浙江大学 文本 Investigation of the subunit composition and the pharmacology of the mitochondrial ATP-dependent K+ channel in the brain_生理学

BRAIN SCIENCE ODIRECT RESEARCH ELSEVIER Brain Research 994 (2003)27-36 Research report Investigation of the subunit composition and the pharmacology of the mitochondrial ATP-dependent K*channel in the brain Zsombor Lacza James A.Snipes",Bela Kis",Csaba Szab Gary Grover David W.Busiia M-SonihPha Acoepted Sepeember 2003 app the cular eight SUR2 were e 2 splice variant or a sition and cell surface mac eymATF-depend 1.Introduction omponent the ponse in most t not al mote can elim ms[5.23.281 uin an Th eficial effects of is d by numerous physiolog- NC27157 USA channels

Research report Investigation of the subunit composition and the pharmacology of the mitochondrial ATP-dependent K+ channel in the brain Zsombor Laczaa,b,*, James A. Snipes a , Be´la Kisa , Csaba Szabo´ b,c, Gary Groverd , David W. Busijaa a Department of Physiology/Pharmacology, Wake Forest University School of Medicine, Winston-Salem, NC, USA b Institute of Human Physiology and Clinical Experimental Research, Semmelweis University, Budapest, Hungary c Inotek Pharmaceuticals Corporation, Beverly, MA, USA d Bristol-Myers-Squibb Pharmaceutical Research Institute, Princeton, NJ, USA Accepted 8 September 2003 Abstract Selective activation of mitoKATP channels can protect the brain or cultured neurons against a variety of anoxic or metabolic challenges. However, little is known about the subunit composition or functional regulation of the channel itself. In the present study, we sought to characterize the mitoKATP channel in the mouse brain using overlapping approaches. First, we determined that mitochondria contain the pore￾forming Kir6.1 and Kir6.2 subunits with Western blotting, immunogold electron microscopy and the identification of mitochondrial transport sequences. In contrast, we found no evidence for the presence of either known sulfonylurea receptors (SUR1 or SUR2) in the mitochondria. However, the ATP-dependent K (KATP) channel inhibitor glibenclamide specifically binds to mitochondria in both neurons and astrocytes, and small molecular weight SUR2-like proteins were concentrated in mitochondria. In addition to mice, similar results were found in rats and pigs. Second, live respiring mitochondria were stained with the membrane potential sensitive dye MitoFluorRed and visualized by confocal microscopy. We investigated the effects of pharmacological closing and opening of the channel with glibenclamide and the specific mitoKATP openers diazoxide and BMS-191095. Closing of the channel inhibited the energization of the mitochondria, which was reversed by the application of the mitoKATP openers. We also found that blocking mitochondrial peroxynitrite formation with FP15 has a similar effect to blocking the mitoKATP channels. We conclude that brain mitochondria contain functional KATP channels. The pore-forming subunit of the channel can be either Kir6.1 or Kir6.2, and the SUR subunit may be a SUR2 splice variant or a similar protein. D 2003 Elsevier B.V. All rights reserved. Theme: B Topic: Membrane composition and cell surface macromolecules Keywords: ATP-dependent potassium channel; Ischemic preconditioning; Mitochondrial nitric oxide synthase; mtNOS; Sulfonylurea receptor 1. Introduction The paradoxical phenomenon of ischemic precondition￾ing, in which a brief episode of ischemia protects the tissues against a subsequent lethal injury, was first de￾scribed in the heart [22] and later extended to brain and other organs [5,23,28]. The beneficial effects of ischemic preconditioning can be simulated by numerous physiolog￾ical stimuli and pharmacological agents, and the activation of ATP-dependent K+ (KATP) channels appears to be a prominent component of the response in most, but not all, models of preconditioning. Thus, agonists of KATP can promote preconditioning, while KATP antagonists can elim￾inate or reduce the degree of preconditioning when given simultaneously with agonists or during brief periods of ischemia or other physiological stimuli [6,7,17]. The typ￾ical KATP channel, composed of four Kir and four SUR subunits, is located principally in the plasma membrane (sKATP) and several variants of the Kir and SUR subunits are known to exist [1]. However, very little is known about the differences between mitoKATP and the other KATP channels. 0006-8993/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2003.09.046 * Corresponding author. Department of Physiology/Pharmacology, Health Sciences, Wake Forest University School of Medicine, Medical Center Blvd., Winston-Salem, NC 27157, USA. Tel.: +1-336-716-4367; fax: +1-336-716-0237. E-mail address: zlacza@mac.com (Z. Lacza). www.elsevier.com/locate/brainres Brain Research 994 (2003) 27 – 36

2Lam时al/Aran Revareb9H03动-j6 Several recent studies bave provided strong evidence was centrifuged for 5 min at 19000 rpm and the layer that selective activation of mitoK channels can protect containing the purified mitochondria was collected.The the brain or cultured neurons against a varicty ol'anoxic preparation was washed with high ionie strength buflier or metabolic challenges.However.litle is known about contsining 125 mM K'.All procedures were performed the subumit composition or fianetional regulation of the on ice Brain mitochondria were also collected fiom brain mitoKArr channel.In one paper,Bajgar et al.3] sodium thiopenthal(I0圆mgke i.p）ave-anestheti2ed identifiod a putarive 55-kDa Kir subunit and a putative adult Wistar rats and newborn piglets using similar 63-kDa SuIR subunit from rt hrain,hut the precise nature appaches. of the sbunits was not deemined hecause of the lack of The purity of the mitochondrial preparation was tested hy specific antibodies against Kir and SUR subunits.Addi. two independent methods.First,electron miemscopic ohser- tionally,Zhou et al.[33]found the Kir6.2 sabunit local- vations showed very little contamination from broken mi- ized to mitochondria in rat brain slices,although other tochondria or lysosomes.Second.Western blottingg showed mitoKarr subunits were not investigated.Lastly.we have that the mitochondra-marker cytochrome c oxidase was at recently sbown that Kir6.1 and SUR2 subunits are kast eight times enriched in the milochondria preparations. enriched in ra brain milochondria compared to whole In contrast.the endoplasmic reticulum marker calreticulin brain using a Western blot approach [29].However.we was markodly reduced to insignificar kvels in the purified did not examine the Kir6 2 ard SURI subunits,did not mitocbondria (data not shown). evaluste the intramitochondrial distribution of subunits, and did pot examine the predominant mitoKam subumits 2.2 Ceil culture in other laboratory species.In temms of mitochandrial function,we are unawure of any previous studies which Primary rut cortical neurons were cultured from E18 have directly examined effects of pharmacological stimuli Wistar rat fetuses bused on the method of Iampson et al. on K furction in live brain mitochoodria. 10].Brains from the fetuses were collected and placed in In the present study.we sought to characterize the ice-cold Dalbecco's Modified Eagle's Medium (DMEM. mitoK charnel in the mouse brain using four over Giboo,Grand Island.NY.USA)The cortical pieces were lapping approaches.First.we determined whether the washed twice in DMEM supplemented with penicillin Kir6.1 or Kir6.2 or SURI or SUR2 subunits were (Sigm2 100 U/ml)and streptomycin (Sigma,100 ug/ml) concentrated in mitochondria using Westem blotting. then were incubated with dispase I (2 Uml,Roche,Mann- immunogold electron micmscopy,and confocal microso- heim,Germany)for 35 min at 37 'C.After stopping the py.Second,we investigaed the effects of pharmcolog- enzymatic reaction by washing the cells twice with DMEM, ical closing and opening of the channel in isolated live cells were dissociated by two series of gentle trituration and mitochandria with the SUR antagonist gliberclamide and plated oato poly-D-lysine coated coverslips for confocal mitoK ists diazoxide and BMS-191095.Third,we microscopic analysis.After cell attachment,the plating examined hether mitochondrial produced peroxynitrite is medium was replaced by Neurobasal medium (Gibcol endogenous opener of brain mitoK Finally,we Supplemented with B27 (Gibeo,2%)r-elutamine (Saema. compared expression of mitok subunits among mice, 0.5 mM).B-meremptocthanol (Giheo,55 pM)and potssium rars and pigs chlride (Sigma,25 mM).Culnures were grown at 37 C in humidified atmosphere containing 5%Co.in air,and medium was chaneed on every third day.Cultures consisted 2.Materiak and methods of more than 98%of neurons verified by positive immu- nostaining for microtubule-associated protein-2 (Becton- 21.Mitochoudrto isolanow Dickinso)ad negntive immunostaining for glial fihrillary acidic protein (Chemicon.Temecula,CA.USA).Experi- lsolated milochordna preparations were collcted by ments were performed in 7-9-day-old cultures,a time discontinuous percoll gradient purilication as described period during which neurons express NMDA,AMPA and previously [141.All procedures were approved by the kainate recepiors and are vuinerable to hypoxia and glucose Animal Care and Use Committee of Wike Forest Uni- dkprivation [21J. versity.Wild type mice from the e57hl6 strain (Jackson, Rat eerebral astrocyte cultures were prepared from Har Harbor,ME]were overanesthetinod with halothgne nconatal Wistar rats [15].Meninges were removed and and deenpitated.The briet exposure 1o halothane was cortieal pieees were mechanically dissocated in astrocyte followed hy the preparation peoeodure in a helothane-free culture modium (DMEM supplemented with 10%fetal evironment ensuring the eliminatian of the anesthetic in bovine serum and antibiotics).Dissociated cells were the final preparation.The brain was removed and homog- seeded into cell culture flasks.In order to obtain type I enized in sucrose buffer containing 12%percoll (Amer- astroglia,confluent cultures were shaken at 37 "C over- sham Pharmacia.Uppsala,Sweden)and layered on top of night.The parity of astrocytes was chocked by immunos- a discontinuous percoll gradient (24/40%).The grodient taining for GFAP.and the cclls were used at passage 1

