p53-dependent regulation of mitochondrial energy production by the rela … · p53-dependent...

Tumor and Stem Cell Biology

p53-Dependent Regulation of Mitochondrial EnergyProduction by the RelA Subunit of NF-kB

Ren�ee F. Johnson, Ini-Isab�ee Witzel, and Neil D. Perkins

AbstractAberrant activity of the nuclear factor kappaB (NF-kB) transcription factor family, which regulates cellular

responses to stress and infection, is associated with many human cancers. In this study, we define a function ofNF-kB in regulation of cellular respiration that is dependent upon the tumor suppressor p53. Translocation ofthe NF-kB family member RelA to mitochondria was inhibited by p53 by blocking an essential interaction withthe HSP Mortalin. However, in the absence of p53, RelA was transported into the mitochondria and recruited tothe mitochondrial genome where it repressed mitochondrial gene expression, oxygen consumption, and cellularATP levels. We found mitochondrial RelA function to be dependent on its conserved C-terminal transactivationdomain and independent of its sequence-specific DNA-binding ability, suggesting that its function in this settingwas mediated by direct interaction with mitochondrial transcription factors. Taken together, our findingsuncover a new mechanism through which RelA can regulate mitochondrial function, with important implica-tions for how NF-kB activity and loss of p53 can contribute to changes in tumor cell metabolism and energyproduction. Cancer Res; 71(16); 5588–97. �2011 AACR.

Introduction

Aberrant activation of the nuclear factor kappaB (NF-kB)transcription factor family together with their activators, theIkB kinases (IKK), is associated with many human diseases,including cancer, in which it has been shown to regulate manytumor cell characteristics, including survival, proliferation,and metastasis (1, 2). There are 5 members of the NF-kBfamily RelA (p65), RelB, c-Rel, p50/p105, and p52/p100 thatform homo- and heterodimers (3). Although there are con-siderable data supporting the role of NF-kB, particularly RelA,in the regulation of cancer cell survival, there is little informa-tion on the possible roles of NF-kB in the regulation of cellmetabolism.

Interest in altered tumor cell metabolism was instigatedby Otto Warburg's hypothesis that a defect in cellularrespiration and subsequent shift to glycolysis was the initi-ating step in tumorigenesis (4). One mechanism that canlead to this state is loss of the tumor suppressor p53, whichresults in decreased oxygen consumption and increasedglycolysis (5). Interestingly, this glycolytic effect can be

dependent on the NF-kB family member RelA, which inthe absence of p53 resulted in enhanced expression of theglucose transporter Glut3 (6).

RelA, together with IKK subunits, has also been identified asa mitochondrial protein (7–10). Mitochondria contain anapproximately 17-kb circular genome that codes for 13 pro-teins, 22 tRNAs, and 2 ribosomal RNAs (SupplementaryFig. S2A), all of which are involved in oxidative phosphoryla-tion (11). There are 3 mitochondrial transcripts. Transcriptionof the light strand produces a single mRNA encoding ND6 and8 tRNA sequences. The heavy strand produces an RNA encod-ing the 2 mitochondrial rRNA sequences, 12S and 16S rRNA,together with a transcript encoding the remaining 12 mRNAs,14 tRNAs, and the 2 rRNAs, which then undergo subsequentRNA splicing to generate separate RNA species (12). NF-kBcan repress mitochondrial gene expression, including cyto-chrome B and cytochrome C oxidase mRNA levels, followingTNF or TNF–related apoptosis-inducing ligand (TRAIL) sti-mulation (8, 9), although the mechanism through which thiswas accomplished, or whether this was a direct effect ofspecific NF-kB subunits, was not established. Moreover, themechanism and consequences of NF-kB mitochondrial loca-lization on oxidative phosphorylation and ATP production,together with how these might contribute to the switch toglycolysis observed in cancer cells, have not been clearlydefined.

Materials and Methods

CellsRelA�/� mouse embryo fibroblast (MEF) cells [provided by

Professor Ron Hay (University of Dundee)] were reconstituted

Authors' Affiliation: Institute for Cell and Molecular Biosciences, New-castle University, Medical School, Newcastle Upon Tyne, United Kingdom

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: Neil D. Perkins, Institute for Cell and MolecularBiosciences, Newcastle University, Medical School, Catherine CooksonBuilding, Framlington Place, Newcastle Upon Tyne, United Kingdom.Phone: 0191-2228866; Fax: 0191-2227424; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-10-4252

�2011 American Association for Cancer Research.

CancerResearch

Cancer Res; 71(16) August 15, 20115588

by lentiviral infection with vector alone (null) or with humanRelA as described previously (13). H1299wtp53 cells have beenpreviously described (14, 15). H1299 cells were purchaseddirectly from the American Type Culture Collection, and olderstocks of H1299 cells were verified against the new cells at theHealth Protection Agency by microsatellite geneotyping. U-2OS and PC3 cells were purchased directly from the EuropeanCollection of Cell Cultures and were grown in Dulbecco'smodified Eagle's medium as previously described (14, 15). Allcells were grown to a maximum confluency of 70% and weresplit 1:5 (U-2 OS, PC3, and H1299 cells) or 1:10 (RelA�/� MEFcells) every 3 to 4 days for a maximum of 25 passages. It shouldbe noted that the RelA�/� MEF cells used in these studiespossess a mutant form of p53 (16). In the text, "early" refers topassages 4 to 13 whereas "late" refers to passages 16 to 25, withcells in the "transition" period not being used. Individualexperiments used a mix of cells with different passage num-bers within these "early" and "late" periods.

Isolation of mitochondrial proteinsMitochondrial proteins were isolated from the cells using

the ProteoExtract Cytosol/Mitochondria Fractionation Kit(Calbiochem) following the manufacturer's instructions. Mito-chondria were further purified by centrifugation through a19% (v/v) Percoll gradient (Sigma).

