Pim-1 preserves mitochondrial morphology byinhibiting dynamin-related protein 1 translocationShabana Dina, Matthew Masona, Mirko Völkersa, Bevan Johnsona, Christopher T. Cottagea, Zeping Wangb,Anya Y. Joyoa, Pearl Quijadaa, Peter Erhardtc, Nancy S. Magnusonb, Mathias H. Konstandina, and Mark A. Sussmana,1

aSan Diego State Heart Institute, San Diego State University, San Diego, CA 92182; bSchool of Molecular Biosciences, Washington State University, Pullman,WA 99164; and cBoston Biomedical Research Institute, Watertown, MA 02472

Edited by Eric N. Olson, University of Texas Southwestern Medical Center, Dallas, TX, and approved March 1, 2013 (received for review July 31, 2012)

Mitochondrial morphological dynamics affect the outcome ofischemic heart damage and pathogenesis. Recently, mitochon-drial fission protein dynamin-related protein 1 (Drp1) has beenidentified as a mediator of mitochondrial morphological changesand cell death during cardiac ischemic injury. In this study, we re-port a unique relationship between Pim-1 activity and Drp1 regu-lation of mitochondrial morphology in cardiomyocytes challengedby ischemic stress. Transgenic hearts overexpressing cardiac Pim-1display reduction of total Drp1 protein levels, increased phosphor-ylation of Drp1-S637, and inhibition of Drp1 localization to the mi-tochondria. Consistent with these findings, adenoviral-inducedPim-1 neonatal rat cardiomyocytes (NRCMs) retain a reticular mito-chondrial phenotype after simulated ischemia (sI) and decreasedDrp1 mitochondrial sequestration. Interestingly, adenovirus Pim-dominant negative NRCMs show increased expression of Bcl-2homology 3 (BH3)-only protein p53 up-regulated modulator of ap-optosis (PUMA), which has been previously shown to induce Drp1accumulation at mitochondria and increase sensitivity to apoptoticstimuli. Overexpression of the p53 up-regulated modulator of apo-ptosis–dominant negative adenovirus attenuates localization ofDrp1 to mitochondria in adenovirus Pim-dominant negative NRCMspromotes reticular mitochondrial morphology and inhibits cell deathduring sI. Therefore, Pim-1 activity prevents Drp1 compartmentali-zation to the mitochondria and preserves reticular mitochondrialmorphology in response to sI.

Ischemic heart disease is a leading cause of death in the Westernworld. Many simultaneous cellular changes occur during cardiac

ischemia and inevitably cause cell death. Mitochondria are es-pecially sensitive to ischemic insult and trigger the initiation ofapoptotic signaling (1). Recently, mitochondrial morphologicaldynamics have been reported to play a role in ischemia/reperfusioninjury (2, 3). Mitochondria constantly undergo either fusion or fis-sion and present either a reticular elongated or punctate frag-mented phenotype, respectively (4, 5). Prolonged simulatedischemic injury causes extensive fragmented mitochondrialmorphology associated with dynamin-related protein 1 (Drp1)translocation to the mitochondria.In mammalian cells, the profission protein Drp1 cycles be-

tween the cytosol and scission events on mitochondria, where itinteracts with accessory proteins to couple GTP hydrolysis andouter mitochondrial membrane fission (6). In addition to steady-state activity, Drp1 is also associated with early apoptosis (4, 7–10). Cellular stress increases Drp1 shuttling, which leads tofragmented mitochondrial networks, whereas prolonged or se-vere insult promotes stabilization of Drp1 to the mitochondria,which increases the likelihood of progression to apoptotic celldeath (7, 8, 11–13). Phosphorylation of Drp1-S637 prevents trans-location to mitochondria, whereas overexpression of Drp1K38A

(a dominant-negative mutation lacking GTPase activity) preventsDrp1 translocation, attenuates a fragmented mitochondrial phe-notype, and decreases cell death (4, 14). Collectively, theseobservations suggest that Drp1 phosphorylation increases cell vi-ability via inhibition of mitochondrial fission.

The serine/threonine kinase Pim-1 is a critical downstream ef-fector of Akt-mediated cardioprotective signaling (15–18). Pim-1is antiapoptotic and proproliferative, and recent findings indicatethat it also exerts beneficial effects in part by preserving mito-chondrial integrity (15). Activation of Pim-1 in vitro and in mousemodels of hypertrophy, infarction, and ischemia/reperfusion in-jury enhances cardiomyocyte survival through inhibition of in-trinsic mitochondrial apoptotic pathways. The possibility thatPim-1 acts on Drp1 localization and activity has never been in-vestigated in the context of ischemic cardiac injury. In this study,Pim-1 is shown to maintain mitochondrial networks by seques-tration of Drp1 from the mitochondria after ischemic injury.

