cally label insulin secretory granules identified by anti-insulin. Immunoelectron microscopy showed Kir6.2antibodies specifically label perimeter membrane re-gions of the secretory granule. Relatively little or nolabeling of other structures, including the plasma mem-brane, was found. Our results demonstrate that theinsulin secretory granule is the major site of KATP

channels of the endocrine pancreas. Diabetes 52:767–776, 2003

Amajor question in insulin secretion is the

cellular site of action of sulfonylureas, whichare taken daily by millions of diabetic subjectsto correct hyperglycemia. One site of sulfonyl-

urea action is at the cytoplasmic face of the plasmamembrane sulfonylurea receptor (1–5) of the ATP-sensi-tive potassium channel (KATP channel) (6). Hundreds ofKATP channels are localized to the plasma membrane ofinsulin-secreting �-cells (7). The pancreatic KATP channelcomprises four regulatory sulfonylurea receptor (SUR1)subunits and four potassium pore-forming (Kir6.2) sub-units (8–10). The plasma membrane KATP channel acts like

an on/off switch. When on, potassium ions flow outthrough the channel electrically hyperpolarizing the �-cell,putting a brake on the signal flow controlling insulinsecretion (11). When off, the KATP channel, which isinhibited by sulfonylureas or by the increased ATP/ADPratio from glucose metabolism (12,13), removes thisbrake, allowing initiation of insulin release by elevatingintracellular calcium (14–17), which triggers the exocy-totic fusion of insulin granule and �-cell membrane.

The calcium signal, however, is insufficient for regulat-ing insulin release in response to glucose under typicalconditions (11,18). While pharmacologically either open-ing or inhibiting the plasma membrane KATP channel,maneuvers that elevate not only intracellular calcium butalso glucose metabolism are necessary to confer glucosedose dependency to stimulated insulin release. By anunknown mechanism, glucose metabolism provides a sec-ond controlling parameter that, more distal in the path-way, enhances the response of the insulin secretorygranule to the calcium signal, allowing amplification ofinsulin release by elevated glucose metabolism. Study ofthe amplification pathway has implicated adenine nucleo-tides as one of the factors coupling glucose metabolic andinsulin exocytotic rates (16,18,19). Designed for ATP andADP sensing, KATP channels could confer graded glucosedose dependency in the amplification pathway in additionto their role as on/off switch at the plasma membrane,elevating intracellular calcium. Because the response ofthe insulin secretory granule is involved, the distal cellularsite for the KATP channels might be the secretory granule.In this study, we used novel live cell confocal imaging, aswell as classical immunohistochemical techniques, to testfor pancreatic KATP channels localized to insulin secretorygranules. We report on numerous lines of evidence indi-cating not only that pancreatic KATP channels localize tothe secretory granule membrane but that it is their majorsite in the �-cell of the endocrine pancreas.

RESEARCH DESIGN AND METHODS

Islet isolation, infection, and culture. Murine islets were isolated fromBALB/c or C57BL/6 mice pancreata by collagenase digestion followed byseparation on a Ficoll gradient and purified by hand under a stereo micro-scope (20). Recombinant Ad.Ins-C-GFP virus was used as described for isletinfection (21). Islets were washed using serum-free culture medium and thenthe medium was nearly completely removed by pipeting, leaving �50 �l. Theexpression levels shown were typical after 1–3 days of infection. Native andinfected islets were cultured in RPMI-1640 media containing 7.5 mmol/lglucose and 10% FCS and incubated at 37°C in 5% CO2. Live cells were usedfor the experiments reported here, as determined by the LIVE/DEAD Viability/Cytotoxicity Kit (Molecular Probes, Eugene, OR), granule exocytotic capacity,or both.Fluorescent sulfonylurea labeling. Green glibenclamide BODIPY FL andred glibenclamide BODIPY TR (Molecular Probes) were each used to achieve

From the 1Department of Cell Biology and Physiology, University of Pitts-burgh School of Medicine, Pittsburgh, Pennsylvania; and the 2Department ofMolecular Genetics and Biochemistry, University of Pittsburgh School ofMedicine, Pittsburgh, Pennsylvania.

Address correspondence and reprint requests to Peter Drain, BiomedicalScience Tower South, Room 323, 3500 Terrace St., Pittsburgh, PA 15261.E-mail: [email protected].

Received for publication 30 August 2002 and accepted in revised form 10December 2002.

gKATP, granule ATP-sensitive potassium channel; KATP channel, ATP-sensi-tive potassium channel; KD, apparent dissociation constant.

DIABETES, VOL. 52, MARCH 2003 767

a final concentration of 40 nmol/l (22). Islets were washed using Krebs-Ringerbuffer (in 7.5 mmol/l glucose) and then labeled by superfusion with thefluorescent glibenclamide within 60 min at 4°C until cellular fluorescence wasreadily detectable. The islets were then washed twice with Krebs-Ringerbuffer and imaged.Confocal fluorescence microscopy. Islets were placed into an opticalrecording chamber (Harvard Apparatus, Holliston, MA) at 37°C. Single-photonconfocal microscopy was performed using an Olympus Fluoview 300 ConfocalLaser Scanning head with an Olympus IX70 inverted microscope (Olympus,Melville, NY). Excitation of GFP, green glibenclamide BODIPY FL, or greenAlexaFluor 488 conjugated to secondary antibody was by the 488 nm Argonlaser line and emission detected using sharp cutoff 510IF long-pass andBA530RIF short-pass filters for PMT 1. Excitation of red glibenclamideBODIPY TR or red AlexaFluor 594 conjugated to secondary antibody was bythe 543-nm green HeNe laser line and emission detected using a sharp cutoffBA610IF long-pass filter for PMT 2. Coimaging was done by sequentialexcitation, and simultaneous detection of emission which showed no cross-talk was detectable under these excitation-detection conditions.DNA constructs and expression

