Ca2+-dependent permeability transition regulation in rat brain mitochondria by 2',3'-cyclic nucleotides and 2',3'-cyclic nucleotide 3'-phosphodiesterase

 doi:10.1152/ajpcell.00006.2009 296:1428-1439, 2009. First published Apr 8, 2009;Am J Physiol Cell Physiol

Baburina, Rolf Stricker, Yuri Evtodienko and Georg Reiser Tamara Azarashvili, Olga Krestinina, Anastasia Galvita, Dmitry Grachev, Yulia2',3'-cyclic nucleotide 3'-phosphodiesterase rat brain mitochondria by 2',3'-cyclic nucleotides and Ca2+-dependent permeability transition regulation in

You might find this additional information useful...

42 articles, 10 of which you can access free at: This article cites http://ajpcell.physiology.org/cgi/content/full/296/6/C1428#BIBL

1 other HighWire hosted article: This article has been cited by

  [PDF]  [Full Text]  [Abstract]

, November 27, 2009; 284 (48): 33097-33106. J. Biol. Chem.E. K. Jackson, J. Ren and Z. Mi

Extracellular 2',3'-cAMP Is a Source of Adenosine

including high-resolution figures, can be found at: Updated information and services http://ajpcell.physiology.org/cgi/content/full/296/6/C1428

can be found at: AJP - Cell Physiologyabout Additional material and information http://www.the-aps.org/publications/ajpcell

This information is current as of December 14, 2009 .  

http://www.the-aps.org/.American Physiological Society. ISSN: 0363-6143, ESSN: 1522-1563. Visit our website at a year (monthly) by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2005 by the

is dedicated to innovative approaches to the study of cell and molecular physiology. It is published 12 timesAJP - Cell Physiology

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

Ca2�-dependent permeability transition regulation in rat brain mitochondriaby 2�,3�-cyclic nucleotides and 2�,3�-cyclic nucleotide 3�-phosphodiesterase

Tamara Azarashvili,1,2 Olga Krestinina,1,2 Anastasia Galvita,1 Dmitry Grachev,1,2 Yulia Baburina,2

Rolf Stricker,1 Yuri Evtodienko,2 and Georg Reiser1

1Institut fur Neurobiochemie, Otto-von-Guericke-Universitat Magdeburg, Medizinische Fakultat, Magdeburg, Germany;and 2Institute of Theoretical and Experimental Biophysics, Russian Academy of Science, Pushchino, Moscow, Russia

Submitted 8 January 2009; accepted in final form 2 April 2009

Azarashvili T, Krestinina O, Galvita A, Grachev D, BaburinaY, Stricker R, Evtodienko Y, Reiser G. Ca2�-dependent permeabil-ity transition regulation in rat brain mitochondria by 2�,3�-cyclicnucleotides and 2�,3�-cyclic nucleotide 3�-phosphodiesterase. Am JPhysiol Cell Physiol 296: C1428–C1439, 2009. First published April8, 2009; doi:10.1152/ajpcell.00006.2009.—Recent evidence indicatesthat 2�,3�-cyclic nucleotide 3�-phosphodiesterase (CNP), a markerenzyme of myelin and oligodendrocytes, is also present in neural andnonneural mitochondria. However, its role in mitochondria is stillcompletely unclear. We found CNP in rat brain mitochondria andstudied the effects of CNP substrates, 2�,3�-cyclic nucleotides, onfunctional parameters of rat brain mitochondria. 2�,3�-cAMP and2�,3�-cNADP stimulated Ca2� overload-induced Ca2� release frommitochondrial matrix. This Ca2� release under threshold Ca2� loadcorrelated with membrane potential dissipation and mitochondrialswelling. The effects of 2�,3�-cyclic nucleotides were suppressed bycyclosporin A, a potent inhibitor of permeability transition (PT). PTdevelopment is a key stage in initiation of apoptotic mitochondria-induced cell death. 2�,3�-cAMP effects were observed on the func-tions of rat brain mitochondria only when PT was developed. Thisdemonstrates involvement of 2�,3�-cAMP in PT regulation in rat brainmitochondria. We also discovered that, under PT development, thespecific enzymatic activity of CNP was reduced. Thus we hypothesizethat suppression of CNP activity under threshold Ca2� load leads toelevation of 2�,3�-cAMP levels that, in turn, promote PT developmentin rat brain mitochondria. Similar effects of 2�,3�-cyclic nucleotideswere observed in rat liver mitochondria. Involvement of CNP in PTregulation was confirmed in experiments using mitochondria fromCNP-knockdown oligodendrocytes (OLN93 cells). CNP reduction inthese mitochondria correlated with lowering the threshold for Ca2�

overload-induced Ca2� release. Thus our results reveal a new functionfor CNP and 2�,3�-cAMP in mitochondria, being a regulator/promotorof mitochondrial PT.

oligodendrocyte mitochondria; 2�,3�-cyclic nucleotide 3�-phosphodi-esterase; permeability transition; calcium transport

THE ENZYME 2�,3�-CYCLIC NUCLEOTIDE 3�-phosphodiesterase (CNP,EC 3.1.4.37) accounts for �2–5% of the total protein in thecentral nervous system myelin and 0.5–1% of peripheral ner-vous system myelin (36). CNP catalyzes the hydrolysis of2�,3�-cyclic nucleotides to form the corresponding 2�-mono-phosphates (3). CNP was shown to be an integral protein ofmyelin of oligodendrocytes in the central nervous system andof peripheral myelin in Schwann cells (36, 39). The majority ofstudies investigating the role of CNP were focused exclusivelyon the expression of CNP in oligodendrocytes and Schwann

cells and the involvement of CNP in myelinogenesis. However,there is increasing evidence showing that this enzyme ispresent in a variety of other cell types. CNP-like enzymeactivity was found in nonmyelin membrane preparations de-rived from spleen, liver, thymus, adrenal glands, kidney, heart,and skeletal muscle. The levels of the enzymatic activity,however, are significantly lower than those in the centralnervous system (12, 15, 40). CNP activity was revealed in ratliver mitochondria, specifically in the outer and inner mito-chondrial membranes (12). In cultured adrenal medullary chro-maffin cells, which are not myelin associated, CNP was foundto be located in mitochondria by means of immunofluorescencestaining (27).

The in vivo biological roles of CNP are still largely un-known, and the possible function of CNP in mitochondria iseven more enigmatic. Two CNP isoforms, CNP1 (46 kDa) andCNP2 (48 kDa), were found (16, 23) that are encoded by asingle gene. The two isoforms are due to the presence of twoalternative translation start sites (28). Several posttranslationalmodifications of CNP, such as phosphorylation, isoprenylation,and acylation, with unclear functional consequences are known(1, 2, 10, 36, 38, 39). A recent study of transfected oligodendro-cyte cells (OLN93), overexpressing CNP1 and CNP2, showedthat CNP2 can be translocated to mitochondria due to the presenceof a mitochondrial targeting signal at the NH2 terminus, which iscleaved on import into mitochondria (24). The translocation ofCNP2 is regulated via phosphorylation of the targeting signal byprotein kinase C. However, there are no data on the function ofCNP in mitochondria.

CNP hydrolyzes not only 2�,3�-cyclic nucleotides, but alsooligonucleotides containing a 2�,3�-cyclic terminus, showingbase preference for purine over pyrimidine (23). 2�,3�-CyclicNADP is also known to be a substrate for CNP, and the Km

values for cyclic NADP and for 2�,3�-cAMP were found to bethe same (35). 2�,3�-Cyclic NADP is usually used for detectionof CNP activity. It should be noted that the enzymatic activityof CNP was established in vitro, and, up to now, there are nodata on the effect of the CNP substrates on mitochondrialfunction. However, association of CNP with the inner seg-ments of photoreceptors in the retina, which contain mitochon-dria, suggested that the electrochemical gradient maintained byphotoreceptors may be regulated, in part, by CNP utilizing2�,3�-cyclic nucleotides (14). 2�,3�-Cyclic phosphate-contain-ing products could be a result of RNA damage. It was recentlyshown that CNP-like 2�,3�-cyclic phosphodiesterase activity ofclostridium thermocellum polynucleotide kinase-phosphatasemight take part in RNA repair (22).

In the present study, the CNP substrates 2�,3�-cAMP and2�,3�-cNADP were shown to be able to stimulate cyclosporin A

Address for reprint requests and other correspondence: G. Reiser, Institut furNeurobiochemie, Otto-von-Guericke-Universitat Magdeburg, MedizinischeFakultat, Leipziger Strasse 44, 39120 Magdeburg, Germany (e-mail: [email protected]).