ResultsDrp1 Localizes to the Mitochondria After Simulated Ischemia. Neo-natal rat cardiomyocytes (NRCMs) displayed 70% reticular mi-tochondrial morphology in full media (Fig. 1A). A time course ofNRCMs treated with simulated ischemia (sI) demonstrated thatmitochondrial fragmentation occurs quickly in response to an is-chemic environment, transitioning from ∼30% of the cells dis-playing a fragmented phenotype to ∼90% by 30 min sI (Fig. 1A).Fragmented mitochondrial morphology has been previously as-sociated with sI, and DRP1K38A was able to prevent this phe-nomenon, but cellular localization of endogenous Drp1 during sIwas not evaluated (3). Fragmentedmitochondrial morphology wasconsistent with Drp1 mitochondrial localization (Fig. 1B), whichpersisted during simulated ischemia reperfusion (sI/R; Fig. 1C). Toconfirm mitochondrial fragmentation was a result of Drp1 mito-chondrial compartmentalization, Drp1 was inhibited by mdivi-1, aselective pharmacological inhibitor of Drp1 activity. Incubation ofNRCMswith 50 μmol/L inhibitor mitochondrial division inhibitor-1 (mdivi-1) led to the formation of enlarged, globularmitochondriaconsistent with reports of Drp1 down-regulation (Fig. S1A). Re-ticular mitochondria morphology was retained at a 2.5-fold higherlevel (P < 0.01) when NRCMs were treated with mdivi-1 duringsI compared with DMSO-treated controls (Fig. S1B), confirmingthat mitochondrial localization of Drp1 causes mitochondrialfragmentation. Prolonged Drp1 mitochondrial association is im-plicated in cell death, and sI/R injury in NRCMs caused a 3.54-foldincrease in TUNEL-positive cells (Fig. S1C). However, NRCMsincubated with mdivi-1 prevented sI/R-related cell death (Fig. S1C).Corroborating in vitro results of Drp1 localization after ische-

mic injury, in vivo ischemia/reperfusion injury resulted in Drp1translocation to the mitochondria after 50 min ischemia and15 min reperfusion by 2.3-fold (P < 0.01; Fig. 1D). To augment

Author contributions: S.D., M.M., M.V., M.H.K., and M.A.S. designed research; S.D., M.M.,M.V., B.J., C.T.C., Z.W., A.Y.J., P.Q., N.S.M., and M.H.K. performed research; P.E. contrib-uted new reagents/analytic tools; S.D., M.M., M.V., M.H.K., and M.A.S. analyzed data; andS.D. and M.A.S. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: heartman4ever@icloud.com.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1213294110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1213294110 PNAS | April 9, 2013 | vol. 110 | no. 15 | 5969–5974

CELL

BIOLO

GY

these results by showing that mdivi-1 is effective in protecting theadult heart, age- and sex-matched mice were subjected to 50 minleft anterior descending artery (LAD) ligation followed by 24 hreperfusion with and without mdivi-1 treatment. Total area at riskwas similar between control and treatment groups, at 59% and56%, respectively (Fig. S1F). Phthalocyanine-blue dye and 2,3,5-triphenyltetrazolium chloride (TTC) staining indicated a smallerinfarct size per area at risk in mice injected with mdivi-1, 0.61 ±0.04 DMSO versus 0.43 ± 0.03 (P < 0.001) mdivi-1, showing thattotal infarct size decreased from 36% (DMSO) to 24% in mdivi-1-treated hearts (Fig. S1 D and E).

