a germ line mutation in the death domain of dapk-1 inactivates

A Germ Line Mutation in the Death Domain of DAPK-1Inactivates ERK-induced Apoptosis*

Received for publication, June 13, 2006, and in revised form, December 18, 2006 Published, JBC Papers in Press, January 23, 2007, DOI 10.1074/jbc.M605649200

Craig Stevens‡, Yao Lin‡, Maria Sanchez‡, Eliana Amin‡, Ellen Copson§, Helen White¶, Vicky Durston¶,Diana M. Eccles§, and Ted Hupp‡1

From the ‡Cancer Research UK p53 Signal Transduction Group, University of Edinburgh, South Crewe Road, Edinburgh EH4 2XR,the §University of Southampton, Southampton SO16 6YD, and the ¶National Genetics Reference Laboratory (Wessex),Salisbury District Hospital, Salisbury, Wiltshire SP2 8BJ, United Kingdom

p53 is activated genetically by a set of kinases that are compo-nents of the calcium calmodulin kinase superfamily, includingCHK2, AMP kinase, and DAPK-1. In dissecting the mechanismofDAPK-1 control, a novelmutation (N1347S) was identified inthe death domain of DAPK-1. The N1347S mutation preventedthe death domain module binding stably to ERK in vitro and invivo. Gel filtration demonstrated that the N1347Smutation dis-rupted the higher order oligomeric nature of the purifiedrecombinant death domain miniprotein. Accordingly, theN1347S death domain module is defective in vivo in the forma-tion of high molecular weight oligomeric intermediates aftercross-linking with ethylene glycol bis(succinimidylsuccinate).Full-length DAPK-1 protein harboring a N1347S mutation inthe death domain was also defective in binding to ERK in cellsand was defective in formation of an ethylene glycol bis(succin-imidylsuccinate)-cross-linked intermediate in vivo. Full-lengthDAPK-1 encoding the N1347S mutation was attenuated intumornecrosis factor receptor-induced apoptosis.However, theN1347S mutation strikingly prevented ERK:DAPK-1-depend-ent apoptosis as defined by poly(ADP-ribose) polymerase cleav-age, Annexin V staining, and terminal deoxynucleotidyl trans-ferase-mediated dUTP nick end labeling imaging. Significantpenetrance of the N1347S allele was identified in normalgenomic DNA indicating the mutation is germ line, not tumorderived. The frequency observed in genomic DNA was from 37to 45% for homozygous wild-type, 41 to 47% for heterozygotes,and 12 to 15% for homozygous mutant. These data highlight anaturally occurringDAPK-1mutation that alters the oligomericstructure of the death domain, de-stabilizes DAPK-1 binding toERK, and prevents ERK:DAPK-1-dependent apoptosis.

The tumor suppressor protein p53 is a stress-activatedDNA-binding protein and transcription factor that can induce a set ofgene products implicated in growth arrest, apoptosis, redoxbalance, and cellular repair pathways (1). Because p53 ismutated or inactivated frequently in human cancers, mucheffort is centered on determining the mechanisms whereby

mutations inactivate the p53 protein, determining which geneproducts mediate the tumor suppressor activity of the protein,and identifying the enzymes that activate the protein as a tumorsuppressor.It is important to determine whether the p53 “activating” or

“inhibitory” enzymes are also themselves mutation targets thatstimulate cancer development. One key paradigm developedfor p53 is that its activity in unstressed cells is held in check byan ubiquitin-dependent degradation pathway that promotesthe rapid turnover of the protein. A set of E32 ligases that canturnover p53 by promoting its ubiquitination include the ring-finger-containing proteins MDM2, COP-1, CHIP, and PirH2(2). A promoter polymorphism in the MDM2 gene enhancescancer incidence thus highlighting the importance of identify-ing genetic changes that may alter disease incidence (37). Asecond paradigm centers around the concept that the stressesor microenvironmental changes that activate p53 and relievethe protein of this negative regulation are quite distinct andinclude agents such as hypoxia, acidification, glucose starva-tion, oncogene activation, genome instability, microtubule dys-regulation, type I interferon, exposure to irradiation, and viralinfection (3). The enzymes that play a role in linking these dis-tinct stresses to p53 activation are beginning to be defined andinclude distinct protein kinase types, including casein kinase 2and ATM (4). Recent studies have interestingly shown thatmembers of the calciumcalmodulin kinase superfamily, includ-ing CHK2, AMPK, and DAPK-1 share the common feature ofbeing genetic components of signaling pathways that activatethe tumor suppressor activity of p53 (5–7). These calcium cal-modulin kinase superfamily members differ in that they areactivated by distinct stresses and provide a common evolution-ary link between cellular stresses and p53 activation. Further,ATM, CHK2, LKB-AMP kinase, and DAPK-1 activities can beattenuated by point mutation or by gene silencing in humancancers.The mechanisms driving p53 activation by phosphorylation

is being defined by biochemical and genetic studies. In response

* This work was supported by a Programme Grant from Cancer Research UK.The costs of publication of this article were defrayed in part by the pay-ment of page charges. This article must therefore be hereby marked“advertisement” in accordance with 18 U.S.C. Section 1734 solely to indi-cate this fact.

1 To whom correspondence should be addressed: Tel.: 44-131-777-3500; Fax:44-131-777-3520; E-mail: [email protected].

2 The abbreviations used are: E3, ubiquitin-protein isopeptide ligase; ATM,ataxia telangiectasia mutated; TUNEL, terminal deoxynucleotidyl transfer-ase-mediated dUTP nick end labeling; AMPK, AMP kinase; TNF, tumornecrosis factor; TNFR, TNR receptor; HA, hemagglutinin; CMV, cytomega-lovirus; PBS, phosphate-buffered saline; PARP, poly(ADP-ribose) polymer-ase; MAPK, mitogen-activate protein kinase; ERK, extracellular signal-reg-ulated kinase; MEK, MAPK/ERK kinase; CHK2, checkpoint kinase-2; DAPK-1,death-activated protein kinase-1; RSK, p90 ribosomal S6 kinase.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 282, NO. 18, pp. 13791–13803, May 4, 2007© 2007 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

MAY 4, 2007 • VOLUME 282 • NUMBER 18 JOURNAL OF BIOLOGICAL CHEMISTRY 13791

by guest on February 19, 2018http://w

ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

to ionizing radiation, p53 is activated by anATM-CHK2 signal-ing pathway that senses DNA strand breaks and leads to directphosphorylation of p53 in its transactivation domain, p53acetylation, and gene induction (8–12). The LKB-AMPKtumor suppressor pathway responds to metabolic stresses byactivating p53 and inducing phosphorylation at theATMsite atSer-15 (5), although the molecular mechanism whereby LKB-AMPK fully activates p53 is not defined (13). Finally, the tumorsuppressor protein DAPK-1 is a component of an oncogenicsignaling pathway that mediates p53 activation (7). The bio-chemical mechanism whereby DAPK-1 activates p53 is unde-fined, but biological studies have shown that DAPK-1 caninducemyosin light chain phosphorylation (14), autophagy (15,16), and/or antagonize FAK (17). Thus, three distinct physio-logical stresses known to activate p53 are now linked to threesignaling components of the calcium calmodulin kinase super-family that have the common feature of being genetic compo-nents of tumor suppressor pathways ATM-CHK2, LKB-AMPK, and DAPK-1.In an attempt to begin to define the mechanism underlying

how the tumor suppressor protein DAPK-1 functions at amolecular level, we have cloned the DAPK-1 gene and haveidentified a previously unidentified germ line mutation in thedeath domain. The death domain of DAPK-1 was identifiedoriginally in a genetic screen to be one of four functionaldomains required for DAPK-1 to exert its growth suppressiveactivity (18). The death domain is thought to be a signalingmodule contained in pro-apoptotic proteins, including Fas,TNFR, FADD, and UNC5H2 (19–21). One recent report hasindicated that the death domain of DAPK-1 forms a dockingsite required for its interaction with ERK (22), identifying thefirst binding protein for the death domain of DAPK-1. Ouranalysis on the effects of the novel germ linemutation on deathdomain function indicates the mutation alters death domainoligomerization and attenuates ERK docking and associatedapoptotic signaling. These data highlight a post-translationalmechanism whereby the DAPK-1 tumor suppressor activitycan be quenched in response to signal transduction events. Fur-ther, the relatively high penetrance of the death domain muta-tion in germ line DNA from at least one population set identi-fies a signaling pathway mutation that might affect DAPK-1activity in apoptotic diseases of the immune system, ischemicinjury, and cancer.

