impact, a protein preferentially expressed in the mouse brain, … · 2005-06-02 · impact, a...

22
Pereira et al. IMPACT, a GCN2 inhibitor in mammals 1 IMPACT, a protein preferentially expressed in the mouse brain, binds GCN1 and inhibits GCN2 activation Cátia M. Pereira 1 , Evelyn Sattlegger 2, Hao-Yuan Jiang 3 , Beatriz M. Longo 4# , Carolina B. Jaqueta 4 , Alan G. Hinnebusch 2 , Ronald C. Wek 3 , Luiz E. A. M. Mello 4 and Beatriz A. Castilho 1 * 1 Departamento de Microbiologia, Imunologia e Parasitologia and 4 Departamento de Fisiologia, Universidade Federal de São Paulo, São Paulo, SP, 04023-062, Brasil, 2 Laboratory of Gene Regulation and Development, National Institute of Child Health and Development, NIH, Bethesda, MD, 20892-2427, U.S.A.,and 3 Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, Indianapolis, IN, 46202, U.S.A. Running title: IMPACT, a GCN2 inhibitor in mammals *To whom correspondence should be addressed: Rua Botucatu, 862 3º andar, São Paulo, SP 04023- 062, Brasil Phone: (55)(11) 5576-4537; Fax: (55)(11) 5572-4711; E-mail: [email protected] Translational control directed by the eIF2α kinase GCN2 is important for coordinating gene expression programs in response to nutritional deprivation. The GCN2 stress response is conserved from yeast to mammals and is critical for resistance to nutritional deficiencies and for the control of feeding behaviors in rodents. IMPACT, a mouse imprinted gene, has sequence similarities with the yeast YIH1 protein, an inhibitor of GCN2. YIH1 competes with GCN2 for binding to a positive regulator, GCN1. In this work, we present evidence that IMPACT is the functional counterpart of YIH1. Overexpression of IMPACT in yeast lowers both basal and amino acid starvation-induced levels of phosphorylated eIF2α , as previously described for YIH1. Overexpression of IMPACT in mouse embryonic fibroblasts inhibits phosphorylation of eIF2α by GCN2 under leucine starvation, abolishing expression of its downstream target genes, ATF4(CREB-2) and CHOP(GADD153). IMPACT binds to the minimal yeast GCN1 segment required for the interaction with yeast GCN2 and YIH1, and to mouse native GCN1. At the protein level, IMPACT was detected mainly in the brain. IMPACT is abundant in the majority of hypothalamic neurons. Scattered neurons expressing this protein at higher levels were detected in other regions, such as the hippocampus and the piriform cortex. The abundance of IMPACT is inversely correlated with eIF2α (P) levels in different brain areas. These results taken together thus suggest that IMPACT ensures constant high levels of translation and low levels of ATF4 and CHOP in specific neuronal cells under amino acid starvation conditions. ________________________________________ The control of protein synthesis plays an important role in diverse physiological conditions, as part of homeostatic mechanisms and long term memory formation, and in pathological conditions such as diabetes, brain ischemia, epilepsy and other neurodegenerative disorders (1-8). At the cellular level, many signaling networks that affect the rate of protein synthesis involve the phosphorylation of Ser 51 in the α subunit of eukaryotic translation initiation factor 2 (eIF2α) by a family of protein kinases that are each activated by different cellular stress conditions. The heterotrimeric factor eIF2 is responsible for binding the initiator methionyl-tRNAi Met , in a GTP dependent mode, and to deliver it to the 40S ribosomal subunit. When the initiator AUG codon is encountered, eIF2 is released in a GDP bound form, with the subsequent formation of the 80S elongating ribosome. The exchange of GDP to JBC Papers in Press. Published on June 2, 2005 as Manuscript M408571200 Copyright 2005 by The American Society for Biochemistry and Molecular Biology, Inc. by guest on March 30, 2020 http://www.jbc.org/ Downloaded from

Upload: others

Post on 23-Mar-2020

5 views

Category:

Documents


0 download

TRANSCRIPT

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

1

IMPACT, a protein preferentially expressed in the mouse brain, binds GCN1 andinhibits GCN2 activation

Cátia M. Pereira1, Evelyn Sattlegger2♦♦♦♦, Hao-Yuan Jiang3, Beatriz M. Longo4#, Carolina B.Jaqueta4, Alan G. Hinnebusch2, Ronald C. Wek3, Luiz E. A. M. Mello4 and Beatriz A. Castilho1*

1Departamento de Microbiologia, Imunologia e Parasitologia and 4Departamento de Fisiologia,Universidade Federal de São Paulo, São Paulo, SP, 04023-062, Brasil, 2Laboratory of Gene Regulation

and Development, National Institute of Child Health and Development, NIH, Bethesda, MD, 20892-2427,U.S.A.,and 3Department of Biochemistry and Molecular Biology, Indiana University School of Medicine,

Indianapolis, IN, 46202, U.S.A.

Running title: IMPACT, a GCN2 inhibitor in mammals

*To whom correspondence should be addressed: Rua Botucatu, 862 3º andar, São Paulo, SP 04023-062, Brasil Phone: (55)(11) 5576-4537; Fax: (55)(11) 5572-4711; E-mail: [email protected]

Translational control directed by the eIF2ααααkinase GCN2 is important for coordinatinggene expression programs in response tonutritional deprivation. The GCN2 stressresponse is conserved from yeast to mammalsand is critical for resistance to nutritionaldeficiencies and for the control of feedingbehaviors in rodents. IMPACT, a mouseimprinted gene, has sequence similarities withthe yeast YIH1 protein, an inhibitor of GCN2.YIH1 competes with GCN2 for binding to apositive regulator, GCN1. In this work, wepresent evidence that IMPACT is the functionalcounterpart of YIH1. Overexpression ofIMPACT in yeast lowers both basal and aminoac id s tarvat ion- induced leve l s o fphosphorylated eIF2αααα, as previously describedfor YIH1. Overexpression of IMPACT inmouse embryonic fibroblasts inhibitsphosphorylation of eIF2αααα by GCN2 underleucine starvation, abolishing expression of itsdownstream target genes, ATF4(CREB-2) andCHOP(GADD153). IMPACT binds to theminimal yeast GCN1 segment required for theinteraction with yeast GCN2 and YIH1, and tomouse native GCN1. At the protein level,IMPACT was detected mainly in the brain.IMPACT is abundant in the majority ofhypothalamic neurons. Scattered neuronsexpressing this protein at higher levels were

detected in other regions, such as thehippocampus and the piriform cortex. Theabundance of IMPACT is inversely correlatedwith eIF2αααα(P) levels in different brain areas.These results taken together thus suggest thatIMPACT ensures constant high levels oftranslation and low levels of ATF4 and CHOPin specific neuronal cells under amino acidstarvation conditions.________________________________________

The control of protein synthesis plays animportant role in diverse physiological conditions,as part of homeostatic mechanisms and long termmemory formation, and in pathological conditionssuch as diabetes, brain ischemia, epilepsy andother neurodegenerative disorders (1-8). At thecellular level, many signaling networks that affectthe rate of protein synthesis involve thephosphorylation of Ser5 1 in the α subunit ofeukaryotic translation initiation factor 2 (eIF2α)by a family of protein kinases that are eachactivated by different cellular stress conditions.