Drp1 Phosphorylation Is Regulated by Pim-1. Previous studies haveshown that phosphorylation of Drp1-S637 reduces mitochondrialfragmentation, Drp1 accumulation at the mitochondria, and apo-ptotic cell death (19, 20). Because we have established that Pim-1kinase preserves mitochondrial integrity and function (15), weinvestigated whether Pim-1 expression could affect pDrp1-S637

levels. Hearts isolated from mice overexpressing cardiac Pim-1(PimWT mice) displayed elevated levels of Drp1-S637 phosphor-ylation by 2.7-fold (P < 0.01; Fig. 2 A and B), and total Drp1protein levels were reduced by 50% (P < 0.01) compared withnontransgenic controls (Fig. 2C). In contrast, Pim-1-dominantnegative (PDN) mice displayed a twofold decrease (P < 0.05) inphospho-Drp1 and a 1.63-fold increase (P < 0.05) in total Drp1protein levels, suggesting that Pim-1 may influence Drp1 phos-phorylation and total protein levels (Fig. 2 D–F). Furthermore,

using fractionated heart lysates from PDN mice, immunoblotanalysis revealed a 2.2-fold increase (P < 0.05) of Drp1 associatedat the mitochondria compared with control (Fig. 2 G and H).

Drp1 Translocation Is Regulated by Pim-1. Pim-1 preserves mito-chondrial integrity, whereasmitochondria isolated fromPDNmiceshow disrupted structural integrity upon calcium-induced swelling(15). Therefore, we investigated the effect of Pim-1 on Drp1 lo-calization. NRCMs transduced with an EGFP-Pim-1 (adPimWT)adenovirus were exposed to sI followed by subcellular fraction-ation. Immunoblot analysis revealed a twofold (P < 0.01) decreasein Drp1 expression in the mitochondrial compartment (Fig. 3A),whereas total amounts of Drp1 protein decreased by 40% (P <0.05) in Pim-1-infectedNRCMs (Fig. S2A). Furthermore, NRCMsoverexpressing a kinase dead Pim-1 adenovirus, EGFP-PDN(adPDN), illustrate a 1.7-fold (P < 0.05) increase in Drp1 accu-mulation in themitochondria fractionwithout being subjected to sI(Fig. 3B). Whole-cell Drp1 protein levels were increased by 1.8-fold (P < 0.01; Fig. S2B) in adPDN-treated NRCMs comparedwith EGFP, corroborating the in vivo results that loss of Pim-1 up-regulates Drp1 protein levels.Differences in Drp1 localization resulting from Pim-1 expres-

sion and sI were assessed in cytosolic and mitochondrial sub-cellular fractions. Drp1 expression was decreased at baselineconditions in adPimWT NRCM cytosol, whereas the mitochon-drial fraction exhibited similar amounts of Drp1 (Fig. S2C). Fur-thermore, Drp1mitochondrial translocation upon sI was inhibited

A

B C

D

Fig. 1. Drp1 Localizes to the mitochondria after sIin vitro. (A) Tom20-labeled NRCMs showing thetransition from reticular to punctate mitochondriaphenotypes after either sI or sI/R insults. Quantifi-cation of mitochondrial phenotype in NRCMs sub-jected to sI or sI/R. FM, full media; GF, glucose-freemedia; *, significant compared with 30 min GFcontrol; #, not significant versus sI 30. (B) Immuno-blot of Drp1 after 30 min of sI versus control, fullmedia M199 with 10% FBS NRCMs. Ctrl, control; **,significant compared with control. (C) Immunoblotof Drp1 localization in NRCMs after sI/R (30 min/30min) versus 30 min sI. (D) Immunoblot of Drp1 lo-calization in adult murine hearts subjected toa sham LAD procedure, 50 min LAD ligation, or 50min/15 (120) min LAD ligation/reperfusion. **, sig-nificant compared with sham. *P < 0.05; **P < 0.01;***P < 0.001.

5970 | www.pnas.org/cgi/doi/10.1073/pnas.1213294110 Din et al.

in adPimWT NRCMs relative to control samples (Fig. S2C).Conversely, adPDN NRCMs displayed an increase in Drp1 ex-pression within the cytosolic and mitochondrial fractions atbaseline (Fig. S2D). Drp1 shuttled into the mitochondria after sItreatment in both control as well as adPDNNRCMs, resulting insimilar Drp1 expression in the mitochondrial fraction (Fig.S2D). Cycloheximide chase analysis demonstrated that Pim-1does not alter Drp1 half-life (16 h) compared with EGFP controlin NRCMs, demonstrating that Pim-1 does not target Drp1degradation (Fig. S3). Proteasome inhibition by MG132 causedan accumulation of Drp1 at 16 h in the presence of cyclohexi-mide, validating Drp1 degradation through the proteasome(Fig. S3).To substantiate that Pim-1 prevents Drp1 translocation from