EXPERIMENTAL PROCEDURES

Sequencing from Genomic DNA

Potential study participants from the Southampton popula-tion with previously diagnosed truncatingmutations of BRCA1were identified from the data base of the Wessex ClinicalGenetics Service, Southampton. Genomic DNA was obtainedfor 116BRCA1mutation carriers, 20male and 96 female. Age attime of screening for BRCA1mutations ranged between 29 and57 years. The clinical histories of all subjects were reviewed toascertain age at first and subsequent malignancies. 59 gene car-riers had developed breast cancer, and a further 14 had ovariancancer. 10 patients had undergone prophylactic bilateraloophorectomy (5 after developing breast cancer), and 3

patients had undergone bilateral risk reducingmastectomy. Allmalignancies occurred in female subjects; male subjects weretherefore excluded from all analyses of cancer incidence. 102anonymous genomic DNA samples (46 male, 56 female, agerange 16–82), referred by theWessex Clinical Genetics Servicefor genetic screening of non-neoplastic conditions, wereobtained to provide an unmatched control group. Genotypingwas performed using PyrosequencingTM technology. Ampli-cons were generated in a 50-�l reaction volume with 15 pmoleach ofDAPK-Forward (5�-gtg cct tct cgc cat ga-3�) andDAPK-Reverse (biotin-5�-ata ggc ctc ctg gcc att-3�), as well as ReverseDAPK-SNP (codon 1347) (biotin-5�-ttg gga gcc ccg tta-3�), 0.2mM dNTPs, 1.5 mM MgCl2, 1� Buffer II (500 mM potassiumchloride and 100mMTris-HCl, pH 8.3) (Applied Biosystems), 1unit of AmpliTaq Gold (Applied Biosystems) using 10 ng ofgenomic DNA. PCR conditions were 94 °C for 7 min; 50 cycleswith denaturation at 94 °C for 30 s, annealing at 60 °C for 30 s,and elongation at 72 °C for 30 s; 1 cycle at 72 °C for 7min; and afinal hold at 15 °C. Thermocycling was performed using a PTC-0225 DNA Engine Tetrad (MJ Research). Single-stranded bio-tinylated PCR products were prepared for Pyrosequencingusing a Vacuum Prep Tool (Biotage AB). 3 �l of StreptavidinSepharoseTM HP (Amersham Biosciences) was added to 37 �lof binding buffer (10 mM Tris-HCl, pH 7.6, 2 M NaCl, 1 mMEDTA, 0.1% Tween 20) and mixed with 20 �l of PCR productand 20 �l of high purity water for 10 min at room temperatureusing aVariomagMonoshaker (Camlab). The beads containingthe immobilized templates were captured onto the filter probesafter applying the vacuum, and then washed with 70% ethanolfor 5 s, denaturation solution (0.2 MNaOH) for 5 s, andwashingbuffer (10 mM Tris-acetate, pH 7.6) for 5 s. The vacuum wasswitched off, and the beads were released into a PSQ 96 wellplate containing 45 �l of annealing buffer (20 mM Tris-acetate,2 mM MgAc2, pH 7.6), 0.3 �M DAPK sequencing primer (5�-GTGCCTTCTCGCCATGA-3�). The samples were heated to80 °C for 2 min and then allowed to cool to room temperature.Pyrosequencing reactions were performed according to themanufacturer’s instructions using the PSQ 96 SNP Reagent Kit(Biotage AB), which contained the enzyme and substrate mix-ture and nucleotides. Assays were performed using the nucleo-tide dispensation order AGTACGCGC. The sample genotypewas determined using SNP Software (Biotage AB).

Plasmids and Site-directed Mutagenesis

The C terminus of DAPK-1 (amino acids 1313–1431) wascloned into theGateway system vector pDONR221 (Invitrogen)using the following primers: Fwd 5�-GGGGACAAGTTTGTA-CAAAAAAGCAGGCTGGAAACTGAGTCGCCTGCTGG-ACCCG-3� and Rev 5�-GGGGACCACTTTGTACAAGAAAG-CTGGGTGTCACCGGGATACAACAGAGCTAAT-3�. Thisvector was used as a template to construct DAPK-1 polymor-phic variants using the in vitro mutagenesis systemQuikChange (Stratagene) as recommended by the manufac-turer. HA-tagged ERK andMEK(EE) expression vectors were agift of Steve Keyse (Dundee University, UK) HA-taggedDAPK-1 was a gift of Adi Kimchi (Weizmann Institute, Israel),and this allele (Asn-1347:Leu-1392) was mutated to derive the“founding” allele (Asn-1347:Phe-1392) and the “SNP” allele

Mutation in the Death Domain of DAPK-1

13792 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 282 • NUMBER 18 • MAY 4, 2007


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

(Ser-1347:Phe-1392). The sequences of the primers used formutagenesis are (bases which introduce amino acid change areunderlined): amino acid 1347 S�N (Fwd 5�-CAAAGTACAA-CACCAATAACGGGGCTCCCAAG-3� and Rev 5�-CTTGG-GAGCCCCGTTATTGGTGTTGTACTTTG-3�) and aminoacid 1392 F�L (Fwd 5�-GGATGCCGCAGACCTTTTGCTG-AAGGCATCC-3� andRev 5�-GGATGCCTTCAGCAAAAGGTC TGC GGC ATCC-3�). For expression of glutathioneS-transferase or His-tagged fusion protein in Escherichia coli,pDONR221-DAPK-1 vectors were recombined with thepDEST15/pDEST17 vector (Invitrogen) as recommended bythe manufacturer. HA-tagged ERK was described previously,and p55-TNFR expression construct was obtained from DavidDornan (Genentech). For expression in mammalian cells, thepDONR-DAPK-1 vectorswere used as a template for PCRclon-ing into the p3XFLAG-myc-CMV-26 expression vector(Sigma), which has an N-terminal 3�FLAG tag and a C-termi-nal myc tag. The primers were designed to include EcoR1 andBglII restriction sites, and PCR products were cloned into theFLAG-myc vector at the same sites. The primers used were:Fwd 5�-TTGAATTCAAAACTGAGTCGCCTGCTGGAC-CCG-3� and Rev 5�-TTAGATCTATCCGGGATACAA-CAGAGCTAATGGA-3�.

Immunoblotting

HCT116 wt cells were grown in McCoy’s medium (Invitro-gen) supplemented with 10% fetal calf serum (Invitrogen) at37 °C in a 5% CO2/H2O-saturated atmosphere. HEK293 cellswere grown in Dulbecco’s modified Eagle’s medium (Invitro-gen) supplementedwith 10% fetal calf serum (Invitrogen). Cellsfor transient transfection were plated out 24 h before transfec-tion at �1.5 � 106 cells per 100-mm dish or 5 � 105 cells per60-mmdish. For Lipofectamine 2000 transfection (Invitrogen),2 �l of Lipofectamine was used for every 1 �g of DNA trans-fected. Cells were harvested after a further incubation of 24–36h. Cells were lysed in ice-cold extraction buffer (50mMTris (pH7.4), 150 mM NaCl2, 5 mM EDTA, 0.5% Nonidet P-40, 5 mMNaF, 1mM sodiumvanadate, 1�protease inhibitormixture) for30 min and centrifuged at 13,000 rpm for 15 min to removeinsoluble material. The protein content of cell extracts wasmeasured using Bio-Rad reagent (Bio-Rad). Typically, 50 �g ofcell extract was immunoblotted. Samples were resolved bydenaturing gel electrophoresis, typically 4–12% precast gels(Novex) and electrotransferred to Hybond C-extra nitrocellu-lose membrane (Amersham Biosciences), blocked in PBS-10%nonfat milk for 30 min, then incubated with primary antibodyovernight at 4 °C in PBS-5% nonfat milk-0.1% Tween-20. Afterwashing (3 � 10 min) in PBS-Tween 20, the blot was incubatedwith secondary antibody, either horseradish peroxidase-conju-gated anti-rabbit or anti-mouse antibody (Dako, 1:5000), for 1 h atroom temperature in PBS-5% nonfat milk-0.1% Tween 20. Afterwashing (3 � 10 min) in PBS-Tween 20, proteins were visualizedby incubation with ECL reagent (Sigma). Equal protein loadingwas confirmedwith Ponceau S staining. FLAGantibody (M2) andFLAGM2-conjugated agarose were purchased from Sigma. Mycantibody (9E10) was purchased from Cancer Research UK. HAand PARP antibodies were purchased from Cell Signaling. Theanti-HA antibody is from Upstate (07–221), the anti-ERK anti-

body (9102), anti-phospho-ERK (9101), and anti-PARP antibody(9542) are fromCell Signaling.