Measurement of ATP levelsATP levels from 5,000 cells per sample were measured

using the CellTitre-Glo Luminescent Cell Viability Assay(Promega) following the manufacturer's instructions. Lumi-nescence was measured following 2-hour incubation with aFLUOstar Omega microplate reader (BMG Labtech). Resultswere expressed as average relative luciferase units (RLU) percell.

Measurement of oxygen consumptionOxygen consumption from 50,000 cells per sample was

measured in a 96-well BD Oxygen Biosensor system (BDBiosystems) following the manufacturer's instructions.Fluorescence was measured after a 5-hour incubation witha FLUOstar Omega microplate reader at excitation of485 nm and emission of 630 nm. Data were expressed asnormalized relative fluorescence (NRF) relative to a blankmeasurement for each individual well in the absence of cellsand media.

Other assaysDNA/siRNA transfections, immunoprecipitations, Western

blotting, chromatin immunoprecipitation (ChIP), and quan-titative real-time reverse transcriptase PCR were carried outas described (14, 17, 18). All experiments were conducted witha minimum of 3 repeats.

Data analysis and statisticsIn all figures, oxygen consumption data are shown as mean

� SD NRF and expressed relative to the control samples, ATPdata are shown as mean � SD RLUs/cell and expressedrelative to the control sample, mRNA data are shown as mean

� SD normalized relative to 18s mRNA, and ChIP data areshown as mean � SD normalized to input and expressedrelative to the Gal4 control. All Western blottings are repre-sentative of a minimum of 3 individual experiments. Bandintensity was quantified using NIH ImageJ software and isshown in Supplementary Table S1.

Statistical comparisons between groups were analyzed byStudent's t test, or where appropriate, the paired t test.Comparisons between more than 2 groups were analyzedusing one-way ANOVA followed by a Tukey–Kramer multiplecomparisons test. Significance was determined at P < 0.05. Forall data, * indicates P < 0.05, ** indicates P < 0.01, and ***indicates P < 0.001.

Results

Regulation of oxygen consumption and ATP levelsby RelA

To investigate any potential role for RelA as a regulator ofoxidative phosphorylation and cellular energy production, theconsequences of RelA depletion by RNA interference wereexamined in the U-2 OS osteosarcaoma cell line. These cells, incommon with many other transformed and immortalized celllines, have a basal level of IKK and NF-kB activity (19), so theseexperiments were carried out without an additional NF-kBactivating stimulus. Interestingly, the effect of RelA knock-down varied dependent upon the time spent in culture, withincreased passage of U-2 OS cells resulting in a switch from adecrease to an increase in oxygen consumption upon RelAdepletion (Fig. 1A). For ease of interpretation, the controllevels in earlier and later passage cells have both been normal-ized to 1 but absolute levels are given in the figure legend.Addition of the antibiotic oligomycin, which inhibits themitochondrial Hþ-ATP synthase, strongly inhibited cellularoxygen consumption in late-passage U-2 OS cells, confirmingthat these measurements derive frommitochondrial oxidativephosphorylation (Fig. 1B). Moreover, under these conditions,no increase in oxygen consumption following RelA knock-down was seen, further confirming NF-kB regulation of cel-lular respiration (Fig. 1B).

To determine whether these changes in oxidative phos-phorylation resulted in changes in cellular ATP levels, weexamined U-2 OS cells following RelA knockdown. As withoxygen consumption, the effect seen proved dependent uponthe time the cells had spent in culture. RelA knockdown inearlier passage human U-2 OS osteosarcoma cells inhibitedATP production, whereas RelA knockdown in later passagecells resulted in increased ATP production (Fig. 1C). Similarresults were seen with knockdown of RelA in PC3 prostatecancer cells, confirming the generality of this observation(Supplementary Fig. S1A). Endogenous RelA has previouslybeen shown to be mitochondrially localized in these cells (9).

Cancer cells derive the majority of their ATP productionfrom glycolysis. We were therefore interested in determiningwhether these changes in ATP levels resulted from the effectsseen on oxygen consumption. Interestingly, treatment of laterpassage U-2 OS cells with oligomycin did not significantlychange ATP levels in control cells, suggesting that, as

RelA and Energy Production

www.aacrjournals.org Cancer Res; 71(16) August 15, 2011 5589

predicted, ATP production comes primarily from glycolysis.However, oligomycin treatment did prevent the increase inATP seen upon RelA knockdown, indicating that this effectresults from RelA repression of oxidative phosphorylation(Fig. 1D). Consistent with this, we saw no significant RelA-dependent change in glucose consumption in late-passage U-2OS cells (Supplementary Fig. S1B).

Large fluctuations in fundamental aspect of cell metabo-lism, such as in ATP levels and oxygen consumption, are not tobe expected. Nonetheless, to confirm the functional signifi-cance of these effects, we investigated downstream effects ofthese changes by analyzing the energy-sensing protein, 50-AMP–activated protein kinase a (AMPKa; ref. 20). Impor-tantly, RelA knockdown differentially altered activatoryTyr172 phosphorylation of AMPKa, in a manner correlatingwith the effects of RelA on total ATP levels (SupplementaryFig. S1C), indicating that these changes were physiologicallysignificant.