Several recent studies have provided strong evidence that selective activation of mitoKATP channels can protect the brain or cultured neurons against a variety of anoxic or metabolic challenges. However, little is known about the subunit composition or functional regulation of the brain mitoKATP channel. In one paper, Bajgar et al. [3] identified a putative 55-kDa Kir subunit and a putative 63-kDa SUR subunit from rat brain, but the precise nature of the subunits was not determined because of the lack of specific antibodies against Kir and SUR subunits. Addi￾tionally, Zhou et al. [33] found the Kir6.2 subunit local￾ized to mitochondria in rat brain slices, although other mitoKATP subunits were not investigated. Lastly, we have recently shown that Kir6.1 and SUR2 subunits are enriched in rat brain mitochondria compared to whole brain using a Western blot approach [29]. However, we did not examine the Kir6.2 and SUR1 subunits, did not evaluate the intramitochondrial distribution of subunits, and did not examine the predominant mitoKATP subunits in other laboratory species. In terms of mitochondrial function, we are unaware of any previous studies which have directly examined effects of pharmacological stimuli on KATP function in live brain mitochondria. In the present study, we sought to characterize the mitoKATP channel in the mouse brain using four over￾lapping approaches. First, we determined whether the Kir6.1 or Kir6.2 or SUR1 or SUR2 subunits were concentrated in mitochondria using Western blotting, immunogold electron microscopy, and confocal microsco￾py. Second, we investigated the effects of pharmacolog￾ical closing and opening of the channel in isolated live mitochondria with the SUR antagonist glibenclamide and mitoKATP agonists diazoxide and BMS-191095. Third, we examined whether mitochondrial-produced peroxynitrite is an endogenous opener of brain mitoKATP. Finally, we compared expression of mitoKATP subunits among mice, rats and pigs. 2. Materials and methods 2.1. Mitochondria isolation Isolated mitochondria preparations were collected by discontinuous percoll gradient purification as described previously [14]. All procedures were approved by the Animal Care and Use Committee of Wake Forest Uni￾versity. Wild type mice from the c57bl/6 strain (Jackson, Bar Harbor, ME) were overanesthetized with halothane and decapitated. The brief exposure to halothane was followed by the preparation procedure in a halothane-free environment ensuring the elimination of the anesthetic in the final preparation. The brain was removed and homog￾enized in sucrose buffer containing 12% percoll (Amer￾sham Pharmacia, Uppsala, Sweden) and layered on top of a discontinuous percoll gradient (24/40%). The gradient was centrifuged for 5 min at 19 000 rpm and the layer containing the purified mitochondria was collected. The preparation was washed with high ionic strength buffer containing 125 mM K+ . All procedures were performed on ice. Brain mitochondria were also collected from sodium thiopenthal (100 mg/kg i.p.) over-anesthetized adult Wistar rats and newborn piglets using similar approaches. The purity of the mitochondrial preparation was tested by two independent methods. First, electron microscopic obser￾vations showed very little contamination from broken mi￾tochondria or lysosomes. Second, Western blotting showed that the mitochondria-marker cytochrome c oxidase was at least eight times enriched in the mitochondria preparations. In contrast, the endoplasmic reticulum marker calreticulin was markedly reduced to insignificant levels in the purified mitochondria (data not shown). 2.2. Cell culture Primary rat cortical neurons were cultured from E18 Wistar rat fetuses based on the method of Hampson et al. [10]. Brains from the fetuses were collected and placed in ice-cold Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco, Grand Island, NY, USA). The cortical pieces were washed twice in DMEM supplemented with penicillin (Sigma, 100 U/ml) and streptomycin (Sigma, 100 Ag/ml) then were incubated with dispase I (2 U/ml, Roche, Mann￾heim, Germany) for 35 min at 37 jC. After stopping the enzymatic reaction by washing the cells twice with DMEM, cells were dissociated by two series of gentle trituration and plated onto poly-D-lysine coated coverslips for confocal microscopic analysis. After cell attachment, the plating medium was replaced by Neurobasal medium (Gibco) supplemented with B27 (Gibco, 2%), L-glutamine (Sigma, 0.5 mM), h-mercaptoethanol (Gibco, 55 AM) and potassium chloride (Sigma, 25 mM). Cultures were grown at 37 jC in humidified atmosphere containing 5% CO2 in air, and medium was changed on every third day. Cultures consisted of more than 98% of neurons verified by positive immu￾nostaining for microtubule-associated protein-2 (Becton￾Dickinson), and negative immunostaining for glial fibrillary acidic protein (Chemicon, Temecula, CA, USA). Experi￾ments were performed in 7 – 9-day-old cultures, a time period during which neurons express NMDA, AMPA and kainate receptors and are vulnerable to hypoxia and glucose deprivation [21]. Rat cerebral astrocyte cultures were prepared from neonatal Wistar rats [15]. Meninges were removed and cortical pieces were mechanically dissociated in astrocyte culture medium (DMEM supplemented with 10% fetal bovine serum and antibiotics). Dissociated cells were seeded into cell culture flasks. In order to obtain type I astroglia, confluent cultures were shaken at 37 jC over￾night. The purity of astrocytes was checked by immunos￾taining for GFAP, and the cells were used at passage 1. 28 Z. Lacza et al. / Brain Research 994 (2003) 27–36

Z.Larma t al.hraw Reumareh 94 (2003)27i6 2.3.Florescent m成royy pared to the effect of the K*ionophore nigericin (10 umol)dissolved in 1%ethanol.MitoFluorRed fhoores Freshly isolated mitochondria were dispersed in buffer cence was comparod betwoen preparations with or without containing 125 mmol KCI,2 mmoll K2HPO4,5 mmol/ the peroxynitrile decomposaion catalyst FP15 (100 umol/ I MgCl 10 mmol/l HEPES,10 HmoMl EGTA at pl 7.0 1):mitochondrial NO production was simultancously amd plated on poly-D lysine coated coverslips.Mitocbon- monitored by DAF-FM (7 umoi/l.Molecular Probes dria were energized by the addition of malate (5 mmol/l) flooreseenee.All dngs were obcamad from Sigmn (St. imd glutamate (5 mmoll)and were visualized using a Lonis.MO)unles stated otherwise. Zeiss scamning confocal microscope with differential in Culturod cells were visualized using a protocnl similar to terference contrast optics (DIC)and a fluorescein'rhoda- that used with mitochoodria The nourons or nstrncytes were mine fiher set at room temperature.The flucropbores used incubated with 1 uM BODIPY-gliberxclamide and 500 nM were MitoFlorRed and BODIPY-glibenclamide (Molecu- MitoTrackerRed (Molecular Probes)for 20 min at 37 "C. lar Probes,Eugene,OR)Five minutes nfter enerpizntion Following incubation,the cells were washed and immedi- the mitochondria were challeneed with 100 umol/I glihen- ately visualized under the confecal microscope. clamide and after 3 min with additional application of the mitoKAr openers 100 pmol1 diazoxide or 10 umol/ 2.4.Hester bloting I BMS191095 (Bristol-Myers-Squibh,Princeton,N)[9] The drugs were dissolved in 1%dimethyl-sulfoxide Equal amounts of protein were sparatod on a%1o (DMSO).The effect of the mitoKAm openers was com- 20%gradient mini gel,and transferred to PVDF mem DIC glybendlamide mitochondra composite isolsted mitochondna astrocyte m mitochondrac ponel A in trarsntined Ight with IXC opoics.ponel B wih BODIPY-glberclartide i preen.poel C wih Mro TrackerRed and parel D is the vottgoaie inage of B aa C.Yiluw polue inliss cu-localialion uf the gllnchenie u he mihomria igrb.The miull ruw (pandh E-H)owas cultured Ecuo色4 ih smnler labdling.The glil1ksn国s loculed to自2 cell merabron2d由xaad色