the cytosol, mitochondrial morphology was assessed after sI onNRCMs transduced with adEGFP or adPimWT adenovirus.Before simulated ischemia, EGFP and PimWT NRCMs dis-played 80% and 77% reticular mitochondrial morphology, re-spectively (Fig. 3C and Fig. S4A). However, after 15 min of sItreatment, 22% reticular mitochondria remained in EGFPNRCMs, whereas PimWT NRCMs retained 65% of their re-ticular morphology (P < 0.01; Fig. 3C). At 30 min of sI, PimWTmaintain 40% reticular mitochondria compared with 11% forEGFP (P < 0.05; Fig. 3C and Fig. S4A). adPDN NRCMs basallyexpress Drp1 at the mitochondria and display a striking 80%fragmented phenotype compared with EGFP control, suggestingthat adPDN NRCMs may be more susceptible to cell death (P <0.01; Fig. 3D and Fig. S4A). Pim-1 overexpression in siDRP-treated NRCMs did not increase reticular mitochondrial mor-phology after sI, demonstrating the necessity of Drp1 for PimWTmitochondrial effects (Fig. S4B).PimWT mice and adPimWT NRCMs exhibited an increase in

Drp1-S637 phosphorylation, thereby suggesting a potential site onDrp1 for Pim-1-mediated phosphorylation (Fig. 2B and Fig.S2E). Adenoviral overexpression of Drp1 in NRCMs, followedby immunoprecipitation of Pim-1, displayed Pim-1 and Drp1 as-sociation before and during sI treatment (Fig. 3E), which wasconfirmed by proximity ligation assay in vitro (Fig. 3F). Further-more, phosphorylation of Drp1 by Pim-1 kinase is demonstratedby in vitro kinase assay (Fig. 3G). Recombinant Drp1 was used assubstrate, and γ-32p-ATP incorporation was measured with

recombinant active Pim-1 and recombinant kinase dead Pim-1.Mutation of serine 637 of Drp-1 to an alanine residue that cannotbe phosphorylated (S637A) induces mitochondrial translocationofDrp-1 and fragmentation, whereasmutation to phosphomimeticaspartic acid (S637D) inhibits translocation (21). The relevance ofthe serine 637 residue for Pim-1 mediated effects on Drp-1 sub-cellular localization and mitochondrial phenotype was assessedby coexpression of Drp1-S637A with either Pim-1-mCherry ormCherry alone as a control. Drp1-S637A localized to mitochondriaand induced fragmentation when coexpressed with mCherry con-trol (Fig. S5 A and B). Coexpression of Pim-1-mCherry togetherwith Drp1-S637A appeared comparable to results observed usingthe mCherry control, supporting the postulate that inhibition ofDrp-1 action by Pim-1 involves phosphorylation on serine 637 ofDrp1 (Figs. 2 and S2E).

Drp1 Translocation Is Mediated by PUMA After Loss of Pim. TheBH3-only protein p53 up-regulated modulator of apoptosis(PUMA) is reported to be required for Drp1 accumulation tothe mitochondria in HeLa cells (22). Proximity ligation assay(PLA) analysis with Drp1 and endogenous PUMA in NRCMsdemonstrated protein–protein interaction at basal conditions(Fig. S6A). PUMA protein levels decreased by 35% in PimWT-transgenic mice (P < 0.05; Fig. 4A), and total PUMA expressiondecreased 50% after sI, as shown by immunoblot, suggesting thatPim-1 reduces Drp1 localization by inhibiting PUMA (P < 0.01;Fig. 4B). In contrast, analysis of mitochondrial fractions fromadEGFP and adPDN NRCMs showed a 2.2-fold increase inPUMA expression at the mitochondria mediated by PDN (P <0.05; Fig. 4C). Collectively, these results support the premise thatinhibiting PUMA may rescue PDN-induced Drp1 localization tomitochondria and blunt fragmented mitochondrial morphology.To test this hypothesis, NRCMs were infected with adPDN aloneor dually infected with adPDN and PUMA dominant-negativeadenovirus (adPumaDN), a BH3 domain defective PUMA. PLAanalysis in adPumaDN NRCMs with Drp1 exhibited interactionbetween the two proteins (Fig. S6B). This result was confirmedby immunoprecipitation of adPumaDN NRCMs with Drp1 be-fore and after sI, which also showed PumaDN and Drp1 in-teraction in vitro (Fig. S6C). Immunoblot analysis after cellfractionation displayed a 1.7-fold increase in Drp1 mitochondrial

A

B C

D G

E F H

Fig. 2. Drp1 Phosphorylation is regulated by Pim-1. (A–F) Immunoblot and quantification of pDrp1S637 and total Drp1 in murine PimWT and PDN whole-heartlysates. (G–H) Immunoblot of Drp1 localization in isolated mitochondria from PimDN (PDN) and FVB nontransgenic (NTG) control hearts harvested from age-and sex-matched mice. *P < 0.05; **P < 0.01; ***P < 0.001.