Green fluorescent protein within C-peptide of mouse insulin. Live cellinsulin granule fluorescent labeling with the Adlox recombinant, designatedsimply “Adlox.Ins-C-GFP,” was used as described (21).Emerald GFP fused separately to COOH-termini of Mouse SUR1 and

Kir6.2. Overlap PCR was used to insert in-frame immediately after the codingsequence of mouse SUR1 or mouse Kir6.2 in pCS2 (23), with an eight-glycinecodon linker between coding segments. The simian CMV promoter of pCS2was used to express SUR1-GFP or Kir6.2-GFP fusions by incubating 4 �g DNAin 200 �l OptiMem media (Gibco/BRL, Gaithersburg, MD) in one tube and 8 �lLipofectamine 2000 (Invitrogen, Carlsbad, CA) in 200 �l OptiMem media in aseparate tube for 5 min at room temperature. The tubes were then combined,incubated for 20 min at room temperature, and added to islets.KATP channel antibodies and confocal immunofluorescence micros-

copy. We raised rabbit anti-mouse SUR1 antibodies against the 21 COOH-terminal amino acid residues of mouse SUR1 (KPEKLLSQKDSVFASFVRADK[24]), and they were affinity purified with the peptide. We raised rabbitanti-mouse Kir6.2 antibodies against either mouse Kir6.2 residues 377–390 (14COOH-terminal residues KAKPKFSISPDSL [25]) or Kir6.2 residues 329–344(DYSKFGNTIKVPTPLC), and each affinity purified by using the correspondingpeptide. The SUR1 and Kir6.2 antibodies detected bands of the predictedmolecular weights by Western analysis, and their Western and immunohisto-chemical signals were blocked by the cognate peptides used to generate theantibodies. Guinea pig anti-insulin was obtained from Dako (Carpinteria, CA).AlexaFluor 594 goat anti-rabbit IgG(H�L) and AlexaFluor 488 anti-guinea pigIgG(H�L) were from Molecular Probes. Freshly purified islets, naıve orinfected by Ad.Ins-C-GFP for 24–48 h, were fixed with 2% paraformaldehydein PBS for 20 min, washed in PBS three times, and blocked by incubation in2% BSA in PBS (pH 7.5) overnight at 4°C. The islets were then incubated withprimary antibodies (as indicated, each at 5 �g/ml) in blocking buffer overnightat 4°C. The islets were washed three times with PBS, incubated with labeledsecondary antibodies in blocking buffer for 2 h at room temperature, and thenwashed three times with PBS. Confocal microscopy to detect the immunoflu-orescence was as described above.Immunoelectron microscopy. Islets were prepared for electron microscopyas described (21). Briefly, islets were cryofixed, cut into ultrathin (70–100 nm)sections with a Reichert Ultracut U ultramicrotome/FC4S cryo-attachment,and then lifted on a 2.3 mol/l sucrose droplet on Formvar-coated copper grids.Sections were washed with PBS and PBG buffer (PBS, 0.5% BSA, and 0.15%glycine), blocked with 5% normal goat serum PBG (30 min), labeled withrabbit anti-Kir6.2 “COOH-terminus” at 1:100 in PBG for 1 h, washed in PBG,labeled with goat anti-rabbit gold-conjugated secondary antibodies (each at adilution of 1:25 for 1 h), washed in PBG and PBS, fixed in 2.5% glutaraldehyde/PBS, washed in PBS, and then washed in double-distilled H2O. Sections werepoststained in 2% neutral uranyl acetate (7 min), washed in double-distilledH2O, stained in 4% uranyl acetate (2 min), and then embedded in 1.25%methylcellulose. Labeling was observed on a JEOL JEM 1210 electronmicroscope (Tokyo) at 80 kV.

RESULTS

Glibenclamide, a potent insulin secretagogue, binds

granule-like structures in islet cells. While covisualiz-ing the arrival of secretagogue with release of insulingranules, we were struck by the observation that thesulfonylurea secretagogue glibenclamide, at low concen-trations, localized strongly to the secretory granule itself.

We found that either green fluorescent glibenclamideBODIPY FL or red fluorescent glibenclamide BODIPY TRat 40 nmol/l applied to freshly isolated mouse islet cellslocalized to punctate cytoplasmic structures with spatialand dynamic properties reminiscent of insulin secretory

FIG. 1. Fluorescent sulfonylureas bind punctate structures within cellsof the endocrine pancreas. A: Green glibenclamide BODIPY FL labelssubcellular cytoplasmic structures that resemble secretory granuleswithin a single islet cell. Z sections (left to right and down) of a singleisolated cell were taken every 0.5 �m. B: Red glibenclamide BODIPY TRlabels similar granule-like structures within a pair of islet cells. Every0.5 �m (left to right and down), z sections are shown of a pair ofisolated cells. The plasma membrane show no detectable fluorescenceafter labeling by either the green or red fluorescent sulfonylurea. Eachwas used at 40 nmol/l on freshly isolated live cells gently dissociatedfrom mouse islets in RPMI-1640 culture medium with 7.5 mmol/lglucose. PlanApo, 60X oil, NA 1.4. Bar equals 2 �m.