2.3. Fluorescent microscopy Freshly isolated mitochondria were dispersed in buffer containing 125 mmol/l KCl, 2 mmol/l K2HPO4, 5 mmol/ l MgCl2, 10 mmol/l HEPES, 10 Amol/l EGTA at pH 7.0 and plated on poly-D-lysine coated coverslips. Mitochon￾dria were energized by the addition of malate (5 mmol/l) and glutamate (5 mmol/l) and were visualized using a Zeiss scanning confocal microscope with differential in￾terference contrast optics (DIC) and a fluorescein/rhoda￾mine filter set at room temperature. The fluorophores used were MitoFluorRed and BODIPY-glibenclamide (Molecu￾lar Probes, Eugene, OR). Five minutes after energization the mitochondria were challenged with 100 Amol/l gliben￾clamide and after 3 min with additional application of the mitoKATP openers 100 Amol/l diazoxide or 10 Amol/ l BMS191095 (Bristol-Myers-Squibb, Princeton, NJ) [9]. The drugs were dissolved in 1% dimethyl-sulfoxide (DMSO). The effect of the mitoKATP openers was com￾pared to the effect of the K+ ionophore nigericin (10 Amol/l) dissolved in 1% ethanol. MitoFluorRed fluores￾cence was compared between preparations with or without the peroxynitrite decomposition catalyst FP15 (100 Amol/ l); mitochondrial NO production was simultaneously monitored by DAF-FM (7 Amol/l; Molecular Probes) fluorescence. All drugs were obtained from Sigma (St. Louis, MO) unless stated otherwise. Cultured cells were visualized using a protocol similar to that used with mitochondria. The neurons or astrocytes were incubated with 1 AM BODIPY-glibenclamide and 500 nM MitoTrackerRed (Molecular Probes) for 20 min at 37 jC. Following incubation, the cells were washed and immedi￾ately visualized under the confocal microscope. 2.4. Western blotting Equal amounts of protein were separated on a 4% to 20% gradient mini gel, and transferred to PVDF mem￾Fig. 1. Isolated mitochondria, astrocytes and neurons visualized by a fluorescent tagged sulfonylurea, BODIPY-glibenclamide. The top row shows isolated mitochondria: panel A in transmitted light with DIC optics, panel B with BODIPY-glibenclamide in green, panel C with MitoTrackerRed, and panel D is the composite image of B and C. Yellow color indicates co-localization of the glibenclamide and the mitochondria signals. The middle row (panels E – H) shows a cultured astrocyte with similar labeling, the majority of the green glibenclamide signal is co-localized with mitochondria. The bottom row (panels I – L) shows cultured neurons with similar labeling. The glibenclamide signal is located to the cell membrane and the mitochondria. Z. Lacza et al. / Brain Research 994 (2003) 27–36 29

2 Lacra al/m是aHN0为-6 hrane.After hlocking with 3%bovine senum albumin, tential specific dye MitoFluorRed was readily taken up hy primary antihodies against the KAreceptor suhunits the isolated organelles,indicating n viable,respiring Kir6.1.Kir6.2.SUR1 or SUR2 (Santa Cruz Bicloch. preparation.The green Duorescent sullonylurea BOD- Santa Cruz.CA)were applicd in a dilution of 1:500 IPY-glibenclamide (1 uM)specifically bound to the followed by horsecradish peroxidase conjugaled sccondry mitochondria.indicaling the presence ol'sulfooylurea iby.Chemilumi3cee锋as used lo￥Bsualize the receptors (Fig.1). bands Specificity of the method was testod by omniting Cytoplasm.endoplasmnic reticula or mclei were larpely the prinary sntibody fiom the proeedure or preadsorp- devoid of staining in cultured astrocyies and neues, tion with the respective blocking pepeides,which resulted showing the relative specificity of the method (Fig 1). in the disnppearanee of the specific hands (dats not This observation indicated the presenee of sulfonylrea shownk binding sites in neuronal and astroglial mitochondria. Quantitative measurements of mitochondrial enrich- However,sulfonylurea binding may be indeperdent of ment were performed in pooled preparations loaded in KArr channels.so we tested the prescnce of the specific increasing procein concentrations and we used an expo- channel-forming subunits via Westem blotting. sure of the membrane.which showed the best resolution on the ligh址t and d虹k bands also.compared the results to a parallel blot using an anti-cytochrome c amibody asa control for a known mitochondrial provein (eighl-fold SUR2 mitochandrial enrichment). 2周 2.5.Elecman microscopy 128 Freshly isolated mitochondria samples were fixed for 3 h in 4%paraformaldehyde/005%glutaraldehyde.The sam ples were embedded in LR-white.were sectioned,and were collected an ucoated nickel grids.Bovine serum albumin (2%)and 1:30 dilution of nomal goat senim (homologous senum to secoedary nntibody)was applied to hlock unspe- cific hinding The sections were incubated ovemight with primary antihodies agninst Karchannel suhunits (see above)in a dilation of 1:100 at 4 C.Immunogold labeling was achieved hy a 4-h incubation with 6-nm gold-labeled sepondary antibody.The specificity of the method was 19 tested by the omission of the primary antibody from the prooddare,which resultod in the climination of gold par- ticles on the sections.Post-soctionine,en-bloc stauning and bnan osmium postfisation were omitted to srvoid imerlerence with the lnbeling 26 Prdeoics Mnl 152 100 The amino acid sequences of the Kir and SUR proteins 7亚 were ohtained from the Swisspro ditabase (www.expasy. org).Prediction of the presence of mitochoodria transport 知 tags was perfoxmed using the iPsort sofware based on the 35 calculations of Lopez ct al.20]. 18 3.Results kCa bcn p111n 3..enlificulicn术Amek库nni初he nitochoudrtal membranes chnrel uhirits The trp ponel chous botting wih an ati- Single or small groups of mitochondria were visaalized ambody.while the lwer ponel shows bioming oth an anti-Kn.I nsbody.Eqeal anenis uf eririn wer:leald in cach bae,aigrificarl following isolation or in cultured neurons or astroglia errichrnest can he at tha -1%0-ard -1-tth wit他实anning comfocal micruscopy.The membrane p sUR2d白：-S0D1k6.Ib

brane. After blocking with 3% bovine serum albumin, primary antibodies against the KATP-receptor subunits Kir6.1, Kir6.2, SUR1 or SUR2 (Santa Cruz Biotech, Santa Cruz, CA) were applied in a dilution of 1:500 followed by horseradish peroxidase conjugated secondary antibody. Chemiluminescence was used to visualize the bands. Specificity of the method was tested by omitting the primary antibody from the procedure or preadsorp￾tion with the respective blocking peptides, which resulted in the disappearance of the specific bands (data not shown). Quantitative measurements of mitochondrial enrich￾ment were performed in pooled preparations loaded in increasing protein concentrations and we used an expo￾sure of the membrane, which showed the best resolution on the light and dark bands also. We compared the results to a parallel blot using an anti-cytochrome c antibody as a control for a known mitochondrial protein (eight-fold mitochondrial enrichment). 2.5. Electron microscopy Freshly isolated mitochondria samples were fixed for 3 h in 4% paraformaldehyde/0.05% glutaraldehyde. The sam￾ples were embedded in LR-white, were sectioned, and were collected on uncoated nickel grids. Bovine serum albumin (2%) and 1:30 dilution of normal goat serum (homologous serum to secondary antibody) was applied to block unspe￾cific binding. The sections were incubated overnight with primary antibodies against KATP-channel subunits (see above) in a dilution of 1:100 at 4 jC. Immunogold labeling was achieved by a 4-h incubation with 6-nm gold-labeled secondary antibody. The specificity of the method was tested by the omission of the primary antibody from the procedure, which resulted in the elimination of gold par￾ticles on the sections. Post-sectioning, en-bloc staining and osmium postfixation were omitted to avoid interference with the labeling. 2.6. Proteomics tools The amino acid sequences of the Kir and SUR proteins were obtained from the Swissprot database (www.expasy. org). Prediction of the presence of mitochondria transport tags was performed using the iPsort software based on the calculations of Lopez et al. [20]. 3. Results 3.1. Identification of KATP channel subunits in the mitochondrial membranes Single or small groups of mitochondria were visualized following isolation or in cultured neurons or astroglia with scanning confocal microscopy. The membrane po￾tential specific dye MitoFluorRed was readily taken up by the isolated organelles, indicating a viable, respiring preparation. The green fluorescent sulfonylurea BOD￾IPY-glibenclamide (1 AM) specifically bound to the mitochondria, indicating the presence of sulfonylurea receptors (Fig. 1). Cytoplasm, endoplasmic reticula or nuclei were largely devoid of staining in cultured astrocytes and neurons, showing the relative specificity of the method (Fig. 1). This observation indicated the presence of sulfonylurea binding sites in neuronal and astroglial mitochondria. However, sulfonylurea binding may be independent of KATP channels, so we tested the presence of the specific channel-forming subunits via Western blotting. Fig. 2. Western blotting shows the mitochondrial enrichment of KATP channel subunits. The top panel shows blotting with an anti-SUR2 antibody, while the lower panel shows blotting with an anti-Kir6.1 antibody. Equal amounts of protein were loaded in each lane, significant mitochondrial enrichment can be observed at the f130- and f30-kDa SUR2 and the f50-kDa Kir6.1 bands. 30 Z. Lacza et al. / Brain Research 994 (2003) 27–36