The heterotrimeric factor eIF2 is responsiblefor binding the initiator methionyl-tRNAiMet, in aGTP dependent mode, and to deliver it to the 40Sribosomal subunit. When the initiator AUG codonis encountered, eIF2 is released in a GDP boundform, with the subsequent formation of the 80Selongating ribosome. The exchange of GDP to

JBC Papers in Press. Published on June 2, 2005 as Manuscript M408571200

Copyright 2005 by The American Society for Biochemistry and Molecular Biology, Inc.

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

2

GTP on eIF2, to allow for further rounds ofinitiation, is catalyzed by the guanine nucleotideexchange factor, eIF2B. Phosphorylation of eIF2αcan have a profound inhibitory effect on overallprotein synthesis because eIF2α (P) is acompetitive inhibitor of eIF2B, which is limitingin cells (9). Concomitant with this globaltranslation inhibition, eIF2α phosphorylation canlead to preferential translation of mRNAsencoding stress-related proteins. Thus, eIF2αphosphorylation can regulate both general andspecific translation.

There are four known eIF2α kinases inmammals [reviewed in (10)]: GCN2, activated byamino acid starvation through the binding ofuncharged tRNA to its regulatory region;PEK/PERK, an ER transmembrane protein,activated by endoplasmic reticulum (ER) stress;PKR, activated by double stranded RNA producedduring viral infection; and HRI, present mainly inreticulocytes, activated by heme deprivation.

GCN2, the sole eIF2α kinase present in theyeast Saccharomyces cerevisiae (11), is found inmice in three isoforms differing only in their N-terminal sequences (12). The most abundant andubiquitously expressed isoform contains all thefeatures of the yeast counterpart, including an N-terminal domain (NTD), which is required in vivofor the activation of the kinase domain through itsinteraction with the activator GCN1 (13,14); apseudo-kinase domain followed by the kinaseregion; a region with similarity to histidyl-tRNAsynthetases, implicated in the recognition ofuncharged tRNA which then signals for theactivation of the kinase domain; and a C-terminaldomain (CTD), involved in the interaction ofGCN2 with the ribosomes (9,15,16). GCN1 isrequired for the activation of GCN2, and it isthought that GCN1 acts as a chaperone to transportuncharged tRNA’s that enter the A site ofribosomes, to the tRNA binding domain of GCN2(13,17). GCN1 works in concert with GCN20 toform a complex with GCN2 on the ribosome,required for the activation of the kinase (13,16,17).

In yeast, GCN2 is required for growth underamino acid starvation conditions. Its activation byhigh levels of uncharged tRNA’s whichaccumulate in these conditions leads to eIF2αphosphorylation and thus to the translation ofGCN4, a transcriptional activator of hundreds of

genes involved in amino acid biosynthesis (18,19).Mammalian GCN2 has been shown to be requiredfor adaptation to amino acid deprivation in miceand is activated under conditions of lowavailability of amino acids (20,21). While there isno GCN4 ortholog in mammalian cells, the levelsof a related transcriptional activator, ATF4, isinduced by eIF2α phosphorylation by amechanism of translation reinitiation similar tothat described for yeast GCN4 (22,23). ATF4,also known as CREB-2, enhances the expressionof additional bZIP transcriptional regulators,including CHOP/GADD153 and ATF3, thattogether contribute to the expression of a largenumber of genes involved in metabolism, redoxchemistry and apoptosis (21,24,25).

Mice lacking GCN2 are viable; however, theGCN2-deficient animals display aberranttranslation in the liver, enhanced skeletal muscleloss and increased morbidity in reponse to aminoacid deprivation (26). Recently, GCN2 has beenshown to be directly involved in the feedingbehavior of mice. Phosphorylation of eIF2α in theanterior piriform cortex is observed immediatelyfollowing intake of diets poor in essential aminoacids (27). Interestingly, contrary to wild typeanimals which tend to avoid meals lacking evenone of the essential amino acids, GCN2-/- miceare defficient in this aversive behavior (28,29).

GCN2 NTD has a sequence motif, called GIdomain, that is also present in the N-terminal halfof the yeast protein YIH1 and its mammalianortholog, IMPACT (14). IMPACT was originallyidentified in a screen for imprinted genes inmouse. Both YIH1 and IMPACT contain in the C-terminal half a conserved sequence (Ancientdomain) found also in bacterial proteins (14,30).

Because of the similarity between the GIdomains of YIH1 and GCN2, it has been proposedthat YIH1 acts as an inhibitor of the activation ofGCN2 mediated by GCN1, through competionwith GCN2 for GCN1 binding. Indeed, recent datain yeast have demonstrated that the binding ofGCN2 to GCN1 can be reduced by overexpressionof YIH1 and that this leads to reduced eIF2αphosphorylation, indicative of impaired GCN2activation (31). The in vivo evidence in yeastclearly indicated that YIH1 inhibits GCN2activation. However, no condition was foundwhere this inhibitory action would be

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

3

physiologically relevant to yeast cells, since adeletion of YIH1 has no apparent phenotype.Because YIH1 was also found to bind to G-actin,it has been proposed that localized action of YIH1in yeast cells might regulate the activity of GCN2in regions where protein synthesis must bemaintained at high levels, such as near thegrowing bud (31).

Given the relevance of GCN2 in mammalianmetabolism and behavior, and the involvement ofeIF2α phosphorylation in several pathologicalconditions, we decided to study the function ofIMPACT. We show here that IMPACT is themammalian functional counterpart of YIH1.IMPACT binds to GCN1 and acts as an inhibitorof mouse GCN2. We also demonstrate thatIMPACT is preferentially expressed in the brain inmice and is especially abundant in thehypothalamus. The levels of IMPACT correlateinversely with the basal levels of eIF2α(P) in alltissues examined. Our results strongly suggest thatIMPACT acts as an inhibitor of GCN2 in themammalian brain. These findings have profoundphysiological implications in the control ofphosphorylation of eIF2α, and consequently in theexpression of ATF4, in different brain areas andspecific neuronal cells.

MATERIALS AND METHODS

Yeast methods. Standard yeast methods wereemployed (32). Yeast strain H1511 (MATα, ura3-52, trp1-63, leu2-3, leu2-11, GAL2+) (33) wasgrown in synthetic complete medium lackingamino acids to select for plasmids, andsupplemented with 2% glucose or 10% galactoseas carbon source. The plasmid encoding YIHfused to GST for expression in yeast under thegalactose inducible promoter GAL1, has beendescribed previously (31). The plasmid encodingIMPACT fused to GST under the control of theGAL1 promoter was constructed by theintroduction of BglII and HindIII sites throughPCR in the cloned sequence of IMPACT presentin plasmid pBE435 (see below), and cloning intoBamHI-HindIII sites of vector pES128-9-1described previously (13). The response of yeaststrains to amino acid starvation was studied eitherby scoring for growth on solid medium lackinghistidine and containing 3-aminotriazole (3AT) at

the concentration indicated, or by growing cellsfor 4h in liquid culture containing 30 mM 3ATand measuring the levels of eIF2αphosphorylation by immunoblot analysis asoutlined in (31).

Animals.This study was conducted underprotocols approved by the Animal Care and UseEthic Committee of the Universidade Federal deSão Paulo, and in accordance with the Guide forCare and Use of Laboratory Animals adopted bythe National Institutes of Health. Male Swissalbino mice (20-30 g) were decapitated and brainand other tissues removed as quickly as possible,washed in PBS and immediately processed.