RelA binds to mitochondrial DNA and regulatesmitochondrial transcription

We were interested in whether mitochondrially localizedRelA was a cause of these effects on oxygen consumption andATP levels. We confirmed the presence of RelA in mitochon-drial protein preparations from U-2 OS cells, with higher levelsbeing seen in later passage cells (Fig. 2A). Because RelA is aDNA-binding protein, one mechanism through which it couldexert its effects on oxygen consumption and ATP levels is theregulation of mitochondrial gene expression. We thereforeexamined the binding of RelA to mitochondrial DNA (mtDNA)by ChIP. Mitochondrial gene transcription occurs bidirection-ally from 3 promoters located in the D-loop region of thegenome (Supplementary Fig. S2A). Consistent with theincreased localization of RelA to mitochondria in later passageU-2 OS cells (Fig. 2A), significant binding of RelA to the D-looppromoter region of the mitochondrial genome was observedonly in late-passage cells (Fig. 2B). Primers to the cytochrome

U-2 OS cellsA B

C DU-2 OS cells

Late-passage U-2 OS cells

Late-passage U-2 OS cells

1.5

1

0.5

O2consumption

(averageNRF)

ATPlevels

(averageRLU/cell)

ATPlevels

(averageRLU/cell)

O2consumption

(averageNRF)

0 0

0.4

0.8

1.2

1.6

0

0.4

0.8

1.2

1.6

0

0.4

0.8

1.21.2

1.8

0.6

0

1.6

siRNA: Control RelA Control RelA siRNA: Control RelA Control RelA

LatePassage: Early

siRNA: Control RelA Control RelA

Late (15–25)Passage: Early (5–14)


Treatment: Control Oligomycin

Treatment: Control Oligomycin

1.5

1

0.5

0

Figure 1. RelA regulates cellular oxygen consumption and ATP production. A, RelA knockdown in U-2 OS cells significantly decreased oxygenconsumption (P < 0.01, n ¼ 8) in early-passage cells but increased oxygen consumption (P < 0.001, n ¼ 8) in late-passage cells. The control levels in earlierand later passage cells have both been normalized to 1. Actual values (NRF/cell) for control early-passage cells were 0.25 � 0.10 and for later passagecells were 0.13 � 0.05. B, RelA knockdown in late-passage U-2 OS cells increased oxygen consumption (P < 0.001) that was blocked by 4-hour incubationwith the Hþ-ATP synthase inhibitor oligomycin (0.5 mg/mL). C, RelA knockdown in U-2 OS cells significantly decreased ATP levels (P < 0.01, n ¼ 9) inearly-passage cells but increased ATP levels (P < 0.01, n ¼ 8) in late-passage cells. The control levels in earlier and later passage cells have both beennormalized to 1. Actual values (RLU/cell) for control early-passage cells were 145.91 � 37.12 and for later passage cells were 113.61 � 26.74. D,RelA knockdown in late-passage U-2 OS cells increased ATP production (P < 0.01) that was blocked by 4-hour incubation with the Hþ-ATP synthaseinhibitor oligomycin (0.5 mg/mL).

Johnson et al.

Cancer Res; 71(16) August 15, 2011 Cancer Research5590

B region, which is immediately adjacent to the D-loop, con-firmed this result (Supplementary Fig. S2B). Because of theresolution allowed by ChIP, this result does not imply thatRelA is specifically binding to sites within the cytochrome Bgene. Interestingly, there was a RelA-dependent decrease inthe binding of POLRMT, the mitochondrial RNA polymerase,to mtDNA that was observed only in late-passage cells(Fig. 2C). To assay what effect this change in POLRMT bindingmay have on mitochondrial gene expression, quantitative PCR(qPCR) analysis was carried out using primers targeted tospecific mitochondrially encoded genes. Significantly, a

decrease in cytochrome B RNA and protein levels was alsoobserved in later passage cells, which was partially reversedupon RelA depletion (Fig. 2D; Supplementary Fig. S2C). Nosignificant effect of RelA knockdown on cytochrome B RNAlevels was seen in the earlier passage cells. Similar, late-passage specific effects were seen for cytochrome C oxidaseI and cytochrome C oxidase III (Supplementary Fig. 2D; datanot shown).

Taken together, these results suggest that RelA effects onoxidative phosphorylation and ATP levels in earlier passagecells are not mediated directly through mitochondria and

Figure 2. RelA binds to themitochondrial genome andregulates mitochondrial geneexpression. A, mitochondrialextracts were isolated from U-2OS cells at early- and late-passage numbers and equivalentlevels of protein (4 mg) wereresolved by SDS-PAGE beforeWestern blotting for the indicatedproteins. Controls formitochondria (VDAC), cytoplasm(a-tubulin), and organellecontamination (BiP) were used.B and C, ChIP analysis of RelA(B, P < 0.05, n ¼ 8/group)or POLRMT (C, P < 0.05,n¼ 4/group) binding to the D-loopregion of the mitochondrialgenome in U-2 OS cells. D, RelAknockdown inducesexpression of the mitochondrialgene cytochrome B in late-passage U-2 OS cells (P < 0.01,n ¼ 6/group). E, ChIP analysis ofRelA binding to the D-loop regionof the mitochondrial genome inU-2 OS cells transfected with 5 mgof Rous sarcoma virus (RSV)expression plasmid encoding anMTS fused to the N-terminalof RelA, RelA DBM, RelA RHD, orcontrol MTS alone (MTS-RSV;P < 0.001, n ¼ 7/group). F,expression of an MTS-taggedRelA or MTS-tagged RelA DBMplasmid significantly decreasedcytochrome B mRNA levels in U-2OS cells (P < 0.01, n ¼ 7/group).Data were normalized relative to18s mRNA and expressed relativeto the MTS-RSV control.