2 Larza t al.Araw Reumreh 994 (2003)22-36 31 Western blotting was performed on full hrain tissue 32.Funcronal m8 astreme是s imd isolated mitochondria homogenates with antibodies taisod against the putalive Kare channel subunits Kit6.1. Isolalod live mitochoodria were energizod by the nddition Kir 6.2.SURI and SUR2.Labeling with the anti Kir6.1 of malate and glutamate in a K-bufer and were visualized amtibody showed a specifie band m -50 kDa (Fig 2) using confocal fooreseent micresoopy.Mitochondria were this protein was markodly [-times)ennched in the loaded with the membrane patental sensitive dye Mito mitochondria preparions compared to the fill tissue FloorRed (1 BM).Appliemtion of glibenclamide (100 HM) The molecalar weight of the protein comesponds well eaused a logs of MitoFluorRed stainine (Fig 4).Addirional with the estimated molecular weight of Kir6.1 (47 kDa) applicntion of mitoK openers diazoxide (100 pM.=18) Anocher,-50 kDa-hand wns recognized by a Kir.2 or BMS-l9109g(10μM为=6 reversod the effect,the antibody in the mitochondria prepurations,however.this mitocbondrial fluorescent signal increused (Fig.4)Vehicle protein did not show mitochondrial enrichment (not of the drugs (1%DMSO)had no effoct on mitocbondrial shownl. volume or fluoreseence (n-6.not shown).To test wbether The anti-SURI antibody showed a specific -40-kDa the effect of the Karp opemers was altributable to K' band.though without apparent mitocbondrial enrichment currents,the K'ionophore nigcricin was used.Application (not shown).Labeling with the anti-SUR2 antibody.bow- of'nigericin (10 jM.n-4)rewersed the elledt of'glibenea- ever,showed two signilicantly (-8 times)eriched bands mide similarly to the Kam openers,indicating that the at 30 and -130 kDa (Fig.2)in bram.In full brain tissuc caanges in MitoFluorRed fluoreseence were indoed the samples,an additional ~175 kDa band was observed with consequence of K influx (Tig.41. anti-SUR2 lbeling.and only this protein had the expected We also observed profound changes in the DIC image molecular weight derived from the publisbed sequence (174 during the above mentioned treatments:closing of the kDal channels with glibenclamide resulted in the loss of contrast Ultrastructural localization of the K channel sub- and an overall brightening of the pocture and this effect was units was tested with immunogold electron microscopy. completely reversed by diazoxide.BMS-191095 cr niger- Specific SUR2 and Kir6.I labeling was found in the icin (Fig.4).The observed changes in light scatter.which irner mitochondrial membrane (Fig.3),while labeling are directly related to K'currents,provide additional with SURI or Kins2 showed no appurent mitochondrial evidence for the presence of functional Kaxrp chanmneks in locnlizntion. the mitochondria. Examination of the published sequences of all four Isolated respiring mitochondia had a strong green fluo. putative subunits showed that both Kir6.I and Kir6.2 rescent signal in the presence of the NO-sensitive probe have an N-terminal micocbondria transport tag,indicating DAF-FM.Diaminofluoresceins have recently heen shown that these proteins are tangeted to the mitochondrial to react with ONOO in cultured glial cells,and therefore, membranes.In contrast,neither of the two known SUR the fluorescence increase can be attributable to either NO or sequences had a transport tag in the N-terminal region ONOO prcduction [25].The DAF FM signal was com Pe.3.Electron microgrzphs of isolned rinechondra with immunogold labeling Parel A shows labeling with an arti-Kim.I arnbody.nm gold pemcles (amall biack dub)ar:doperod is he nntrix ta iner menbnne Pud B shons sinibe labeling wih i uri-SUR2 antibuly.Bkgruand sunng b milecbosdridl stracture

Western blotting was performed on full brain tissue and isolated mitochondria homogenates with antibodies raised against the putative KATP channel subunits Kir6.1, Kir 6.2, SUR1 and SUR2. Labeling with the anti Kir6.1 antibody showed a specific band at f50 kDa (Fig. 2), this protein was markedly (f 8 times) enriched in the mitochondria preparations compared to the full tissue. The molecular weight of the protein corresponds well with the estimated molecular weight of Kir6.1 (47 kDa). Another, f 50 kDa-band was recognized by a Kir6.2 antibody in the mitochondria preparations, however, this protein did not show mitochondrial enrichment (not shown). The anti-SUR1 antibody showed a specific f 40-kDa band, though without apparent mitochondrial enrichment (not shown). Labeling with the anti-SUR2 antibody, how￾ever, showed two significantly (f 8 times) enriched bands at f 30 and f 130 kDa (Fig. 2) in brain. In full brain tissue samples, an additional f 175-kDa band was observed with anti-SUR2 labeling, and only this protein had the expected molecular weight derived from the published sequence (174 kDa). Ultrastructural localization of the KATP channel sub￾units was tested with immunogold electron microscopy. Specific SUR2 and Kir6.1 labeling was found in the inner mitochondrial membrane (Fig. 3), while labeling with SUR1 or Kir6.2 showed no apparent mitochondrial localization. Examination of the published sequences of all four putative subunits showed that both Kir6.1 and Kir6.2 have an N-terminal mitochondria transport tag, indicating that these proteins are targeted to the mitochondrial membranes. In contrast, neither of the two known SUR sequences had a transport tag in the N-terminal region. 3.2. Functional measurements Isolated live mitochondria were energized by the addition of malate and glutamate in a K+ -buffer and were visualized using confocal fluorescent microscopy. Mitochondria were loaded with the membrane potential sensitive dye Mito￾FluorRed (1 AM). Application of glibenclamide (100 AM) caused a loss of MitoFluorRed staining (Fig. 4). Additional application of mitoKATP openers diazoxide (100 AM, n = 18) or BMS-191095 (10 AM, n = 6) reversed the effect; the mitochondrial fluorescent signal increased (Fig. 4). Vehicle of the drugs (1% DMSO) had no effect on mitochondrial volume or fluorescence (n = 6, not shown). To test whether the effect of the KATP openers was attributable to K+ currents, the K+ ionophore nigericin was used. Application of nigericin (10 AM, n = 4) reversed the effect of glibenca￾mide similarly to the KATP openers, indicating that the changes in MitoFluorRed fluorescence were indeed the consequence of K+ influx (Fig. 4). We also observed profound changes in the DIC image during the above-mentioned treatments: closing of the channels with glibenclamide resulted in the loss of contrast and an overall brightening of the picture and this effect was completely reversed by diazoxide, BMS-191095 or niger￾icin (Fig. 4). The observed changes in light scatter, which are directly related to K+ currents, provide additional evidence for the presence of functional KATP channels in the mitochondria. Isolated respiring mitochondria had a strong green fluo￾rescent signal in the presence of the NO-sensitive probe DAF-FM. Diaminofluoresceins have recently been shown to react with ONOO
in cultured glial cells, and therefore, the fluorescence increase can be attributable to either NO or ONOO
production [25]. The DAF-FM signal was com￾Fig. 3. Electron micrographs of isolated mitochondria with immunogold labeling. Panel A shows labeling with an anti-Kir6.1 antibody, 6-nm gold particles (small black dots) are dispersed in the matrix and inner membrane. Panel B shows similar labeling with an anti-SUR2 antibody. Background staining is negligible in both preparations. We chose to omit post-staining of the samples to avoid interference with the labeling, hence the low contrast of the mitochondrial structure. Z. Lacza et al. / Brain Research 994 (2003) 27–36 31

12 21am时ad/Arain Resnareb [M0-6 5 min 8 min 10 min eontrol glibenclamide +diaroxide 1 jum control +glibenclamide +BMS-191095 G W 1 wum control glibcnclamide nigericin K 1 um cordrinns (A.trarmetted lighel the minchondria are murd,fr and have a bright MimFluorited igal (D,floreceat ireape)AppEcation af 100 urol gisenclamide couses nsane change in lighet scamter,the DIC image lses cortrast and the orgarelles are hardly vissaliaed (B)ond the ned fluaorescence 适ppean (E队Aulititsal年icalin t decoride [l00 molill neves te poovict:erl on lign scar (C》atd fuot3s:nngetas wair (F.5iila起 effect can he achieund hy the applratien oe 10 umold MS-191095.a stmctrally dfterer mitok an opene [-1)Applicarion of he k'-iorophar nierrici (10 gmnol en also reverse the edect of glibenelmide.indiein the rok of x"cueeents n the mechonim (-LL

Fig. 4. KATP channels regulate mitochondrial membrane potential. The timescale shows minutes after energization of the mitochondria. Under baseline conditions (A, transmitted light), the mitochondria are round, fat, and have a bright MitoFluorRed signal (D, fluorescent image). Application of 100 Amol/l glibenclamide causes instant change in light scatter, the DIC image loses contrast and the organelles are hardly visualized (B), and the red fluorescence disappears (E). Additional application of diazoxide (100 Amol/l) reverses the previous effect on light scatter (C) and fluorescence increases again (F). Similar effect can be achieved by the application of 10 Amol/l BMS-191095, a structurally different mitoKATP opener (G – I). Application of the K+ -ionophore nigericin (10 Amol/l) can also reverse the effect of glibenclamide, indicating the role of K+ currents in the mechanism (J – L). 32 Z. Lacza et al. / Brain Research 994 (2003) 27–36