INSULIN SECRETORY GRANULE KATP CHANNEL

768 DIABETES, VOL. 52, MARCH 2003

granules (Fig. 1). At up to 10-fold higher concentration, thefluorescent BODIPY dyes by themselves did not localize tosubcellular structures, indicating that the glibenclamidemoiety directed the fluorescent moieties to the puncta.Furthermore, application of unlabeled 4,000 nmol/l gliben-clamide largely prevented labeling of the granules by 40nmol/l fluorescent glibenclamides. The results suggesthigh-affinity sulfonylurea receptors localized to the insulinsecretory granule, consistent with previous findings(26,27).Islet glibenclamide-labeled puncta are insulin secre-

tory vesicles. We demonstrated that the major receptorsite for the glibenclamides was localized to insulin secre-tory granules by covisualizing red glibenclamide togetherwith the green fluorescent “Ins-C-GFP” reporter of insulingranules (21) expressed in the same islets. Application of40 nmol/l red glibenclamide onto islets expressing Ins-C-GFP resulted in colocalization of the red and green fluo-rescent reporters at the same cytoplasmic punctatestructures (Fig. 2). As before, the red glibenclamide boundto subcellular puncta. Green Ins-C-GFP fluorescence ex-pressed in some of the �-cells of the same islet colocalizedwith red glibenclamide patterns. We studied this moreclosely at the single-cell level (Fig. 3). Insulin secretorygranules identified by green Ins-C-GFP overlapped exten-sively with the red fluorescent glibenclamide. Obviouscolocalization observed as yellow fluorescence was foundby optically sectioning throughout the �-cell. Furthermore,the fluorescent glibenclamide resulted in high-intensitysignals at the granule relative to other cell structures,suggesting that a substantial number of high-affinity sulfo-nylurea receptors are natively expressed at the insulinsecretory granule. Little or no labeling of the �-cell mem-brane, or any other membranes than those of the secretorygranules, was detected. The green glibenclamide had beenpreviously shown to bind SUR of bovine monocytes andmuscle cells with a subnanomolar apparent dissociationconstant (KD) �40 nmol/l (22). SUR1 is likely the onlyreceptor of the endocrine pancreas competent to appre-ciably bind 40-nmol/l fluorescently labeled glibenclamides.The results suggest that the high-affinity sulfonylureareceptor at the insulin granule is SUR1 (previously shownto be a subunit of the pancreatic KATP channel), which alsoresides on the �-cell plasma membrane (1–7,12,13).Expressed SUR1-GFP trafficks to insulin secretory

granules. The results provide strong evidence for high-affinity sulfonylurea receptors localized to insulin secre-tory granules but do not directly identify the receptor asSUR1. The high intensity of the fluorescent glibenclamidesignals at the insulin granule suggests fluorescent tagsdirectly on SUR1 should be detectable. To test this, wedetermined whether SUR1 tagged with GFP localizesgreen fluorescence to insulin granules. We used the CMVpromoter to express a COOH-terminal GFP fusion to themouse SUR1 (“SUR1-GFP”) subunit of the mouse pancre-atic KATP channel (23). We determined whether islet cellsexpressing green SUR1-GFP puncta could be colabeledwith the red glibenclamide. Figure 4 demonstrates thecolabeling at granule-like structures observed. Small andlarge size classes of punctate structures were labeled bySUR1-GFP and red glibenclamide. The small granulesexhibited more uniform diameters averaging �400 nm and

FIG. 2. The insulin granule reporter green Ins-C-GFP colocalizes withred fluorescent glibenclamide receptor sites within cells of the endo-crine pancreas. A: Cells within an islet infected with Ad.Ins-C-GFPafter 2 days of expression. A subset of cells within the islet showsgreen fluorescent insulin granules containing green fluorescent Ins-C-GFP. The subset is due in part to adenovirus typically infecting aminority of islets cells and in part to the Ins-C-GFP transcribed by theinsulin II gene promoter expressing only in insulin-secreting �-cells,which are �60% of the total islet cells. B: Glibenclamide BODIPY TR(red fluorescence). C: Merge of A and B. Note perimeter labelingreflects labeled granules lined up along cell membrane. Plasma mem-brane labeling would be at the diffraction limit, resulting in far thinner,perfectly uniform fluorescent lines that would not exhibit variablethickness. PlanApo, 60X oil, NA 1.4. Bar equals 2 �m.

X. GENG AND ASSOCIATES


are bona fide insulin secretory granules, consistent withprevious measurements (21). Colocalization of greenSUR1-GFP and red fluorescent glibenclamide at secretorygranules was evident by the small yellow puncta in mergedimages. Little or no fluorescence was detectable at the cellperimeter, consistent with a dramatically lower density ofKATP channel protein at the plasma membrane (7). Thelarge punctate structures strictly correlated with long-termculturing, which was done to achieve substantial expres-sion of the fluorescent KATP channel subunit-GFP report-ers. Large puncta were found in naive islets, as well asislets expressing transgenes, and were detectable by Ins-C-GFP or anti-insulin, indicating a close relationship toinsulin granules. The large puncta might plausibly belysosomal intermediates containing insulin-dense cores(28) or nascent insulin granules that failed to mature (29).Further study of the large puncta was obviated by theirabsence in fresh islets where the anti-insulin and anti–KATP channel subunit antibodies always identified morecharacteristic, small-sized insulin secretory granules (seebelow). Previously, we had demonstrated that the GFPused in the construction of SUR1-GFP, expressed alone,resulted in uniform green fluorescence in the cytoplasm,which accumulates in the nucleus (21). Clearly, SUR1protein sequences directed the GFP fluorescent moiety notto these compartments but to insulin secretory granules.