Cell Culture: Cross-linking of Proteins,and Immunoprecipitation

16 h post-transfection, HCT116wt cells expressing FLAG-myc-taggedDAPK-1 proteinswere cross-linkedwith 2mMEGS(Pierce) for 1 h at 37 °C in a 5% CO2/H2O-saturated atmo-sphere. Cross-linking was terminated by washing cells in PBSand harvesting. For immunoprecipitation of cross-linked pro-teins, 30 �l of FLAG M2 beads (Sigma) were incubated over-night at 4 °C with rotation, together with 500 �l of cell extract(�1mg) prepared in extraction buffer (50mMTris (pH7.4), 150mM NaCl2, 5 mM EDTA, 0.5% Nonidet P-40, 5 mM NaF, 1 mMsodium vanadate, 1� protease inhibitor mixture). The beadpellets were then washed five times in lysis buffer before resus-pension in 100 �l of FLAG peptide elution buffer (0.25 �g/�lFLAGpeptide (Sigma) resuspended inTBS (50mMTris, pH7.5,150mMNaCl) andmixed with constant rotation for 2 h. Elutedproteins were then resuspended in 3� SDS-loading buffer andanalyzed by denaturing gel electrophoresis and immunoblot-ting with antibodies specific to the myc tag. For immunopre-cipitation of exogenous HA-ERK from HCT116 wt cells, HAantibody (Cell Signaling) was incubated with 30 �l of washedProtein G beads (Sigma) overnight at 4 °C with constant rota-tion, together with cell extract (�1 �g) diluted to a volume of500�l in extraction buffer (50mMTris (pH 7.4), 150mMNaCl2,5 mM EDTA, 0.5% Nonidet P-40, 5 mM NaF, 1 mM sodiumvanadate, 1� protease inhibitor mixture). The bead pelletswere then washed five times in extraction buffer before beingresuspended in 3� SDS-loading buffer and analyzed by dena-turing gel electrophoresis and immunoblotting. To monitorPARP cleavage, HEK293 cells or HCT116 wt cells were seededand transfected 24 h later with vectors encoding the respectivedeath domain modules (5 �g, Asn-1347:Phe-1392 or Ser-1347:Phe-1392), empty vector control, together with p55-TNFR (0.5�g), orDAPK-1, ERK, andMEK(EE) vectors (1�g of each). 16 hpost-transfection cells were harvested and immunoblotted forPARP cleavage of FLAG-tagged death domain expression.

Apoptosis Assays

TUNEL Assay—HCT116wt or HEK293 cells were seededdirectly onto cover slips in 6-well plates. 18 h later cells weretransfected with the indicated amounts of expression vectorsfor a further 24 h. Following transfection cells were fixed in 1%paraformaldehyde for 10 min at room temperature. Apoptoticcells were labeled usingApoptag Plus Fluorescein In SituApop-tosis Detection Kit S7111 (Chemicon) according to the manu-facturer’s instruction and viewed by fluorescence microscopy.Annexin V staining—HCT116wt cells were seeded into

6-well plates. 18 h later cells were transfectedwith the indicatedamounts of expression vectors for a further 24 h. To include anyfloating cells, media was collected into fluorescence-activatedcell sorting tubes and centrifuged at 1700 rpm for 4min to leavea cell pellet. Media was then discarded from tubes. Petri disheswere washed with 2 ml of PBS, PBS was discarded, and then 1.5ml of trypsin was added to each plate and the mixture wasincubated until cells became detached. Following cell detach-




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

ment 1.5 ml of McCoy’s 5A medium supplemented with 10%fetal calf serumwas added to each plate to stop the trypsin. Thecell suspensions were then added to each of the relevant cellpellets from the floating cells. Samples were centrifuged at 1700rpm for 4 min, pellets were resuspended in 1 ml of McCoy’s 5Amedium supplemented with 10% fetal calf serum, and the mix-ture was incubated for 5 min. Samples were centrifuged for afurther 4 min at 1700 rpm prior to resuspension in 1 ml ofice-cold PBS. Samples were centrifuged for a further 4 min at1700 rpm. Apoptotic cells were detected using TACS AnnexinV-FITC Apoptosis Detection Kit (R&D Systems) according tothemanufacturer’s instruction and analyzed by flow cytometry.PARPCleavage—HEK293 cells were seeded into 6-cmplates.

18 h later cells were transfected with the indicated amounts ofexpression vectors (DAPK-1, ERK, andMEK (EE)) for a further24 h. Cells were lysed in ice-cold TNN buffer (150mMNaCl, 50mM Tris, pH 8, 0.1% Nonidet P-40, phosphatase inhibitor mix-ture, 5mMNaF, and 1mM sodiumorthovanadate) for 30min onice prior to centrifugation at 13,000 rpm for 15 min at 4 °C.Cleared lysates were then resolved by SDS-gel electrophoresis,and proteins were detected using Rabbit polyclonal antibody toPARP (Cell Signaling #9542).

Gel Filtration

To examine effects of the death domain mutation on its oli-gomerization in vitro, a Superdex 200 10/300 GL (Tricon) highperformance gel filtration column was equilibrated with buffer(PBS, 5% glycerol, 1 mM benzamidine pre-filtrated with a Mil-lipore filter 0.22 �m). Sample containing the purified recombi-nant death domain protein (800 �g) was injected and elutedinto fractions of 1 ml each. The amount of protein eluted intoeach fraction was then quantified using Bradford assay. Theequipment employed in the gel filtration was AKTA fast pro-tein liquid chromatography (Amersham Biosciences) and frac-

tion collector FRAC-950 (AmershamBiosciences). The software used wasUNICORN version 4.10 (AmershamBiosciences).Molecularweightmark-ers were obtained from Sigma.

Enzyme-linked ImmunosorbentAssay for Measuring ERK: DeathDomain Stability

A 96-well microtiter plate (Corn-ing Inc.) was coated with purifiedDAPK death domain proteindiluted in 0.1 M Na2HCO3, pH 8.0,and incubated overnight at 4 °C.Each well was washed 6� with PBScontaining 0.1% Tween 20 (PBS-T)followed by incubation for 1 h atroom temperature with gentle agi-tation in PBS-T supplemented with3% bovine serum albumin. Thewells were washed 6� with PBS-Tprior to incubation with appropriateamounts of purified ERK protein(Upstate) diluted in PBS-T 3%

bovine serum albumin for 1 h at room temperature. After 1 hincubation the plate was washed again 6� with PBS-T andincubated with antibody specific to ERK (p44/42 MAPK anti-body #9102, Cell Signaling) for 1 h at room temperature. Fol-lowing a further 6� washes with PBS-T wells were incubatedwith secondary rabbit horseradish peroxidase antibodies fol-lowed by further washing and ECL. The results were quantifiedusing Fluoroskan Ascent FL equipment (Labsystems) and ana-lyzed with Ascent Software version 2.4.1 (Labsystems).

RESULTS

Asn 3 Ser Mutations at Codon 1347 in the Death DomainAttenuate ERK Binding—Upon cloning and sequencing of theDAPK-1 gene cloned from total RNA of a tumor cell line, weidentified twomutations in the death domain that differed fromthe original wild-type sequence (Fig. 1). These two mutationsinclude an Asn3 Ser mutation at codon 1347 and an Leu3Phemutation at codon 1392. To determine whether the Asn3Ser mutation at codon 1347 or the Leu 3 Phe mutation atcodon 1392 alters the function of the death domain, biochem-ical characterization was initiated. The only direct biochemicalfunction defined for the death domain of DAPK-1 is that itfunctions as an ERK docking site leading to phosphorylation atamino acid 735 (22). Mutation of the ERK docking site onDAPK-1 from 1392-LXL-1394 to 1392-AXA-1394 attenuatesERK binding (22). This latter characterization utilized the Leu-1392 allele ofDAPK-1 that differs from the genewe have clonedharboring the Phe-1392 codon. It is possible that the L to Fmutation at codon 1392 alters ERK binding either positively ornegatively, andwe first determinedwhether the L to Fmutationaltered ERK binding. The transfection of the minimal deathdomain alleles, Asn-1347:Leu-1392 or Asn-1347:Phe-1392, didnot alter the ability of the death domain to co-immunoprecipi-tate with transfected HA-tagged ERK (Fig. 2A, lanes 2 and 3

FIGURE 1. Location of SNP mutations in the death domain of DAPK-1. The functional domains of DAPK-1 areas indicated; kinase domain, an inhibitory autophosphorylation site, calmodulin domain, ankyrin repeats, anERK phosphorylation site and docking site (22), and the death domain. The mutations described in this reportwithin the death domain at codons 1347 and 1392 that differ from the originally cloned wild-type allele arehighlighted as indicated: N � S and L � F.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

versus 1). These data suggest thatthe original identification of ERK asa death domain binding protein atthe LXL motif is not altered by thePhe-1392 substitution.We subsequently evaluatedwheth-

er the Asn3 Ser mutation at codon1347 alters ERK binding despite thefact that this mutation lies out withthe LXL-containing ERK dockingsite. The transfection of the isolateddeath domain containing the fourpermutations at codons 1347 and1392 did not alter the steady-statelevels of the death domain (Fig. 2B,lanes 2–5 versus 1). However, theco-transfection of ERK with the iso-lated death domain variants differ-entially alters ERK binding afterco-immunoprecipitation (Fig. 2C,lanes 2–5 versus 1). The Ser-1347mutation on either the Leu-1392 orPhe-1392 background attenuatedERK binding (Fig. 2C, lanes 3 and 5versus 2 and 4) identifying this Asnto Ser mutation as having a specificfunctional defect.The studies above utilized death

domain transfected into humancells, and it was important to deter-mine whether the mutation directlyattenuates ERK binding to the deathdomain in vitro. As such, theN1347S death domain miniproteinvariants were purified from recom-binant E. coli expression strains(Fig. 2D), and binding activity toERK was measured in a quantitativeenzyme-linked immunosorbentassay. The addition of increasingamounts of ERK resulted in a dose-dependent increase in complex for-mation to the Asn-1347 deathdomainmodule, whereas ERK bind-ing was attenuated when using thedeath domain encoded by the Ser-1347 allele (Fig. 2E). These dataindicate that the death domainN1347S mutation directly alters theERK binding interface.The Asn3 SerMutation at Codon