214mt/mw&nmh9420到X-话 untrested FP15 FP15+diazoxide B H Fig.S.Mrochondrial ONOC prodection is sequired or the buildap of membene potestia Te frst oolrin shews uneaed mitochordra which hve brizht DAF-FM (A)anl MiuFlurRodl (DI fuoncencr (u7 The cetler poluam hhows inother goup of miltchurdi which were pronaled with the ONOO" ocaeger 1715 1100 uM.-71.Panel l shows that mtochondris do roe exhiht DAF'-FM fuorescence and carot lad Mitol heeRed n te peecence of FP15.Alilius ofh mKdanel opetr daouide (100 doef uf NO pod(C bullouee:(F)The DIC:imges (G-l)ernfimm that simiar fields of mitechdra wene stadind aad he lec nf fluomocrnce not due o the loes of socas in the picam. plctely abolished in the presence of'the ONOO-decompo- the mitochondria with FP15 abo resulted in reduead Mito- sation emalyst FP15 indscating that milochondrally fiomod Foor代d fuoresecnee,which was restored b方he applic NO is readily converted to ONOO-(Fig 5).Treatmer of 1io箱of'dizoxde(r-6)Fig5S线. SUR2 --1231 kf%s SUR2 30D1 6m,1 -500 a and130-kDa SLR2s are sinikrly espeessol ana milchodrielly enicod in Te:fferetn spovies

pletely abolished in the presence of the ONOO
decompo￾sition catalyst FP15 indicating that mitochondrially formed NO is readily converted to ONOO
(Fig. 5). Treatment of the mitochondria with FP15 also resulted in reduced Mito￾FluorRed fluorescence, which was restored by the applica￾tion of diazoxide (n = 6) (Fig. 5). Fig. 5. Mitochondrial ONOO
production is required for the buildup of membrane potential. The first column shows untreated mitochondria, which have bright DAF-FM (A) and MitoFluorRed (D) fluorescence (n = 7). The center column shows another group of mitochondria, which were pretreated with the ONOO
scavenger FP15 (100 AM, n = 7). Panel B shows that mitochondria do not exhibit DAF-FM fluorescence and cannot load MitoFluorRed in the presence of FP15. Addition of the mitoKATP channel opener diazoxide (100 AM) does not have an effect of NO production (C), but restores the red fluorescence (F). The DIC images (G – I) confirm that similar fields of mitochondria were studied and the loss of fluorescence is not due to the loss of focus in the picture. Fig. 6. Species comparison of the expression patterns of mitoKATP subunits. Respective bands of Western blots show that the f50-kDa Kir6.1 and the f30- and f130-kDa SUR2s are similarly expressed and mitochondrially enriched in three different species. Z. Lacza et al. / Brain Research 994 (2003) 27–36 33

34 2am/Arai Reurareb0动-j6 3.3.Species compartsous for the cooventiccal SUR subunits,were detected in that study either. In addition to mouse brain samples (n-8).mitochonris Although the cxpocted.large molecular weight SUR were also prepired fiom rat (n-4)and piglet (n-3)brains ns were not found in1 he mitochondri填，wo other Western blotting with ami-Kirf.I snd ant-SUR2 antibodes bards of'-30 and -130 kDa were picked up by the snti- showed similar results scen with the mouse preparations SUR2 antibody.These protcins showed prominent mito- (Fig.6)Thus,there is im enchmend of Kir6.1 and SUR2 chondrial enrichment by Westem blotting and immnogold related hands in mitochondria compared to whole hrain in electron mieroscopy Furthermore,the specifle SUR2 anti- all three species. body kcaliged the signal to the inner mitochondrial mem- brane.Examining the published sequence of the 174-kDa SUR2 confirmed that no mitochondrial transport tag is 4.Discussion present in the N-terminus of the protemn,supporting the idea that SUR2 is not involved in mitoKAr While it is There are three main findings in this study.First.Kir6.I possible that the two SUR-2 bands represent fragments of ad a SUR2 related subun appear to be the predomninant the larger SUR-2 protein.we do not think that this is the subtypes of the K in brain mitochondris Socond,phar- case since the anlibody is direcled against a short amino acid macological agems are able to open and close KArp and Seuence of the C-terminal of'the protein.Thus,it woukl be change membrane potental and morphology of isolated nlikely theat both fingments would be deteeted using the mitochoodria.Third,NO/peroxynitrite appears to be an SUR 2 antibody. emdogenous opener of mitoKAr Thus,taken together.our A 28 kDa SUR protein was identified in liver mitochon resalts indicate an important role of KAr in the regulation dria by lbeled glibenclamide blocking in arother stdy of mitochondrial fimction. (32].At the present time,the naturre of these two SUR2 Sabunit characterizatioo was achieved by immunolog- specific bands is unclear.but they may represent krown ical probing of the two types of obligatory subunits of a smaller molecular weight SUR subunits or may represent fumctioeal KAv charnel.Application of a selective Kir6.I new but undescribed SUR variants. antibody showed a mitochoedria specific protein with In our pharmacological experiments.we chose to exam- hoth Westem blotting and electron microscopy.This ine sequential responses of individnal or small groups of protein is located in the inner mitochondrial memhrane, mitochondria using high optical mgnification in order to and its estimated molecular weight (~50 kDa)comre- eliminate some of the technical concerns that may exist sponds well with the expected molecular weight derived hen ocher preparatios,such as tissue slices or suspended from the known sequence (47 kDa).The published mitocbondria,are studied.For example,cellular events may sequence of mouse Kirt.I has an N-terminal mitochoodria affect mitochoodrial responses for intact cells and hulk transport tag and we found Kir6.I immunoreactivity and measurements of saspended mitochondria may not detect mitochandrial concertration in two other speries (rat and beterogeneous responses of mitochondria.We are unaware pigk thereby establishine the eenerality of our firdings in of any prevoous studies of'isolated mitochondrin usine the the mouse.In contrast to Kirf,1.Kirf.2 was not enriched sClp0省 in the mitchondrin preparations compared to the fall Direct visualintion of indivilual mitochondria allowed hrain tissue.Althouph a mitochondria-specific transport sto directly characterize the disappemranee and reloading tng was identified in the N-terminaal region of the se- of the fluorescent probe in respoese to glibenclamide and quence,this channel is undouhtedly expressed in the diazoxide or BMS 191095,respecrively.The fluorescent plasma memhrane ns well as the inner mitochondrial intensity of mitochondria selective probes may he affected membrane [27,30,31].Therefore,it is remsonahle to hy- by loading conditions.redox potential,ApH,Av and pothesize that the Kir 62 subamit coetributes to mitoK probably other,yet undetermined factors [4].In the present compositon to a lesset extent than the Kir 6.I subunit or study,we chose MitoFluorRod to monitor mitocbondrial that the Kir 6.2 subunit cam substilute for the Kir 6.I membetne potential becsusse of the advantages of this dye, subunit 8uch as rapid loading and photostabilily.Pilot expcriments mmmnological p4obH笔w1hli-SURI arxd anti-SU代2 with more ck4 ssical小s like thodmine1.2,32月8owed ibodics laild to show any mitochondrial protcins at the Simalar results,so i seems unlikely that the changes that we espected molecular weight of-175 kDa.These observa- ohserved refleet nonspecinie eflicts of the prohe Further- tions may refleet that nenther one of the well-deserihed,large more changes in liphe seattering independent of the fhur- SUR pmteins are present in the mitochondria.However,our csecent probe was nlso prominent after.the blockede and experiments with BODIPY-glibenclamide showed that mi- opening of the milokcrr channels.and this effect directly tochondria specifically bind sulfonylureas,and this com reflects K'currents 32]. pourd was successfully used to isola Kar channels from Recent studies challenged the specificity of diazoxide rat beain miochondria in another study 31 However.no as a mitoKArp opener bassed on measurements of flavo- large molecular weight bands.such as would be expected protein or TMRE fluoreseence [11,161.These studies