The SUR1-GFP results, in addition to the fluorescentglibenclamide results, suggest that native SUR1 might behighly trafficked to insulin granules.Native SUR1 localizes to insulin secretory granules

by immunofluorescence. To directly test for native SUR1localized to insulin secretory granules, we took an immu-nohistochemical approach. Antibodies against a COOH-terminal peptide of mouse SUR1 were used with anti-insulin antibodies to probe freshly purified mouse islets.Figure 5 shows extensive colocalization of the anti-SUR1and anti-insulin antibodies at secretory granules. Colocal-ization is evident by extensive yellow patterns of fluores-cence in the merged image (n � 34 islets in sevenexperiments). Note that in these islets freshly preparedfrom the animal, only the small uniformly sized punctathat best approximate insulin granules are observed. Pre-incubation with the peptide that generated the anti-SUR1antibodies, or omission of the SUR1 antibodies, resulted inno detectable fluorescence. The results suggest nativeSUR1 are highly localized to insulin secretory granules, ascompared with the �-cell membrane. Because the SUR1subunit has thus far always associated with the pore-forming subunit Kir6.2 in the KATP channel (6,8–10), theresults might be further supported by testing for a granulesite for Kir6.2.

FIG. 3. Single-cell detail of Ins-C-GFP colocalized with red fluorescent glibenclamide. A: Green Ins-C-GFP identifies insulin secretory granulesof a cell within an islet. B: Glibenclamide BODIPY TR (red fluorescence) localization to puncta. Most of the red glibenclamide overlaps atpunctate structures with the green Ins-C-GFP, but there are green granules that are not also red and there are red structures without greenIns-C-GFP. C. Merge of the red and green images; yellow intensity indicates punctate structures colocalizing red glibenclamide and the greeninsulin secretory granule reporter Ins-C-GFP. D: Optical sections, 2 �m apart, including the section exploded in the previous panels. PlanApo,60X oil, NA 1.4.



Expressed Kir6.2-GFP trafficks to secretory granules.

We tested for targeting to granules of expressed Kir6.2-GFP fusions of the mouse pancreatic KATP channel (23).Figure 6 shows that simultaneous expression of greenfluorescent Kir6.2-GFP and labeling by 40 nmol/l red glib-enclamide results in colabeling indicated by yellow punctain islet cells. Extensive colabeling was observed through-out optical sections in all cells (n � 47 islets in fiveexperiments). The average fluorescence diameter was0.435 � 0.231 �m (n � 283 granules in seven cells of fiveislets in three experiments). A minority of granulesshowed one but not the other label, consistent with highKir6.2-GFP expression and low diffusion of red gliben-clamide, or low Kir6.2-GFP expression and complete dif-fusion. No plasma membrane fluorescence was detectable.As mentioned, exogenously expressed GFP alone is nottargeted to the insulin granule (21). Localization to secre-tory granule of pore-forming Kir6.2-GFP fusions, but notGFP alone, suggests that native KATP channel subunitstraffick to insulin secretory granules.Native Kir6.2 localizes to insulin secretory granules.

We tested for native Kir6.2 localized to the granule bylabeling with either of two distinct antibodies specific to

the COOH-terminal cytoplasmic domain of mouse Kir6.2and colabeling by anti-insulin. Figure 7 shows that nativeKir6.2 is localized to the insulin secretory granule. Anantibody to the C-terminus of Kir6.2 directed red fluores-cent secondary antibody to insulin granules, identified bythe anti-insulin labeled by green fluorescent secondaryantibody (n � 24 islets in five experiments). Native Kir6.2in the �-cell membrane was not detectable under theseconditions, in which granule Kir6.2 protein labeling wasstrong. A second antibody directed against another pep-tide of the COOH-terminal cytoplasmic domain of Kir6.2showed highly similar results (data not shown). The Kir6.2peptides used to generate the antibodies prevented alllabeling by their respective antibodies, as did omission ofthe primary antibodies. Given that Kir6.2 is an integralmembrane protein, antibodies against Kir6.2 might labelspecifically at the membrane perimeter of the secretorygranule.Native Kir6.2 resides in the membrane of the secre-

tory granule. For the immunoelectron microscopy exper-iments, we prepared 70- to 100-nm thin sections of mouseislets. Figure 8 shows that gold particle–conjugated sec-ondary antibodies against anti-Kir6.2 COOH-terminus anti-

FIG. 4. Exogenously expressed SUR1-GFP colocalizes with red fluorescent glibenclamide at insulin secretory granules. A: Localization ofSUR1-GFP (green fluorescence) to punctate structures of a cell within an islet. B: Localization of acutely applied glibenclamide BODIPY TR (redfluorescence). C: Merge of green SUR1-GFP and red glibenclamide. Yellow intensity indicates punctate structures colocalizing the greenSUR1-GFP and the red glibenclamide. Images were collected sequentially with 488-nm line excitation followed by 543-nm line excitation and thenmerged offline. D: Optical sections, 2 �m apart. Note that large puncta were detected by SUR1-GFP only after long-term culturing needed toobtain substantial fluorescent signal. The large puncta were also labeled consequent to long-term culturing of naıve as well as transgenic isletsby anti-insulin antibodies, indicating a close relationship with insulin granules. The small puncta best approximate insulin granule size. PlanApo60X oil, NA 1.4.



bodies localized around dense cores, decorating theperimeter of secretory granules, while little or none(�0.02) localized to plasma membrane, mitochondria,nucleus, or other cellular structures (n � 11 islets in three

experiments). The gold particle labeling was associatedwith the membrane of the secretory granule and not thedense core. The results show that native Kir6.2 subunitsare largely localized to the membranes of insulin secretorygranules in �-cells.