Am J Physiol Cell Physiol 296: C1428–C1439, 2009.First published April 8, 2009; doi:10.1152/ajpcell.00006.2009.

0363-6143/09 $8.00 Copyright © 2009 the American Physiological Society http://www.ajpcell.orgC1428

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

(CsA)-sensitive, Ca2� overload-induced Ca2� release in rat brainmitochondria. 2�,3�-cAMP and 2�,3�-cNADP shortened the lagtime before Ca2� release, increased the rate of Ca2� efflux fromthe mitochondrial matrix, and activated Ca2�-induced, CsA-sen-sitive swelling of mitochondria. Similar effects of the 2�,3�-cyclicnucleotides were observed in rat liver mitochondria. These find-ings indicate involvement of 2�,3�-cyclic nucleotides in regulationof mitochondrial permeability transition (PT). The oligodendro-cyte cell line OLN93 endogenously expresses CNP. We generatedCNP knockdown OLN93 cells. Interestingly, in mitochondriaisolated from CNP knockdown OLN93 cells, the reduced CNPlevel correlated with lowered threshold of Ca2� concentration tostimulate Ca2� efflux. Thus the present results suggest that CNPmight participate in regulation of Ca2� fluxes in mitochondria andprobably in regulation of PT via hydrolysis of 2�,3�-cyclic nucle-otides.

MATERIALS AND METHODS

Isolation of rat brain mitochondria. Adult male Wistar rats (230–250 g), which were used for obtaining tissues, were purchased fromHarlan Winkelmann (Borchen, Germany). All animal procedures havebeen approved by the ethics committee of the German federal state ofSachsen-Anhalt and are in accordance with the European Coummu-nities Council Directive (86/609/EEC). Rat brains were rapidly re-moved (within 30 s) and placed in ice-cold solution, containing 0.32M sucrose, 0.5 mM EDTA, 0.5 mM EGTA, 0.02% bovine serumalbumin (fraction V), and 10 mM Tris �HCl (pH 7.4). All solutionsused were ice cold; all manipulations were carried out at �4°C. Thetissue was homogenized in a glass homogenizer; the ratio of braintissue to isolation medium was 1:10 (wt/vol). The homogenate wascentrifuged at 2,000 g for 3 min. The pellet of mitochondria wasobtained by centrifugation of the 2,000 g supernatant at 12,500 g for10 min. At the next step, in representative experiments, mitochondriawere purified on Percoll gradient (15%:23%:40%), according topublished procedures (34). Rat brain mitochondria were suspended inice-cold solution containing 0.32 M sucrose and 10 mM Tris �HCl (pH7.4) and were additionally washed by centrifugation at 11,500 g for 10min. The protein concentrations in the stock mitochondrial suspen-sions were 25–30 mg/ml. Mitoplast and outer membranes of rat brainmitochondria were obtained according to Rice and Lindsay (31).

Isolation of mitochondria from cultured oligodendrocyte cellsOLN93. Isolation of mitochondria from oligodendrocytes was carriedout at 4°C, according to a published procedure (25) with modifica-tions. Confluent monolayers of OLN93 cells, cultured in T-75 flasks,were washed with phosphate-buffered saline, collected by scraping inbuffer, containing 0.32 M sucrose, 1 mM EDTA, 1 mM EGTA, 0.02%bovine serum albumin (fraction V), and 10 mM Tris �HCl (pH 7.4),and pelleted at 500 g for 5 min. Cells were resuspended in the sameice-cold buffer and disrupted with a Dounce homogenizer. The cellhomogenate was centrifuged twice at 500 g for 5 min to removeunbroken cells and nuclei, and twice at 3,000 g for 4 min to removetrapped peroxisomes and other organelles. The supernatant obtainedwas centrifuged at 12,000 g for 10 min. The mitochondrial pellet wassuspended in ice-cold medium containing 0.32 M sucrose and 10 mMTris �HCl (pH 7.4) and centrifuged at 12,000 g for 10 min. Theisolated mitochondria were resuspended in the same ice-cold medium.

Evaluation of mitochondrial functions. The mitochondrial mem-brane potential was measured as described earlier (4) by determining thedistribution of tetraphenylphosphonium (TPP�) in the incubation me-dium with a TPP�-selective electrode, and Ca2� transport was deter-mined with a Ca2�-sensitive electrode (Nico). Oxygen consumption ratewas detected with a Clark-type O2 electrode in the 2-ml cell volume.Mitochondria (1.0 mg protein/ml) were incubated in the medium con-taining 100 mM KCl, 100 mM sucrose, 10 mM Tris �HCl, 0.4 mM

K2HPO4, and 2 �M rotenone, pH 7.4, at 37°C. Succinate (5 mMpotassium succinate) was used as mitochondrial respiratory substrate.

PT opening in rat brain mitochondria was induced by thresholdCa2� load (each addition of Ca2� contained 80–120 nM Ca2�/mgprotein). For Ca2� loading of mitochondria isolated from OLN93cells, each addition was 70 nM Ca2�/mg protein at 0.7 mg protein/mlin the chamber. Rat liver mitochondria (1.0 mg protein/ml) wereincubated in the medium containing 2 mM K2HPO4 (instead of 0.4K2HPO4), and PT opening was induced by addition of 50 nMCa2�/mg protein two times. All experiments were performed in anopen chamber.

For measuring mitochondrial parameters, we used succinate assubstrate in the presence of rotenone that prevented oxidation ofNAD-dependent substrates. We used this condition for the followingreasons. First, under Ca2�-induced PT pore opening, strong NADHoxidation occurs that is followed by irreversible NAD� depletion (11,41). These effects could interfere with PT pore opening and preventATP-induced PT pore closing. Second, we wanted to avoid possibleeffect of 2�,3�-cAMP on the well-known regulation of NAD-depen-dent dehydrogenases in mitochondria by 3�,5�-cAMP (8, 29).

Swelling of rat liver mitochondria was measured as a change inscattering of the mitochondrial suspension absorbance at 540 nmusing Specord M-40 spectrophotometer at 30°C. Standard incubationmedium for swelling assay contained 125 mM KCl, 10 mM Tris, 2mM KH2PO4, 10 mM succinate, 0.5 �M oligomycine, and 0.5 �Mrotenone. The concentration of protein in the cuvette was 0.5 mg/ml.The swelling was initiated by addition of 60 �M Ca2�.

Detection of CNP activity in isolated mitochondria. CNP activitywas detected according to published procedures (9). Aliquots of themitochondrial suspension (rat brain mitochondria or OLN93 cells)containing 40 �g of protein were taken from the chamber undercontrol conditions and after PT opening and were then solubilized inLaemmli buffer. Solubilized mitochondria were separated by electro-phoresis in a 12.5% polyacrylamide gel and were transferred tonitrocellulose membranes. CNP activity was detected on membraneafter blotting. The membrane was preincubated for 1 h in 50 mM MES(2-(N-morpholino)ethanesulfonic acid) buffer, pH 6.1, containing 30mM MgCl2, 0.1% Triton X-100, 3 M guanidinium chloride, 1 mMEDTA, 1 mM dithiothreitol, and 5% glycerol and then for 1 h in thesame buffer without guanidinium chloride. The preincubation stage isessential for recovery of enzyme activity after SDS electrophoresis.The blot was then stained for activity by using nitro blue tetrazolium[3,3�(3,3�-dimetoxy-4,4�-diphenylene)2,2�-di-p-nitrophenyl-5,5�-diphenylditetrazolium] chloride and phenazine methosulfate to detectNADPH produced in enzyme-linked reaction for CNP. The coloredreaction product was immobilized with diluted agarose. Equal vol-umes of 1% agarose in water at 55°C and reaction medium (200 mMMES buffer, pH 6.1, containing 60 mM MgCl2, 0.2% Triton X-100,0.1 mM 2�,3�-cyclic NADP, and 4 mg/ml glucose-6-phosphate, 0.4mg /ml nitro blue tetrazolium, 0.04 mg/ml phenazine methosulphate,and 0.7 units of glucose-6-phosphate dehydrogenase) were rapidlymixed and poured on the nitrocellulose membrane. Colored reactionwas visible within 15 min and was stopped with 10% acetic acid. CNPactivity in the bands (intensity of colored band) on the blot wasevaluated by determining the optical density of the colored product.