1347 Disrupts Death Domain Oligo-merization—It was surprising thatthe mutation at codon 1347 dis-rupted ERK binding despite thefact that this lies outside the LXLERK binding site at codon 1392(22). These data suggest that alter-

FIGURE 2. The N1347S mutation in the death domain of DAPK-1 attenuates ERK binding. A, ERK binding tocodon 1392 death domain variants. HA-tagged ERK and FLAG-tagged death domain expression vectors weretransfected into cells as indicated under “Experimental Procedures.” ERK was immunoprecipitated with ananti-HA antibody and immunoblotted for (i) total ERK protein (HA-immunoblot) and (ii) immunoprecipitateddeath domain using an anti-FLAG IgG immunoblot. The input death domain was quantified using an anti-FLAGIgG, as indicated (top panel). B, steady-state levels of the Asn-1347 and Ser-1347 death domain modules.Vectors encoding the indicated Asn-1347 and Ser-1347 alleles of the death domain were transfected into cellsand immunoblotted to examine for general changes in steady state levels. C, examination of ERK binding tocodon 1347 death domain mutants. HA-tagged ERK and FLAG-tagged Asn-1347 and Ser-1347 alleles of thedeath domain expression vectors were transfected into cells as indicated under “Experimental Procedures.”Total ERK and death domain were immunoblotted, and the amount of proteins present after immunoprecipi-tation with an anti-HA antibody was determined by immunoblotting with an anti-HA IgG for ERK and ananti-FLAG IgG for the death domain. The top panel represents total protein in lysate by direct blotting, and thebottom panel represents protein purified in complex by co-precipitation. D, purification of recombinant deathdomain miniproteins. Death domain miniproteins were expressed and purified using nickel-chelate chroma-tography as indicated under “Experimental Procedures.” Purified proteins were resolved using SDS-PAGE andstained with Coomassie Blue (lane 2, Asn-1347 death domain; lane 4, mutant Ser-1347 death domain).E, quantitation of complex formation between ERK and death domain miniproteins. Purified death domainminiproteins were adsorbed on the solid phase and increasing amounts of ERK proteins were titrated intoreactions and added to microtiter wells to allow complex formation. The amount of ERK bound is representedas relative light units as a function of increasing ERK protein titrated.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

ations in death domain conformation might alter ERK bind-ing by as yet undefined determinants. The death domainmodules Fas and TNFR are capable of forming oligomers insolution and gave rise to the suggestion that they may formoligomeric scaffolds in vivo (24). However, this is not alwaysthe case, because the death domain of the neurotrophinreceptor remains monomeric in solution (25). We investigatedtherefore whether or not the DAPK-1 death domain can formoligomers and whether the Asn 3 Ser codon 1347 mutationalters this property.When the codon 1347 amino acid sequences in the death

domain of DAPK-1 are highlighted with respect to the putativeoligomeric structure, then it is striking that the localization isconfined to a putative interdomain interface between twomonomeric subunits (Fig. 3, A and B). Further, an inactivatingmutation in the death domain of murine Fas (lpr) resides at theN-terminal position of helix 3, which is distal to the position ofthe codon 1347 DAPK-1 death domain mutation at the C-ter-minal position in helix 3 (Fig. 3F). It therefore remained possi-ble that defects in ERK binding due to the codon 1347mutationmight relate to alterations in the oligomeric structure of thedeath domain due to helix 3 distortions. As such, the deathdomain encoded by the Asn-1347:Phe-1392 allele was firsttransfected into cells without co-transfected ERK to determinewhether themodule is, in fact, able to form oligomers. Potentialoligomers were fixed in vivo by the use of the membrane per-meable bi-functional cross-linking reagent EGS. After 24 h,transfected cells were incubatedwith EGS for various times andincreasing concentrations (data not shown) and lysates immu-noblotted to evaluate oligomerization status of the deathdomain. Under limiting cross-linking conditions, the majorityof the death domain remains monomeric (Fig. 3C, lane 2). Incontrast cells incubatedwith EGS exhibit significant cross-link-ing of the death domain as defined by higher molecular weightbands (Fig. 3C, lane 3 versus lane 2). These data are consistentwith the possibility that death domain of DAPK-1 can formoligomers (Fig. 3B).The respective death domain modules were subsequently

transfected into cells to determine whether the Asn 3 Sermutation confers changes in oligomerization. The Asn3 Sermutation might enhance oligomerization and preclude ERKdocking, or it might block quaternary or tertiary structure andreduce the extent of oligomerization that might be importantfor ERK interactions. After 24 h, cells were incubated with EGSunder conditions where the majority of the death domain

encoded by the Asn-1347 allele was assembled into higherorder oligomers and immunoblotted to evaluate oligomeriza-tion status of the Asn 3 Ser mutant protein. Death domainmodules containing the Leu or Phe mutation in the Asn-1347background displayed similar assembly into oligomeric inter-mediates and loss of the monomeric species (Fig. 3D, lanes 2and 4). By contrast, the protein modules containing the Ser-1347 mutation in either the Leu-1392 or Phe-1392 backgroundexhibited attenuated assembly into the multi-protein com-plexes with significant monomeric isoforms remaining (Fig.3D, lanes 3 and 5 versus 2 and 4).These data provide a mechanism to suggest how ERK bind-

ing might be attenuated by the Ser-1347 mutation as the deathdomain module becomes defective in its assembly into a multi-protein complex in cells. In fact, when the highly purified Asn-1347 death domainminiproteinwas applied to gel filtration, theprotein formed relatively high molecular weight oligomers(�600-kDa relative molecular mass) with an apparent equilib-rium between a lower molecular mass oligomer with an appar-ent molecular mass of �150 kDa (Fig. 3E, left panel). By con-trast, the death domain miniprotein encoded by the Ser-1347allele produced a population of molecules with a shift in theequilibrium between the large and smaller oligomers and themajority eluted later on the gel filtration column (Fig. 3E, rightpanel). The correlation between higher oligomeric structure ofthe death domain (Fig. 3E) and ERK binding (Fig. 2E) suggeststhat the ERK docking interface in the death domain is not justdefined by a simple linear epitope.Asn 3 Ser Mutation Attenuates the Biological Activity of

Full-lengthDAPK-1—The studies above evaluated the effects ofthe DAPK-1 mutation on the isolated activity of the deathdomain but do not address the effects of the mutation on full-length DAPK-1 function. Expression vectors encoding full-length HA-tagged DAPK-1 (Asn-1347:Phe-1392 allele and Ser-1347:Phe-1392) were developed and used to evaluate DAPK-1function in ERK binding and in oligomerization. The transfec-tion of either gene into cells resulted in similar steady-statelevels of DAPK-1 produced (Fig. 4A, lanes 2 and 3 versus 1).Further, the co-transfection of ERK into cells (Fig. 4B, lanes4–6) did not alter steady-state levels of DAPK-1 protein (Fig.4A, lanes 5 and 6). The immunoprecipitation of DAPK-1 pro-tein (Asn-1347:Phe-1392) using an anti-DAPK-1 antibody (Fig.4C, lanes 2 and 5) resulted in significant co-precipitation ofERK (Fig. 4D, lanes 5 versus 4). However, the immunoprecipi-tation of DAPK-1 encoded the Ser-1347:Phe-1392 allele using