A B

C D

E F

Cytoplasm

p12

p20

p12

p20

Mitochondria

RelA

Relativeenrichment

(ChiP assay)

Relativeenrichment

(ChiP assay)

Relativeenrichment

(ChiP assay)

Relativecytochrome

B mRNAlevel

Relativecytochrome

B mRNAlevel

MT

S R

SV

MT

S R

SV

RS

V R

elA

RS

V R

elA

DB

M

RS

V R

elA

RH

D

MT

S R

elA

MT

S R

elA

MT

S R

elA

RH

D

MT

S R

elA

RH

D

MT

S R

elA

DB

M

MT

S R

elA

DB

M

3

2

1

0

12

8

4

0

20

15

10

1.81.61.41.2

10.80.60.40.2

0

2.52

1.51

0.50

0.08

0.06

0.04

0.02

3

2

1

0

VDAC

α-Tubulin

Bip

U-2 OS cells

D-loop

D-loopα-POLRMT

D-loopα-RelA

*

**

***

** ** ** **

**

**

Antibody: Control RelA

Passage: Early

Control RelA

Late

siRNA: Control RelA

Passage: Early

Control RelA

Late

siRNA: Control RelA

Passage: Early

Control RelA

Late



most probably result from regulation of nuclear gene expres-sion. However, as the cells continue to grow in culture, aprocess generally associated with acquisition of a moretransformed phenotype, RelA actively represses oxidativephosphorylation, at least in part through repression of mito-chondrial gene expression.

Mitochondrial RelA function requires its C-terminaltransactivation domain

To learn more about RelA function in mitochondria, wecreated a series of fusion proteins in which a mitochondrialtargeting sequence (MTS) was fused to the N-terminus ofRelA, thereby allowing the effects of mitochondrial RelA to bedissociated from nuclear and other cellular effects. In additionto full-length human RelA, MTS-tagged versions of a C-term-inal deleted RelA, encoding the amino-terminal DNA bindingand dimerization Rel homology domain (RHD), together witha specific DNA-binding mutation (DBM) of full-length RelA,were created. ChIP analysis of a non-MTS–tagged form of theRelA DBM mutant with primers to the IkBa promoter, aknown NF-kB target (21), confirmed the loss of sequence-specific DNA-binding ability (Supplementary Fig. S3A). Wes-tern blot analysis confirmed that all MTS-tagged proteins weretargeted to mitochondria (Supplementary Fig. S3B). Interest-ingly, all MTS-tagged forms of RelA, including the DBMmutant, were seen to bind mtDNA by ChIP analysis, withsignificantly higher levels of binding seen with the C-termin-ally deleted RelA RHD (Fig. 2E). This result suggests that RelAdoes not require direct binding to kB elements within themtDNA and that recruitment can result indirectly throughinteraction with mitochondrial transcription factors.

Consistent with the data seen upon depletion of endogen-ous RelA (Fig. 2D), expression of exogenous, MTS-tagged RelA

in U-2 OS cells, resulted in repression of cytochrome B andcytochrome C oxidase III RNA levels (Fig. 2F; SupplementaryFig. S3C) as well as a significant decrease in oxygen consump-tion and ATP levels (Supplementary Fig. S3D and E). None ofthese effects were seen upon expression of nontagged, wild-type RelA, confirming that MTS tagging can distinguishmitochondrial from other cellular effects of RelA. Repressionof oxygen consumption by exogenously expressedMTS-taggedRelA but not wild-type RelA was confirmed in p53-null humannon–small-cell lung carcinoma H1299 cells (SupplementaryFig. S3F). Consistent with these data and our previous analysis(Fig. 2C), expression of MTS-tagged RelA resulted in adecrease in POLRMT binding to mtDNA (SupplementaryFig. S3G).

Analysis of the MTS-tagged RelA mutants produced asurprising result. Although MTS-tagged RelA RHD bindsmtDNA more efficiently than any other RelA construct(Fig. 2E), it failed to repress mitochondrial gene expression,oxygen, and ATP levels together with POLRMT binding(Fig. 2F; Supplementary Fig. 3C–E and G). In contrast,MTS-tagged RelA DBM exhibited all of these effects almostto the same levels as MTS-tagged wild-type RelA. Therefore,DNA binding and/or recruitment to the mitochondrial gen-ome are not sufficient in itself to regulate oxidative phosphor-ylation and there is a requirement for the C-terminal RelAtransactivation domain (TAD).

p53 regulates RelA mitochondrial localization andfunction

A number of positive and negative feedback loops betweenthe p53 tumor suppressor and NF-kB pathways have beenidentified (22), and p53 suppresses NF-kB–dependent tumor-igenesis in a murine lung cancer model (23). Furthermore, p53

A B CCytoplasm Mitochondria

H1299 cells

WB:RelA

p53

Plasmid: Plasmid:pc

DNA

p53

pcDNA

p53

pcDNA

p53

pcDNA

p53

VDAC

α-Tubulin

BiP

WB:RelA

p53

VDAC

α-Tubulin

BiP

p53–/– MEF cells

Treatment:

Relativeenrichment

(ChiPassay)

4

3

2

1

0Control

H1299wtp53 cells

IPTG

*

Cytoplasm MitochondriaCytochrome B

-RelA

Figure 3. RelA mitochondrial localization is p53 dependent. A and B, equivalent levels of mitochondrial extracts (8 mg) from H1299 cells (A) and p53�/� MEFcells (B) transfected with 5 mg of pcDNA3 or pcDNA3-p53 expression plasmid were resolved by SDS-PAGE and Western blotting (WB) to determineRelA protein levels. C, ChIP analysis showing significant RelA binding to the cytochrome B region, which was inhibited upon induction of p53, inH1299wtp53 cells prepared with and without IPTG induction of p53 (P < 0.05, n ¼ 4/group).