Electrophoresis and immunoblotting of mitochondrial proteins. Forimmunoblotting, mitochondrial proteins solubilized in Laemmlibuffer were separated under denaturing conditions on 10% or 12.5%SDS-PAGE gels and transferred to nitrocellulose membranes. Preci-sion Plus Pre-stained Standards from Bio-Rad Laboratories (Hercules,CA) were used as markers. After overnight blocking, the membranewas incubated with the appropriate primary antibody. Monoclonalanti-CNP antibody was obtained, as described (30). Anti-SOD2 (Mn-superoxide dismutase 2, antibody 13533, Abcam, 1:10,000 dilution),anti-VDAC (voltage-dependent anion channel) antibody, (Ab-5, Cal-biochem, Schwalbach, Germany, 1:2,000 dilution), and �-tubulin Iantibody (Sigma, 1:20,000 dilution) were used for immunoblotting.

C14292�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

Immunoreactivity was detected using the appropriate secondary anti-body conjugated to horseradish peroxidase. Peroxidase activity wasdetected with ECL chemiluminescence reagents (Pierce Chemical).

Cell culture and transfection with small interfering RNA. OLN93cells, an oligodendrocyte cell line (32), kindly provided by C. Richter-Landsberg (University of Oldenburg, Germany), were grown at 37°Cin a 5% CO2 atmosphere in Dulbecco’s modified Eagle’s medium(Biochrom KG, Berlin, Germany) supplemented with 10% fetal calfserum (Biochrom, Berlin, Germany), containing 2 mM glutamine, 100�g/ml penicillin, and 100 �g/ml streptomycin. To knock down theendogenous CNP, OLN93 cells were transiently transfected with thechemically synthesized pool of small interfering RNAs (siRNAs)targeting CNP (Dharmacon, Chicago, IL), or with the nonsilencingsiRNA served as a scrambled control (Qiagen). Cells (50% confluent)were transfected with rat CNP siRNA by using magnet-assistedtransfection, according to the manufacturer’s protocol (IBA). CNPknockdown was assessed by Western blot at 48 h after transfection.

Statistical analysis. For statistical analysis, relative levels of pro-tein density were expressed as means � SD from at least three to fourindependent experiments with samples of mitochondria isolated atdifferent days. Several replica experiments were done with eachsample of mitochondria. The total number of observations or samplesize was 10 for rat brain mitochondria and 12 for rat liver mitochon-dria. Statistical significance was evaluated using the Student’s t-testand ANOVA with Bonferroni post hoc comparison. A value of P �0.05 was accepted as significant.

RESULTS

Effects of 2�,3�-cAMP on mitochondrial functions in ratbrain mitochondria. The effect of the CNP substrate 2�,3�-cAMP on mitochondrial functions is still unclear. We assayedthe influence of 2�,3�-cAMP on functions of rat brain mito-chondria by simultaneous registration of Ca2� flux, membranepotential change (�M), and O2 consumption, using selectiveCa2�, TPP�, and oxygen electrodes. The recordings for thethree parameters, Ca2� influx/efflux, �M measured by TPP�

distribution, and respiration, under threshold Ca2� load, arepresented separately, in Fig. 1, A–C, and average data in D. Foraverage data of control values, we obtained the followingmeans and standard deviations: Ca2� influx rates after firstCa2� pulse (Vin.1

Ca2�, 224 � 32 nmol �min1 �mg protein1); Ca2�

efflux rate (VoutCa2�

, 117 � 24 nmol �min1 �mg protein1); lagtime for Ca2� efflux (lag time, 230 � 61 s); oxygen consump-tion rate in state 2 (Vst.2

O2 , 12.9 � 3.2 ng-atom O �min1 �mgprotein1); oxygen consumption rate in state 3 after Ca2�

addition (Vst.3O2 , 53.6.9 � 4.9 ng-atom O �min1 �mg protein1);

and ATP-induced membrane repolarization, calculated asTPP� influx rate (Vin

TPP�, 0.24 � 0.04 �M/min).

Figure 1, A–C, demonstrates typical results with rat brainmitochondria in the absence and in the presence of 2�,3�-cAMP (5

Fig. 1. Effect of 2�,3�-cAMP on mitochondrial functions in rat brain mitochondria. Isolated rat brain mitochondria (1.0 mg protein/ml) were incubated in theelectrode chamber under conditions described in MATERIALS AND METHODS. Oxygen consumption rate, Ca2� concentration in the incubation medium, and changesin mitochondrial membrane potential were recorded simultaneously in an open chamber. The time scale represents the time after rat brain mitochondria wereadded to the chamber. Two additions of Ca2� (80 �M each) were applied. Arrows show the times at which CaCl2 and ATP were added. 2�,3�-cAMP (5 �M)was applied 20 s after addition of rat brain mitochondria to the incubation medium. Recordings are shown for the same samples for Ca2� fluxes (A),tetraphenylphosphonium (TPP�) concentration (membrane potential) (B), and oxygen consumption (C). In all panels, the solid trace shows control rat brainmitochondria, and the dashed trace shows rat brain mitochondria treated with 2�,3�-cAMP. D: the summary of effects of 2�,3�-cAMP on rat brain mitochondriafunctions. The following parameters of Ca2�-induced permeability transition (PT) pore opening were calculated from the curves, for which examples are givenin A–C: oxygen consumption rate in state 2 (Vst.2

O2 , ng-atom O �min1 �mg protein1), Ca2� influx rates (VinCa2�

, nmol �min1 �mg protein1), after first (Vin.1Ca2�

) andafter second calcium pulse (Vin.2

Ca2�), lag time for Ca2� efflux (lag time, s), Ca2� efflux rate (Vout

Ca2�, nmol �min1 �mg protein1), and ATP-induced membrane

repolarization, calculated as TPP� influx rate (VinTPP�

, �M/min). Values are given relative to the control value of 100 and represent the means � SD from threeindependent experiments. *P � 0.05 vs. control. A detailed description of the parameter analysis is given in our laboratory’s previous publications (4, 5). Vout

TPP�,

TPP� efflux rate.

C1430 2�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

�M). As Ca2� fluxes in Fig. 1A show, rat brain mitochondriarapidly accumulated Ca2� after the first addition of 80 nmolCa2�/mg protein. The rate of uptake (Vin.1

Ca2�) was similar in the

absence and in the presence of 2�,3�-cAMP. The second Ca2�

addition in this case exceeds the threshold Ca2� concentration.After the second Ca2� pulse, reduced Ca2� accumulation withthe following spontaneous Ca2� efflux from the rat brainmitochondria matrix was observed. In the 2�,3�-cAMP-treated rat brain mitochondria, this Ca2� uptake rate (Vin.2

Ca2�)

was decreased compared with control rat brain mitochondria.Moreover, the ability to retain Ca2� at the threshold Ca2�-

induced Ca2� release differed significantly in control and in2�,3�-cAMP-treated rat brain mitochondria. Therefore, wemeasured the lag time (see explanation in Fig. 1A). The lagtime was �4 min for control and �1.4 min for 2�,3�-cAMP-treated rat brain mitochondria. Thus, in 2�,3�-cAMP-treated ratbrain mitochondria, the Ca2� loading-induced Ca2� releasewas accelerated with a three times shorter lag time. In addition,the rate of the Ca2�-stimulated Ca2� efflux was higher in2�,3�-cAMP-treated rat brain mitochondria, as shown by theparameter Vout

Ca2�in Fig. 1. Simultaneous registration of �M in

the same preparations is shown in Fig. 1B. After the firstaddition of Ca2�, a transient depolarization of the inner mem-brane took place. After the second Ca2� addition, a sustaineddepolarization occurred. Inhibition of Ca2� uptake after thesecond Ca2� addition (see Fig. 1A) was accompanied byCa2�-induced decline in �M (see Fig. 1B) in both control and2�,3�-cAMP-treated rat brain mitochondria.

The ability of ATP to initiate repolarization of the innermembrane of rat brain mitochondria in the absence and in thepresence of 2�,3�-cAMP was also checked. In control rat brainmitochondria, we observed rapid Ca2� reaccumulation andrestoration of the membrane potential, when ATP was addedafter PT induction. However, reaccumulation of Ca2� wasslower in the presence of 2�,3�-cAMP (Fig. 1A). The rate ofATP-induced repolarization of rat brain mitochondria wasdecreased in the presence of 5 �M 2�,3�-cAMP (Fig. 1B). Thus2�,3�-cAMP was able to accelerate threshold Ca2�-inducedCa2� efflux and to slow down both ATP-induced restoration of�M and the rate of ATP-induced Ca2� influx.

The respiration of rat brain mitochondria was also measuredunder the same conditions. The results are shown in Fig. 1C.No significant changes in oxygen consumption rates wereobserved in the presence and in the absence of 2�,3�-cAMPbefore Ca2� addition. However, after the threshold Ca2� load,oxygen consumption was increased in the presence of 2�,3�-cAMP. The latter demonstrates an uncoupling-like effect,which is related to promotion of the PT development.