Din et al. PNAS | April 9, 2013 | vol. 110 | no. 15 | 5971

CELL

BIOLO

GY

translocation (P < 0.05) compared with EGFP control; however,overexpression of PumaDN significantly attenuates this accumu-lation by 2.4-fold (P < 0.01; Fig. 4D).Protein analysis of whole-cell lysates confirmed no change in

total Drp1 protein levels, attributing the effect of PumaDNsolely to localization of Drp1 (Fig. 4E). To support PumaDNinhibition of Drp1 translocation, mitochondrial morphology wasalso evaluated after PumaDN rescue of the PDN-associatedfragmented mitochondrial phenotype. Addition of PumaDN toNRCMs with PDN promoted a 60% reticular mitochondrialmorphology compared with adPDN alone (Fig. 4F). ExtendedDrp1 localization to the mitochondria can cause cell death;therefore, we assessed whether PumaDN could prevent apoptosisafter sI. NRCMs infected with adPDN and subjected to 30 min sIresulted in 27% cell death compared with control (P < 0.05),whereas NRCMs dually infected with adPDN and adPumaDNdisplayed only 13% cell death (P < 0.001; Fig. 4G).

DiscussionMitochondria constantly undergo the process of fission and fu-sion, sustaining a homeostatic balance within the cell. Why mi-tochondria undergo fusion and fission is not well understood, butit is hypothesized that fission and fusion allow for complementa-tion of mitochondrial mutations and elimination of dysfunctionalmitochondria, thereby preserving functional mitochondrial pop-ulations (23). Mitochondrial fission and fusion classically involvethe participation of five individual proteins: Drp1, Mfn1, Mfn2,Fis1, and Opa1. Fusion involves the GTPase Opa1, Mfn1, andMfn2, whereas Fis1 and Drp1 promote mitochondrial fission (2).Disruption of mitochondrial morphological dynamics sensitizes

cells to apoptotic cell death (9, 24, 25), and recent studies havereported that disturbances within the balance of these two pro-cesses can implicate the outcome of cardiovascular disease andneurodegenerative disease (20). In this study, we show that is-chemic challenge in myocytes promotes translocation of Drp1 to

A B

C D

E F G

Fig. 3. Pim-1 overexpression affects Drp1 Localization in NRCMs. (A) Immunoblot of Drp1 localization in NRCMs transduced with adEGFP control virus oradEGFP-PimWT after sI. (B) Protein analysis of Drp1 compartmentalization in adEGFP- or adPDN-transduced NRCMs after fractionation. (C) Analysis andquantification of mitochondrial morphology in NRCMs infected with adPimWT or adEGFP subjected to up to 30 min of sI. *, significant to EGFP by two-wayANOVA (D) Quantification of adPDN and adEGFP NRCMs maintaining reticular mitochondria before and after 30 min of sI. **, significant compared with EGFPcontrol. (E) Immunoprecipitation of Drp1-infected NRCMs with Pim-1 followed by immunoblot of Drp1 and Pim-1. (F) Proximity ligation assay with Pim-1 andDrp1 in cultured HeLa cells. Each red dot represents a single protein–protein interaction. (G) Phosphorylation of Drp1 by Pim-1 kinase in vitro. Recombinant Drp1wasused as substrate in an in vitro kinase assay that measures γ-32p-ATP incorporation. His-Pim-1 (wild-type, kinase dead) was expressed in Escherichia coli and affinity-purified for use in the kinase assay. Phosphorylated proteins were separated by SDS/PAGE and visualized by autoradiography. *P < 0.05; **P < 0.01.