DISCUSSION

The main finding of this study is the demonstration by avariety of approaches that the major site for the pancreaticKATP channel is the insulin secretory granule. That thefluorescent glibenclamides labeled native sulfonylurea re-ceptors localized to insulin secretory granules was furthersupported by the observation that red fluorescent gliben-clamide colocalized with the insulin fluorescent reporter,Ins-C-GFP. We have shown that Ins-C-GFP labels secre-tory granules that can be mobilized and exocytosed byinsulin secretagogues (21). Our results corroborate previ-ous evidence for a sulfonylurea receptor at the insulinsecretory granule. For example, early work by Hellmanet al. (30) showed evidence for dramatic uptake of gli-benclamide into islets. Another study combined 3H-glibenclamide labeling, autoradiography, and electronmicroscopy to determine an insulin granule receptor forthe sulfonylurea (26).

We extended our results by using a quite differentsecond approach. Expression in islet �-cells of the pancre-atic high-affinity sulfonylurea receptor fused to GFP(SUR1-GFP), but not the GFP expressed alone, localizedfluorescence to the same intracellular punctate structureslabeled by red fluorescent glibenclamide. Anti-SUR1 anti-body immunohistochemistry was a third approach thatindicated insulin secretory granules as the major site forSUR1. The three results each support the conclusion thatnative SUR1 is targeted to insulin secretory granules.Previous observations are consistent with our findings.The pancreatic SUR1 protein has an apparent molecularweight of 140 kDa with more highly glycosylated formsoccurring (8). Ozanne et al. (27) showed, by densitygradient centrifugation and 3H-glibenclamide labeling, that�90% of the glibenclamide mapped to intracellular struc-tures, with the majority associated with insulin secretorygranules. They showed evidence for 3H-glibenclamidebinding to intracellular receptors with nanomolar affinityand for 3H-glibenclamide cross-linking to 140- and 170-kDaproteins. Additional results implicate a sulfonylurea recep-tor protein of 65-kDa in zymogen as well as insulingranules (31), which might have also contributed to thegranule fluorescence reported here.

A secretory granule SUR1 suggested its Kir6.2 counter-part in the pancreatic KATP channel would also be targetedto secretory granules. We showed that expressed Kir6.2-GFP fusions localized to insulin granules labeled by redfluorescent glibenclamide. Second, two different anti-Kir6.2 antibodies indicated insulin secretory granules asthe major �-cell site for native Kir6.2. Third, immunoelec-tron microscopy with the anti-Kir6.2 antibodies demon-strated a membrane location of dense-core granules forKir6.2. Taken together, our results provide the first evi-dence specifically identifying SUR1 and Kir6.2 localized toinsulin secretory granules. Given that SUR1 and Kir6.2assemble into KATP channels in the endoplasmic reticulum(32–34), we propose that pancreatic KATP channels are

FIG. 5. Anti-SUR1 peptide antibodies localize at secretory granulesidentified by anti-insulin. A: Anti-insulin antibodies (green fluores-cence) identifiy secretory granules in a �-cell gently dissociated fromfreshly purified islets. B: Anti-SUR1 peptide antibodies identify similarpattern of punctate structures in the same cell. C: Merge of greeninsulin secretory granule labeling and red SUR1 labeling. PlanApo, 60Xoil, NA 1.4.



then targeted by unknown mechanisms to insulin secre-tory granules. Once inserted into the ER membrane, theKATP channel, like other proteins, is expected to maintainits membrane topology for all cellular membranes. Thus,the SUR1 COOH-terminus and Kir6.2 NH2- and COOH-termini are cytoplasmic at granule as well as plasmamembranes.Implications of granule KATP channels of the endo-

crine pancreas. Nearly all studies on glucose-stimulatedinsulin secretion signaling cite solely a plasma membrane�-cell KATP channel. Early on, Gylfe et al. (35) reviewednumerous published data on islet uptake characteristics ofradioactive sulfonylurea compounds. The five sulfonylureacompounds studied, including tolbutamide, showed rapidsaturated binding with little uptake, suggesting interac-tions with the surface of the �-cells. Glibenclamide, how-ever, did not rapidly reach uptake equilibrium, but ratherprogressively accumulated in substantial quantity withinthe islets (30). While discussing this exceptional mode ofinteraction of glibenclamide, Gylfe et al. neverthelessconcluded that the insulin-releasing action of these drugsis due to binding with the �-cell membrane, because mostof the sulfonylurea compounds were not appreciablyinternalized yet act as insulin secretagogues. The overallmodel presented, that cell-surface interaction of sulfonyl-

ureas decreases potassium permeability, leading to Ca2�

influx and insulin release, seemed reasonable at the timeand explains much of what we currently know of first-phase insulin release. Like most studies on the action ofsulfonylureas on insulin secretion, these studies fail toaddress any other cellular domain but the plasma mem-brane. Yet the experimental results are nearly alwaysinterpreted to mean that the isulinotropic action of sulfo-nylurea compounds is exclusively at the plasma mem-brane, and not internalized. Our results indicate that thisview is incomplete and cellular mechanisms of pancreaticKATP channels might be productively studied at the in-sulin secretory granule membrane, as well as the plasmamembrane.