Johnson et al.


effects on glycolysis can be NF-kB dependent (6). We observeda decrease in p53 levels associated with increased growth of U-2 OS cells in culture, (Supplementary Fig. S4A). Therefore, weinvestigated whether p53 activity might also regulate RelAmitochondrial function. Significantly in both p53-null H1299cells and in p53�/� MEF cells, used to avoid any issues withbackground levels of wild-type p53, reexpression of p53reduced RelA levels in mitochondria (Fig. 3A and B) whereas,induction of p53 expression by isopropyl-b-D-thiogalactoside(IPTG) treatment of H1299wtp53 cells (15; SupplementaryFig. S4B) resulted in loss of RelA binding to mtDNA asdetermined by ChIP analysis (Fig. 3C).We next determined the consequences of the ability of p53

to regulate RelA mitochondrial localization. Consistent withp53-induced loss of mitochondrial RelA, siRNA depletion ofendogenous RelA no longer resulted in an increase in cyto-chrome B or cytochrome C oxidase III RNA levels in eitherH1299 or p53�/� MEF cells in which p53 was reexpressed(Fig. 4A and B, Supplementary Fig. 4C–H). qPCR and Westernblot analysis confirmed the effects of p53 on RelA did notresult from changes in RelA protein or RNA levels (Supple-mentary Fig. S4B, D, E, G, and H).These observations suggested that RelA-dependent changes

in oxidative phosphorylation and ATP levels would also be p53

dependent. Indeed, induction of p53 expression in H1299wtp53

cells abolished the increase in oxygen consumption and ATPlevels seen upon siRNA depletion of endogenous RelA (Fig. 4Cand D). Moreover, a complete reversal of RelA function wasobserved, with depletion of endogenous RelA now reducing theincrease in oxygen consumption and ATP levels seen uponinduction of p53. Importantly, similar contrasting effects ofRelA on ATP levels were observed betweenwild-type or p53�/�

MEF cells. Here, RelA knockdown in wild-type cells resulted ina decrease in ATP levels, whereas in the absence of p53, RelAknockdown had the opposite effect (Supplementary Fig. S4I).

Endogenous RelA is localized to mitochondria throughan interaction with Mortalin

Although our results showed clear effects on RelA mito-chondrial localization, the mechanism through which this isachieved and how this might be perturbed by p53 was notclear. RelA lacks a defined N-terminal MTS, suggesting trans-port to mitochondria through a distinct, rate-limited, mechan-ism to proteins exclusively or predominantly localized in thisorganelle. If RelA possessed a classical MTS, a disproportion-ate level of protein could become localized to this organelle.

An indication of how this occurs resulted from separateanalysis of proteins binding the region of threonine 505 (T505)

Figure 4. RelA regulation ofmitochondrial energy productionis p53 dependent. A and B, theeffect of RelA knockdown oncytochrome B mRNA levelswas examined in H1299 cells(A; P < 0.05, n ¼ 6/group) andp53�/� MEF cells (B; P < 0.05,n ¼ 5/group) followingtransfection with 5 mg of pcDNA3or pcDNA3-p53 expressionplasmid. Data were normalizedrelative to 18s mRNA andexpressed relative to the pcDNA3control. C and D, oxygenconsumption (C; P < 0.001,n ¼ 4/group) and ATP production(D; P < 0.01, n ¼ 4/group) weremeasured after RelA knockdownin H1299wtp53 cells with andwithout IPTG induction of p53(200 mmol/L for 24 hours).

A*

3

2

1

0

4

3

2

1

0

1.8

1.2

0.6

0

3

2

1

0

*

**

* ***

* *

B

C D

Relativecytochrome

B mRNAlevel

O2consumption

(averageNRF)

ATPlevels

(averageRLU/cell)

Relativecytochrome

B mRNAlevel

H1299 cells p53–/– MEF cellsControl

siRNA:Control RelA Control RelA

p53

H1299wtp53 cells

Control IPTG

siRNA:

Treatment:

Control RelA RelAControl

H1299wtp53 cells

Control IPTG

siRNA:

Treatment:

Control RelA RelAControl

Control


p53



phosphorylation in the RelA TAD. Previously, we identifiedphosphorylation of RelA at T505 as an important regulatorysite that can help determine cell fate by inhibiting antiapop-totic function of RelA (17, 18). To identify proteins binding tothe evolutionarily conserved T505 region, we carried outpeptide affinity chromatography with either a T505 regionpeptide or a scramble control peptide (Fig. 5A). Analysis of theeluted proteins revealed a 75-kDa protein that specifically

bound the T505 peptide, identified by mass spectrometry asMortalin (Fig. 5B). Western blot analysis of column fractionsconfirmed that Mortalin was present only in the samples fromthe T505 peptide columns (Fig. 5C). Importantly, immuno-precipitation of endogenous RelA showed binding to endo-genous Mortalin in HeLa, U-2 OS, and HEK 293 cells (Fig. 5Cand D). We next analyzed the effect of mutating the RelA T505residue to alanine using RelA�/� immortalized MEF cells

A

E F G

C D

B

RelA (p53) 1

Rel homology domain 428

480T505

519

521

551 kDa191

97

645139

2819

14

HeLa cells

HeLa cells

HeLa cells

U-2 OS cells HEK 293 cells

Relativeenrichment

(ChiPassay)

Cell type: RelA reconstituted MEF cells

Cytoplasm Mitochondria

WB:Nul

l

T505

A

Rel

A

Nul

l

T505

Am

lgG

Mor

talin

mlg

GM

orta

lin

Mor

talin

mlg

G

Rel

A

Nul

l

T505

A

Rel

A

RelAWB:

RelA

Mortalin VDAC

α-Tubulin

BiP

Null RelA T505A

4

3

2

1

0

D-loopα-RelA

Mortalin

Mortalin

Mortalin

RelA

Mortalin

**

RelA

Fr 3

–4

Inpu

t

rlgG

Inpu

t

mlgG

Mor

talin

Inpu

t

mlgG

Mor

talin

mlgG

RelA

Mor

talin

Fr 5

–6

Fr 3

–4

Fr 5

–6

Mortalin

Input Null

RelA reconstituted MEF cells

T505ARelA

T505 peptide

IP:

IP:

IP:

IP:

Scramble

T505

pep

tide

Scr

ambl

e

TA1TA2

Figure 5. The HSP70 family member Mortalin is a RelA-binding protein. A, the RelA T505 region in the TAD is highly conserved across species. A peptidespanning this region (T505 peptide) and a control scramble peptide (scramble) were used to identify binding partners of this region by peptide affinitychromatography. B, a 75-kDa protein from HeLa whole-cell extracts that bound to the T505 peptide was identified by mass spectrometry as Mortalin. C,Western blotting (WB) of eluted column fractions confirming Mortalin binding to the RelA T505 motif peptide (top) and coimmunoprecipitation of endogenousRelA with Mortalin from HeLa cell extracts (bottom). D, coimmunoprecipitation of endogenous RelA with Mortalin from U-2 OS and HEK 293whole-cell extracts. Mouse IgG (mIgG) was used as a control in the immunoprecipitation (IP). E, in RelA�/� MEF cells reconstituted with RelA or a RelA T505Amutant, Mortalin was immunoprecipitated fromwhole-cell lysates, followed by subsequent Western blotting. F, mitochondria were isolated from RelA�/�MEFcells reconstituted with RelA or RelA T505A mutant, followed by Western blot analysis. G, ChIP analysis of RelA binding to the D-loop region of themitochondrial genome in RelA�/� MEF cells reconstituted with RelA or RelA T505A mutant (P < 0.01, n ¼ 5/group).

Johnson et al.


reconstituted with wild-type and mutant forms of RelA, toovercome background interactions from wild-type RelA. Con-firming the specificity of this interaction, T505A-mutated RelAno longer coimmunoprecipitated with Mortalin (Fig. 5E).Mortalin, also known as mitochondrial HSP70, is a member

of the HSP-70 family that plays an important role in a numberof cellular processes, including the stress response, cellularproliferation, intracellular trafficking, antigen processing, dif-ferentiation, and tumorigenesis (24, 25). Mortalin is known toact as a molecular chaperone facilitating protein unfoldingand transport across the mitochondrial inner membranethrough interaction with translocase of inner membrane(TIM) proteins (26–28), although it can also be found inthe cytoplasm of certain cell types (29). We were thereforeinterested in whether this interaction with Mortalin provideda basis for RelA mitochondrial localization and function.Consistent with this hypothesis, the RelA T505A mutant

also displayed reduced mitochondrial localization (Fig. 5F)and failed to bind mtDNA, as determined by ChIP analysis

(Fig. 5G; Supplementary Fig. S5A). Furthermore, siRNA knock-down of Mortalin in U-2 OS and H1299 cells resulted inreduced levels of mitochondrial RelA (Fig. 6A; SupplementaryFig. S5B) and reduced binding of endogenous RelA to mtDNAas determined by ChIP analysis (Fig. 6B), whereas overexpres-sion of Mortalin resulted in increased levels of endogenousmitochondrial RelA (Fig. 6C). Taken together, these dataindicated that mitochondrial RelA localization is mediatedby an interaction between the T505 region of the RelA TADand Mortalin.

Importantly, these data provided a potential explanation forthe ability of p53 to exclude RelA from mitochondria, becauseit has been previously reported that Mortalin also binds p53and can sequester it in the cytoplasm (30, 31). Consistent withthis hypothesis, induction of p53 expression by IPTG treat-ment of H1299wtp53 cells resulted in reduced RelA binding toMortalin (Fig. 6D).

These results indicate that RelA mitochondrial localizationand function are dependent upon the p53 status of the cell,

Figure 6. The HSP70 familymember Mortalin regulates RelAmitochondrial localization. A,Mortalin is required for RelAlocalization in mitochondria.H1299 cells were treated with aMortalin siRNA and mitochondriawere prepared. The presence ofRelA in the mitochondria wasdetermined by Western blotting(WB). B, knockdown of Mortalin inthe reconstituted RelA MEF cellssignificantly reduced the bindingof RelA to the cytochrome B(P < 0.05, n¼ 4/group) and D-loopregions (P < 0.05, n ¼ 3/group). C,H1299 cells were treated with apcDNA3 control (Cont) or apcDNA3-Mortalin (Mort)expression plasmid, andmitochondria were preparedfollowed by Western blot analysis.D, p53 inhibits the association ofRelA with Mortalin.Coimmunoprecipitation of RelA,Mortalin, and p53 was carried outin H1299wtp53 extracts preparedwith and without IPTG induction ofp53.

A B

C D

Cytoplasm

siRNA:

WB:

Con

trol

Mor

talin

Con

trol

Mor

talin

Plasmid:

IP:

IP:

pcD

NA

Mor

talin

pcD

NA

Mor

talin

RelA

Mortalin Relativeenrichment

(ChiPassay)

α-RelA2.5

2

1.5

1

0.5

0

VDAC

α-Tubulin

H1299 cells

H1299 cells

BiP

WB:

mlgG

mlgG

Mor

talin

mlgG

Mor

talin

p53

Mor

talin

Contro

l

+ IP

TG

RelA

Mortalin

VDAC

α-Tubulin

BiP

siRNA: Cont

Conrol

Conrol

+IPTG

+IPTG

inputs

H1299wtp53 cells

H1299wtp53 cells

Mort Cont Mort

RelA

Mortalin

Mortalin

Mortalin

Gene: Cytochrome B D-loop

Mitochondria

Cytoplasm Mitochondria

* *



with high p53 levels inhibiting RelA translocation to thisorganelle through inhibition of RelA association with Morta-lin. As a consequence, p53 prevents RelA-dependent repres-sion of mitochondrial gene expression and oxidativephosphorylation. However, RelA and p53 can also coopera-tively promote oxidative phosphorylation through othermechanisms, most likely involving regulation of nuclear geneexpression (Fig. 7).