FIGURE 3. The N1347S mutation in the death domain of DAPK-1 inhibits oligomerization of the module. A and B, predicted location of the codon 1347mutation (in green) based on the structure modeled with p75 Netrin receptor (A) and UNC5H2 (B). C, the Asn-1347 death domain forms oligomers in vivo. Cellswere transfected with expression vectors encoding the FLAG-tagged death domain and subject to in vivo cross-linking with EGS or Me2SO control for 1 h to limitthe extent of potential cross-linking of the death domain. Cells were harvested, lysed, and immunoblotted with anti-FLAG IgG to define changes in theassembly of the death domain into intermediate structures. D, the S1347death domain exhibits defects in oligomerization in cells. Cells were transfected withexpression vectors encoding the FLAG-tagged Asn-1347 variants (lanes 2 and 4) or Ser-1347 death domain variants (lanes 3 and 5) as indicated and subjectto in vivo cross-linking with EGS or Me2SO control under conditions where the majority of the death domain encoded by the Asn-1347 allele has beenassembled into oligomers (lanes 2 and 4). Cells were harvested and immunoprecipitated/immunoblotted as indicated under “Experimental Procedures” todefine changes in the assembly of the death domain into intermediate structures. E, the Ser-1347 death domain exhibits defects in oligomerization in vitro. Thepurified death domain miniproteins (as in Fig. 2D) were injected onto a gel filtration column as indicated under “Experimental Procedures.” Elution of the deathdomain proteins (Asn-1347 death domain module is in the left panel, and the Ser-1347 mutant death domain elution is depicted in the right panel) wasmeasured as a function of elution volume (in milliliters, as indicated) and integrated to molecular weight markers, as indicated by the arrows. F, the location ofthe DAPK-1 mutation in helix-3 of the death domain in relation to the Fas mutation lrp, which is known to denature the Fas module (27). The sequence of thehelix 3 region from mouse Fas, TNFR, and human DAPK-1 are indicated. The amino acids in red highlight the position of the mutation in Fas and TNFR,respectively (V238N and L1337N). The green amino acid in DAPK-1 (N1347S) highlights the position of the mutation relative to helix 3 (underlined).




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

an anti-DAPK-1 antibody (Fig. 4C, lanes 3 and 6) did not resultin significant co-precipitation of ERK (Fig. 4D, lane 6 versuslanes 4 and 5). These data are consistentwith the inability of theminimal death domain (encoded by the Ser-1347 allele) to bindto ERK in a co-precipitation assay and indicates that the singlepoint mutation in full-length DAPK-1 is sufficient to prevent itassembling into an ERK multi-protein complex in vivo.Although theN1347Smutation in theminimal death domain

revealed differences in oligomerization in vivo (Fig. 3), whetherthis mutation confers changes in full-length DAPK-1 assemblyin vivo requires examination. The DAPK-1 expression vector(Asn-1347:Phe-1392 allele) was first transfected into cells todetermine whether the module is able, in fact, to form oli-gomers. Potential oligomers were fixed in vivo, as above (Fig. 3)by the use of the membrane permeable cross-linking reagentEGS. After 24 h, transfected cells were incubated with EGS, andlysates were immunoblotted to evaluate the oligomerizationstate of full-length DAPK-1. Under these limiting cross-linkingconditions, where the majority of DAPK-1 remained uncross-linked, a significant pool of full-lengthDAPK-1 (Asn-1347:Phe-1392 allele) was cross-linked to a single high molecular weightspecies as well as exhibiting a “smear” into higher molecularweight cross-linked adducts (Fig. 4E, lane 5 versus 2). By con-trast, the full-length DAPK-1 vector encoding the Ser-1347:Phe-1392 allele did not exhibit any detectable cross-linking invivo (Fig. 4E, lane 6 versus 5). The immunoprecipitation of thecross-linkedDAPK-1 adducts resulted in significant proteolysisof the DAPK-1 oligomers (Fig. 4F), thus it was difficult to iden-

tify selected proteins in complex with the cross-linked full-length DAPK-1 protein. Nevertheless, these data highlight theeffect the Ser-1347 mutation has on full-length DAPK-1 multi-protein complex assembly in cells. Presumably this defect willbemanifest inDAPK-1 signaling assays, and a range of DAPK-1assays were developed to determine which intracellular path-ways are affected by the N1347S mutation.In defining the function for the death domain, it was origi-

nally suggested that themini-module is required forDAPK-1 toinduce cell death, because small peptides from this region act asdominant negative effectors of DAPK-1 apoptotic function(18), and the death domain can attenuate TNF-mediated apo-ptosis (19). Further, the lpr mutation in the murine deathdomain of Fas denatures the module and reduces intracellularsignaling induced by the TNF pathway (26, 27). We examinedwhether the mutation at codon 1347 attenuates the biologicalfunction of the death domain. A previous report has shown thatthe death domain of DAPK-1 (a Asn-1347:Leu-1392 allele) canattenuate p55-TNFR-dependent cell death (19). We evaluatedwhether the Asn 3 Ser codon 1347 mutation in the death

FIGURE 4. The N1347S mutation in full-length DAPK-1 attenuates ERKbinding. A–D, full-length DAPK-1 binding to ERK. HA-tagged ERK and full-length DAPK-1 expression vectors were transfected into HCT116 cells as indi-cated: steady-state levels of DAPK-1 (A) and ERK or DAPK-1 (B) immunopre-cipitated and immunoblotted for DAPK-1 (C) and ERK (D). E and F, defects inthe in vivo cross-linking of full-length DAPK-1 Ser-1347 allele in vivo. Cellswere transfected with expression vectors encoding the full-length DAPK-1alleles (lanes 2–3 and 5– 6) and subjected to in vivo cross-linking with EGS(lanes 4 – 6) or Me2SO control (lanes 1–3). Cells were harvested and immuno-precipitated followed by immunoblotting as indicated under “ExperimentalProcedures” to define changes in the assembly of the full-length DAPK-1 intointermediate structures.

FIGURE 5. The Ser-1347 mutation of DAPK-1 attenuates the pro-apopto-tic TNFR pathway. A and B, the mutant death domain is unable to attenuateTNFR signaling. Vectors encoding the respective death domain modules (5�g, Asn-1347:Phe-1392 or Ser-1347:Phe-1392) were co-transfected with p55-TNFR (0.5 �g) into HEK293 cells in 25-cm2 wells, and cells were harvested 16 hlater and analyzed for PARP cleavage: lane 1, vector controls only; lane 2,Asn-1347 death domain only; lane 3, p55 only; lane 4, p55 and Asn-1347 deathdomain; and lane 5, p55 and Ser-1347 death domain. Lysates were immuno-blotted with antibodies to PARP (A) or FLAG-tagged death domain (B). C, themutant full-length DAPK-1 has an attenuated cooperation with TNFR signal-ing. Vectors encoding the respective full-length DAPK-1 alleles (5 �g, Asn-1347:Phe-1392 or Ser-1347:Phe-1392) were co-transfected with p55-TNFR(0.5 �g) into HEK293 cells in 25-cm2 wells, and cells were harvested 16 h laterand analyzed for PARP cleavage: In C: lane 1, vector control; lane 2, TNFR; lane3, TNFR plus N1347:DAPK-1; and lane 4, TNFR plus S1347:DAPK-1. Lysateswere immunoblotted with antibodies to PARP (C) or HA (D).




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

domain affected the ability of the mini-module to act as a dom-inant-negative effector and block TNF-dependent signaling.Transfection into HEK293 cells of p55-TNFR induced signifi-cant PARP cleavage (Fig. 5A, lane 3 versus 2), which is a classicmolecular indicator of apoptosis being triggered. TheAsn-1347death domain module was able to attenuate PARP cleavage(Fig. 5A, lane 4 versus 3), which is consistent with previousreports showing that the death domain can attenuate p55-in-duced apoptosis (19). By contrast, the Ser-1347 miniproteinwas unable to attenuate p55-TNFR signaling (Fig. 5A, lane 5versus 3) suggesting that the mutation produces an unstablemodule that cannot act as a dominant negative effector of theDAPK-1-dependent TNFR signaling pathway. Transfectedamounts of the death domain modules are depicted in Fig. 5B.Full-length DAPK-1 expression vectors encoding the Asn or

Ser amino acid at codon 1347 were also tested in TNFR-medi-ated pro-apoptotic assays. Transfection of the TNFR-1 alonegave rise to significant increases in PARP cleavage (Fig. 5C,lanes 2 versus 1). The co-transfection of full-length DAPK-1

(N1347) inducedmarginal increasesin the levels of PARP cleavage (Fig.5C, lanes 3 versus 2), whereas thefull-length DAPK-1 encoded by theSer-1347 allele reduced basalTNFR-1-mediated PARP cleavage(Fig. 5C, lane 4 versus 2). Trans-fected amounts of full-lengthDAPK-1 proteins are depicted inFig. 5D. These data together indi-cate that the 1347 codon substitu-tion can alter the pro-apoptoticactivity of DAPK-1. However, thedeath domain module is not knownto recruit ERK in the TNFR-medi-ated signaling pathway, and the keybiochemical defect we have identi-fied in the Ser-1347 allele was inac-tivity in ERK binding. As such, weinvestigated whether the Ser-1347mutation altered the pro-apoptoticactivity of DAPK-1 in an ERK-de-pendent apoptotic cell system.The Asn 3 Ser Mutation in