3.3. Species comparisons In addition to mouse brain samples (n = 8), mitochondria were also prepared from rat (n = 4) and piglet (n = 3) brains. Western blotting with anti-Kir6.1 and anti-SUR2 antibodies showed similar results as seen with the mouse preparations (Fig. 6). Thus, there is an enrichment of Kir6.1 and SUR2 related bands in mitochondria compared to whole brain in all three species. 4. Discussion There are three main findings in this study. First, Kir6.1 and a SUR2 related subunit appear to be the predominant subtypes of the KATP in brain mitochondria. Second, phar￾macological agents are able to open and close KATP and change membrane potential and morphology of isolated mitochondria. Third, NO/peroxynitrite appears to be an endogenous opener of mitoKATP. Thus, taken together, our results indicate an important role of KATP in the regulation of mitochondrial function. Subunit characterization was achieved by immunolog￾ical probing of the two types of obligatory subunits of a functional KATP channel. Application of a selective Kir6.1 antibody showed a mitochondria specific protein with both Western blotting and electron microscopy. This protein is located in the inner mitochondrial membrane, and its estimated molecular weight (f 50 kDa) corre￾sponds well with the expected molecular weight derived from the known sequence (47 kDa). The published sequence of mouse Kir6.1 has an N-terminal mitochondria transport tag and we found Kir6.1 immunoreactivity and mitochondrial concentration in two other species (rat and pig), thereby establishing the generality of our findings in the mouse. In contrast to Kir6.1, Kir6.2 was not enriched in the mitochondria preparations compared to the full brain tissue. Although a mitochondria-specific transport tag was identified in the N-terminal region of the se￾quence, this channel is undoubtedly expressed in the plasma membrane as well as the inner mitochondrial membrane [27,30,31]. Therefore, it is reasonable to hy￾pothesize that the Kir 6.2 subunit contributes to mitoKATP composition to a lesser extent than the Kir 6.1 subunit or that the Kir 6.2 subunit can substitute for the Kir 6.1 subunit. Immunological probing with anti-SUR1 and anti-SUR2 antibodies failed to show any mitochondrial proteins at the expected molecular weight of f 175 kDa. These observa￾tions may reflect that neither one of the well-described, large SUR proteins are present in the mitochondria. However, our experiments with BODIPY-glibenclamide showed that mi￾tochondria specifically bind sulfonylureas, and this com￾pound was successfully used to isolate KATP channels from rat brain mitochondria in another study [3]. However, no large molecular weight bands, such as would be expected for the conventional SUR subunits, were detected in that study either. Although the expected, large molecular weight SUR proteins were not found in the mitochondria, two other bands of f 30 and f 130 kDa were picked up by the anti￾SUR2 antibody. These proteins showed prominent mito￾chondrial enrichment by Western blotting and immunogold electron microscopy. Furthermore, the specific SUR2 anti￾body localized the signal to the inner mitochondrial mem￾brane. Examining the published sequence of the 174-kDa SUR2 confirmed that no mitochondrial transport tag is present in the N-terminus of the protein, supporting the idea that SUR2 is not involved in mitoKATP. While it is possible that the two SUR-2 bands represent fragments of the larger SUR-2 protein, we do not think that this is the case since the antibody is directed against a short amino acid sequence of the C-terminal of the protein. Thus, it would be unlikely that both fragments would be detected using the SUR-2 antibody. A 28-kDa SUR protein was identified in liver mitochon￾dria by labeled glibenclamide blocking in another study [32]. At the present time, the nature of these two SUR2 specific bands is unclear, but they may represent known smaller molecular weight SUR subunits or may represent new but undescribed SUR variants. In our pharmacological experiments, we chose to exam￾ine sequential responses of individual or small groups of mitochondria using high optical magnification in order to eliminate some of the technical concerns that may exist when other preparations, such as tissue slices or suspended mitochondria, are studied. For example, cellular events may affect mitochondrial responses for intact cells and bulk measurements of suspended mitochondria may not detect heterogeneous responses of mitochondria. We are unaware of any previous studies of isolated mitochondria using the present approach. Direct visualization of individual mitochondria allowed us to directly characterize the disappearance and reloading of the fluorescent probe in response to glibenclamide and diazoxide or BMS 191095, respectively. The fluorescent intensity of mitochondria selective probes may be affected by loading conditions, redox potential, DpH, DC and probably other, yet undetermined factors [4]. In the present study, we chose MitoFluorRed to monitor mitochondrial membrane potential because of the advantages of this dye, such as rapid loading and photostability. Pilot experiments with more classical dyes like rhodamine 1,2,3 [29] showed similar results, so it seems unlikely that the changes that we observed reflect nonspecific effects of the probe. Further￾more, changes in light scattering independent of the flur￾osecent probe was also prominent after the blockade and opening of the mitoKATP channels, and this effect directly reflects K+ currents [32]. Recent studies challenged the specificity of diazoxide as a mitoKATP opener based on measurements of flavo￾protein or TMRE fluorescence [11,16]. These studies 34 Z. Lacza et al. / Brain Research 994 (2003) 27–36

2aH/mt系mrh9202无- showed that diazoxide can inhibit the mitochondrial the physiologicnl findings indicating an important role of respiratory chain enzymes and therefore cannot be used mitoK channeks in protection of the central nervous as a selective mitoKarr openet.It has boen known for systemn fiom ischemic stress. ttany years that diacoxide tas other.non-mitochondral elfects,for exampk,as a vasodilator trealment in hyper- tensive patients.The present study shows that diszoxide can open mitoK chanels in inteet respiring mitochon- Acknowledgements dria and this effect is similar to those of'a chemically different mitoK opener or a K"ionophore.The novel This sludy was supported by grants from the Amenican mitokrr opene BMS-191095 wis found to be oqually Heart Association (Mid-Allanbe Grar 99512724.Bugher effective as diazoxide at 10 times lower dose.This Foundation Award 0270114N).the NIH (HL30260. HL46558,HL50587.DK 62372)ad the Hungarimm OTKA compound can induce pharmacological preconditioning (T029169.T-037885.T-037386)nd上TT(2182001,248 in heart iscbemia without opening sarcolemmal or vascu lar Karp channels [9.Since BMS-191095 can get 2003.乙。L.wg5 upponted by an OTKA91ca through the blood-brain barrier (brain/plasma ratio: fellowship. 0.31)it may be a suitable pharmacological lool for the imestigation of mitoKarpin stroke models. References An mleresting observaton of the presem study is the poesibl role of mitochondrially fommed NO ar peroynitrite [1]L.Agsil-Bryar,J Hryan.Molocubr bidlngy nf adcroene trphre- as amn endogenous mitoKAr opener.Previous studies dem t6sse2 ssom chamels,E山ot.Rex.2019网Io1-1J5 orstrated that NO can activate mitoKArr channels [26]and 间CP.Bi5M.P,里Piw,Prulen kisaas and krp-ta山sd ONOO my also open mitoKr through activation of elfectors in the bte phove of ischamtic preendemning.Hoic Res procein kinase C [2].It is also well known that these Cd0L520l12)7=21&. compounds are produced in lange quantities during brain [S]R.Raigar,S.Sochararar,A.1 Kqwabrwski,K.D.Carlid,P.Pauccl lamificaliun and prupertio of a tuvd inmx luar (nilechadialy ischemia and reoxygenation [12 Moreover.mtNOS is ATP-wnithe porzcomm charnel in hrin.J lunl Che 3760) upregulated after hypoxia,raising he possibality that mito- 33360-33374. chondrial NO formation plays a role in the pathology of hypoxia-reoxygenation [14].The mitochoedrial respiratory Reynelds.MiloTracker labeling tn primtary neuronol and astcyuc chain is a constant source of superoxide,which rendily calare:influno:国iu止立embeun:pokslid arl oilasla reacts with NO so form ONOO [18].Therefore,NO and [5]F.Domoii,J.V.Perciaccante,R.Veltkamp,F.Bari.D.W.Basie ONOO are produced together by respiring mitochondria Miehoedrial putoam cfonael nener diniside prewerves retro- [8].The NO-dependent fluorescent poobe DAF has heen ral-vascelr fancrion after cerehral ischemia i newbom pies.Smoke shown to be reactive to boch NO and ONOO [13,25].The 3019992713-2718. presemt study is the first to document that endogenously 间dA8.cia,H Fnrke M.Moank.K Nicher.P I风，Nmm pronection by ATPdependeat poussium chonnels in rot teocorticol produced reactive mitrogen speeis activate the mitoK bnin slios daring bypoxi Nemnnci.Lett 233 (199%)13-16. channel,which my serve as s self-regulatory proeess ofthe K D.Giarlid,片k的m的y.II.N Merray.RI.nr orpanelle.However,the exact mechanism of mitochoedral betudu.AJ.D'AuNJ.La,A.Snih GJ.Gn.C山当- NO production is not yet disoovered and its importance in physiolgical mitoK civation cannot be estimmtod ar ATP-milive K+chainada.Puoochl mochuian of canfienrccticn 0r.Rs.1(I9T力1072-1t2 ths time. [81 P.Ghafuarifir,U.Scesk.S.D.Kl:is,C.Bichker,Mohoudrial Large-scale clinical studies with various phamacolog- itni3-gaih号nme cmubtirn cauo cytochmne e rdeaoe句m ical agents during the recent few years showed that isolared rehechosdria.Eviderce or moninochondrial perosynine results gained in rodent stroke models cannot be easily wuuu,1BlCn274(9093115-3113 9]Gl.(imer.A.J D'Alozn.K D.Grld.B Bygv.NJ.Lodge.P& adaplod to human use.Therefore,in the search for new Skpl,R B.Du:benzis,TA.H:M.A.Smith.P.Pusek,K.5. therapeutic targets,it is cssential to prove that the Arwal,Pumrcologic chorachrgation nf BMS-191IN5,a mitochna- proteclive mochanism Bs oqually cllective in large amimals dnal K(ATP)opeer wi no perpheral vasocilaor or cartiac action as well as rodents.The very lirst study,which descnbod ativily,J.Eap.Thet.297 (2001) the protective ellect of mitokcrr chanel opening in 11%4-1142 brain.was condueted in pigs [5 Sinee then similar [I0川R王Hamesen.J.Ma.SA Deedwyle,Conrebidood and kappe ioid receptrex rod定trun K aeat vi由tivatien af G results were puhlished in dogs,mice,adult and newbom rats with signifcant imgrovements in ncumlogical score 236-2361. [19,24,28]coupled to a marked decresse in neurunal cell IIPLHh,Mk：LM.amc,U.BL.Da,TPh death or infarct volume (adult mouse:-85%:adult nek-independert tryex of diamide and 5-hydmtydocanrate in the rat:-50%newbom rat:-20%).The present study. he3L.1.sL.542120K21535-14L 12]C.baoocha.Bragghl and dek sidks uf nitre caide is bclanit brin showing a similar enhancement of K channel subumnits 可小y.Trerds Nremo闭17)1-lw in mituchundna from brains of several species,reinforces [13]D.Joud'houi.Inesed uic oxile-depeadem nimusylaton of 4.5