Discussion

This study shows endogenous RelA binding to the mito-chondrial genome and links this to regulation of oxidativephosphorylation. Such direct regulation of mitochondrial geneexpression by normally nuclear transcription factors canpotentially provide a mechanism to link the regulation ofmitochondrial function with other cellular signaling pathways.We propose that this process contributes to the switch toglycolysis seen in cancer cells and acts as a complementarymechanism to the RelA-dependent increase in glycolytic geneexpression seen upon loss of p53 reported by Kawauchi andcolleagues (ref. 6; Fig. 7). We anticipate that pathway will act inparallel to other nuclear transcription factors, such as Stat3,which can also exhibit nontranscriptional mitochondrial func-tions that contribute to tumorigenesis (32).

Our data suggest that p53 prevents RelA mitochondriallocalization by inhibiting its interaction with Mortalin(Fig. 6D). Because p53 has itself been shown to bind to RelAand form a transcriptionally active complex upon stimulationwith TNFa or replication stress (33), as well as interactdirectly with Mortalin (30, 31), this could be achieved throughcompetitive binding and sequestration. However, othermechanisms are possible and our data also suggest that RelAmitochondrial import is regulated by posttranslational mod-

ification. This conclusion is implied from our identification ofMortalin as a protein binding the T505 region of RelA (Fig. 5),with the T505A mutation resulting in disruption of the RelA/Mortalin complex and inhibition of RelA mitochondrial loca-lization (Fig. 5E–G). Although this suggests that phosphoryla-tion at this site may be required to promote this interaction, apurely structural effect cannot be ruled out.

It is probable that the physiologic role of RelA in mitochon-dria is selective and that it will not regulate mitochondrialfunction in all cell types at all times. We propose that this ismost likely to occur during NF-kB–dependent processes invivo, coupling the physiologic response to the need forincreased or decreased energy and ATP levels. For example,IKK/NF-kB signaling has been shown to affect mitochondrialfunction and biogenesis in skeletal muscle (34), whereasGuseva and colleagues showed that TRAIL activation ofNF-kB DNA binding in mitochondria in LNCAP and PC3prostate cancer cells resulted in repression of mitochondrialgene expression (9). We suggest that tumor cells (and the celllines derived from them) that have become dependent uponNF-kB for their ability to grow and survive have subverted NF-kB mitochondrial function, where it contributes to the switchfrom oxidative phosphorylation to glycolysis, in part throughfulfilling a negative regulatory role on mitochondrial geneexpression.

Our results provide another mechanism linking the p53 andNF-kB pathways that will contribute to the ability of p53 tosuppress the oncogenic characteristics of NF-kB (22) andinhibit NF-kB–dependent tumor growth (23). We proposethat p53 loss during tumorigenesis removes an importantcontrol on RelA regulation of mitochondrial function, provid-ing a pathway through which NF-kB can contribute towardcancer cell malignancy by fundamentally altering cellularmetabolism.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Ron Hay and Aichi Msaki for supplying reagents, Phil Pasdois andAndrew Halestrap for technical advice, and Sonia Rocha, Derek Mann, ChrisParaskeva, and all members of the NDP Laboratory as well as former colleaguesat the universities of Bristol and Dundee for their advice and support.

Grant Support

R.F. Johnson and I. Witzel were funded by CRUK grants C1443/A12750,C1443/A4201 and C1443/A6721.

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received November 24, 2010; revised June 21, 2011; accepted June 26,2011; published OnlineFirst July 8, 2011.

References1. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and

cancer. Cell 2010;140:883–99.2. Kim HJ, Hawke N, Baldwin AS. NF-kB and IKK as therapeutic targets

in cancer. Cell Death Differ 2006;13:738–47.

+p53

+ Mortalin

Ox. Phos. Ox. Phos.

Nucl

ear

effe

cts

Nucl

ear

effe

cts

–p53RelA

Excludedfrom

mitochondria

Translocationto

mitochondria

Figure 7.Model of p53 and Mortalin-dependent RelA regulation of cellularenergy production. p53 can act to exclude RelA from mitochondria andtogether RelA and p53 promote oxidative phosphorylation. However,during tumorigenesis, upon loss of p53, RelA associates with Mortalin andenters mitochondria where it represses mitochondrial gene expressionand thus oxidative phosphorylation (Ox. Phos.). This then contributes tothe switch to glycolysis observed in cancer cells.

Johnson et al.


3. Hayden MS, Ghosh S. Signaling to NF-kB. Genes Dev 2004;18:2195–224.

4. Warburg O. On the origin of cancer cells. Science 1956;123:309–14.5. Matoba S, Kang JG, Patino WD, Wragg A, Boehm M, Gavrilova O,

et al. p53 regulates mitochondrial respiration. Science 2006;312:1650–3.

6. Kawauchi K, Araki K, Tobiume K, Tanaka N. p53 regulates glucosemetabolism through an IKK-NF-kB pathway and inhibits cell trans-formation. Nat Cell Biol 2008;10:611–8.

7. Bottero V, Rossi F, Samson M, Mari M, Hofman P, Peyron JF.Ikb-a, the NF-kB inhibitory subunit, interacts with ANT, themitochondrial ATP/ADP translocator. J Biol Chem 2001;276:21317–24.

8. Cogswell PC, Kashatus DF, Keifer JA, Guttridge DC, Reuther JY,Bristow C, et al. NF-kB and I kB a are found in the mitochondria.Evidence for regulation of mitochondrial gene expression by NF-kB. JBiol Chem 2003;278:2963–8.