Whenever we observed intense fluorescence localized tosecretory granules, little if any cell membrane labeling wasdetected. The relative density of the pancreatic KATPchannel in the secretory granule membrane is likely highrelative to the �-cell membrane. The observation is con-sistent with electrophysiological estimates of the plasmamembrane KATP channel, where hundreds are thought toreside (7). On one hand, the KATP channels spread acrossthe �-cell plasma membrane would be largely undetect-able by the confocal microscopy conditions used here, butare readily detectable by electrophysiology. On the other

FIG. 6. Exogenously expressed Kir6.2-GFP trafficking to secretory granules colabeled by red glibenclamide BODIPY TR. A: Localization ofKir6.2-GFP (green fluorescence) labels punctate structures in a cell within an islet. B: Localization of acutely applied glibenclamide BODIPY TR(red fluorescence) to punctate structures within the same cell. C: Merge of previous images where yellow reflects colocalization. D: Opticalsections, 2 �m apart, including the section exploded in the previous panels. The large puncta are associated with long-term culture, which wasroutinely done to obtain significant punctate fluorescence from Kir6.2-GFP. PlanApo, 60X oil, NA 1.4.



hand, the electrophysiology that detects KATP channels atthe cell membrane is incapable of testing for KATP chan-nels at granule membranes. Interestingly, glibenclamide isdistinguished from other sulfonylureas, not only by its

exceptional ability to be internalized within �-cells (30,35),but also by its superior secretory efficacy (3). The obser-vations call for sulfonylurea drug design in the future tospecifically target the intracellular secretory granule sites.

FIG. 7. Anti-Kir6.2 peptide antibodies localize at secretory granules identified by anti-insulin. A: Anti-insulin antibodies (green fluorescence)labeling in a cluster of �-cells within a freshly purified islet. B: Anti-Kir6.2 peptide antibodies (red fluorescence) applied to the same cells. C:Merge of the green insulin and red Kir6.2 images where yellow shows areas of obvious colocalization of insulin granules and Kir6.2. D: Localizationof insulin secretory granules by anti-insulin (green fluorescence) in a �-cell dissociated from a freshly purified islet. E: Localization of anti-Kir6.2antibodies (red fluorescence) in same �-cell. F. Merge of previous images where yellow shows obvious colocalization of green insulin and redKir6.2 labeling. PlanApo, 60X oil, NA 1.4.



Our results suggest there are two roles for pancreaticKATP channels that depend on cellular position. KATPchannels localized to the �-cell membrane play a well-established role in coupling glucose metabolism to plasmamembrane excitability that switches calcium influx on forrelease. The calcium influx triggers exocytosis of primedinsulin granules docked at the � plasma membrane, ob-served as first-phase release (36). KATP channels localizedto the insulin granule membrane might also play a role incoupling glucose metabolism in a dose-dependent mannerto granule transport and exocytosis, observed as second-phase release (see below). Transgenic studies show thatinactivation of the pore-forming Kir6.2 subunit rendersglucose ineffective on insulin secretion from isolated is-lets, with a very small first-phase possibly remaining (37).Transgenic inactivation of SUR1 also disrupts the re-sponse of both phases to glucose, where only a dramati-cally delayed and blunted release was found that might ormight not be mechanistically related to wild-type second-phase release (38). Even with calcium chronically elevatedby the SUR1 inactivation, second- as well as first-phaseresponse to elevated glucose was disrupted, consistentwith granule KATP (gKATP) channels mediating an amplify-ing action of glucose on insulin release.

gKATP channels are designed and positioned to conferglucose dose dependency to the second phase of insulinrelease, where calcium is unlikely rate limiting (14,16,39,40). There is evidence that the ATP/ADP ratio exertscontrol strength over signal flow, coupling glucose metab-olism and insulin exocytosis rates. In particular, increasesin ATP/ADP ratios that signal elevated blood glucoselevels accelerate steps distal to calcium influx through theplasma membrane, including insulin granule translocationand exocytosis (16,18,19,41). With ATP sites on Kir6.2 thatinhibit activity and ADP sites on SUR1 that antagonize theinhibition, gKATP channels are designed as ATP and ADPsensors of glucose metabolism. The increased ATP/ADPratio could inhibit granule membrane potassium conduc-tance of the gKATP channel, leading to changes in granulemembrane potential, ion composition, or both. Verdugoand colleagues (42,43) have reported evidence that secre-tory granules require potassium channels for appropriaterelease. The gKATP channel might not only function as achannel as it does in the plasma membrane. The gKATPchannel might play additional roles due to different ligand,

phospholipid, and protein interactions between the twomembrane domains. Thus, via increases in ATP and de-creases in ADP that combine with changes in intracellularcalcium and phosphoinositides, glucose metabolism couldalter gKATP channel activities, involving interactions withlipids, motors, and associated proteins that mediate trans-port, recruitment, and exocytosis of the granule at the�-cell membrane (44–50). Whereas the minority of KATPchannels in the plasma membrane provide on/off switchregulation for Ca2� influx that mediates first-phase releaseand is permissive for second-phase release (11,14–18), themajority of KATP channels in secretory granule membranesare situated to accelerate, in a graded way with glucosemetabolism, granule translocation, priming, and exocyto-sis rates that otherwise limit second-phase insulin release.

ACKNOWLEDGMENTS

This study was supported by the Department of CellBiology and Physiology of University of Pittsburgh Schoolof Medicine, the Juvenile Diabetes Research Foundation(grant no. 4-1999-845), and the National Science Founda-tion (grant no. MCB 9817116).

The authors thank Drs. Meir Aridor, Bob Bridges, RayFrizzell, Linton Traub, and Massimo Trucco for helpfuldiscussions, and Rita Bottino, Bala Balamurugan, andYigang Chen for generously providing highly pure mouseislets. This study is dedicated to the memory of Paul F.Drain, who taught us measure and patience.