5972 | www.pnas.org/cgi/doi/10.1073/pnas.1213294110 Din et al.

the mitochondria, resulting in a fragmented mitochondrial phe-notype (Figs. 1 and 2), and that Pim-1 kinase prevents thesechanges. Drp1 has been known to be associated with mitochon-drial fission, and its persistent sequestration at the mitochondriais associated with cell death. The mechanism by which Drp1causes fission is not completely understood.In the past, Pim-1 has been shown to have multiple protective

effects at the mitochondria (15, 26, 27), but a relationship be-tween Pim-1 and mitochondrial morphologic dynamics has notbeen examined previously. Pim-1 overexpression preventedDrp1 accumulation at mitochondria and decreased total Drp1levels (Fig. 3A and Fig. S2 A and C). Conversely, the PDNmutant significantly increased total protein levels and mito-chondrial Drp1 localization at baseline, but Drp1 translocationafter sI did not increase relative to control samples (Fig. 3Band Fig. S2 B and D). Collectively, these findings support thepostulate that Pim-1 inhibition of Drp1 translocation is the

predominant mechanism of Drp1 regulation by Pim-1. Drp1translocation was significantly reduced by Pim-1 overexpressionduring sI, and Drp1 accumulation at mitochondria is promotedby PDN under basal condition. Because mitochondria alreadydisplay a fragmented phenotype during sI before reperfusion,small increases of mitochondrial Drp1 promote fragmentedmitochondrial morphology.Protein–protein interaction between Pim-1 and Drp1 occurs

under both basal and ischemic conditions (Fig. 4E), which is con-sistent with prior reports of Pim-1 binding to and stabilizing proteins(19), raising the possibility that Pim-1 promotes Drp1 cytosolic se-questration by association.Drp1mitochondrial shuttling is inhibitedby phosphorylation at S637, with Drp1-S637 dephosphorylationmodulated by calcineurin (21, 28). Dephosphorylation of Drp1-S637

prevents cytosolic sequestration of Drp1 and promotes mitochon-drial fission, which is consistent with our observation of increasedlevels of phospho-Drp1 in Pim-1 overexpressing transgenic hearts.

A B C

D E

F G

Fig. 4. PUMAmediates the effect of Pim-1 on Drp1 translocation. (A) Immunoblot of PUMA in nontransgenic (NTG) and PimWT transgenic mice. Immunoblotof NRCM lysates infected with either (B) adPimWT and adEGFP or (C) adPDN and adEGFP adenovirus for PUMA. (D) Immunoblot of Drp1 protein levels at themitochondria after rescue with adPumaDN. *, significant compared with EGFP control; #, significant compared with adPDN. (E) Drp1 protein expression intotal NRCM lysates after infection with adEGFP, adPDN, adPumaDN, and adPDN + adPumaDN. (F) Morphological assessment of reticular mitochondria in theaforementioned groups by Tom20 staining. **, significant compared with EGFP; ##, significant compared with PDN. (G) Cell death analysis with propidiumiodide of NRCMs subjected to simulated ischemia. *, significant compared with respective baseline control; #, significant compared with EGFP sI; $, significantcompared with PDN sI. #P < 0.05; *P < 0.05; ##P < 0.01; $$P < 0.01; **P < 0.01; $$$P < 0.001; ***P < 0.001.

Din et al. PNAS | April 9, 2013 | vol. 110 | no. 15 | 5973

CELL

BIOLO

GY

Although Drp1-S637 is a phosphorylation target for Pim-1, addi-tional candidate residues on Drp1 are also likely phosphorylated toalter Drp1 function and localization, which will be assessed infuture studies.Drp1-mediated fission events primarily occur through multiple

protein–protein interactions that form multimeric complexes(29). It was recently reported that PUMA is necessary for Drp1to accumulate at the mitochondria; however, a direct interactionbetween PUMA and Drp1 remained unexplored (22).Our findings suggest a pivotal role for PUMA and Drp1 mito-

chondrial localization. PUMA is an upstream activator of theapoptotic cell death pathway through the mitochondria andmediates the cell death response during ischemic injury (30).Mechanistically, PUMA’s proapoptotic activity is linked to se-questration of Bcl-2 and Bcl-xL that causes a displacement of Baxand/or Bak and activation of the proapoptotic functions of theseproteins (31, 32). A loss of Pim kinase activity promoted an in-crease of Drp1 at the mitochondria and a greater number offragmented mitochondria; however, addition of PumaDN was ableto rescue a fragmented mitochondrial phenotype. This suggeststhat PUMA may play a role during Drp1 scission events and thatthe BH3 domain of PUMA may be necessary for the properfunctioning of either Drp1 assembly or GTPase activity.Several lines of evidence support the cardioprotective activity

of Pim-1, including enhanced regeneration of the myocardiumafter myocardial infarction and preservation of mitochondrialstructure during cardiac challenge. In this study, we demonstratean additional beneficial effect of Pim-1 on preventing mitochon-drial fission throughDrp1 cytosolic sequestration. Interferencewithmitochondrial structural dynamics can prevent apoptotic cell death,as the intrinsic pathway of apoptosis occurs at the mitochondria.