The bar graph in Fig. 1D gives a quantitative analysis of theparameters exemplified in the traces in Fig. 1, A–C. Thesemean data obtained from several experiments demonstrate therelative effects of the nucleotide 2�,3�-cAMP on mitochondrialfunctions (100% corresponds to control). 2�,3�-cAMP did notaffect rat brain mitochondria functions in normal conditionsbefore the threshold Ca2� load, that is, the Vst.2

O2 and the Vin.1Ca2�

.However, 2�,3�-cAMP clearly affected the rat brain mitochon-dria functions at the second Ca2� addition (threshold Ca2�

load). Vin.2Ca2�

was decreased. The time of Ca2� retention (lagtime) was decreased twofold, and the Vout

Ca2�was increased by

30%. In addition, the ATP-induced �M restoration (TPP�

reaccumulation determined by VinTPP�

) was diminished twofold

in the presence of 2�,3�-cAMP, demonstrating a lowered repo-larization ability of the inner membrane. Thus 2�,3�-cAMP wasable to promote the threshold Ca2�-induced PT pore openingand to resist the ATP-induced PT pore closing.

Efficiency of 2�,3�-cAMP action and effects of 2�,3�-cNADPon mitochondrial functions in rat brain mitochondria. Next weevaluated the effects of different concentrations of 2�,3�-cAMP,and the effect of combined application of 2�,3�-cAMP and CsA, aswell as of 2�,3�-cNADP and CsA. As shown in Fig. 2A, withaddition of CsA, a specific inhibitor of PT, no Ca2�-induced Ca2�

efflux was seen in the presence of 2�,3�-cAMP. This inhibition ofCa2� efflux by CsA supports the conclusion of participation ofPT. The concentration dependence of the efficiency of 2�,3�-cAMP to accelerate Ca2� release was also checked. The maximalefficiency of 2�,3�-cAMP was observed at a concentration of �5�M, as measured by Vout

Ca2�and lag time of Ca2� efflux. However,

a marked change was seen already at 0.5 �M 2�,3�-cAMP. Theconcentrations of 2�,3�-cAMP examined were in the rangefrom 0.1 to 50 �M (Fig. 2B). At 10 �M 2�,3�-cAMP, the VoutCa2�

reached a plateau. After further increasing the 2�,3�-cAMPconcentration up to 50 �M, a slight increase of the Vout

Ca2�and

decrease of the lag time of Ca2� efflux were observed. Thiscould be due to a superposition of an unspecific Ca2� bindingeffect of 2�,3�-cAMP.

We also checked whether 2�,3�-cNADP (5 �M) and 2�,3�-cGMP (5 �M) were able to affect the threshold Ca2�-inducedCa2� release. Figure 2 demonstrates the effects of 2�,3�-cNADP and 2�,3�-cGMP used at the same concentration as2�,3�-cAMP (5 �M). In Fig. 2C, trace 1 demonstrates Ca2�

fluxes in control rat brain mitochondria at threshold Ca2� load.In the presence of 2�,3�-cNADP, after the threshold Ca2� load,a decreased rate of Ca2� influx (Vin.2

Ca2�) and a shortened lag time

of Ca2� efflux were seen (Fig. 2C, trace 2). The lag time in2�,3�-cNADP-treated rat brain mitochondria was about twotimes shorter than in control rat brain mitochondria (comparetraces 1 and 2). No acceleration of Ca2� efflux was observedin the presence of 2�,3�-cGMP (trace 3). CsA was found toprevent also the 2�,3�-cNADP-induced acceleration of Ca2�

release in the presence of threshold Ca2� load (trace 4).The average data obtained with 5 �M 2�,3�-cNADP, 2�,3�-

cAMP, and 2�,3�-cGMP are shown in Fig. 2D, which presentslag time and Ca2� efflux under Ca2� overload (Vout

Ca2�). There

was no significant effect of 2�,3�-cGMP on Ca2� efflux in thepresence of threshold Ca2� concentrations; also, similarly,2�,3�-cCMP was not effective (data not shown).

Effects of 2�,3�-cAMP and 2�,3�-cNADP on mitochondrialfunctions in rat liver mitochondria. The paper reporting CNPactivity in rat liver mitochondria appeared already 27 yearsago, where the CNP activity was detected using 2�,3�-cNADPas substrate (12). However, so far, there are no data yet on theeffect of CNP substrates on rat liver mitochondria functions.Therefore, we next used rat liver mitochondria to study theeffect of 2�,3�-cNADP and 2�,3�-cAMP on stimulation of Ca2�

release after addition of threshold Ca2� concentration. Thishelped to evaluate the effect of the cyclic nucleotides onanother type of mitochondria, besides rat brain mitochondria.Figure 3A shows that, in control rat liver mitochondria as wellas in 2�,3�-cAMP- and 2�,3�-cNADP (5 �M)-treated rat livermitochondria, after the first and the second Ca2� addition,mitochondria were able to accumulate Ca2�, but after a lagtime (the time of calcium retention), spontaneous Ca2� efflux

C14312�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

took place. The lag time was �2.5 min for control rat livermitochondria (Fig. 3A, trace 1). 2�,3�-cAMP shortened the lagtime of Ca2� release to 1 min, while the lag time with2�,3�-cNADP was 0.5 min (Fig. 3A, traces 2 and 3). 2�,3�-cNADP was found to be more effective than 2�,3�-cAMP inaccelerating the Ca2� efflux.

Just after the second pulse of Ca2�, dissipation of �M

was observed as TPP� efflux (Fig. 3B). The rate of TPP�

efflux (VoutTPP�

) was used to measure the rate of depolarization ofthe inner membrane. This gave the following order of potency:2�,3�-cNADP-treated rat liver mitochondria � 2�,3�-cAMP-treated rat liver mitochondria � control rat liver mitochondria.This depolarization of the inner membrane of rat liver mito-chondria correlated with the corresponding rates of Ca2� effluxafter the second addition of calcium (compare Fig. 3, A and B).After the �M dissipation was almost completed, a significantincrease in Vout

Ca2�was observed that indicated PT opening. The

Ca2� efflux and �M dissipation were not observed in thepresence of CsA (data not shown). Respiratory activity of the ratliver mitochondria was increased after the second Ca2� pulse(Fig. 3C, traces 1–3) showing uncoupling-like effect.

The average data obtained with 2�,3�-cNADP and 2�,3�-cAMP (5 �M) on the parameters obtained under Ca2� over-load, that is, Vin.2

Ca2�, lag time, Vout

Ca2�, and �M decrease (Vout

TPP�)

in rat liver mitochondria, are shown in Fig. 3D. For averagedata of control values, we obtained the following means andstandard deviations: Vin.2

Ca2�, 263 � 39 nmol �min1 �mg pro-

tein1; VoutCa2�

, 164 � 35 nmol �min1 �mg protein1; lag timefor Ca2� efflux (lag time), 192 � 26 s; Vst.2

O2 , 19.9 � 3.8ng-atom O �min1 �mg protein1; Vst.3

O2 : 113.7 � 22.7 ng-atomO �min1 �mg protein1; and ATP-induced membrane repolar-ization, calculated as Vin

TPP�, 0.35 � 0.04 �M/min.

The data presented in Fig. 3D show that 2�,3�-cNADPclearly stimulated the rate of �M dissipation and shortenedthe lag time for the Ca2�-induced Ca2� efflux in Ca2�-overloaded rat liver mitochondria, while, in rat brain mitochon-dria, 2�,3�-cAMP was more effective.

Effects of 2�,3�-cAMP and 2�,3�-cNADP on mitochondrialswelling in rat brain and liver mitochondria. Induction ofmitochondrial swelling is a prominent feature of PT. Therefore,we examined the influence of 2�,3�-cAMP and 2�,3�-cNADPon Ca2�-induced swelling of isolated rat brain mitochondriaand rat liver mitochondria to corroborate our hypothesis thatthe cyclic nucleotides are involved in PT regulation. Swellingwas initiated by addition of Ca2� to mitochondria incubated instandard medium (see MATERIALS AND METHODS). The decreaseof light scattering at 540 nm that indicates swelling wasaccelerated in the presence of 5 �M 2�,3�-cAMP or 5 �M2�,3�-cNADP. The swelling was more obvious in rat livermitochondria than in rat brain mitochondria.

The data in Fig. 4 (A, typical curves, and B, average data)show that the relatively small swelling of rat brain mitochon-dria is significantly enhanced by 2�,3�-cAMP and 2�,3�-cNADP. We used the half time for reaching the maximum ofthe rat brain mitochondria swelling as characteristic parameterfor quantification.