Extract preparation. Extracts of whole brain,brain parts and other tissues were prepared inbuffer containing 20 mM Hepes-KOH, pH 7.5,150 mM NaCl, 1% Triton X-100, 10% glycerol, 1mM EDTA, 1 mM PMSF, 4 µg/ml aprotinin, 2µg/ml pepstatin, 100 mM NaF and 10 mMtetrasodium pyrophosphate. All experiments usedindependent pools of extracts obtained from 4animals.

RNA isolation and cDNA synthesis. TotalRNA from mouse whole brain, cortex,hippocampus and hypothalamus was obtained byTrizol extraction as described by the manufacturer(Life Technologies, Inc.). For cDNA cloning, RT-PCR was performed using the following primerp a i r s : B C 3 6 7 ( 5 ’ -GGAATTCATGGCTGAAGAGGAAGTAGGGAACAGCC- 3 ’ ) a n d B C 3 8 2 ( 5 ’ -GCGGCCGCTTAATGATCATTCTTCTTCTTGTCTTTC-3’), for IMPACT; and BC456 (5’-GGGGATCCATGGCCGTTAAGAGCC-3’) andB C 4 5 7 ( 5 ’ -GGCTCGAGGCTGAGGATCATGTCC-3’) formouse GCN1 (mGCN1; GI:51827542). Fordetection of transcripts in the brain parts, thefollowing oligonucleotides were used: BC380 (5’-TGGGCTTCCTCATGAAGTTTCAGATCG-3’)and BC387 (5’-TGCTGGGCAGGCCAGCTTCC-3’) , for mGCN1; and BC366 (5’-CAACATACCCAGATGTAGTTCCCGAAATAGA- 3 ’ ) a n d B C 3 7 8 ( 5 ’ -GAGGATGTCACACGAGCCAGGAGAG-3’),for βGCN2.

IMPACT and mGCN1 cloning and proteinpurification. A 1 kb DNA fragment comprisingthe complete open reading frame of IMPACT was

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

4

obtained by RT-PCR from whole brain RNA usingthe primers described above. The IMPACT cDNAwas inserted as an E c oRI-NotI fragment inpET28a plasmid (Novagen), generating plasmidpBE435. The N-terminal His-tagged recombinantprotein was expressed in E. coli Rosetta (DE3)grown in LB with 100 µg/ml of kanamycin and100 µg/ml chloramphenicol, after induction with0.1 mM isopropyl-β-D-thiogalactopyranoside(IPTG) at 230C overnight. Purification of theprotein was performed essentially as described bythe manufacturer of the Ni-NTA resin (Qiagen).Briefly, the cells were harvested, resuspended inlysis buffer (10% sucrose, 0.2 M NaCl, 50 mMTris-HCl pH 7.5), frozen, incubated withlysozyme (1 mg/ml) for 30 minutes on ice andbriefly sonicated. The cell lysates were centrifugedand the supernatant was applied to Ni-NTA resinequilibrated with binding buffer (500 mM NaCl,20 mM Tris-HCl pH 8.0, 5 mM imidazole). Thecolumn was washed with washing buffer (500 mMNaCl, 20 mM Tris-HCl pH 8.0, 20 mMimidazole). The His-tagged protein was elutedwith elution buffer (500 mM NaCl, 20 mM Tris-HCl pH 8.0, 1 M imidazole) and dialyzed against20 mM Tris-HCl pH 7.5, 10 mM 2-mercaptoethanol. The mouse GCN1 (mGCN1)sequence (gi51827542) encoding residues 2204 to2651 was obtained by RT-PCR and cloned as a 1.3kb BamHI-XhoI fragment in plasmids pET28a andpGEX6p3. Expression was obtained in E. coliBL21(DE3) for the His-tagged protein and inDH5α for the GST fusion. Extracts were preparedas described above. The insoluble recombinantproteins present in the bacterial extract pellet weresolubilized by 8M urea. The His6-mGCN1 proteinused for immunization was purified on Ni-NTAresin in the presence of urea, as described by themanufacturer (Qiagen), followed by preparativeSDS-PAGE and elution from the gel. The GST-mGCN1 protein was purified from the ureasolubilized pellet by preparative SDS-PAGEfollowed by elution of the protein from thepolyacrilamide gel slice.

Mouse embryonic fibroblast (MEF) celltransfection and amino acid starvationconditions. For overexpression of IMPACT inMEF cells, an E c oRI-NotI fragment encodingIMPACT was isolated from plasmid pBE435, andplaced under the control of the CMV promoter in

plasmid pCI-neo (Promega), which contains theSV40 origin of replication, originating plasmidpBE514. MEF cells immortalized by the simianvirus 40 large T antigen were cultured inDulbecco’s modified Eagle’s medium (DMEM;BioWhitaker) supplemented with 1 mMnonessential amino acids, 100 U/ml penicillin, 100µg/ml streptomycin and 10% fetal bovine serumand transfected using Lipofectamine (Invitrogen),as described previously (25). The non-transfectedand transfected MEF cells were subjected toamino acid starvation by culturing in DMEMwithout leucine (BioWhitaker) for the indicatednumber of hours. Lysates were prepared asdescribed and equal amounts of proteins wereanalyzed by immunoblot using antibodies specificto eIF2α (P), CHOP, ATF4, or β-actin, asdescribed (25). Additionally wild-type GCN2+/+and GCN2-/- MEF cells were subjected to leucinestarvation, and lysates were analyzed byimmunoblot.

Preparation of mono-specific antibodiesagainst IMPACT and mGCN1. Polyclonalantibodies were produced by immunizing rabbitswith His6-IMPACT and His6-mGCN1 purifiedrecombinant proteins. Monospecific antibodieswere obtained by incubating the immune sera withHis6-IMPACT or GST-mGCN1 immobilized onnitrocellulose membranes. One milligram ofpurified protein was submitted to preparative SDS-PAGE, and transferred to a Hybond-C membrane.A strip of the membrane containing the protein, asvisualized by Ponceau staining, was blocked with5% non-fat milk in phosphate-buffered saline(PBS), followed by incubation with antiserum at a1:10 dilution in PBS, for 3 hours. After washingwith PBS, bound antibodies were eluted with 0.1M glycine pH 2.5 and 1 mM EGTA for 10minutes. The pH was immediately neutralized byadding an equal volume of 0.1 M Trisma base. Immunoblot analysis. For immunoblot analysisof IMPACT expression and eIF2αphosphorylation, Laemmli sample buffer wasadded to samples, and after boiling for 3 min, theproteins were separated on 12% SDS-PAGE andtransferred to nitrocellulose membrane (Hybond-CExtra, Amersham) at 1 A for 1 h using the bufferconditions described (34). The membrane wasblocked with 5% non-fat milk in phosphate-buffered saline (PBS) for anti-IMPACT or in Tris-

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

5

buffered saline (TBS) for anti-eIF2α(P), with0.1% Tween 20 for 1 h at room temperature, andincubated overnight at 40C with anti-IMPACTmonospecific antibodies (1: 500 in PBS, 5% non-fat milk and 0.1% Tween 20) or anti-eIF2α(P)(Biosource, 1:1000 in TBS, 1% BSA and 0.1%Tween 20). After three washings with 0.1%Tween 20 in PBS or TBS, bound antibodies weredetected with Protein A-horseradish peroxidase(HRP) (Amersham), diluted 1:4000 in PBS or withgoat anti-rabbit IgG-HRP (1:2000) (Santa Cruz),for anti-IMPACT and anti-eIF2α(P) immunoblots,respectively. After incubation for 1 h at roomtemperature and washings in PBS or TBS, thebound antibodies were detected using the ECLchemiluminescence method (Amersham). Afterstripping the membranes by incubation with 100mM 2-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl, pH 6.7 at 50ºC for 30 minutes, the samemembrane was incubated with anti-actinantibodies (1:300 dilution, Sigma) or anti-eIF2α(Biosource, 1:1000) and processed as above.eIF2α(P) levels in yeast whole cell extracts weredetermined as described (31). The conditions forimmunoblots of MEF’s for the detection of ATF4,CHOP, eIF2α(P) and β-actin were essentially asdescribed previously (25).