showed that diazoxide can inhibit the mitochondrial respiratory chain enzymes and therefore cannot be used as a selective mitoKATP opener. It has been known for many years that diazoxide has other, non-mitochondrial effects, for example, as a vasodilator treatment in hyper￾tensive patients. The present study shows that diazoxide can open mitoKATP channels in intact respiring mitochon￾dria and this effect is similar to those of a chemically different mitoKATP opener or a K+ ionophore. The novel mitoKATP opener BMS-191095 was found to be equally effective as diazoxide at 10 times lower dose. This compound can induce pharmacological preconditioning in heart ischemia without opening sarcolemmal or vascu￾lar KATP channels [9]. Since BMS-191095 can get through the blood – brain barrier (brain/plasma ratio: 0.31), it may be a suitable pharmacological tool for the investigation of mitoKATP in stroke models. An interesting observation of the present study is the possible role of mitochondrially formed NO or peroxynitrite as an endogenous mitoKATP opener. Previous studies dem￾onstrated that NO can activate mitoKATP channels [26] and ONOO
may also open mitoKATP through activation of protein kinase C [2]. It is also well known that these compounds are produced in large quantities during brain ischemia and reoxygenation [12]. Moreover, mtNOS is upregulated after hypoxia, raising the possibility that mito￾chondrial NO formation plays a role in the pathology of hypoxia-reoxygenation [14]. The mitochondrial respiratory chain is a constant source of superoxide, which readily reacts with NO to form ONOO
[18]. Therefore, NO and ONOO
are produced together by respiring mitochondria [8]. The NO-dependent fluorescent probe DAF has been shown to be reactive to both NO and ONOO
[13,25]. The present study is the first to document that endogenously produced reactive nitrogen species activate the mitoKATP channel, which may serve as a self-regulatory process of the organelle. However, the exact mechanism of mitochondrial NO production is not yet discovered and its importance in physiological mitoKATP activation cannot be estimated at this time. Large-scale clinical studies with various pharmacolog￾ical agents during the recent few years showed that results gained in rodent stroke models cannot be easily adapted to human use. Therefore, in the search for new therapeutic targets, it is essential to prove that the protective mechanism is equally effective in large animals as well as rodents. The very first study, which described the protective effect of mitoKATP channel opening in brain, was conducted in pigs [5]. Since then similar results were published in dogs, mice, adult and newborn rats with significant improvements in neurological score [19,24,28] coupled to a marked decrease in neuronal cell death or infarct volume (adult mouse: f85%; adult rat:f 50%; newborn rat: f 20%). The present study, showing a similar enhancement of KATP channel subunits in mitochondria from brains of several species, reinforces the physiological findings indicating an important role of mitoKATP channels in protection of the central nervous system from ischemic stress. Acknowledgements This study was supported by grants from the American Heart Association (Mid-Atlantic Grant 99512724, Bugher Foundation Award 0270114N), the NIH (HL30260, HL46558, HL50587, DK 62372) and the Hungarian OTKA (T-029169, T-037885, T-037386) and ETT (218/2001, 248/ 2003). Z. L. was supported by an OTKA postdoctoral fellowship. References [1] L. Aguilar-Bryan, J. Bryan, Molecular biology of adenosine triphos￾phate-sensitive potassium channels, Endocr. Rev. 20 (1999) 101 – 135. [2] C.P. Baines, J.M. Pass, P. Ping, Protein kinases and kinase-modulated effectors in the late phase of ischemic preconditioning, Basic Res. Cardiol. 96 (2001) 207 – 218. [3] R. Bajgar, S. Seetharaman, A.J. Kowaltowski, K.D. Garlid, P. Paucek, Identification and properties of a novel intracellular (mitochondrial) ATP-sensitive potassium channel in brain, J. Biol. Chem. 276 (2001) 33369 – 33374. [4] J.F. Buckman, H. Hernandez, G.J. Kress, T.V. Votyakova, S. Pal, I.J. Reynolds, MitoTracker labeling in primary neuronal and astrocytic cultures: influence of mitochondrial membrane potential and oxidants, J. Neurosci. Methods 104 (2001) 165 – 176. [5] F. Domoki, J.V. Perciaccante, R. Veltkamp, F. Bari, D.W. Busija, Mitochondrial potassium channel opener diazoxide preserves neuro￾nal – vascular function after cerebral ischemia in newborn pigs, Stroke 30 (1999) 2713 – 2718. [6] d.A.S. Garcia, H. Franke, M. Pissarek, K. Nieber, P. Illes, Neuro￾protection by ATP-dependent potassium channels in rat neocortical brain slices during hypoxia, Neurosci. Lett. 273 (1999) 13 – 16. [7] K.D. Garlid, P. Paucek, V. Yarov-Yarovoy, H.N. Murray, R.B. Dar￾benzio, A.J. D’Alonzo, N.J. Lodge, M.A. Smith, G.J. Grover, Cardi￾oprotective effect of diazoxide and its interaction with mitochondrial ATP-sensitive K+ channels. Possible mechanism of cardioprotection, Circ. Res. 81 (1997) 1072 – 1082. [8] P. Ghafourifar, U. Schenk, S.D. Klein, C. Richter, Mitochondrial nitric-oxide synthase stimulation causes cytochrome c release from isolated mitochondria. Evidence for intramitochondrial peroxynitrite formation, J. Biol. Chem. 274 (1999) 31185 – 31188. [9] G.J. Grover, A.J. D’Alonzo, K.D. Garlid, R. Bajgar, N.J. Lodge, P.G. Sleph, R.B. Darbenzio, T.A. Hess, M.A. Smith, P. Paucek, K.S. Atwal, Pharmacologic characterization of BMS-191095, a mitochon￾drial K(ATP) opener with no peripheral vasodilator or cardiac action potential shortening activity, J. Pharmacol. Exp. Ther. 297 (2001) 1184 – 1192. [10] R.E. Hampson, J. Mu, S.A. Deadwyler, Cannabinoid and kappa opioid receptors reduce potassium K current via activation of G(s) proteins in cultured hippocampal neurons, J. Neurophysiol. 84 (2000) 2356 – 2364. [11] P.J. Hanley, M. Mickel, M. Loffler, U. Brandt, J. Daut, K(ATP) chan￾nel-independent targets of diazoxide and 5-hydroxydecanoate in the heart, J. Physiol. 542 (2002) 735 – 741. [12] C. Iadecola, Bright and dark sides of nitric oxide in ischemic brain injury, Trends Neurosci. 20 (1997) 132 – 139. [13] D. Jourd’heuil, Increased nitric oxide-dependent nitrosylation of 4,5- Z. Lacza et al. / Brain Research 994 (2003) 27–36 35