In isolated rat liver mitochondria, the effects of 2�,3�-cAMPand 2�,3�-cNADP on swelling were more prominent, as shown inFig. 4, C and D. In the presence of 500 nM 2�,3�-cNADP (trace2) or 500 nM 2�,3�-cAMP (trace 3), the rate of swelling wasgreatly accelerated. 2�,3�-cAMP and 2�,3�-cNADP induced rapid,large amplitude swelling of rat liver mitochondria within 5 min. Inrat liver mitochondria, the cyclic nucleotides accelerated the

Fig. 2. Effect of 2�,3�-cAMP, 2�,3�-cNADP,and 2�,3�-cGMP on Ca2�-induced Ca2� effluxin rat brain mitochondria. A: effects of 2�,3�-cAMP and cyclosporin A (CsA) on Ca2�-in-duced Ca2� efflux and PT pore opening. Re-cordings represent Ca2� fluxes in the absence orpresence of 2�,3�-cAMP (5 �M), or the com-bined application of 2�,3�-cAMP and CsA (1�M). B: effect of different concentrations of2�,3�-cAMP, as indicated, on relative Vout

Ca2�and

lag time in rat brain mitochondria. C: traces ofthe effect of 2�,3�-cNADP (5 �M), in the ab-sence and in the presence of CsA (1 �M), and2�,3�-cGMP (5 �M) on Ca2� fluxes in rat brainmitochondria: trace 1, control rat brain mito-chondria; trace 2, rat brain mitochondria treatedwith 2�,3�-cNADP; trace 3, rat brain mitochon-dria treated with 2�,3�-cGMP; trace 4, rat brainmitochondria treated with CsA and 2�,3�-cNADP. Experimental conditions are as in Fig1. 2�,3�-cAMP, CsA, 2�,3�-cNADP, and 2�,3�-cGMP were applied 20 s after the addition of ratbrain mitochondria to the incubation medium.Arrows show the times at which CaCl2 (80 �Meach) was applied in the experiments. D: sum-mary of effects of 5 �M 2�,3�-cNADP, 5 �M2�,3�-cAMP, and 5 �M 2�,3�-cGMP on Ca2�-induced Vout

Ca2�and lag time in rat brain mitochon-

dria. For analysis of the parameters, see legendto Fig. 1. Values are given relative to the controlvalue of 100 and represent the means � SDfrom three independent experiments. *P � 0.05vs. control.

C1432 2�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

mitochondrial swelling �3.5 times compared with control Ca2�-induced swelling (Fig. 4D). Mitochondrial swelling elicited by2�,3�-cAMP and 2�,3�-cNADP was prevented by CsA (traces 4and 5 in Fig. 4C), confirming the participation of PT.

CNP localization in submitochondrial fractions and influ-ence of threshold Ca2� on CNP activity in rat brain mitochon-dria. Here, we also determined the enzymatic CNP activity inisolated rat brain mitochondria after SDS-PAGE and Westernblot on nitrocellulose membrane. This method is used accord-ing to the procedure published before (9). A suspension ofpurified rat brain mitochondria was added to the incubationmedium in the chamber with installed electrodes to measure PTdevelopment. From there, the samples were taken for CNPenzyme activity assay. The CNP activity was measured beforeand after Ca2� threshold loading that leads to PT development.The enzymatic activity was decreased in the presence ofthreshold Ca2� load, as shown in Fig. 5, A and B. Equal levelsof CNP in rat brain mitochondria were found both in controlcondition and after Ca2� threshold loading (Fig. 5, C and D).

We then examined the localization of CNP in fractions of ratbrain mitochondria, such as mitoplasts (the inner membranewith matrix) and outer membranes. As shown in Fig. 5E, CNPwas found in both fractions. The fractions tested were alsostained for SOD2 (marker for mitoplasts) and for VDAC(marker for outer membranes). The faint staining of mitoplastswith anti-VDAC antibody revealed small amounts of VDAC inthe mitoplast fraction. This can be explained by the presence ofcontact sites in the mitoplasts that usually contain VDAC.SOD2 was found mainly in mitoplasts, and only a negligible

trace amount of SOD2 was detected in the outer membranes.We also checked for the presence of CNP in mitochondriaisolated from several other tissues, like rat liver, heart, andpancreas, as well as mitochondria isolated from C6 gliomacells. In all mitochondria isolated from these tissues and cells,CNP was detected by immunoblotting (data not shown).

Ca2�-dependent PT in mitochondria isolated from CNPsiRNA-treated oligodendrocytes. Next, we studied the questionof whether the CNP content was related to the induction of PT.Therefore, RNA interference studies were carried out. Aftertesting the CNP expression in different types of cultured cells(HEK293 human embryonic kidney; N2A mouse neuroblas-toma, and OLN93, a rat oligodendrocyte line; Fig. 6A), wefound that endogenous expression of CNP detected with theanti-CNP-antibody was highest in the OLN93 cells. OLN93cells were chosen for transfection with siRNA targeting CNP.Nontargeting scrambled siRNA served as a control. Forty-eighthours after transfection, the level of CNP expression wasassayed by Western blot. Transfection with CNP-targetingsiRNA significantly reduced the CNP protein expression, asmeasured in total OLN93 cell lysate (Fig. 6B). Mitochondriawere isolated from OLN93 wild-type cells, as well as fromCNP knockdown OLN93 cells. Scrambled siRNA did notaffect CNP expression. In mitochondria isolated from CNPknockdown OLN93 cells, the level of CNP was reduced to�20% of control level (Fig. 6C). Reduction of CNP expressionin mitochondria correlated with decreased enzymatic CNPactivity, which was not affected in mitochondria isolated fromOLN93 cells transfected with scrambled siRNA (Fig. 6D).

Fig. 3. Effect of 2�,3�-cNADP and 2�,3�-cAMP on rat liver mitochondria functions.Experimental conditions are as in Fig. 1.2�,3�-cAMP (5 �M) and 2�,3�-cNADP (5�M) were applied 20 s after the addition ofrat liver mitochondria to the incubation me-dium. Arrows in A–C show the times atwhich CaCl2 (80 �M) was applied. Record-ings for the same samples are shown of Ca2�

fluxes (A), the membrane potential (B), andoxygen consumption (C). In all panels, trace1, control rat liver mitochondria; trace 2, ratliver mitochondria treated with 2�,3�-cAMP;trace 3, rat liver mitochondria treated with2�,3�-cNADP. D: summary of effects of 5�M 2�,3�-cNADP and 5 �M 2�,3�-cAMP onVst.2

O2 , VinCa2�

, lag time, VoutCa2�

, and rate of depo-larization Vout

TPP�(see traces in B) in rat liver

mitochondria. Lag time was significantly re-duced for 2�,3�-cNADP, but not for 2�,3�-cAMP. Values are given relative to the con-trol value of 100 and represent the meanvalues � SD from three independent exper-iments. *P � 0.05 vs. control.

C14332�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

We also investigated whether the reduction in CNP level inmitochondria was correlated with the Ca2� release ability. Wemeasured the threshold Ca2�, which is the Ca2� concentrationdetermined as the amount of added Ca2�, leading to Ca2�

efflux from the matrix of mitochondria. Figure 7A shows thatthreshold Ca2� concentration in mitochondria isolated fromwild-type OLN93 cells was achieved after four additions ofCa2� (70 nmol Ca2�/mg protein each), giving a thresholdCa2� for OLN93 mitochondria of 280 nmol/mg protein in thegiven experiment. A similar threshold was found in mitochon-dria isolated from scrambled siRNA-treated OLN93 cells (datanot shown). The calcium threshold was evaluated and ispresented on the basis of multiple Ca2� additions. For thecalculation, we used three additions (4 � 70 nmol) or fouradditions (4 � 70 nmol), and then, for the last addition byinterpolation, we determined the exact amount of Ca2� loadthat was accumulated before Ca2� efflux developed. Thatallowed presentation of the statistical data.

Mitochondria isolated from CNP knockdown OLN93 cellsreleased Ca2� after the third addition (also 70 nmol Ca2�/mgprotein each), having a lower threshold calcium concentrationequal to 210 nmol Ca2�/mg protein (Fig. 7B). The average dataon Ca2�-induced Ca2� efflux in mitochondria isolated fromwild-type and knockdown cells are shown in Fig. 7E. Nonoticeable changes in Vin.1

Ca2�were observed in all kinds of

isolated mitochondria. However, a reduced Ca2� threshold tostimulate Ca2�-induced Ca2� release was found for CNP knock-down mitochondria. These results demonstrate a correlation be-tween reduced CNP content and lowered Ca2� capacity.