Pull-down assays. GST-pulldown assays wereperformed as described previously (13). GST-yGCN1[2052-2428]R2259 and GST-yGCN1[2052-2428]A2259 fusion proteins andGST were purified from E. coli carrying plasmidspES123-B1, pES164-2A and pGEX2T,respectively (31). Briefly, E. coli cells (DH5α)carrying the especified plasmid, were grown in LBmedium containing ampicillin (100 µg/ml) to 0.8A600nm and the induction of recombinant proteinexpression was obtained by incubation with 0.1mM IPTG, at 30ºC for 4 hr. Cells were collected,ressuspended in PBS and the extract prepared asdescribed before. The GST fusion proteins werepurified from the soluble fraction of the cellextract on glutathione-Sepharose (AmerhamBiotec), as described by the manufacturer. Thepurified proteins were dialyzed against 20 mMTris-HCl, pH 8.0 and 1 mM EDTA. For the pull-down assays, the purified proteins (20 µg) wereimmobilized on 20 µl of glutathione-Sepharosebeads and incubated with 500 µg of brain extractprepared as described above, in a total volume of

200µl in 30 mM Tris-HCl pH 7.5, 50 mM KCl,10% glycerol, 0.1 mM PMSF, 5 mM β -mercaptoethanol for 2 hours at 4ºC. After washingwith PBS, the beads were resuspended in 20 µl 4X Laemmli’s sample buffer, boiled for 3 min andproteins were separated on 12% SDS-PAGE. Aftertransfer to a Hybond-C membrane, the proteinswere stained with Ponceau S. Western blot usinganti-IMPACT antibodies was performed asdescribed above.

Co-immunoprecipitation of mGCN1 andIMPACT. Brain extracts (5 mg) were pre-clearedby incubation with 20 µl protein A-agarose beadsand 1 µl pre-immune serum, in the bufferdescribed for the preparation of extracts frommouse tissues. The supernatant was then incubatedovernight at 4ºC with 20 µl protein A-agarosebeads-bound mono-specific anti-mGCN1antibodies, an irrelevant IgG or buffer only. Thebeads were washed three times with the samebuffer and the bound material resolved on SDS-PAGE followed by immunoblots using anti-IMPACT or anti-mGCN1 antibodies.

Immunohistochemistry for IMPACT. Animalswere deeply anesthetized with a thionembutaloverdose (100 mg/kg) and perfused through theheart with 50 ml saline followed by 300 mlparaformaldehyde 4% at 4°C. The brains wereremoved and cryoprotected in 30% sucrose in PBSfor 24 h. Coronal sections 40 µm thick werecollected, washed with PBS for 30 min, incubatedin blocking buffer [PBS containing 0.3% Triton X-100 (Amresco) and 0.5% normal goat serum(Vector Laboratories, Inc.)] for 30 min, followingincubation with anti-IMPACT mono-specificantibodies (diluted 1:100 in blocking buffer) atroom temperature for 24 h. The sections were thenwashed with PBS for 30 min., incubated with goatanti-rabbit biotinylated IgG (1:200) for 2 h,washed again with PBS for 30 min and incubatedwith ABC (avidin-biotin complex, Vector EliteABC kit) for 90 min. The bound antibodies werede tec t ed wi th a n i cke l - in t ens i f i eddiaminobenzidine tetrahydrochloride reaction.

RESULTS

Overexpression of IMPACT in yeast cellscauses a Gcn- phenotype. In order to determinewhether IMPACT is functionally related to YIH1,

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

6

we overexpressed this protein in yeast cells.Overexpression of YIH1 imparts a Gcn- phenotypeto yeast cells (31), which can be identified by theinability to grow in the presence of 3-aminotriazole (3AT), an inhibitor of the HIS3enzyme necessary for the biosynthesis of histidine.In order to grow in the presence of 3AT, cells haveto synthesize GCN4, a transcriptional activator ofamino acid biosynthetic genes and salvaging ofnutrients. GCN4 is under translational regulationand is synthesized in conditions that lower theamounts of active ternary complex, such as wheneIF2α is phosphorylated. Thus, cells unable tophosphorylate eIF2α will not translate GCN4, andtherefore will not overcome the inhibitory actionof 3AT, visible by impaired growth. To assess theeffect of IMPACT on the ability of cells to grow in3AT, IMPACT was expressed in yeast under thecontrol of a galactose-inducible promoter, as afusion with glutathione-S-transferase (GST). Cellsoverexpressing GST-YIH1, also under the controlof the galactose-inducible promoter, were used ascontrol. Cells overexpressing IMPACT displayeda Gcn- phenotype, that is impaired growth in the3AT containing medium, as did cellsoverexpressing YIH1 (Fig. 1A). The levels of bothproteins were similar, as shown by immunoblotsperformed in extracts prepared from galactose-induced cultures using antibodies directed againstGST (Fig. 1B).

The Gcn- phenotype of IMPACToverexpression in yeast is due to inhibition ofeIF2αααα phosphorylation. The Gcn- phenotype is anindication of a defect in eIF2α phosphorylationand GCN4 translational derepression, such aswhen the cells are deleted for the eIF2α kinaseGCN2. We then investigated the levels of eIF2αphosphorylation in cells overexpressing IMPACT.There is a basal level of eIF2α(P) even underoptimal growth conditions, which is thensignificantly elevated when cells are grown underamino acid starvation conditions or elicited by theaddition of 3AT. As shown in Figure 2, the basallevels of eIF2α phosphorylation are decreased incells overexpressing IMPACT (compare lanes 2-3with lane 4), similar to the effect observed forYIH1 overexpression (compare lanes 1 with lanes2-3). The overexpression of IMPACT also hindersthe activation of GCN2 under starvationconditions (lanes 6-7 versus lanes 8-10), to the

same extent as YIH1 overexpression, loweringeIF2α phosphorylation to approximately half ofthe eIF2α (P) levels found in the GSToverexpression. These data thus indicate thatIMPACT inhibits GCN2 activation in yeast,probably by interacting with and sequesteringyGCN1.