2amNa/ms9HN0利-6 caminofluorescel by8ud版起tiots for the seasurement o 24月NRpe,K.Shinia,B.K线1Spes,ZL2LD.Bs色 nite ucide,Frov Rualic Biul Mod 33 (2002)676-684 Adivuion uf mil:hoerial ATP-octpilive polzoium derecls pro- 1可之1a1 xkar.JP Ipimea,1.e.N.Ry可3，Dw erts naimnal cell derh aer ischeris in neanotal res,Nermoc Bua.Modeorial airi anide syuhbe is cosituivdly utive 113272004205-212. and is firctiorlly pmegulaled is bypeoia Froe Radlic Bird.Med [2明S Koychowihiry.A.山he.G Kelhaff,0.Wolf.T.F.Hem,Ox 312001)19-1615 dative sress in ghal coltares:dctoction by DAF2 ruorescence used 间L.mib日FCe1TPk,CS.PdKD.Psi5w,A. tal u maoue proayailrite nlar llan nire coide,Gili 33 Kimonoczy.Iitteets nf argnine vaoresn aad.atrepertin mn alil 2p)10-114. cell volame memsured s 3MG spc.Am.J.PlosoL 264 (1993) 26]N.Sueki T.Sulo.A.Chkr.B.O'Rouke.E Marbur.Activuion of C603-C80R. ATP-deperdet potasim hy niric nxide, [o]C.L.Lawrerre,B.Bilues Gi.C.Rodkigo,N.B.Sunden,The KATP Cirulalin1012000439-445 charnel operer daomcide pmorects candoc myoestes dirrg merahohc [27]I.Sehorzuryrn.A.Cher,N.Sasski.I Fmor,T.Smo,nC Johs.I O Rouke.E.Marten,Molkccuker comnposison of milecond ATP. 1aidt5▣BBg.】.3al.14a01)535-542 eraitie patrain chamnds prohod by vial Kir gee trarfer,J.Mol 117]G.Lebuffe.PT.Schumacker.Z.H.Shoo.T.Anderson.H Iwase. Cell Ca122N3123-190 H.T.L.Vandkn,ROS und NO uige curly pooeraliitung:tlalup- 2鬥J,G Shale,E.A.Pok,ELGL,MJd▣馬，JC. ship to mtochondral K AIP chinel,Am.J.Phyxiol,Hfeon ('nc Traecren.IM Itedraond,WA.lotmgartrer,Paarrarningicaly in- P%e.28420G1H29月=3g duced precondzioting with diazoxcie:a rovel approech %o beait [18]L.Lsudet,F.0.Sorino,C.Smha,Bioleggy of miric ooke aigreling. prtectian Arn.Therac.Sarg ?2 (2001)1849-1854. m.C0d.2w2m57-N52 2yKhm.乙.L3xn,N.n★，Tl0nck,1.Ses,nW [19]D.Lu.C.LE,R.Wan,W.W.Aryoung.M.P.Mulbue,Adivuio o Buija.SuK(ATP)upetT.diouid:poluss Isunal d raitochotdrel Al1'-ipandene pataorem chonels prnteets neumns adfer mddle cerhol artery ncchwirn in the rat,Am I P'ayiol,Heart agairst ischertia-isduood dexh by a mechantsm iavohing suppoes CE.s0L232002)H105-HI011 ene uf Beo trarabcasun and rytuchre e rlepe,1.Cercb.Bhd 〔3 M Suaiki,RA5T日iH.U:ammN.&anotn,Y.Oh 1wMrh.222)431-443 Sekne.M.TanagTwa.T Ogrr.S.Seiro.E.Marton,1.Nabaya 201 M.F.Lopez.B.S.KabtaL E Chemukebiayt,A.Laarev,A.I.Sho- Fuictiotral polei of curdi ud vpcuke ATP.enive potsum spolov.A Bodanoa.M.Robnson,High-huglput peotiling o chatnes clarified by Kir6.2-krockout.mke.Cin.Res.85 (2001) the manchondrial prutome tning affinity fractaration and alote 51-5,2 m日1e5h3sk21120127-344回 L3】MS1,N.Swl.T Mik.N.Salcmeto,Y《hr-Selaine.M. 21]M.P Mutbus,Y.Zlan4.S.Booe,Gruwh fuclos poevetl milochun- TH4h45c出u,E.Maben且Nalayu,Rul of sarcoletamal dral dyshneso,lras nf calciam homnaoeauis and cell irjery,but no K(ATP)chnrel in cardiop四ciagar同mn ATP deplerion t hippecampal netroas deprived of glacose.Exp. 0nn32J.lin.Imes10922)S09=516 unl1211931-t3. [32]A.Soewezyk,B.NSkolajek,S.Pkul.M.J.Naloc Pelzoiam den- nel openens indace michondnal matrte vohme charges va acti- ria:a ddlay of kthl ocll inury n bebcmic myosmlium,Circulelion valus o ATPoctsive K+clam:L Pol.J.Piansacdl.45 (1993) 1441961124-114. 437-448 [23]I.Nokapo.H.Notase.S.Akera,Y.Kamoda M.Yarashit T. 33]M Zhou,O.Tandka.M.Seliguch,K.Sokabe,M.Anzal I.Izurida Salaki,A-dpdnl polaoian charnel me适ake neurupndecti图 T.huEK.Kawalan H.Ab,Localicni图sf起TP.aalive by caerical pmeandroning witt I-amepmpinmc acid in Dril hip- potagdrm chrrel aihne [Kim IAiK(ATP1)in rar hearn.Hoin pe球s,kuoL1.3000033-36 Re线Mol Brai山Rs.74199明15-2s

diaminofluorescein by oxidants: implications for the measurement of intracellular nitric oxide, Free Radic. Biol. Med. 33 (2002) 676 – 684. [14] Z. Lacza, M. Puskar, J.P. Figueroa, J. Zhang, N. Rajapakse, D.W. Busija, Mitochondrial nitric oxide synthase is constitutively active and is functionally upregulated in hypoxia, Free Radic. Biol. Med. 31 (2001) 1609 – 1615. [15] L. Latzkovits, H.F. Cserr, J.T. Park, C.S. Patlak, K.D. Pettigrew, A. Rimanoczy, Effects of arginine vasopressin and atriopeptin on glial cell volume measured as 3-MG space, Am. J. Physiol. 264 (1993) C603 –C608. [16] C.L. Lawrence, B. Billups, G.C. Rodrigo, N.B. Standen, The KATP channel opener diazoxide protects cardiac myocytes during metabolic inhibition without causing mitochondrial depolarization or flavopro￾tein oxidation, Br. J. Pharmacol. 134 (2001) 535 – 542. [17] G. Lebuffe, P.T. Schumacker, Z.H. Shao, T. Anderson, H. Iwase, H.T.L. Vanden, ROS and NO trigger early preconditioning: relation￾ship to mitochondrial KATP channel, Am. J. Physiol, Heart Circ. Physiol. 284 (2003) H299 – H308. [18] L. Liaudet, F.G. Soriano, C. Szabo, Biology of nitric oxide signaling, Crit. Care Med. 28 (2000) N37 – N52. [19] D. Liu, C. Lu, R. Wan, W.W. Auyeung, M.P. Mattson, Activation of mitochondrial ATP-dependent potassium channels protects neurons against ischemia-induced death by a mechanism involving suppres￾sion of Bax translocation and cytochrome c release, J. Cereb. Blood Flow Metab. 22 (2002) 431 – 443. [20] M.F. Lopez, B.S. Kristal, E. Chernokalskaya, A. Lazarev, A.I. Shes￾topalov, A. Bogdanova, M. Robinson, High-throughput profiling of the mitochondrial proteome using affinity fractionation and automa￾tion, Electrophoresis 21 (2000) 3427 – 3440. [21] M.P. Mattson, Y. Zhang, S. Bose, Growth factors prevent mitochon￾drial dysfunction, loss of calcium homeostasis, and cell injury, but not ATP depletion in hippocampal neurons deprived of glucose, Exp. Neurol. 121 (1993) 1 – 13. [22] C.E. Murry, R.B. Jennings, K.A. Reimer, Preconditioning with ische￾mia: a delay of lethal cell injury in ischemic myocardium, Circulation 74 (1986) 1124 – 1136. [23] I. Nakagawa, H. Nakase, S. Aketa, Y. Kamada, M. Yamashita, T. Sakaki, ATP-dependent potassium channel mediates neuroprotection by chemical preconditioning with 3-nitropropionic acid in gerbil hip￾pocampus, Neurosci. Lett. 320 (2002) 33 – 36. [24] N. Rajapakse, K. Shimizu, B. Kis, J. Snipes, Z. Lacza, D. Busija, Activation of mitochondrial ATP-sensitive potassium channels pre￾vents neuronal cell death after ischemia in neonatal rats, Neurosci. Lett. 327 (2002) 208 – 212. [25] S. Roychowdhury, A. Luthe, G. Keilhoff, G. Wolf, T.F. Horn, Oxi￾dative stress in glial cultures: detection by DAF-2 fluorescence used as a tool to measure peroxynitrite rather than nitric oxide, Glia 38 (2002) 103 – 114. [26] N. Sasaki, T. Sato, A. Ohler, B. O’Rourke, E. Marban, Activation of mitochondrial ATP-dependent potassium channels by nitric oxide, Circulation 101 (2000) 439 – 445. [27] J. Seharaseyon, A. Ohler, N. Sasaki, H. Fraser, T. Sato, D.C. Johns, B. O’Rourke, E. Marban, Molecular composition of mitochondrial ATP￾sensitive potassium channels probed by viral Kir gene transfer, J. Mol. Cell. Cardiol. 32 (2000) 1923 – 1930. [28] J.G. Shake, E.A. Peck, E. Marban, V.L. Gott, M.V. Johnston, J.C. Troncoso, J.M. Redmond, W.A. Baumgartner, Pharmacologically in￾duced preconditioning with diazoxide: a novel approach to brain protection, Ann. Thorac. Surg. 72 (2001) 1849 – 1854. [29] K. Shimizu, Z. Lacza, N. Rajapakse, T. Horiguchi, J. Snipes, D.W. Busija, MitoK(ATP) opener, diazoxide, reduces neuronal damage after middle cerebral artery occlusion in the rat, Am. J. Physiol, Heart Circ. Physiol. 283 (2002) H1005 – H1011. [30] M. Suzuki, R.A. Li, T. Miki, H. Uemura, N. Sakamoto, Y. Ohmoto￾Sekine, M. Tamagawa, T. Ogura, S. Seino, E. Marban, H. Nakaya, Functional roles of cardiac and vascular ATP-sensitive potassium channels clarified by Kir6.2-knockout mice, Circ. Res. 88 (2001) 570 – 577. [31] M. Suzuki, N. Sasaki, T. Miki, N. Sakamoto, Y. Ohmoto-Sekine, M. Tamagawa, S. Seino, E. Marban, H. Nakaya, Role of sarcolemmal K(ATP) channels in cardioprotection against ischemia/reperfusion in￾jury in mice, J. Clin. Invest. 109 (2002) 509 – 516. [32] A. Szewczyk, B. Mikolajek, S. Pikula, M.J. Nalecz, Potassium chan￾nel openers induce mitochondrial matrix volume changes via acti￾vation of ATP-sensitive K+ channel, Pol. J. Pharmacol. 45 (1993) 437 – 443. [33] M. Zhou, O. Tanaka, M. Sekiguchi, K. Sakabe, M. Anzai, I. Izumida, T. Inoue, K. Kawahara, H. Abe, Localization of the ATP-sensitive potassium channel subunit (Kir6. 1/uK(ATP)-1) in rat brain, Brain Res. Mol. Brain Res. 74 (1999) 15 – 25. 36 Z. Lacza et al. / Brain Research 994 (2003) 27–36