Then we tested the ability of 2�,3�-cAMP to affect the Ca2�

retention in mitochondria. The effect of 2�,3�-cAMP on PTinduction in mitochondria isolated from wild-type OLN93 cellsis demonstrated by comparison of Fig. 7, A and C, where Ca2�

efflux occurred already after the third Ca2� addition. Thepresence of 2�,3�-cAMP leads to further stimulation of Ca2�

release after Ca2� overloading in CNP knockdown mitochon-dria, when we compare Fig. 7, B and D.

DISCUSSION

The physiological role of 2�,3�-cyclic nucleotides in biolog-ical systems is scarcely understood. 2�,3�-Cyclic nucleotidesand oligonucleotides containing a 2�,3�-cyclic terminus arehydrolyzed by CNP to produce 2�-nucleotides. CNP wasshown to be able to hydrolyze the terminal cyclic phosphate ofRNA without influencing the internucleotide linkages (36, 39).Besides being a major protein in the central nervous systemand highly expressed in oligodendrocytes and Schwann cells,CNP was also found outside myelin. Earlier, CNP activity wasdiscovered in rat liver mitochondria (12) and later in mitochon-dria in cultured adrenal cells (27). Nevertheless, the CNPfunction in mitochondria has not yet been determined, and theinfluence of the CNP substrates on the mitochondrial functionhas not yet been investigated.

We found CNP in mitochondria isolated from brain, liver,heart, and from C6 glioma cells. In all mitochondrial samplestested here, CNP was visualized on Western blots stained bythe highly specific monoclonal anti-CNP antibody. In our

Fig. 4. Effect of 2�,3�-cNADP and 2�,3�-cAMP on rat brain andliver mitochondria swelling. A: effect of 2�,3�-cNADP and2�,3�-cAMP on rat brain mitochondria swelling. Swelling wasinitiated by addition of 100 �M Ca2� to rat brain mitochondriaincubated in standard medium (see MATERIALS AND METHODS).Light scattering at 540 nm in rat brain mitochondria wasmeasured. Trace 1, control-calcium-induced swelling of ratbrain mitochondria; trace 2, effect of 5 �M 2�,3�-cNADP; trace3, effect of 5 �M 2�,3�-cAMP. B: average results of half-time(T1/2) for reaching the maximal swelling of rat brain mitochon-dria in the presence of 2�,3�-cNADP and 2�,3�-cAMP. C: effectof 2�,3�-cNADP, 2�,3�-cAMP, and CsA on rat liver mitochon-dria swelling. Trace 1, control-calcium-induced swelling of ratliver mitochondria; trace 2, effect of 2�,3�-cNADP (500 nM);trace 3, effect of 2�,3�-cAMP (500 nM); trace 4, effect of2�,3�-cAMP (500 nM) in the presence CsA (1 �M); and trace5, effect of 2�,3�-cNADP (500 nM) in the presence of CsA (1�M) on calcium-induced swelling of rat liver mitochondria.D: average results of T1/2 for reaching maximal Ca2�-inducedswelling of rat liver mitochondria in the presence of 2�,3�-cNADP and 2�,3�-cAMP. Values represent the means � SDfrom three independent experiments. *P � 0.05 vs. control.OD540 nm, 540-nm optical density.

C1434 2�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

experiments, we, for the first time, studied the effect of 2�,3�-cAMP and 2�,3�-cNADP and other 2�,3�-cyclic nucleotides onmitochondrial functions. We found that the cyclic nucleotidesdid not affect Ca2� transport and other functions in mitochon-dria having the PT pore closed, before threshold Ca2� loading.However, 2�,3�-cAMP and 2�,3�-cNADP enhanced PT devel-opment, as seen on Ca2� transport, membrane potential dissi-pation, and swelling of both rat brain and liver mitochondria inthe presence of threshold Ca2� concentrations. The cyclicnucleotides were found to be effective in the low micromolarconcentration range. The effects of 2�,3�-cAMP on PT devel-opment were observed even at 0.5 �M, and maximal efficiencywas seen at 5 �M. 2�,3�-Cyclic NADP is also a substrate for2�,3�-cyclic nucleotide 3�-phosphohydrolase (20). 2�,3�-cNADP was able to shorten the lag phase before Ca2� effluxand to increase the rate of Ca2� release from both kinds ofmitochondria, rat brain and liver mitochondria, when they wereloaded with threshold Ca2� concentrations. 2�,3�-cGMP (Fig.2D) and 2�,3�-cCMP (not shown) were not effective in induc-tion of threshold Ca2�-induced Ca2� efflux. The potencysequence in rat brain mitochondria is as follows: 2�,3�-cAMP � 2�,3�-cNADP �� 2�,3�-cGMP, 2�,3�-cCMP. Inter-estingly, the weak capacity of 2�,3�-cCMP to activate Ca2�-induced Ca2� release in Ca2� overloaded mitochondria is in

parallel with the earlier reported lower capacity of CNP tohydrolyze 2�,3�-cCMP compared with the cAMP analog (13).

The 2�,3�-cAMP- and 2�,3�-cNADP-induced stimulation ofCa2� efflux from mitochondria and collapse of �M were CsAsensitive, confirming involvement of PT in the process. Mito-chondrial PT is manifested as a sudden opening of nonselectivemegachannel (pore) in the inner mitochondrial membrane,which increases the permeability to solutes with molecularmass up to 1,500 Da in response to mitochondrial Ca2�

overload and/or oxidative stress (7, 18). PT is a complexprocess with many known inducers, modulators, and inhibitors.The Ca2�-induced, CsA-sensitive PT system is a multiproteincomplex formed in the contact sites between outer and innermembranes of mitochondria to increase permeability of theinner membrane. Main events of PT opening are collapse of themembrane potential, Ca2�-induced Ca2� efflux, induction ofmitochondrial swelling, and release of apoptotic factors (7, 19,43). We observed that even 500 nM 2�,3�-cAMP or 2�,3�-cNADP was able to accelerate the Ca2�-stimulated, large-amplitude swelling of liver mitochondria that was prevented byCsA. Moreover, the CNP level was not changed in Ca2�-loaded mitochondria, but the enzymatic CNP activity wasdecreased after Ca2� threshold loading, which prevented hy-drolysis of 2�,3�-cAMP. Consequently, this increases the effi-

Fig. 5. Detection of 2�,3�-cyclic nucleotide 3�-phosphodiesterase (CNP) activity in rat brainmitochondria and CNP localization in submito-chondrial fractions. Detection of CNP activity(see MATERIALS AND METHODS) in rat brain mi-tochondria is shown under control conditionsand in the presence of threshold Ca2� in themedium. A: detection of CNP activity. B: rela-tive level of CNP activity under control condi-tions and in rat brain mitochondria loaded withthreshold Ca2�. OD in control was taken as100%. C: detection of CNP level in the samesamples in control (in the absence of Ca2� in themedium) or in the presence of threshold Ca2�

concentration. D: relative levels of CNP undercontrol conditions and in Ca2� overloaded ratbrain mitochondria. Values represent means � SDfrom three independent experiments. *P � 0.05.E: CNP localization in submitochondrial fractions.Isolated rat brain mitochondria were subfraction-ated, as described in MATERIALS AND METHODS.Isolated outer membranes and mitoplasts wereseparated by electrophoresis in 12.5% SDS-PAGE(20 �g of protein/lane) with following Westernblot. The membrane was treated with anti-CNPantibody. The same membranes were stained withanti-voltage-dependent anion channel (VDAC)and anti-SOD2 (Mn-superoxide dismutase 2) anti-bodies, respectively. Experiments were repeatedthree times with similar results. IB, immunoblot.

C14352�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

ciency of the action of 2�,3�-cAMP in a feedback cycle.Therefore, we propose that, in living cells, inhibition of CNPactivity under threshold Ca2� accumulation in mitochondriacontributes to elevating the 2�,3�-cAMP level. 2�,3�-cAMP, inturn, seems to work as a second messenger by promoting themitochondrial Ca2� efflux and �M dissipation.

Functional importance of CNP in mitochondria was obtainedby RNA interference experiments in OLN93 cells, whichcontain a high endogenous level of CNP. OLN93 cells weretransfected with siRNA targeting CNP. As a result, the endog-enous CNP protein expression level was reduced. Mitochon-dria isolated from CNP knockdown OLN93 cells possessedreduced level of CNP and decreased enzymatic CNP activitycompared with mitochondria isolated from wild-type OLN93cells. Mitochondria isolated from scrambled nontargetingsiRNA OLN93 cells served as a control. Lowered level of CNPand reduced CNP activity correlated with facilitation of activationof Ca2� efflux from mitochondria. That process was furtherincreased in the presence of added 2�,3�-cAMP. These resultsallow us to suppose that the CNP level and activity in mitochon-dria are important for the regulation of PT development.