Overexpression of IMPACT in mouseembryonic fibroblasts inhibits GCN2 activationupon leucine starvation. GCN2 is required forinduced eIF2α phosphorylation and enhancedexpression of ATF4 and its target gene CHOP inresponse to amino acid limitation in mammaliancells (21,25). In order to determine whetherIMPACT inhibits mammalian GCN2, MEF cellswere transfected with a plasmid expressingIMPACT under the control of the CMV promoter.There were high levels of IMPACT in thetransfected MEF cells as judged by immunoblotanalysis, as opposed to cells carrying only thevector or the non-transfected cells (Fig. 3B).Activation of GCN2 was determined byphosphorylation of eIF2α upon incubation of thecells in medium lacking leucine. The non-transfected cells showed a significant increase ineIF2α(P) that is the result of activation of GCN2,as illustrated by the observation that eIF2α(P)levels were diminished in GCN2-/- cells that weresubjected to leucine starvation (Fig. 3A). Cellscarrying the vector alone showed an increase inthe basal levels of eIF2α(P), in the presence ofleucine, when compared with the non-transfectedcells, suggesting perhaps that the transfectionprocedure and/or high replication rate of theplasmid activates an eIF2α kinase; however, theactivation of GCN2 upon leucine starvation isclearly evident in these cells. MEF cellsoverexpressing IMPACT, although displaying thesame amount of basal eIF2α (P) as the vectorcontaining cells under non-starvation conditions,clearly did not show the same increase in eIF2α(P)upon leucine depletion (Fig. 3B). The signal foreIF2α(P) remained constant through out the 6hours of incubation without leucine. Given thateIF2α(P) signals for the increased expression ofATF4 and CHOP, we also asked whether theoverexpression of IMPACT affected the levels ofthese two proteins. Consistent with earlier reports(25) translational induction of ATF4 expressioncontributed to early expression of this

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

7

transcriptional regulator, within 1 hour of theleucine limitation in the wild-type or vector-transfected MEF cells. Levels of the ATF4-targetgene CHOP were also induced in these cellswithin 3 hours of the leucine starvation.Importantly, overexpression of IMPACT resultedin a minimal production of ATF4 and CHOPunder amino acid starvation, when compared tothe cells transfected with the vector (Fig. 3B).These results clearly show that IMPACT is aninhibitor of the activation of the mammalianGCN2 stress pathway in response to nutrientdeprivation.

IMPACT binds yeast and mammalian GCN1.In order to determine whether IMPACT inhibitsGCN2 by binding to GCN1, as does its yeastortholog, we first addressed whether IMPACTcould bind yGCN1. It has been shown thatresidues 2052-2428 of yGCN1 are sufficient forthe interaction with both yeast GCN2 and YIH1(13). The region comprised by residues 2052-2428of the yeast protein shows 34% identity and 61%similarity to the equivalent region in mGCN1. Themutation R2259A in this region has been shown toabolish the interaction of yGCN1 with bothyGCN2 and YIH1, when present in the completeprotein in vivo and in a GST-yGCN1[2052-2428]fusion in vitro (13,31). The R2259 residue isidentical and neighboring sequences are identicalor highly similar in all sequenced orthologs, asshown in the alignment of Figure 4A. We thenused a purified GST-yGCN1 fusion proteincomprising residues 2052-2428 (GST-yGCN1[2052-2428]R2259) in pull downexperiments with extracts prepared from micebrain (see below). As shown in Figure 4B, brainIMPACT was immobilized on glutathione beadsthrough interaction with GST-yGCN1[2052-2428]R2259. Because only a small fraction ofIMPACT present in the brain extract associatedwith the immobilized GST-yGCN1 (appr. 0.2-0.5%), we used also the yGCN1 mutant whereresidue 2259 was altered from arginine to alanine(GST-yGCN1[2052-2428]A2259) to address thespecificity of the binding. This mutant protein wasunable to bind to IMPACT (Fig. 4B). The smallamount of binding observed for the wild typeGST-yGCN1[2052-2428] could be due to the useof a fragment of a heterologous protein, whichmay bind with low affinity with the endogenousIMPACT present in the extracts. Alternatively, it

is possible that the native IMPACT is present in alarge complex that may not be stable enough to beretained by the GST-yGCN1-beads.

In order to show that IMPACT interacts withmGCN1, we raised antibodies against part of themouse GCN1 protein (residues 2204-2651,containing the putative GCN2/IMPACT-interacting region), to perform a co-immunoprecipitation assay. The completesequence of mouse GCN1 cDNA indicates aprotein of 2806 residues, with a predicted mass of307 kDa. These antibodies recognized a protein ofapproximately 280kDa and, to a lesser degree, aslower migrating protein, the latter also recognizedby the pre-immune serum. The smaller thanpredicted mass is also observed for the yGCN1native protein. As shown in Figure 4C, IMPACTwas immunoprecipitated along with mGCN1 frombrain extracts using the purified antibodies.mGCN1 and IMPACT were not detected in thecontrol where non-related purified IgG was usedin the same amount as the anti-mGCN1 purifiedantibodies. These results taken together providestrong evidence that IMPACT binds specifically toGCN1.

A very small percentage of IMPACT wasfound associated with GCN1 in brain extracts, asdetected by this co-immunoprecipitation assay.This result is not surprising, given that a moreextensive interaction would probably block theamino acid starvation response mediated byGCN2. Interestingly, in the yeast model, the invivo interaction between YIH1 and yGCN1 couldonly be detected when YIH1 was overexpressed(31). Thus, it is possible that the IMPACT-GCN1complexes occur exclusively in a small populationof neuronal cells that expresses high levels ofIMPACT.

IMPACT is preferentially expressed in thehypothalamus. It has been previously shownthrough in situ and Northern hybridizations inmice that the IMPACT mRNA is preferentiallyexpressed in the brain (30). Using highly specificantibodies directed against IMPACT raised in thiswork, we analyzed in detail the abundance of thisprotein in mouse tissues. Immunoblots of extractsobtained from several organs showed that theIMPACT protein is highly expressed in the brain,correlating with the previous data on the mRNAabundance (Fig. 5A). In order to determine theexpression of IMPACT in different brain areas,

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

8

immunohistochemistry was performed in mousebrain slices. As controls, we used non-specific IgGpurified from pre-immune serum in the sameconcentration as the purified antibodies directed toIMPACT. As shown in Figure 6, IMPACT wasfound to be expressed in high levels in somescattered neurons in several brain areas, as shownhere for the hippocampus, where a few neuronsshow strong labeling, and in the cortex, wherelayer II of the piriform cortex showed moreintense labeling. The expression in other neuronalcells in these regions is comparatively very low.On the other hand, IMPACT was found to behighly expressed in the majority of neurons inseveral hypothalamic regions. Strong labeling wasevident around the third ventricular region,including the paraventricular, dorso-medial andposterior hypothalamic nuclei, the pre-optic areaand supra-chiasmatic nuclei, among others. Theseresults are in agreement with data obtained fromthe mouse expres s ion da t a base(http://expression.gnf.org/cgi-bin/index.cgi),where the hypothalamus had 10 times moreIMPACT mRNA than other organs, andapproximately 3 times more than in other brainareas (35). These results were further confirmedand quantitated by Western blots of extractsprepared from the cortex, hippocampus andhypothalamus (Fig. 5B). The abundance ofIMPACT in the hypothalamus relative to the otherbrain parts is clearly evident.

The hypothalamus displays the lowest basallevels of eIF2αααα phosphorylation. Because ourresults showed that IMPACT overexpressioninhibits mouse GCN2, we hypothesized that thehigh levels of IMPACT in the hypothalamus mayinhibit the endogenous GCN2, resulting in lowbasal levels of eIF2α phosphorylation in this brainregion, as compared to other areas with lowIMPACT expression. We then determined theratio of eIF2α(P) to total eIF2α by immunoblotsof extracts from the cortex, hippocampus andhypothalamus. As shown in Figure 7A, in thehypothalamus, where high amounts of IMPACTare found, the basal levels of eIF2α(P) are muchlower than in the hippocampus and cortex.