9. Guseva NV, Taghiyev AF, Sturm MT, Rokhlin OW, Cohen MB. Tumornecrosis factor-related apoptosis-inducing ligand-mediated activa-tion of mitochondria-associated nuclear factor-kB in prostatic carci-noma cell lines. Mol Cancer Res 2004;2:574–84.

10. Zamora M, Merono C, Vinas O, Mampel T. Recruitment of NF-kB intomitochondria is involved in adenine nucleotide translocase 1 (ANT1)-induced apoptosis. J Biol Chem 2004;279:38415–23.

11. Anderson S, Bankier AT, Barrell BG, de BruijnMH, Coulson AR, DrouinJ, et al. Sequence and organization of the human mitochondrialgenome. Nature 1981;290:457–65.

12. Scarpulla RC. Transcriptional paradigms in mammalian mitochondrialbiogenesis and function. Physiol Rev 2008;88:611–38.

13. O’Shea JM, Perkins ND. Thr435 phosphorylation regulates RelA(p65) NF-kB subunit transactivation. Biochem J 2010;426:345–54.

14. Barre B, Perkins ND. A cell cycle regulatory network controlling NF-kBsubunit activity and function. EMBO J 2007;26:4841–55.

15. Rocha S, Martin AM, Meek DW, Perkins ND. p53 represses cyclin D1transcription through down regulation of Bcl-3 and inducing increasedassociation of the p52 NF-kB subunit with histone deacetylase 1. MolCell Biol 2003;23:4713–27.

16. Campbell KJ, Rocha S, Perkins ND. Active repression of antiapop-totic gene expression by RelA(p65) NF-kB. Mol Cell 2004;13:853–65.

17. Campbell KJ, Witty JM, Rocha S, Perkins ND. Cisplatin mimics ARFtumor suppressor regulation of RelA (p65) nuclear factor-kB trans-activation. Cancer Res 2006;66:929–35.

18. Rocha S, Campbell KJ, Perkins ND. p53- and Mdm2-independentrepression of NF-kB transactivation by the ARF tumor suppressor.Mol Cell 2003;12:15–25.

19. Barre B, Perkins ND. The Skp2 promoter integrates signaling throughthe NF-kB, p53, and Akt/GSK3b pathways to regulate autophagy andapoptosis. Mol Cell 2010;38:524–38.

20. Hardie DG. AMPK: a key regulator of energy balance in the single celland the whole organism. Int J Obes (Lond)2008;32 Suppl 4: S7–12.

21. Ito CY, Kazantsev AG, Baldwin AS. 3 NF-kB sites in the IkBa promoterare required for induction of gene-expression by TNF-A. Nucleic AcidsRes 1994;22:3787–92.

22. Perkins ND. Integrating cell-signalling pathways with NF-kB and IKKfunction. Nat Rev Mol Cell Biol 2007;8:49–62.

23. Meylan E, Dooley AL, Feldser DM, Shen L, Turk E, Ouyang C, et al.Requirement for NF-kB signalling in a mouse model of lung adeno-carcinoma. Nature 2009;462:104–7.

24. Kaul SC, Deocaris CC, Wadhwa R. Three faces of Mortalin: a house-keeper, guardian and killer. Exp Gerontol 2007;42:263–74.

25. Wadhwa R, Taira K, Kaul SC. An Hsp70 family chaperone, Mortalin/MtHsp70/PBP74/Grp75: what, when, and where?Cell Stress Chaper-ones 2002;7:309–16.

26. Moro F, Okamoto K, Donzeau M, Neupert W, Brunner M. Mitochon-drial protein import: molecular basis of the ATP-dependent interactionof MtHsp70 with Tim44. J Biol Chem 2002;277:6874–80.

27. Voos W, Rottgers K. Molecular chaperones as essential mediators ofmitochondrial biogenesis. Biochim Biophys Acta 2002;1592:51–62.

28. Yaguchi T, Aida S, Kaul SC, Wadhwa R. Involvement of Mortalin incellular senescence from the perspective of its mitochondrial import,chaperone, and oxidative stress management functions. Ann N YAcad Sci 2007;1100:306–11.

29. Ran Q, Wadhwa R, Kawai R, Kaul SC, Sifers RN, Bick RJ, et al.Extramitochondrial localization of Mortalin/MtHsp70/PBP74/GRP75.Biochem Biophys Res Commun 2000;275:174–9.

30. Wadhwa R, Takano S, Robert M, Yoshida A, Nomura H, Reddel RR,et al. Inactivation of tumor suppressor p53 by mot-2, a Hsp70 familymember. J Biol Chem 1998;273:29586–91.

31. Wadhwa R, Yaguchi T, Hasan MK, Mitsui Y, Reddel RR, Kaul SC.Hsp70 family member, mot-2/mthsp70/GRP75, binds to the cyto-plasmic sequestration domain of the p53 protein. Exp Cell Res2002;274:246–53.

32. Wegrzyn J, Potla R, Chwae YJ, Sepuri NB, Zhang Q, Koeck T, et al.Function of mitochondrial Stat3 in cellular respiration. Science2009;323:793–7.

33. Schneider G, Henrich A, Greiner G, Wolf V, Lovas A, Wieczorek M,et al. Cross talk between stimulated NF-kB and the tumor suppressorp53. Oncogene 2010;29:2795–806.

34. Bakkar N, Wang J, Ladner KJ, Wang H, Dahlman JM, Carathers M,et al. IKK/NF-kB regulates skeletal myogenesis via a signaling switchto inhibit differentiation and promote mitochondrial biogenesis. J CellBiol 2008;180:787–802.



p53-dependent regulation of mitochondrial energy production by the rela … · p53-dependent...

Documents