Localization of CNP in the inner and outer membranes of ratbrain mitochondria was found, which is in agreement with the

distribution of CNP activity in liver mitochondria (12). SuchCNP localization in mitochondria indicates its possible concen-tration in contact sites, similar to the PT pore localization. It isdifficult to determine which components of the PT pore can betargets for CNP and its substrates, since the exact protein com-position of the pore is not established so far. The PT pore complexwas previously considered as a complex formed by the associationof adenine nucleotide translocase of the inner membrane andcyclophilin D of the matrix, the VDAC and peripheral benzodi-azepine receptor localized in the outer membrane, and additionalproteins, such as proteins of the Bcl-2 family, creatine kinase, andhexokinase. Recent experiments with mitochondria from adeninenucleotide translocator- and VDAC-deficient mice demonstratedthat both proteins are not indispensable structural elements of theunselective pore, but they could still be considered as regulators ormodulators of PT (6, 21). Since facilitation of PT development byCNP substrates in the presence of threshold Ca2� concentrationswas observed, and reduced CNP activity correlated with PTinitiation, a regulatory role of CNP in PT pore opening might besuggested.

An important question is still, what are the sources of theCNP substrates, 2�,3�-cAMP, and other cyclic nucleotides inmitochondria? Oligonucleotides containing a 2�,3�-cyclic ter-

Fig. 6. CNP level and enzymatic activity inoligodendrocyte OLN93 cells and OLN93mitochondria. A: endogeneous CNP proteinexpression in different types of cultured cellswas evaluated by Western blotting, as de-scribed in MATERIALS AND METHODS. Experi-ments were repeated three times with similarresults. Twenty-five micrograms of total celllysates were loaded. B: OLN93 cells weretransfected with either 50 nM scrambled smallinterfering RNA (siRNA) (control siRNA) orCNP siRNAs. At 48 h after transfection, CNPin total cell lysates was determined by Westernblot analysis. Representative data from threeindependent experiments are given, and �-tu-bulin I served as control for specificity ofknockdown. C: OLN93 cells were transfectedwith either CNP siRNAs (50 nM) or scram-bled siRNA (50 nM); after 48 h, mitochondriawere isolated. The reduction of CNP expres-sion in mitochondria was evaluated by West-ern blotting. CNP and SOD2 protein bandswere quantified by densitometry. The histo-gram shows relative units as ratio of CNP toSOD2. Mitochondria isolated from nontrans-fected OLN93 cells were used as reference andassigned the value of 100%. Values representthe means � SD from three independent ex-periments. *P � 0.05. D: CNP activity inmitochondria isolated from OLN93 cells,transfected with siRNA or nontransfected,was detected, as described in MATERIALS AND

METHODS. The activity of CNP in mitochon-dria isolated from nontransfected OLN93cells was taken as 100%. Values representthe means � SD from three independentexperiments. *P � 0.05.

C1436 2�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

minus could be generated as 2�,3�-cyclic intermediate in theenzymatic degradation of RNA, as well as in processing andsplicing reactions for mammalian RNA. In this relation, it isinteresting to mention that RNA could be a substrate for themitochondrial unspecific endonuclease G, which is releasedfrom the inner mitochondrial membrane during the early stageof apoptosis (26). RNA may be a natural substrate for CNP(37). Moreover, recently it was found that mammalian CNP

could function as a tRNA splicing enzyme (33). In addition, itwas reported that CNP is an RNA binding protein that inhibitsprotein synthesis (17).

On the other hand, cAMP can be synthesized in mitochon-dria by adenylyl cyclases. The bicarbonate-regulated, Ca2�-dependent soluble adenylyl cyclase was detected in isolatedmitochondria (42). Although the substrate selectivity of themitochondrial soluble adenylyl cyclase is not completely clear,

Fig. 7. Induction of PT in mitochondria isolated from CNP siRNA-transfected OLN93 cells. Mitochondria isolated from wild-type (wt) OLN93 cells, from scrambledsiRNA-transfected OLN93 cells, and from CNP siRNA-transfected OLN93 cells were incubated in the open electrode chamber, as described in MATERIALS AND METHODS.Ca2� concentration in the incubation medium and changes in mitochondrial membrane potential were recorded. The time scale represents the time after OLN93mitochondria were added to the chamber. Arrows show the times at which Ca2� was added. 2�,3�-cAMP (5 �M) was applied 20 s after addition of mitochondria to theincubation medium. A: mitochondria isolated from wt OLN93 cells (control). B: mitochondria isolated from CNP siRNA-transfected OLN93 cells (control condition).C: mitochondria isolated from wt OLN93 cells in the presence of 5 �M 2�,3�-cAMP. D: mitochondria isolated from CNP siRNA-transfected OLN93 cells in the presenceof 5 �M 2�,3�-cAMP. E: summary of effects of reduction of CNP expression on the following Ca2� parameters: Vin.1

Ca2�after the first Ca2� addition, Ca2� threshold

(amount of Ca2� inducing Ca2� efflux). Values represent the means � SD from three independent experiments. The levels of functional parameters ofmitochondria isolated from wt OLN93 cells loaded with threshold Ca2� were taken as 1.0 (control). *P � 0.05 vs. control.

C14372�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

we can suppose that, at the conditions of our experiments,under threshold Ca2� load, the soluble adenylyl cyclase wasactivated by Ca2� and 3�,5�-cAMP was produced. 2�,3�-cAMPcould be created by transformation of the 3�,5�-cyclic terminusto the 2�,3�-cyclic one that can occur in alkaline conditions.This conversion was shown for the 3�,5�-cNADP to 2�,3�-cNADP conversion, with 2�,3�-cyclic NADP as substrate for2�,3�-cyclic nucleotide 3�-phosphodiesterase (35). It is wellknown that highly alkaline conditions are prevalent in mito-chondria under threshold Ca2� load.

In summary, the results obtained here suggest that 2�,3�-cyclic nucleotides might be able to modulate PT in mitochon-dria, a key process in initiation of apoptotic cell death. Inparticular, 2�,3�-cAMP might probably act as a second mes-senger. In addition, the intramitochondrial 2�,3�-cAMP levelmight be elevated due to inhibition of CNP activity underthreshold Ca2� load.

GRANTS

This study was supported by Russian Foundation for Basic Research GrantsN06-04-48763 and N08-04-00723, and by Bundesministerium fur Bildung undForschung grant 01ZZ0107, and grants RUS 04/004 and RUS 08/002.

REFERENCES

1. Agrawal HC, Sprinkle TJ, Agrawal D. 2�,3�-Cyclic nucleotide-3�-phosphodiesterase in the central nervous system is fatty-acylated bythioester linkage. J Biol Chem 265: 11849–11853, 1990.

2. Agrawal HC, Sprinkle TJ, Agrawal D. 2�,3�Cyclic nucleotide-3�-phos-phodiesterase in peripheral nerve myelin is phosphorylated by a phorbolester-sensitive protein kinase. Biochem Biophys Res Commun 170: 817–823, 1990.

3. Allen FW, Davis FF. A specific phosphodiesterase from beef pancreas.Biochim Biophys Acta 21: 14–17, 1956.

4. Azarashvili T, Grachev D, Krestinina O, Evtodienko Y, Yurkov I,Papadopoulos V, Reiser G. The peripheral-type benzodiazepine receptoris involved in control of Ca2�-induced permeability transition pore open-ing in rat brain mitochondria. Cell Calcium 42: 27–39, 2007.

5. Azarashvili T, Krestinina O, Odinokova I, Evtodienko Y, Reiser G.Physiological Ca2� level and Ca2�-induced permeability transition porecontrol protein phosphorylation in rat brain mitochondria. Cell Calcium34: 253–259, 2003.

6. Baines CP, Kaiser RA, Sheiko T, Craigen WJ, Molkentin JD. Voltage-dependent anion channels are dispensable for mitochondrial-dependentcell death. Nat Cell Biol 9: 550–555, 2007.

7. Bernardi P. Mitochondrial transport of cations: channels, exchangers, andpermeability transition. Physiol Rev 79: 1127–1155, 1999.

8. Boneh A. Regulation of mitochondrial oxidative phosphorylation bysecond messenger-mediated signal transduction mechanisms. Cell MolLife Sci 63: 1236–1248, 2006.

9. Bradbury JM, Thompson RJ. Photoaffinity labelling of central-nervous-system myelin. Evidence for an endogenous type I cyclic AMP-dependentkinase phosphorylating the larger subunit of 2�,3�-cyclic nucleotide 3�-phosphodiesterase. Biochem J 221: 361–368, 1984.