DAPK-1 Inhibits ERK-mediatedApo-ptosis—To evaluate whether thedeath domain mutation alters ERK-mediatedapoptosis,we firsthad to setup a cell-based system that canmeas-ure MEK-dependent activation ofERK, as defined by elevated ERKphosphorylation. The transfection ofactivated MEK (MEK EE) into cellsinduced elevated phosphorylation ofendogenous ERK (Fig. 6A, lane 2 ver-sus 3), relative to total ERK protein(Fig. 6B). Further, the co-transfectionof activated MEK with ERK elevatedphosphorylation of the transfected

ERK protein (Fig. 6A, lanes 4 versus 1). Having defined anassay where ERK can be activated by MEK, a series of assayswere developed that addressed whether the ERK:DAPK-1pro-apoptotic pathway is altered by the N1347S mutation inDAPK-1. The first assay evaluated changes in Annexin Vstaining, which is an indicator of early events in apoptosis.The transfection of DAPK-1, MEK-ERK, or MEK-ERK-DAPK-1 induced differential changes in Annexin staining(Fig. 6C). The DAPK-1 Asn-1347 allele inducedmore AnnexinV staining than the DAPK-1 Ser-1347 allele (Fig. 6C, lanes 2versus 3). Further, MEK-ERK co-transfection with theAsn-1347 allele of DAPK-1 allele induced the most significantchanges in Annexin V staining, relative to the MEK-ERK co-transfection with the Ser-1347 allele of DAPK-1 (Fig. 6C, lanes 5versus6).Consistentwith this, immunoblotting forPARPcleavageas a distinct measure of apoptosis also revealed that the Asn-1347DAPK-1 allele was more effective at inducing a pro-apoptoticpathway in cooperation with MEK-ERK, relative to the DAPK-1Ser-1347 allele (Fig. 6D, lanes 5 versus 6).

FIGURE 6. ERK-mediated apoptosis is inhibited by the N1347S codon mutation in DAPK-1. A and B, acti-vation of ERK by MEK transfection. HCT116 cells were transfected as indicated with ERK, MEK, or a combinationof the two. Total ERK levels and extent of ERK phosphorylation were measured by immunoblotting with apan-specific antibody (B) and phospho-specific antibody (A), respectively. C, the N1347S death domain muta-tion attenuates ERK-DAPK-1-dependent changes in Annexin V staining. HCT116 cells were transfected asindicated with ERK, full-length DAPK-1 (Asn-1347 and Ser-1347 alleles), MEK, a combination of the two, or allthree together. Annexin V staining was measured as indicated under “Experimental Procedures” and is plottedas the percentage of cells in the lower right quadrant as a function of gene transfected. D, the N1347S deathdomain mutation attenuates ERK-DAPK-1-dependent changes in PARP cleavage. HEK293 cells were trans-fected as indicated with ERK, full-length DAPK-1 (Asn-1347 and Ser-1347 alleles), MEK, a combination of thetwo, or all three together. Full-length PARP or proteolyzed PARP was measured by immunoblotting and high-lighted as indicated by the arrows.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

The final assay used to determinewhether the DAPK-1 death domainmutation at codon 1347 alteredMEK-ERK-DAPK-1 pro-apoptoticsignaling was TUNEL, which meas-ures changes in chromatin structureand inducedDNAbreaks. Similar tothat observedwithAnnexinV stain-ing or PARP proteolysis, theco-transfection of MEK-ERK, andAsn-1347 allele of DAPK-1 inducedthe most pronounced DNA cleav-age (Fig. 7D; summarized in Fig.7G). By contrast, the DAPK-1 Ser-1347 allele was defective in theinduction of DNA damage in coop-eration withMEK and ERK (Fig. 7F;summarized in Fig. 7G). Together,these data provide evidence that thedeath domain codon N1347S mis-sense mutation can reduce deathdomain oligomerization, reduceERK binding, and prevent acti-vated ERK-mediated apoptosis.N1347S Allele Is a Novel Germ

Line Mutation in DAPK-1—Ourcloning from a tumor cell line, witha wild-type p53 pathway, of amutant death domain allele at bothcodons 1347 and 1392 promptedthe biochemical characterizationsummarized above to determinewhether the mutations would haveany functional consequences.Although the Leu 3 Phe 1392codon mutation did not reveal anydefects, significant biochemical dif-ferences were observed by theAsn 3 Ser codon 1347 mutationand as such, we set out to determinewhether the codon 1347 mutationwas tumor-derived or a germ line,because the sequence differs fromthe published sequence (Asn-1347:Leu-1392). The alleles might havebeen tumor-derived due to selec-tion pressures in cancer cells, and ifso we would need to examine muta-tion frequency in a range of cancersto define the penetrance. Alterna-tively, the mutations might be germline mutations in patients, and wewould therefore need to sequencenormal genomic DNA to deter-mine whether or not the mutationsare rare or common. Preliminarydata indicated that the Asn to Sermutation at codon 1347 was not

FIGURE 7. ERK-mediated apoptosis as defined using a TUNEL assay is inhibited by the N1347S codonmutation in DAPK-1. A–F, the N1347S death domain mutation attenuates ERK-DAPK-1-dependentchanges in TUNEL staining. HCT116 cells were transfected as indicated with vector only, ERK, full-lengthDAPK-1 (Asn-1347 and Ser-1347 alleles), MEK, a combination of the two, or all three together. Cells werefixed for TUNEL assay as indicated under “Experimental Procedures,” and representative images are high-lighted. G, quantitation of apoptosis frequency using TUNEL from A–F. The percentages of apoptotic cellswere calculated relative to the total number of cells using Image J software (National Institutes of Health).H, summary of two distinct stages where ERK-DAPK-1 axis can contribute to cell survival or apoptosis.DAPK-1 can function as a survival factor (43) or a pro-apoptotic factor (22). When ERK activation inducesphosphorylation of p90RSK kinase, this attenuates DAPK-1-dependent apoptotic function and induces asurvival pathway (40). By contrast, if ERK activation triggers a direct docking-dependent interaction withthe death domain of DAPK-1, then this triggers a pro-apoptotic pathway. The naturally occurring N1347Smutation allows a direct test and independent evidence supporting that the death domain is required forERK-mediated apoptosis.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

tumor-derived, because it was detected upon sequencing nor-mal genomic DNA from selected patient sets (data not shown,see below). Interestingly, upon examination of the mutationfrequency at codon 1392 using normal genomic DNA, we weresurprised to find no evidence for the Leu-1392 allele, whichwasthe original cDNA expressed from the DAPK-1 locus (23). Inover 250 different genomic DNA samples analyzed, the aminoacid encoded by codon 1392 was phenylalanine (Fig. 8A, datanot shown). Upon examination of the death domain sequencein other species, the Asn-1349:Phe-1392 allele was the onlysequence observed (Fig. 8D), suggesting that this might be thefounding member in the human population. Because the Leu-1392 mutation has not yet been observed in our analyses, itmight have originally been a tumor cell line derivedmutation ortheremay be other genetic backgrounds which have harbored aSNP change at this site.To evaluate the frequency of codon 1347 mutation in the

population, genomic DNA was first sequenced from distinctpatient groups to acquire more statistically significant data onthe relative frequency of the alleles in the general population. Inthe first group of patients genomic DNA from normal volun-teers was sequenced to determine the relative frequency of the1347 codon mutation. In the normal control group, patientshomozygous for the amino acid asparagine at codon 1347 were

represented at 37.8%, patients heterozygous for asparagine andserine at codon 1347were observed at a frequency of 47.3%, andpatients homozygous for serine at codon 1347were observed at14.9% (Fig. 8B). In a second set of patients, we also evaluated thepenetrance of the codon 1347 alleles in normal genomic DNAfrom cancer patients with a BRCA1 germ line mutation (Fig.8C). In this sample group, patients homozygous for the aminoacid asparagine at codon 1347 were represented at 45.3%,patients heterozygous for asparagine and serine at codon 1347were observed at a frequency of 41.9%, and patients homozy-gous for serine at codon 1347 were observed at 12.8% (Fig. 8C).The bias against a higher frequency for the homozygousmutantSer-1347 codon in both groups might suggest a selection pres-sure against the double mutant Ser-1347 allele; however, largerpopulation analysis is required for this to be substantiated. Ifthere were selections against the homozygous mutant alleles,then thismight suggest a toxic role for this homozygous allele inDAPK-1 mediated development. Regardless, the data indicatethat there can be a relatively high degree of the 1347 codonmutation in a patient group. Because the founding allele inhumans is apparentlyAsn-1347:Phe-1392 (Fig. 8D), we proposethat the codon Asn-1347:Phe-1392 is the wild-type allele andthe Ser-1347:Phe-1392 allele is SNP allele based on divergencefrom the conserved death domain sequence (Fig. 8D). The SNPis also inactivating based on its inability to function biochemi-cally in ERK-dependent pathways (Figs. 2, 4, 6, and 7).However,whether the SNP alters DAPK-1 interaction in other signalingpathways remains to be determined.