It was possible that the low eIF2α(P) in thehypothalamus could be instead due to lower levelsof GCN2 or GCN1. In the mouse transcriptomemicroarray data, GCN2 was found to be equally

expressed in all the brain areas analyzed here (35).However, the microarray data did not differentiateamong the three isoforms of GCN2 present inmice, all of them known to be expressed in thebrain (12). Isoform beta has complete homology toother GCN2 homologs, including the region at theN-terminus that may interact with GCN1. Isoformalpha lacks the N-terminal 280 amino acidresidues, and is therefore not a target for GCN1binding. Isoform gamma starts at an amino acidcorresponding to position 86 of βGCN2, andcarries 6 additional residues in the N-terminus.This isoform contains part of the GI domain. Toinvestigate the presence of the βGCN2 isoform inthe different brain areas, we performed RT-PCRusing oligonucleotides specific for this isoform.Furthermore, GCN1 mRNA was not reported to bepresent in the transcriptome database (35), and wetherefore carried out the RT-PCR analysis for theGCN1 transcript in the different brain areas. Theresults, shown in Figure 7B, indicate that bothβGCN2 and GCN1 are expressed in thehypothalamus, cortex and hippocampus. Thus, thedifferent levels of basal eIF2α(P) in these areasmight be related to the levels of IMPACT and thusto the differential basal activation of GCN2.

In order to provide another comparativeanalysis, we also quantitated the levels ofe IF2α (P) in the heart, where there is littleIMPACT protein, as determined previously. In theheart, basal eIF2α(P) was found to be very highcompared to the hypothalamus (Figure 7C),therefore providing further support for an inverserelationship between the abundance of IMPACTand eIF2α phosphorylation.

DISCUSSION

It has been recently shown that YIH1 regulatesthe activation of the eIF2α kinase, GCN2, in yeastthrough its interaction with GCN1. In this report,we provided in vivo evidence that IMPACT, themammalian ortholog of YIH1, inhibits both yeastand mouse GCN2, by its ability to bind to GCN1.We suggest then that IMPACT is a negativeregulator of GCN2 in mammals.

As a prediction of this hypothesis, high levelsof IMPACT should lead to a low intrinsicactivation of GCN2 in mouse tissues and therefore

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

9

to lower basal levels of eIF2α phosphorylation, asobserved in the yeast model. We were able toshow that this correlation occurs in all of the testedtissues in mice. In the heart, where almost noIMPACT can be detected relative to the brain,high levels of eIF2α (P) were found. In thehypothalamus, where IMPACT shows the highestexpression, eIF2α(P) levels are very low. OthereIF2α kinases or the activity of eIF2α specificphosphatases may also participate in establishingthe basal levels of eIF2α(P) in the different brainareas and organs studied. However, it isreasonable to assume that differences in eIF2αkinase activity among organs or cells of a normalanimal should rely heavily on GCN2 andPEK/PERK, both of which are intrinsically relatedto metabolic regulation. Low glucose activatesPEK/PERK, and possibly GCN2, in analogy toyeast (36) and from the phenotypes of knock outanimals suggesting an overlap of PEK/PERK andGCN2 in glucose sensing, whereas amino aciddeprivation activates GCN2. Both mechanisms arenecessary for maintaining homeostasis and shouldbe constantly monitored. Thus, GCN2 activity canbe considered as an important contributor to thelevels of phosphorylated eIF2α in mammals undernormal physiological conditions. Under aminoacid starvation conditions, the results shown herestrongly suggest that IMPACT has an importantrole in controlling the levels of eIF2α(P) throughGCN2 inhibition, which may be more relevant inspecific populations of neuronal cells.

The hypothalamus is critically involved withthe maintenance of homeostasis, such as thecontrol of body temperature and the balance offluids and energy, and it is constantly adjusting theorganism’s metabolism and behavior to itsimmediate needs. It is interesting to speculate thatdue to the constant signaling required fromneurons in the hypothalamus, protein synthesismust be maintained at constant high levels evenunder conditions where GCN2 would be activatedin other cell types, such as amino acid starvation.Thus, the mechanism of inhibition of GCN2activation represented by the overexpression ofIMPACT may be important for the function ofneurons in this area. Along these lines, it ispossible that an inhibitor of PEK/PERK may alsobe overexpressed in this brain region. The supra-chiasmatic nuclei (SCN) of the hypothalamus

show the highest expression of IMPACT asdetermined from microarrays (35). We have notquantitated IMPACT nor eIF2α(P) in the SCN asa separate region in immunoblots, butimmunohistochemistry analyses suggestedelevated IMPACT levels in the SCN (data notshown). The SCN are involved in circadianrhythm determination and maintenance. Thepossibility that SCN neurons have an even morestringent control of eIF2α phosphorylation isintriguing.

Recent findings on the activation of GCN2 inthe anterior piriform cortex (APC) upon feeding alow amino acid diet are not discrepant relative toour results showing high levels of IMPACT in thepiriform cortex. Upon a close inspection of theanterior piriform cortex, IMPACT was foundmainly in interneurons in layer II (data notshown), whereas in animals subjected to lowamino acid diet, only a few pyramidal neurons inthe APC show eIF2α(P) labeling (27). Certainly,co-localization studies will be highly relevant.

YIH1 has been shown to bind to G-actin inyeast. It is not clear what role the interactionbetween YIH1 and actin plays in the yeast cells.However, low actin levels lead to an impairmentof the activation of GCN2, suggesting that theresulting higher pool of free YIH1 would interactwith GCN1, preventing activation of GCN2 (31).We were not able to detect the binding of β-actinto purified His6-IMPACT or GST-IMPACT addedto brain extracts (data not shown). It is possiblethat the actin binding site in this protein might behidden in the conformation of the recombinantprotein. Thus, the issue of IMPACT binding toactin in mammals should be further analyzed.

To the best of our knowledge, this work is thefirst to show that eIF2α is differentiallyphosphorylated in the mammalian brain areasunder normal conditions. The intrinsic levels ofeIF2α(P) may be important for determining theexpression of ATF4, a protein that is proposed toplay pivotal roles in neurons. ATF4, besides itsrole in the build up of a stress recovery program,has been recently demonstrated to negativelyregulate long term potentiation (LTP): transgenicmice expressing an inhibitor of ATF4 activityhave enhanced hippocampal based spatial memoryand LTP, and ATF4 is related to the Aplysiamemory suppressor gene ApCREB-2 (37). The

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

10

expression of IMPACT in the hippocampus maybe relevant for the attenuation of the expression ofATF4. However, our observation that IMPACT isoverexpressed exclusively in some interneurons inthis region was surprising. Despite the wealth ofinformation on the functional role ofATF4/CREB-2 in the biochemical pathwaysassociated with learning and memory, there is noinformation regarding the distribution of ATF4over different populations of hippocampalneurons. Thus, further understanding of the role

played by IMPACT in the regulation of basal andamino acid starvation-induced levels of eIF2α(P),and consequently in the levels of ATF4, in specificneuronal cells of the hippocampus may add newinsights into the molecular mechanisms involvedin LTP.

.