Collective findings of cardiac overexpression of Pim-1 offer an ex-cellent solution to use Pim-1 as a therapeutic agent for the treat-ment and intervention of cardiac cell death.

MethodsAdenoviral Constructs, NRCM Culture, Transgenic Animals, and Animal Use. Gen-eration of EGFP, EGFP-PDN, and EGFP-PimWT adenoviral constructs and PimWTand PDN transgenic mice was reported previously (15, 17, 18). NRCM isolationand culture was performed as previously described (15, 18). I/R surgery wasperformed on sex-matched FVB mice 12–14 wk of age; further details areprovided in SI Methods.

sI/R andMitochondrial Morphology Analysis.After adenoviral infection, sI or sI/R treatmentswere done in glucose-freemediawith 5mM sodium cyanide and20mM 2-deoxy-glucose. For sI/R, 10 mM glucose was added after sI treatmentfor indicated times. Cells were fixed then stained with Tom20 (Santa Cruz,1:200), and mitochondrial fragmentation was qualified on a cell-to-cell basis,using a confocal microscope.

Statistical Analysis. Statistical analysis was performed using Student’s t-test.Comparison of more than two groups was performed by one-way ANOVAwith Bonferroni’s post hoc test or by two-way ANOVA. A P value of less than0.05 was considered statistically significant. Error bars represent SEM. Sig-nificance indicators are *P < 0.05, **P < 0.01, and ***P < 0.001.

ACKNOWLEDGMENTS. We thank Dr. Luca Scorrano for providing Drp1-S637A

plasmid. We thank all members of the M.A.S laboratory for helpful discus-sion and comments This work was supported by National Heart, Lung,and Blood Institute Grants R21-HL102714, R01-HL067245, R37-HL091102,P01-HL085577, RC1-HL100891, R21-HL102613, R21 HL104544, and R01HL105759 (to M.A.S.); American Heart Association Pre-Doctoral Fellowship12PRE12060248 (to S.D.); the Rees-Stealy Research Foundation (S.D.); andDeutsche Forschungsgemeinschaft Grants MV 1659 1/1 and KO 3900/1-1 (toM.V. and M.H.K.).

1. Machado NG, Alves MG, Carvalho RA, Oliveira PJ (2009) Mitochondrial involvement incardiac apoptosis during ischemia and reperfusion: Can we close the box? CardiovascToxicol 9(4):211–227.

2. Ong SB, Hausenloy DJ (2010) Mitochondrial morphology and cardiovascular disease.Cardiovasc Res 88(1):16–29.

3. Ong SB, et al. (2010) Inhibiting mitochondrial fission protects the heart against ischemia/reperfusion injury. Circulation 121(18):2012–2022.

4. Cerveny KL, Tamura Y, Zhang Z, Jensen RE, Sesaki H (2007) Regulation of mito-chondrial fusion and division. Trends Cell Biol 17(11):563–569.

5. Chen H, Chan DC (2005) Emerging functions of mammalian mitochondrial fusion andfission. Hum Mol Genet 14:R283–R289.

6. Karbowski M, et al. (2002) Spatial and temporal association of Bax with mitochondrialfission sites, Drp1, and Mfn2 during apoptosis. J Cell Biol 159(6):931–938.

7. Bras M, et al. (2007) Drp1 mediates caspase-independent type III cell death in normaland leukemic cells. Mol Cell Biol 27(20):7073–7088.

8. Cassidy-Stone A, et al. (2008) Chemical inhibition of the mitochondrial division dy-namin reveals its role in Bax/Bak-dependent mitochondrial outer membrane per-meabilization. Dev Cell 14(2):193–204.

9. Estaquier J, Arnoult D (2007) Inhibiting Drp1-mediated mitochondrial fission selec-tively prevents the release of cytochrome c during apoptosis. Cell Death Differ 14(6):1086–1094.

10. Scorrano L (2009) Opening the doors to cytochrome c: Changes in mitochondrialshape and apoptosis. Int J Biochem Cell Biol 41(10):1875–1883.