DISCUSSION

Key genetically defined activators of p53 are members of thecalcium calmodulin kinase superfamily and include ATM-CHK2, DAPK-1, and LKB-AMPK. Mutations exist in genesencoding ATM, CHK2, and LKB that can attenuate the p53response, but DAPK-1 is not known to be mutated in humancancers. However, DAPK-1 is thought to function like a tumorsuppressor in that its gene expression can be attenuated bypromotermethylation (38). DAPK-1was originally identified inan antisense screen as a gene implicated in interferon-�-induced apoptosis in HeLa cells (23). DAPK-1 is now known tobe involved in a TNF-mediated apoptotic pathway (19), sup-pression of metastasis via integrin degradation (17, 39), anoncogene-driven p53-dependent checkpoint pathway (7), andits apoptotic function is either inhibited by a RSK-dependentpathway or stimulated by an ERK-mediated pathway (22, 40).The attenuation of the pro-apoptotic function of DAPK-1 bypromoter methylation is one of the keymechanisms thought toreduce DAPK-1 tumor suppressor activity. However, in renalcancers, the DAPK-1 gene is hypermethylated, yet significantexpression of the protein is retained suggesting post-transla-tional mechanismsmight exist for reducing the specific activityof the protein (28), possibly by RSK (40). We also surprisinglyshow relatively high level expression of DAPK-1 protein inbreast cancers (data not shown) suggesting the protein expres-sion can be readily achieved in vivo, despite the fact that thegene is generally held to be silenced in cancers. In this report,we identify a novel germ line mutation in DAPK-1 in a func-tionally significant region of the protein, the pro-apoptotic

FIGURE 8. The codon 1347 allele in the death domain of DAPK-1 is a com-mon germ line mutation. A–C, mutation frequency at codon 1392 and 1347from genomic DNA. Sequencing of the death domain was performed afteramplification of the death domain using primers as indicated under “Experi-mental Procedures.” The data are tabulated as percentages of samples, whichwere homozygous or heterozygous with respect to the 1347 (B and C) and1392 codons (A). D, primary structure of the death domain. Homology of thethree human death domain modules is highlighted relative to primate andmurine DAPK-1 death domains. The Asn-1347 and Phe-1392 amino acids arepresent in these vertebrates suggesting this Asn-1347:Phe-1392 allele is thelikely founding gene in Homo sapiens and that the Ser-1347:Phe-1392 allele istherefore the mutant version of the DAPK-1 gene.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

death domain, which might have implications for post-transla-tional attenuation of the DAPK-1 tumor suppressor pathway indiseases with modified pro-apoptotic pathways.Functional domains of the DAPK-1 protein have been iso-

lated using a genetic screen searching for mini-domains ofDAPK-1 that act in a dominant negative function overDAPK-1,and fourmodules were found, including the death domain (18).The death domain serves as a docking site for ERK thus cata-lyzing DAPK-1 phosphorylation by ERK at Ser-735 (22), andthe death domain allows, in part, DAPK-1 assembly into anUNC5H2 multiprotein complex (20). The structure of thedeath domain is predicted based on superimposition of theamino acids with homology to other members of the deathdomain, including the neurotrophin receptor, UNC5H2, Fas,and FADD (Fig. 3). The generic death domain contains twosets of �-helices arranged with three helices packed parallelin one plane perpendicular to another packing of three hel-ices. The Asn3 Ser mutation at codon 1347 is predicted tooccur at the edge of a putative linker between helix 3 and 4.In this report, we provide evidence that the DAPK-1 deathdomain is oligomeric in vitro and in vivo and that a germ linemutation can attenuate the extent of its oligomerization andERK docking.Apoptosis plays important physiological functions in pro-

grammed cell death during fetal development, maintaining tis-sue integrity for example as a tumor-suppressingmechanismorin ischemic injury-dependent repair, immune system develop-ment, and as a defense mechanism to combat infection byviruses and micro-organisms. Defects in the apoptoticresponses are therefore responsible for disease development,including cancer progression, ischemic injury responses, andimmune system dysfunction. There is growing realization thatthe apoptotic program is maintained by a proteome-like net-work where various nodes are linked by conserved proteindomainmodules that serve as signatures for allowing cross-talkbetween the various interacting signals. Some examples of theconserved pro-apoptotic domains include: death domains,death effector domain, the caspase-recruitment domains, CIDEdomains (cell-death inducing FDD45-like effectors), and inhib-itors of apoptosis domains. The death domain is a relativelyabundant protein module thought to allow for protein-proteincomplex assembly and propagate signaling, classically, in theNF-�B and TNF family of cytokine receptors. The deathdomain protein interaction module has evolved combinatori-ally into larger proteins implicated in apoptosis that harborcaspase-like domains, IgG-like folds, leucine zippers, TNFRdomains, andRel-homology domains (24).Mice having amuta-tion (named lpr) in the death domain of Fas exhibit a lupus-likelymphoproliferative autoimmune disorder (26). The lprmuta-tion results in Fas death domain denaturation (27), whichexplains why the lpr Fas domain is biochemically inactive.The links between ERK activation, DAPK-1 signaling, and

p53 function are only beginning to be unraveled. ERK isrequired to activate p53 gene expression and maintain p53pathway activation status (44). Further, DAPK-1 activates p53in an oncogene signaling pathway by an undefined mechanism(7). As such, ERK may co-ordinate p53 and DAPK-1 pathwaysto maintain p53 function as a tumor suppressor in response to

oncogenic signaling. Two distinct ERK-dependent pathwayscontrol the specific activity of DAPK-1 (Fig. 7H). The moreclassic ERK-dependent pro-survival pathway can be stimulatedby ERK activation of RSK kinase, which in turn attenuatesDAPK-1 pro-apoptotic activity (40). Opposing this pathway(Fig. 7H), direct interaction of ERK with the death domain ofDAPK-1 can stimulate the pro-apoptotic activity of the ERK:DAPK-1 axis. In support of this model, a naturally occurringpolymorphism in the death domain at codon 1347 reduces ERKbinding and precludes ERK-triggered apoptosis (Fig. 7H). Pre-sumably, this would also reduce the specific activity of p53when its function is being triggered by the DAPK-1 oncogenicsignaling pathway.Although inactivating mutations in the tumor suppressor

p53 gene are one of the most common genetic changes inhuman cancers (29–31), a germ line mutation exists in the p53gene at codon 72 within the transactivation domain-II that canalter apoptotic rates, cancer incidence, and anti-cancer drugresponses (32, 33). Interestingly, a relatively rare germ linemutation exists in the p53 gene at a MAPK kinase phosphoryl-ation site near the p53 transactivation domain at Ser-47. Thismutation can attenuate p53 phosphorylation by MAPK andreduce the induction of apoptosis (34). The penetrance of thePro3 Ser codon 47 germ line mutation is apparently confinedto African-American populations and the frequency of the S47allele in this population is �1–4% (35). This is considerablylower than the codon 72 mutation, which occurs at approxi-mately similar ratios. Loss of heterozygosity at the p53 locus isrelatively frequent in human cancers, and patients heterozy-gous for the codon 72mutation preferentially lose the P72 allelein cancers (36), suggesting a selective advantage in a tumor forretaining the pro-apoptotic R72 allele. Other germ line muta-tions in p53 regulators can also have an impact of p53 pathwayfunction. p53 activity is normally kept low by an MDM2-de-pendent pathway, and germ line mutations in theMDM2 pro-moter that result in enhancedMDM2protein production resultin an attenuated p53 pathway as well as increase in the age ofonset for particular types of cancer (37).It is possible that the death domain mutation in DAPK-1

might confer reduced apoptotic potential to the DAPK-1 path-way in vivo, and this might accelerate diseases where the p53-dependent or p53-independent apoptotic function of DAPK-1maintains tissue integrity. Attenuated apoptosis in injured neu-rons, developing cancers, or in immune cell integrity mightalter biological responses in vivo (41). However, given oursequencing data showing the death domain mutation showssignificant penetrance in one group of patients, the question israised as to whether there have been selection pressures formaintaining the mutant allele in some populations duringhuman evolution. Preliminary sequencing data in a distinctEuropean population has surprisingly found no evidence of theSer-1347 allele (data not shown). DAPK-1-mediated function isalso involved in signaling pathways, including stress-inducedautophagy (16), TNF-mediated apoptosis (19), ERK-dependentapoptosis (22), and ischemic-mediated injury in neuronal cells(42). Maintenance of the mutant Ser-1347 allele in the popula-tion might be linked to responses to infection and immunityrelated to fitness. Attenuated apoptosis conferred by the




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

DAPK-1 death domain mutation might maintain T-cellresponses to some infections and delay loss of cells involved inimmunity or it might attenuate cell death involved in neuronaldevelopment, injury, or repair thus altering tissue integrity.Determining where DAPK-1 functions in normal and diseasedconditions in vivo, and the signaling pathways that activate theprotein under these conditions, will facilitate understanding ofthe effects of the DAPK-1 death domain mutation on tissuephysiology.