REFERENCES

1. Kaufman, R. J., Scheuner, D., Schroder, M., Shen, X., Lee, K., Liu, C. Y., and Arnold, S. M.(2002) Nat. Rev. Mol. Cell. Biol. 3, 411-421

2. Kaufman, R. J. (2002) J. Clin. Invest. 110, 1389-13983. Ron, D. (2002) J. Clin. Invest. 110, 1383-13884. Si, K., Giustetto, M., Etkin, A., Hsu, R., Janisiewicz, A. M., Miniaci, M. C., Kim, J. H., Zhu, H.,

and Kandel, E. R. (2003) Cell 115, 893-9045. Kumar, R., Azam, S., Sullivan, J. M., Owen, C., Cavener, D. R., Zhang, P., Ron, D., Harding, H.

P., Chen, J. J., Han, A., White, B. C., Krause, G. S., and DeGracia, D. J. (2001) J. Neurochem.77, 1418-1421

6. Carnevalli, L. S., Pereira, C. M., Longo, B. M., Jaqueta, C. B., Avedissian, M., Mello, L. E., andCastilho, B. A. (2004) Neurosci. Lett. 357, 191-194

7. Abbott, C. M., and Proud, C. G. (2004) Trends Biochem. Sci. 29, 25-318. Klann, E., and Dever, T. E. (2004) Nat. Rev. Neurosci. 5, 931-9429. Hinnebusch, A. G. (2000) in Translational control of gene expression (Sonenberg, N., Hershey, J.

W. B., and Mathews, M. B., eds), pp. 185-243, Cold Spring Harbor Laboratory Press, ColdSpring Harbor, N. Y.

10. Dever, T. E. (2002) Cell 108, 545-55611. Dever, T. E., Feng, L., Wek, R. C., Cigan, A. M., Donahue, T. F., and Hinnebusch, A. G. (1992)

Cell 68, 585-59612. Sood, R., Porter, A. C., Olsen, D., Cavener, D. R., and Wek, R. C. (2000) Genetics 154, 787-80113. Sattlegger, E., and Hinnebusch, A. G. (2000) EMBO J. 19, 6622-663314. Kubota, H., Sakaki, Y., and Ito, T. (2000) J. Biol. Chem. 275, 20243-2024615. Dong, J., Qiu, H., Garcia-Barrio, M., Anderson, J., and Hinnebusch, A. G. (2000) Mol. Cell 6,

269-27916. Garcia-Barrio, M., Dong, J., Ufano, S., and Hinnebusch, A. G. (2000) EMBO J. 19, 1887-189917. Marton, M. J., Vazquez de Aldana, C. R., Qiu, H., Chakraburtty, K., and Hinnebusch, A. G.

(1997) Mol. Cell. Biol. 17, 4474-448918. Natarajan, K., Meyer, M. R., Jackson, B. M., Slade, D., Roberts, C., Hinnebusch, A. G., and

Marton, M. J. (2001) Mol. Cell. Biol. 21, 4347-4368

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

11

19. Qiu, H., Dong, J., Hu, C., Francklyn, C. S., and Hinnebusch, A. G. (2001) EMBO J. 20, 1425-1438

20. Zhang, P., McGrath, B. C., Reinert, J., Olsen, D. S., Lei, L., Gill, S., Wek, S. A., Vattem, K. M.,Wek, R. C., Kimball, S. R., Jefferson, L. S., and Cavener, D. R. (2002) Mol. Cell. Biol. 22, 6681-6688

21. Harding, H. P., Novoa, I., Zhang, Y., Zeng, H., Wek, R., Schapira, M., and Ron, D. (2000) Mol.Cell 6, 1099-1108

22. Vattem, K. M., and Wek, R. C. (2004) Proc. Natl. Acad. Sci. U S A 101, 11269-1127423. Lu, P. D., Harding, H. P., and Ron, D. (2004) J. Cell. Biol. 167, 27-3324. Harding, H. P., Zhang, Y., Zeng, H., Novoa, I., Lu, P. D., Calfon, M., Sadri, N., Yun, C., Popko,

B., Paules, R., Stojdl, D. F., Bell, J. C., Hettmann, T., Leiden, J. M., and Ron, D. (2003) Mol. Cell11, 619-633

25. Jiang, H. Y., Wek, S. A., McGrath, B. C., Lu, D., Hai, T., Harding, H. P., Wang, X., Ron, D.,Cavener, D. R., and Wek, R. C. (2004) Mol. Cell. Biol. 24, 1365-1377

26. Anthony, T. G., McDaniel, B. J., Byerley, R. L., McGrath, B. C., Cavener, D. R., McNurlan, M.A., and Wek, R. C. (2004) J. Biol. Chem. 279, 36553-36561

27. Gietzen, D. W., Ross, C. M., Hao, S., and Sharp, J. W. (2004) J. Nutr. 134, 717-72328. Hao, S., Sharp, J. W., Ross-Inta, C. M., McDaniel, B. J., Anthony, T. G., Wek, R. C., Cavener, D.

R., McGrath, B. C., Rudell, J. B., Koehnle, T. J., and Gietzen, D. W. (2005) Science 307, 1776-1778

29. Maurin, A.-C., Jousse, C., Averous, J., Parry, L., Bruhat, A., Cherasse, Y., Zeng, H., Zhang, Y.,Harding, H. P., Ron, D., and Fafournous, P. (2005) Cell Metabolism 1, 273-277

30. Hagiwara, Y., Hirai, M., Nishiyama, K., Kanazawa, I., Ueda, T., Sakaki, Y., and Ito, T. (1997)Proc. Natl. Acad. Sci. U S A 94, 9249-9254

31. Sattlegger, E., Swanson, M. J., Ashcraft, E. A., Jennings, J. L., Fekete, R. A., Link, A. J., andHinnebusch, A. G. (2004) J. Biol. Chem. 279, 29952-29962

32. Sherman, F. (1991) Methods Enzymol. 194, 3-2133. Foiani, M., Cigan, A. M., Paddon, C. J., Harashima, S., and Hinnebusch, A. G. (1991) Mol. Cell.

Biol. 11, 3203-321634. Towbin, N., Staehelin, T., and Bordon, J. (1979) Proc. Natl. Acad. Sci. U S A 76, 4350-435435. Su, A. I., Cooke, M. P., Ching, K. A., Hakak, Y., Walker, J. R., Wiltshire, T., Orth, A. P., Vega,

R. G., Sapinoso, L. M., Moqrich, A., Patapoutian, A., Hampton, G. M., Schultz, P. G., andHogenesch, J. B. (2002) Proc. Natl. Acad. Sci. U S A 99, 4465-4470

36. Yang, R., Wek, S. A., and Wek, R. C. (2000) Mol. Cell. Biol. 20, :2706-271737. Chen, A., Muzzio, I. A., Malleret, G., Bartsch, D., Verbitsky, M., Pavlidis, P., Yonan, A. L.,

Vronskaya, S., Grody, M. B., Cepeda, I., Gilliam, T. C., and Kandel, E. R. (2003) Neuron 39,655-669

Present addresses: #CPqGM- FIOCRUZ, R. Waldemar Falcão, 121, Salvador, Bahia, Brasil;♦Institute of Molecular BioSciences, Massey University, Private Bag 102904, North Shore Mail Centre,Albany, Auckland 1311, New Zealand

Acknowledgements: We thank Dr. Jan van’t Riet for anti-S22 antibodies. This work was supported bygrants from Fundação de Amparo `a Pesquisa do Estado de São Paulo (FAPESP) (B.A.C. andL.E.A.M.M.) and from NIH (R.C.W.). C.M.P. and B.M.L. are recipients of post-doctoral fellowshipsfrom FAPESP. B.A.C. and L.E.A.M.M. are supported by CNPq.