11. Chang CR, Blackstone C (2007) Cyclic AMP-dependent protein kinase phosphorylationof Drp1 regulates its GTPase activity and mitochondrial morphology. J Biol Chem282(30):21583–21587.

12. Karbowski M, Youle RJ (2003) Dynamics of mitochondrial morphology in healthy cellsand during apoptosis. Cell Death Differ 10(8):870–880.

13. Parra V, et al. (2008) Changes in mitochondrial dynamics during ceramide-inducedcardiomyocyte early apoptosis. Cardiovasc Res 77(2):387–397.

14. Cribbs JT, Strack S (2007) Reversible phosphorylation of Drp1 by cyclic AMP-dependentprotein kinase and calcineurin regulates mitochondrial fission and cell death. EMBORep 8(10):939–944.

15. Borillo GA, et al. (2010) Pim-1 kinase protects mitochondrial integrity in car-diomyocytes. Circ Res 106(7):1265–1274.

16. Fischer KM, et al. (2009) Enhancement of myocardial regeneration through geneticengineering of cardiac progenitor cells expressing Pim-1 kinase. Circulation 120(21):2077–2087.

17. Muraski JA, et al. (2008) Pim-1 kinase antagonizes aspects of myocardial hypertrophyand compensation to pathological pressure overload. Proc Natl Acad Sci USA 105(37):13889–13894.

18. Muraski JA, et al. (2007) Pim-1 regulates cardiomyocyte survival downstream of Akt.Nat Med 13(12):1467–1475.

19. Cho SG, Du Q, Huang S, Dong Z (2010) Drp1 dephosphorylation in ATP depletion-induced mitochondrial injury and tubular cell apoptosis. Am J Physiol Renal Physiol299(1):F199–F206.

20. Cho DH, Nakamura T, Lipton SA (2010) Mitochondrial dynamics in cell death andneurodegeneration. Cell Mol Life Sci 67(20):3435–3447.

21. Cereghetti GM, et al. (2008) Dephosphorylation by calcineurin regulates translocationof Drp1 to mitochondria. Proc Natl Acad Sci USA 105(41):15803–15808.

22. Wang JX, Li Q, Li PF (2009) Apoptosis repressor with caspase recruitment domaincontributes to chemotherapy resistance by abolishing mitochondrial fission mediatedby dynamin-related protein-1. Cancer Res 69(2):492–500.

23. Dorn GW, 2nd (2013) Mitochondrial dynamics in heart disease. Biochim Biophys Acta1833(1):233–241.

24. Landes T, et al. (2010) The BH3-only Bnip3 binds to the dynamin Opa1 to promotemitochondrial fragmentation and apoptosis by distinct mechanisms. EMBO Rep 11(6):459–465.

25. Wang JX, et al. (2011) miR-499 regulates mitochondrial dynamics by targeting calci-neurin and dynamin-related protein-1. Nat Med 17(1):71–78.

26. Yan B, et al. (2003) The PIM-2 kinase phosphorylates BAD on serine 112 and reversesBAD-induced cell death. J Biol Chem 278(46):45358–45367.

27. Lilly M, Sandholm J, Cooper JJ, Koskinen PJ, Kraft A (1999) The PIM-1 serine kinaseprolongs survival and inhibits apoptosis-related mitochondrial dysfunction in partthrough a bcl-2-dependent pathway. Oncogene 18(27):4022–4031.

28. Cho DH, et al. (2009) S-nitrosylation of Drp1 mediates beta-amyloid-related mito-chondrial fission and neuronal injury. Science 324(5923):102–105.

29. Zhu PP, et al. (2004) Intra- and intermolecular domain interactions of the C-terminalGTPase effector domain of the multimeric dynamin-like GTPase Drp1. J Biol Chem279(34):35967–35974.

30. Yu J, Zhang L (2008) PUMA, a potent killer with or without p53. Oncogene 27(Suppl1):S71–S83.

31. Ren D, et al. (2010) BID, BIM, and PUMA are essential for activation of the BAX- andBAK-dependent cell death program. Science 330(6009):1390–1393.

32. Nickson P, Toth A, Erhardt P (2007) PUMA is critical for neonatal cardiomyocyte apoptosisinduced by endoplasmic reticulum stress. Cardiovasc Res 73(1):48–56.

5974 | www.pnas.org/cgi/doi/10.1073/pnas.1213294110 Din et al.