Acknowledgment—We thank Steve Keyse for his enthusiastic supportand advice.

REFERENCES1. Harris, S. L., Gil, G., Robins, H., Hu,W., Hirshfield, K., Bond, E., Bond, G.,

and Levine, A. J. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 16297–163022. Brooks, C. L., and Gu, W. (2006)Mol. Cell 21, 307–3153. Levine, A. J., Hu,W., and Feng, Z. (2006)Cell Death Differ. 13, 1027–10364. Lane, D. P., and Hupp, T. R. (2003) Drug Discovery Today 15, 347–3555. Jones, R. G., Plas, D. R., Kubek, S., Buzzai, M., Mu, J., Xu, Y., Birnbaum,

M. J., and Thompson, C. B. (2005)Mol. Cell 18, 283–2936. Takai, H., Naka, K., Okada, Y.,Watanabe,M., Harada, N., Saito, S., Ander-

son, C.W., Appella, E., Nakanishi,M., Suzuki, H., Nagashima, K., Sawa,H.,Ikeda, K., and Motoyama, N. (2002) EMBO J. 21, 5195–5205

7. Raveh, T., Droguett, G., Horwitz, M. S., DePinho, R. A., and Kimchi, A.(2001) Nat. Cell Biol. 3, 1–7

8. Craig, A. L., and Hupp, T. R. (2004) Oncogene 23, 8411–84189. Lambert, P. F., Kashanchi, F., Radonovich, M. F., Shiekhattar, R., and

Brady, J. N. (1998) J. Biol. Chem. 273, 33048–3305310. Dornan, D., and Hupp, T. R. (2001) EMBO Rep. 2, 139–14411. Dornan, D., Shimizu, H., Burch, L., Smith, A. J., and Hupp, T. R. (2003)

Mol. Cell. Biol. 23, 8846–886112. Barlev, N. A., Liu, L., Chehab, N. H., Mansfield, K., Harris, K. G., Hala-

zonetis, T. D., and Berger, S. L. (2001)Mol. Cell 8, 1243–125413. Hardie, D. G. (2005) Curr. Opin. Cell Biol. 17, 167–17314. Bialik, S., Bresnick, A. R., and Kimchi, A. (2004) Cell Death Differ. 11,

631–64415. Inbal, B., Bialik, S., Sabanay, I., Shani, G., and Kimchi, A. (2002) J. Cell Biol.

157, 455–46816. Gozuacik, D., and Kimchi, A. (2004) Oncogene 23, 2891–290617. Wang,W. J., Kuo, J. C., Yao, C. C., and Chen, R. H. (2002) J. Cell Biol. 159,

169–17918. Raveh, T., Berissi, H., Eisenstein, M., Spivak, T., and Kimchi, A. (2000)

Proc. Natl. Acad. Sci. U. S. A. 97, 1572–157719. Cohen, O., Inbal, B., Kissil, J. L., Raveh, T., Berissi, H., Spivak-Kroizaman,

T., Feinstein, E., and Kimchi, A. (1999) J. Cell Biol. 146, 141–14820. Llambi, F., Laurenco, F. C., Gozuacik, D., Guix, C., Pays, L., Del Rio, G.,

Kimchi, A., and Mehlen, P. (2005) EMBO J. 24, 1192–120121. Chen, G., and Goeddel, D. V. (2002) Science 296, 1634–1635

22. Chen, C. H.,Wang,W. J., Kuo, J. C., Tsai, H. C., Lin, J. R., Chang, Z. F., andChen, R. H. (2005) EMBO J. 24, 294–304

23. Cohen, O., Feinstein, E., and Kimchi, A. (1997) EMBO J. 16, 998–100824. Boldin, M. P., Mett, I. L., Varfolomeev, E. E., Chumakov, I., Shemer-Avni,

Y., Camonis, J. H., and Wallach, D. (1995) J. Biol. Chem. 270, 387–39125. Liepinsh, E., Ilag, L. L., Otting, G., and Ibanez, C. F. (1997) EMBO J. 16,

4999–500526. Watanabe-Fukunaga, R., Brannan, C. I., Copeland, N. G., Jenkins, N. A.,

and Nagata, S. (1992) Nature 356, 314–31727. Eberstadt, M., Huang, B., Olejniczak, E. T., and Fesik, S. W. (1997) Nat.

Struct. Biol. 4, 983–98528. Wethkamp,N., Ramp,U., Geddert, H., Schulz,W.A., Florl, A. R., Suschek,

C. V., Hassan, M., Gabbert, H. E., and Mahotka, C. (2006) Eur. J. Cancer42, 264–274

29. Hardie, L. J., Darnton, S. J., Wallis, Y. L., Chauhan, A., Hainaut, P., Wild,C. P., and Casson, A. G. (2005) Cancer Lett. 217, 221–230

30. Zalcenstein, A., Stambolsky, P., Weisz, L., Muller, M., Wallach, D., Gon-charov, T. M., Krammer, P. H., Rotter, V., and Oren, M. (2003) Oncogene22, 5667–5676

31. Kakudo, Y., Shibata, H., Otsuka, K., Kato, S., and Ishioka, C. (2005)CancerRes. 65, 2108–2114

32. McGregor, J. M., Harwood, C. A., Brooks, L., Fisher, S. A., Kelly, D. A.,O’Nions, J., Young, A. R., Surentheran, T., Breuer, J., Millard, T. P., Lewis,C. M., Leigh, I. M., Storey, A., and Crook, T. (2002) J. Invest. Dermatol.119, 84–90

33. Sullivan, A., Syed, N., Gasco, M., Bergamaschi, D., Trigiante, G., Attard,M., Hiller, L., Farrell, P. J., Smith, P., Lu, X., and Crook, T. (2004)Oncogene23, 3328–3337

34. Li, X., Dumont, P., Della Pietra, A., Shetler, C., and Murphy, M. E. (2005)J. Biol. Chem. 280, 24245–24251

35. Murphy, M. E. (2006) Cell Death Differ. 13, 916–94036. Brooks, L. A., Tidy, J. A., Gusterson, B., Hiller, L., O’Nions, J., Gasco, M.,

Marin, M. C., Farrell, P. J., Kaelin, W. G., Jr., and Crook, T. (2000) CancerRes. 60, 6875–6877

37. Bond, G. L., Hu, W., Bond, E. E., Robins, H., Lutzker, S. G., Arva, N. C.,Bargonetti, J., Bartel, F., Taubert, H., Wuerl, P., Onel, K., Yip, L., Hwang,S. J., Strong, L. C., Lozano, G., and Levine, A. J. (2004) Cell 119, 591–602

38. Levy, D., Plu-Bureau, G., Decroix, Y., Hugol, D., Rostene, W., Kimchi, A.,and Gompel, A. (2004) Clin. Cancer Res. 10, 3124–3130

39. Inbal, B., Cohen, O., Polak-Charcon, S., Kopolovic, J., Vadai, E., Eisenbach,L., and Kimchi, A. (1997) Nature 390, 180–184

40. Anjum, R., Roux, P. P., Ballif, B. A., Gygi, S. P., and Blenis, J. (2005) Curr.Biol. 15, 1762–1767

41. Velentza, A. V., Schumacher, A. M., and Watterson, D. M. (2002) ExpertOpin. Ther. Targets 6, 497–506

42. Velentza, A. V., Wainwright, M. S., Zasadzki, M., Mirzoeva, S., Schuma-cher, A. M., Haiech, J., Focia, P. J., Egli, M., and Watterson, D. M. (2003)Bioorg Med. Chem. Lett. 13, 3465–3470

43. Jin, Y., and Gallagher, P. J. (2003) J. Biol. Chem. 278, 51587–5159344. Agarwal, M. L., Ramana, C. V., Hamilton, M., Taylor, W. R., DePrimo,

S. E., Bean, L. J., Agarwal, A., Agarwal,M. K.,Wolfman, A., and Stark, G. R.(2001) Oncogene 20, 2527–2536




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

Vicky Durston, Diana M. Eccles and Ted HuppCraig Stevens, Yao Lin, Maria Sanchez, Eliana Amin, Ellen Copson, Helen White,

Apoptosis Inactivates ERK-inducedDAPK-1A Germ Line Mutation in the Death Domain of

doi: 10.1074/jbc.M605649200 originally published online January 23, 20072007, 282:13791-13803.J. Biol. Chem.

10.1074/jbc.M605649200Access the most updated version of this article at doi:

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

http://www.jbc.org/content/282/18/13791.full.html#ref-list-1

This article cites 44 references, 20 of which can be accessed free at


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/lookup/doi/10.1074/jbc.M605649200

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;282/18/13791&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/282/18/13791

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=282/18/13791&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/282/18/13791

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/282/18/13791.full.html#ref-list-1

http://www.jbc.org/

a germ line mutation in the death domain of dapk-1 inactivates

Documents