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

12

Abbreviations used: eIF2α , α subunit of eukaryotic translation initiation factor 2; GST, glutathione-S-transferase; PMSF, phenyl-methylsulfonyl fluoride.

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

13

FIGURE LEGENDS

FIGURE 1. Overexpression of IMPACT in yeast causes a Gcn- phenotype. (A) Serial 10-folddilutions of overnight cultures from strain H1511 expressing GST-IMPACT (two independenttransformants are shown), GST-YIH1 or GST under the control of the galactose-inducible promoter fromplasmids pES234-6-2, pES187-B1 and pES128-9-1, respectively, were grown on minimal mediumcontaining galactose or galactose supplemented with 30mM 3-aminotriazole (3AT). (B) Immunoblot,using anti-GST serum, of whole cell extracts (4µg) prepared from the same strains as in (A), grown ingalactose. The ribosomal S22 protein was used for normalization.

FIGURE 2. IMPACT overexpression in yeast inhibits GCN2. Immunoblots of whole cell extractsprepared from yeast strains described in Figure 1 grown to exponential phase in medium containinggalactose, in the absence or presence of 30 mM 3AT, using antibodies directed to total eIF2α and to thephosphorylated form of eIF2α [eIF2α(P)]. All lanes contain 10 µg of total protein, except for the GST-expressing cells subjected to starvation conditions where 5, 10 and 20 µg were used in the immunoblotanalysis.

FIGURE 3. Inhibition of GCN2 activation by IMPACT in mammalian cells. (A) GCN2+/+ andGCN2-/- MEF cells were subjected to leucine starvation conditions for the indicated number of hours, orno stress (0), and immunoblot analyses were carried out to measure activation of the GCN2 stresspathway. (B) GCN2+/+ MEF cells without transfection, or transfected with the pCI-neo plasmid vectoralone or with the plasmid expressing IMPACT were grown in medium lacking leucine for the indicatednumber of hours, or with no stress (0). Equal amounts of protein lysates were analyzed by immunoblotusing antibodies against IMPACT, CHOP, ATF4 eIF2α(P) and β-actin.

FIGURE 4. IMPACT binds yeast and mammalian GCN1. (A) Conservation of GCN1 sequences.Alignment of the region of GCN1 around the position R2259 (numbering relative to the yeast protein)(arrow) from Mus musculus (Mm) (gi51827542), Homo sapiens (Hs) (gi41149891), Saccharomycescerevisiae (Sc) (gi477122), Neurospora crassa (Nc) (gi32410355) and Arabidopsis thaliana (At)(gi5042415), with identical residues indicated in reverse bold and conserved residues boxed. (B) GSTpull-down. yGCN1[2052-2428] wild type (R2259) and mutant (A2259), fused to GST, and GST alone(20 µg of each) purified from E. coli, were incubated with 20 µl glutathione-Sepharose beads and with500 µg of whole extracts from mouse brains. The proteins associated with the beads (100% of the boundmaterial) (pellet) were subjected to SDS-PAGE and detected by immunoblot with mono-specificantibodies raised against IMPACT (upper panel) and by Ponceau staining (lower panel). The laneslabeled “input” correspond to 1/50th of the extract used in the pull-down. (C) Co-immunoprecipitation.Protein A-agarose beads coupled to mono-specific antibodies raised against mGCN1, to an irrelevant IgG,or no antibodies were incubated with total brain extract (5 mg). Half of the material bound to the beadswas subjected to 12 % SDS-PAGE followed by immunoblot with anti-IMPACT antibodies, and the otherhalf subjected to 6 % SDS-PAGE followed by immunoblot with anti-GCN1 antibodies. The lanes withthe supernatants of the immunoprecipitations in the immunoblots with anti-IMPACT and anti-mGCN1contain 50 µg and 100 µg, respectively, of total protein.

FIGURE 5. Preferential expression of IMPACT in the brain. Immunoblots of extracts from theindicated organs (A) and brain areas (B) using anti-IMPACT antibodies (upper panel); β-actin was usedto normalize the amounts of total protein added to each lane by incubating the same filter with anti-β-actin antibodies after stripping the first antibodies (lower panel). In (B), different amounts of total proteinwere loaded (30, 15, 10 and 5 µg, as indicated by the triangles above the lanes). The ratios ofIMPACT/actin, plotted in the graph, were calculated using values obtained from different total proteinloadings running in parallel, that showed a linear range of signal for both Impact and actin, obtained from

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Pereira et al. IMPACT, a GCN2 inhibitor in mammals

14

at least 3 independent immunoblot analyses identical to the one shown in (B), and normalized to the ratioobtained for the hypothalamus. The standard deviations are shown as bars.

FIGURE 6. Differential expression of IMPACT in neuronal cells. Immunohistochemistry of (Aand B) the dentate girus area of the hippocampus, (D and E) piriform cortex and (G and H) hypothalamususing mono-specific anti-IMPACT antibodies. Control immunohistochemistry of the hippocampus (C),piriform cortex (F) and hypothalamus (I) performed with identical amounts of non-specific IgG isolatedfrom pre-immune serum by affinity purification on protein A-Sepharose. Quantification of theseantibodies relative to anti-IMPACT mono-specific antibodies was performed by dot blots using proteinA-horseradish peroxidase followed by detection with ECL. Panels A, C, D, F, G and I are shown in thesame scale, with the bar in A corresponding to 300 µm; panels B, E and H are in a higher magnification,with the bar in B corresponding to 100 µm. gcl: granule cell layer.

FIGURE 7. IMPACT levels correlate inversely with eIF2αααα(P). Varying amounts of total extracts(30, 15, 10 and 5 µg, as indicated by the triangles above the lanes) of the indicated brain areas (A) ororgans (C) were analyzed for the levels of eIF2α(P) by immunoblots using antibodies specific to thephosphorylated form of eIF2α and, after stripping, the same filters were probed with antibodies againsttotal eIF2α. The ratios eIF2α(P)/eIF2α shown in the graphs were calculated by assigning the value of 1 tothe ratio of the signal intensity of eIF2α (P) over the signal intensity of eIF2α obtained for thehypothalamus. The signal intensity values were obtained within a linear range determined from thedifferent amounts of total protein loads. The results represent data obtained from at least threeindependent immunoblots of different pools of extracts. The standard deviations are shown as bars. In(B), RT-PCR products obtained from total RNA isolated from the cortex (1), hippocampus (2) andhypothalamus (3) using oligonucleotides specific for β-actin, βGCN2 and GCN1; the left lane contains100 bp ladder DNA size standards.

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Beatriz Castilho
Figure 1

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Beatriz Castilho
Figure 2

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Beatriz Castilho
Figure 3

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Beatriz Castilho
Figure 4

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Beatriz Castilho
Figure 5

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Beatriz Castilho
Figure 6

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Beatriz Castilho
Figure 7

Jaqueta, Alan G. Hinnebusch, Ronald C. Wek, Luiz E.A. M. Mello and Beatriz A. CastilhoCátia M. Pereira, Evelyn Sattlegger, Hao-Yuan Jiang, Beatriz M. Longo, Carolina B.

inhibits GCN2 activationIMPACT, a protein preferentially expressed in the mouse brain, binds GCN1 and

published online June 2, 2005J. Biol. Chem. 

  10.1074/jbc.M408571200Access 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

by guest on March 30, 2020

http://ww

w.jbc.org/

Dow

nloaded from