REFERENCES

1. Schwanstecher M, Schwanstecher C, Dickel C, Chudziak F, Moshiri A,Panten U: Location of the sulphonylurea receptor at the cytoplasmic faceof the �-cell membrane. Br J Pharmacol 113:903–911, 1994

2. Aguilar-Bryan L, Nichols CG, Weschler SW, Clement IV JP, Boyd III AE,Gonzalez G, Herrera-Sosa H, Nguy K, Bryan J, Nelson DA: Cloning of the� cell high-affinity sulfonylurea receptor: a regulator of insulin secretion.Science 268:423–426, 1995

3. Ashcroft FM: Adenosine 5�-triphosphate-sensitive potassium channels.Annu Rev Neurosci 11:97–118, 1988

4. Babenko AP, Aguilar-Bryan L, Bryan J: A view of SUR/Kir6.X, KATP

channels. Annu Rev Physiol 60:667–687, 19985. Seino S: ATP-sensitive potassium channels: a model of heteromultimeric

potassium channel/receptor assemblies. Annu Rev Physiol 61:337–362,1999

6. Inagaki N, Gonoi T, Clement JPT, Namba N, Inazawa J, Gonzalez G,Aguilar-Bryan L, Seino S, Bryan J: Reconstitution of IKATP: an inward

FIG. 8. Anti-Kir6.2 antibodies localize to secretory granule perimeters. A: Immunoelectron microscopy of dense core secretory granules withanti-Kir6.2 peptide antibodies labeled with gold particle–conjugated secondary antibody. Mitochondrion in upper left and other structures did notlocalize gold particles. B: Independent experiment also showing anti-Kir6.2 labeling of dense-core secretory granules. Note frequent separationof gold particle labels from dense core, consistent with secretory granule membrane localization of Kir6.2. Cells imaged were from thin sectionsof intact islets. Scale bar is 200 nm. Magnification is �30,000.



rectifier subunit plus the sulfonylurea receptor. Science 270:1166–1170,1995

7. Cook DL, Satin LS, Ashford ML, Hales CN: ATP-sensitive K� channels inpancreatic �-cells: spare-channel hypothesis. Diabetes 37:495–498, 1988

8. Clement IV JP, Kunjilwar K, Gonzalez G, Schwanstecher M, Panten U,Aguilar-Bryan L, Bryan J: Association and stoichiometry of KATP channelsubunits. Neuron 18:827–838, 1997

9. Inagaki N, Gonoi T, Seino S: Subunit stoichiometry of the pancreaticbeta-cell ATP-sensitive K� channel. FEBS Lett 409:232–236, 1997

10. Shyng SL, Nichols CG: Octameric stoichiometry of the KATP channelcomplex. J Gen Physiol 110:655–664, 1997

11. Henquin JC: Triggering and amplifying pathways of regulation of insulinsecretion by glucose. Diabetes 49:1751–1760, 2000

12. Cook DL, Hales CN: Intracellular ATP directly blocks K� channels inpancreatic �-cells. Nature 311:271–273, 1984

13. Ashcroft FM, Harrison DE, Ashcroft SJ: Glucose induces closure of singlepotassium channels in isolated rat pancreatic �-cells. Nature 312:446–448,1984

14. Gembal M, Gilon P, Henquin JC: Evidence that glucose can control insulinrelease independently from its action on ATP-sensitive K� channels inmouse � cells. J Clin Invest 89:1288–1295, 1992

15. Sato Y, Aizawa T, Komatsu M, Okada N, Yamada T: Dual functional role ofmembrane depolarization/Ca2� influx in rat pancreatic �-cell. Diabetes

41:438–443, 199216. Gembal M, Detimary P. Gilon P, Gao ZY, Henquin JC: Mechanism by which

glucose can control insulin release independently from its action onadenosine triphosphate-sensitive K� channels in mouse � cells. J Clin

Invest 91:871–880, 199317. Aizawa T, Sato Y, Ishihara F, Taguchi N, Komatsu M, Suzuki N, Hashizume

K, Yamada T: ATP-sensitive K� channel-independent glucose action in ratpancreatic � cell. Am J Physiol 266:C622–C627, 1994

18. Sato Y, Henquin JC: The KATP channel-independent pathway of regulationof insulin secretion by glucose: in search of the underlying mechanism.Diabetes 47:1713–1721, 1998

19. Detimary P, Van den Berghe G, Henquin JC: Concentration dependenceand time course of the effects of glucose: on adenine and guaninenucleotides in mouse pancreatic islets. J Biol Chem 271:20559–20565, 1996

20. Liu M, Shapiro ME: A new method for isolation of murine islets withmarkedly improved yields. Transplant Proc 27:3208–3210, 1995

21. Watkins S, Geng X, Li L, Papworth G, Robbins PD, Drain P: Imagingsecretory vesicles by fluorescent protein insertion into propetide ratherthan mature secreted peptide. Traffic 3:461–471, 2002

22. Lohrke B, Derno M, Kruger B, Viergutz T, Matthes HD, Jentsch W:Expression of sulphonylurea receptors in bovine monocytes from animalswith a different metabolic rate. Pflugers Arch-Eur J physiol 434:712–720,1997

23. Drain P, Li L, Wang J: KATP channel inhibition by ATP requires distinctfunctional domains of the cytoplasmic C terminus of the pore-formingsubunit. Proc Natl Acad Sci U S A 95:13953–13958, 1998

24. Suzuki M, Fujikura K, Kotake K, Inagaki N, Seino S, Takata K: Immuno-localization of sulphonylurea receptor 1 in rat pancreas. Diabetologia

42:1204–1211, 199925. Suzuki M, Fujikura K, Inagaki N, Seino S, Takata K: Localization of the

ATP-sensitive K� channel subunit Kir6.2 in mouse pancreas. Diabetes

46:1440–1444, 199726. Carpentier JL, Sawano F, Ravazzola M, Malaisse WJ: Internalization of