10. De Angelis DA, Braun PE. Isoprenylation of brain 2�,3�-cyclic nucleotide3�-phosphodiesterase modulates cell morphology. J Neurosci Res 39:386–397, 1994.

11. Di Lisa F, Menabo R, Canton M, Barile M, Bernardi P. Opening of themitochondrial permeability transition pore causes depletion of mitochon-drial and cytosolic NAD� and is a causative event in the death ofmyocytes in postischemic reperfusion of the heart. J Biol Chem 276:2571–2575, 2001.

12. Dreiling CE, Schilling RJ, Reitz RC. 2�,3�-Cyclic nucleotide 3�-phos-phohydrolase in rat liver mitochondrial membranes. Biochim Biophys Acta640: 114–120, 1981.

13. Drummond GI, Iyer NT, Keith J. Hydrolysis of ribonucleoside 2�,3�-cyclicphosphates by a diesterase from brain. J Biol Chem 237: 3535–3539, 1962.

14. Giulian D, Iwanij V, Dean G, Drummond RJ. Localization of 2�,3�-cyclic nucleotide-3�-phosphohydrolase within the vertebrate retina. BrainRes 265: 217–225, 1983.

15. Giulian D, Moore S. Identification of 2�:3�-cyclic nucleotide 3�-phospho-diesterase in the vertebrate retina. J Biol Chem 255: 5993–5995, 1980.

16. Gravel M, DeAngelis D, Braun PE. Molecular cloning and character-ization of rat brain 2�,3�-cyclic nucleotide 3�-phosphodiesterase isoform 2.J Neurosci Res 38: 243–247, 1994.

17. Gravel M, Robert F, Kottis V, Gallouzi IE, Pelletier J, Braun PE.2�,3�-Cyclic nucleotide 3�-phosphodiesterase: a novel RNA-binding pro-tein that inhibits protein synthesis. J Neurosci Res 87: 1069–1079, 2009.

18. Halestrap AP. Calcium, mitochondria and reperfusion injury: a pore wayto die. Biochem Soc Trans 34: 232–237, 2006.

19. Halestrap AP, McStay GP, Clarke SJ. The permeability transition porecomplex: another view. Biochimie 84: 153–166, 2002.

20. Heaton PA, Eckstein F. Diastereomeric specificity of 2�,3�-cyclic nucle-otide 3�-phosphodiesterase. Nucleic Acids Res 24: 850–853, 1996.

21. Juhaszova M, Wang S, Zorov DB, Nuss HB, Gleichmann M, MattsonMP, Sollott SJ. The identity and regulation of the mitochondrial perme-ability transition pore: where the known meets the unknown. Ann N YAcad Sci 1123: 197–212, 2008.

22. Keppetipola N, Shuman S. Characterization of the 2�,3� cyclic phospho-diesterase activities of Clostridium thermocellum polynucleotide kinase-phosphatase and bacteriophage lambda phosphatase. Nucleic Acids Res35: 7721–7732, 2007.

23. Kurihara T, Nishizawa Y, Takahashi Y, Odani S. Chemical, immuno-logical and catalytic properties of 2�:3�-cyclic nucleotide 3�-phosphodies-terase purified from brain white matter. Biochem J 195: 153–157, 1981.

24. Lee J, O�Neill RC, Park MW, Gravel M, Braun PE. Mitochondriallocalization of CNP2 is regulated by phosphorylation of the N-terminaltargeting signal by PKC: implications of a mitochondrial function forCNP2 in glial and non-glial cells. Mol Cell Neurosci 31: 446–462, 2006.

25. Leissring MA, Farris W, Wu X, Christodoulou DC, Haigis MC,Guarente L, Selkoe DJ. Alternative translation initiation generates anovel isoform of insulin-degrading enzyme targeted to mitochondria.Biochem J 383: 439–446, 2004.

26. Low RL. Mitochondrial endonuclease G function in apoptosis andmtDNA metabolism: a historical perspective. Mitochondrion 2: 225–236,2003.

27. McFerran B, Burgoyne R. 2�,3�-Cyclic nucleotide 3�-phosphodiesteraseis associated with mitochondria in diverse adrenal cell types. J Cell Sci110: 2979–2985, 1997.

28. O�Neill RC, Minuk J, Cox ME, Braun PE, Gravel M. CNP2 mRNAdirects synthesis of both CNP1 and CNP2 polypeptides. J Neurosci Res50: 248–257, 1997.

29. Papa S, De Rasmo D, Scacco S, Signorile A, Technikova-Dobrova Z,Palmisano G, Sardanelli AM, Papa F, Panelli D, Scaringi R, Santer-amo A. Mammalian complex I: a regulable and vulnerable pacemaker inmitochondrial respiratory function. Biochim Biophys Acta 1777: 719–728,2008.

30. Reiser G, Kunzelmann U, Steinhilber G, Binmoller FJ. Generation ofa monoclonal antibody against the myelin protein CNP (2�,3�-cyclicnucleotide 3�-phosphodiesterase) suitable for biochemical and for immu-nohistochemical investigations of CNP. Neurochem Res 19: 1479–1485,1994.

31. Rice JE, Lindsay JG. Subcellular fractionation of mitochondria. In:Subcellular Fractionation: A Practical Approach, edited by Graham JMand Rickwood D. Oxford, UK: IRL, 1997, p. 107–142.

32. Richter-Landsberg C, Heinrich M. OLN-93: a new permanent oligo-dendroglia cell line derived from primary rat brain glial cultures. J Neu-rosci Res 45: 161–173, 1996.

33. Schwer B, Aronova A, Ramirez A, Braun P, Shuman S. Mammalian2�,3� cyclic nucleotide phosphodiesterase (CNP) can function as a tRNAsplicing enzyme in vivo. RNA 14: 204–210, 2008.

34. Sims NR. Rapid isolation of metabolically active mitochondria from ratbrain and subregions using Percoll density gradient centrifugation. J Neu-rochem 55: 698–707, 1990.

35. Sogin DC. 2�,3�-Cyclic NADP as a substrate for 2�,3�-cyclic nucleotide3�-phosphohydrolase. J Neurochem 27: 1333–1337, 1976.

36. Sprinkle TJ. 2�,3�-Cyclic nucleotide 3�-phosphodiesterase, an oligoden-drocyte-Schwann cell and myelin-associated enzyme of the nervous sys-tem. Crit Rev Neurobiol 4: 235–301, 1989.

37. Sprinkle TJ, Tippins RB, Kestler DP. Inhibition of bovine and humanbrain 2�:3�-cyclic nucleotide 3�-phosphodiesterase by heparin and polyri-bonucleotides and evidence for an associated 5�-polynucleotide kinaseactivity. Biochem Biophys Res Commun 145: 686–691, 1987.

C1438 2�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from

38. Stricker R, Lottspeich F, Reiser G. The myelin protein CNP (2�,3�-cyclic nucleotide 3�-phosphodiesterase): immunoaffinity purification ofCNP from pig and rat brain using a monoclonal antibody and phosphor-ylation of CNP by cyclic nucleotide-dependent protein kinases. Biol ChemHoppe Seyler 375: 205–209, 1994.

39. Vogel US, Thompson RJ. Molecular structure, localization, and possiblefunctions of the myelin-associated enzyme 2�,3�-cyclic nucleotide 3�-phosphodiesterase. J Neurochem 50: 1667–1677, 1988.

40. Weissbarth S, Maker HS, Raes I, Brannan TS, Lapin EP, Lehrer GM.The activity of 2�,3�-cyclic nucleotide 3�-phosphodiesterase in rat tissues.J Neurochem 37: 677–680, 1981.

41. Zago EB, Castilho RF, Vercesi AE. The redox state of endogenouspyridine nucleotides can determine both the degree of mitochondrialoxidative stress and the solute selectivity of the permeability transitionpore. FEBS Lett 478: 29–33, 2000.

42. Zippin JH, Chen Y, Nahirney P, Kamenetsky M, Wuttke MS, Fis-chman DA, Levin LR, Buck J. Compartmentalization of bicarbonate-sensitive adenylyl cyclase in distinct signaling microdomains. FASEB J17: 82–84, 2003.

43. Zoratti M, Szabo I, De Marchi U. Mitochondrial permeability transi-tions: how many doors to the house? Biochim Biophys Acta 1706: 40–52,2005.

C14392�,3�-CYCLIC NUCLEOTIDES, MITOCHONDRIA Ca2� RELEASE

AJP-Cell Physiol • VOL 296 • JUNE 2009 • www.ajpcell.org

on Decem

ber 14, 2009 ajpcell.physiology.org

Dow

nloaded from