3H-glibenclamide in pancreatic islet cells. Diabetologia 29:259–261, 198627. Ozanne SE, Guest PC, Hutton JC, Hales CN: Intracellular localization and

molecular heterogeneity of the sulphonylurea receptor in insulin-secretingcells. Diabetologia 38:277–282, 1995

28. Schnell Landstrom AH, Westman J, Håkan Borg LA: Lysosomes andpancreatic islet function: time course of insulin biosynthesis, insulinsecretion, and lysosomal transformation after rapid changes in glucoseconcentration. Diabetes 37:309–316, 1988

29. Molinete M, Dupuisw S, Brodsky FM, Halban PA: Role of clathrin in the

regulated secretory pathway of pancreatic �-cells. J Cell Sci 114:3059–3066, 2001

30. Hellman B, Sehlin J, Taljedal IB: Glibenclamide is exceptional amonghypoglycaemic sulphonylureas in accumulating progressively in beta-cell-rich pancreatic islets. Acta Endocrinol 105:385–390, 1984

31. Thevenod F, Anderie I, Schulz I: Monoclonal antibodies against MDR1P-glycoprotein inhibit chloride conductance and label a 65-kDa protein inpancreatic zymogen granule membranes. J Biol Chem 262:24410–24417,1994

32. Sharma N, Crane A, Clement IV JP, Gonzalaz G, Babenko AP, Bryan J,Aguilar-Bryan L: The C terminus of SUR1 is required for trafficking of KATP

channels. J Biol Chem 274:20628–20632, 199933. Zerangue N, Schwappach B, Jan YU, Jan LY: A new ER trafficking signal

regulates the subunit stoichiometry of plasma membrane KATP channels.Neuron 22:537–548, 1999

34. Schwappach B, Zerangue N, Jan YN, Jan LY: Molecular basis for KATP

assembly: transmembrane interactions mediate association of a K� chan-nel with an ABC transporter. Neuron 26:155–167, 2000

35. Gylfe E, Hellman B, Sehlin J, Taljedal B: Interaction of sulfonylurea withthe pancreatic �-cell. Experientia 40:1126–1134, 1984

36. Daniel S, Noda M, Straub SG, Sharp GW: Identification of the dockedgranule pool responsible for the first phase of glucose-stimulated insulinsecretion. Diabetes 48:1686–1690, 1999

37. Miki T, Nagashima K, Tashiro F, Kotake K, Yoshitomi H, Tamamoto A,Gonoi T, Iwanaga T, Miyazaki J-I, Seino S: Defective insulin secretion andenhanced insulin action in KATP channel-deficient mice. Proc Natl Acad Sci

U S A 95:10402–10406, 199838. Seghers V, Nakazaki M, DeMayo F, Aguilar-Bryan L, Bryan J: SUR1

knockout mice: a model for KATP channel-independent regulation ofinsulin secretion. J Biol Chem 275:9270–9277, 2000

39. Komatsu M, Schermerhorn T, Noda M, Straub SG, Aizawa T, Sharp GWG:Augmentation of insulin release by glucose in the absence of extracellularCa2�: new insights into stimulus-secretion coupling. Diabetes 46:1928–1938, 1997

40. Henquin JC, Ishiyama N, Nenquin M, Ravier MA, Jonas JC: Signals andpools underlying biphasic insulin secretion. Diabetes 51 (Suppl. 1):S60–S67, 2002

41. Eliasson L, Renstrom E, Ding WG, Proks P, Rorsman P: Rapid ATP-dependent priming of secretory granules precedes Ca2� induced exocyto-sis in mouse pancreatic � cells. J Physiol 503:399–412, 1997

42. Nguyen T, Chin WC, Verdugo P: Role of Ca2�/K� ion exchange inintracellular storage and release of Ca2�. Nature 395:908–912, 1998

43. Quesada I, Chin WC, Steed J, Campos-Bedolla P, Verdugo P: Mouse mastcell secretory granules can function as intracellular ionic oscillators.Biophys J 80:2133–2139, 2001

44. Donelan MJ, Morfini G, Julyan R, Sommers S, Hays L, Kajio H, Briaud I,Easom RA, Molkentin JD, Brady ST, Rhodes CJ: Ca2�-dependent dephos-phorylation of kinesin heavy chain on beta-granules in pancreatic beta-cells. Implications for regulated beta-granule transport and insulinexocytosis. J Biol Chem 277:24232–42, 2002

45. Koriyama N, Kakei M, Nakazaki M, Yaekura K, Ichinari K, Gong Q,Morimitsu S, Yada T, Tei C: PIP2 and ATP cooperatively prevent cytosolicCa2�-induced modification of ATP-sensitive K� channels in rat pancreatic�-cells. Diabetes 49:1830–1839, 2000

46. Baukrowitz T, Schulte U, Oliver D, Herlitze S, Tucker SJ, Ruppersbrerg JP,Fakler B: PIP2 and PIP as determinants for ATP inhibition of KATP

channels. Science 282:1141–1144, 199847. Baukrowitz T, Fakler B: KATP channels gated by intracellular nucleotides

and phospholipids. Eur J Biochem 267:5842–5848, 200048. Hilgemann DW, Ball R: Regulation of cardiac Na�, Ca2� exchange and

KATP potassium channels by PIP2. Science 273:956–959, 199649. Shyng SL, Nichols CG: Membrance phospholipid control of nucleotide

sensitivity of KATP channel. Science 282:1138–1141, 199850. Simonsen A, Wurmser AE, Emr SD, Stenmark H: The role of phosphoi-

nositides in membrane transport. Curr Opin Cell Biol 13:485–492, 2001



the insulin secretory granule is the major site of katp channels of

Documents