Are there multiple proteolytic to c-Fos, c-Jun and p53 protein degradation in vivo

46at several places(in and out of the chromatin)and undergo phosphorylation at multiple sites.p53is a multifunctional tumor suppressor protein essential for maintaining the integrity of higher eu-caryote genomes[4,5].The functional protein is an homotetramer,the structure of which oscillates be-tween various conformations recognizable by specific sets of monoclonal antibodies[6,7].It is supposed that the various conformers show different biological activities in vivo.p53is also an unstable protein which shows a number of similarities with c-Fos and c-Jun: (i)it can localize both in the cytoplasm and in the nucleus according to the experimental/physiological conditions,(ii)its turnover is regulated upon activa-tion of intracellular signalling cascades,(iii)it can associate with various protein partners,(iv)mutant proteins found in spontaneously occurring tumors are more stable than the wild-type protein and(v)sev-eral catabolic pathways operate on it.Thus,p53is mostly degraded following the ubiquitin/proteasome pathway.However,calpains also seem to contribute to its degradation,at least under specific conditions. Where are c-Fos and c-Jun degraded?The nuclear turnover of c-Fos has long been known to be rapid.Paradoxically,the definite proof that c-Fos can be degraded within the nucleus is still lacking. Efficient destruction within this compartment is,how-ever,very likely,because of the presence of both the proteasome and of enzymes of the ubiquitin pathway [8].However,it cannot be ruled out formally that c-Fos has to return to the cytoplasm,after a post-translational modification for example,for being de-graded at the nuclear periphery.Along this line,(i)it is worth noting that the human immunodeficiency virus Rev regulator is very unstable within the cytoplasm and relatively long-lived within the nucleus[9]and (ii)it has been suggested that interaction between the Mdm2protooncoprotein and p53triggers the nuclear export of the latter for degradation into the cytoplasm [10].Work by the laboratory has shown that c-Fos transport from the cytoplasm,where it is synthesized, into the nucleus,where it plays its part as a transcrip-tion factor,is regulated by extracellular signals[11, 12].In other words,c-Fos can be retained within the cytoplasm where it shows a very short half-life indi-cating that cytoplasmic proteases,which remain to be identified,can operate on c-Fos(Figure2).Retention of c-Jun within the cytoplasm has also been docu-mented[12]but it remains to determine whether the protein is unstable in this compartment.In vitro proteolysis of c-Fos and c-Jun by calpains Milli-and micro-calpain(which are both called the ubiquitous calpains),are abundant cytoplasmic cys-teine proteases activated by calcium.They recognize structural determinants rather than primary motifs and they display a relatively strict specificity for their sub-strates which they usually cleave at only a limited number of sites(for reviews,see[13,14]).So far,only a few substrates for calpains have been identified in vivo.c-Fos and c-Jun being short-lived in the cytoplasm when their transport into the nucleus is inhibited,pos-sible degradation by calpains has been addressed by several groups.It was shown that both proteins are highly susceptible to both milli-and microcalpain in vitro[15–17].Moreover,inhibition of ubiquitous cal-pains by overexpression of their endogenous protein inhibitor,calpastatin,in transient transfection assays leads to2-3-fold higher c-Fos and c-Jun transcription activity in mouse embryofibroblasts([16]and our non-published results).This observation is consistent with the idea of a limited contribution of calpains to c-Fos and c-Jun degradation in living cells although an indirect effect of calpains could not be ruled out.Since calpains degrade their substrates to a limited extent,it remains to characterize the proteolytic system taking over for complete breakdown of c-Fos and c-Jun.In vitro study of Fos and Jun protein degradation has led to a series of interesting observations regard-ing to the mode of action of calpains.PEST sequences are peptide motifs rich in Pro,Glu,Asp,Ser and Thr commonly found in unstable proteins[18,19].They have thus been proposed to constitute instability mo-tifs as well as sites for cleavage by calpains since it was speculated that they could bind calcium,owing to their negative charge,and,thus,present the activat-ing co-factor to the protease at the level of substrates [18].However,analysis of c-Fos deletion mutants,as well as of a variety of calpain-sensitive and calpain-resistant proteins,has shown that PEST motifs are neither necessary nor sufficient for degradation by cal-pains[20].Interestingly,the viral v-FosFBR protein, but not the viral v-FosFBJ and v-JunASV17ones,is resistant to cleavage by calpains[21].Resistance to degradation is not due to the mutations in the Fos part of the protein,but to addition of exogenous motifs de-47 Figure1.Schematic representation of c-Fos and its viral counterparts.v-FosFBJ harbors5point mutations(asteriscs)plus a frameshift which replaces its last48amino acids(aa)by49others.v-FosFBR is a mutated Fos protein fused,at its N-terminus,with310amino acids derived from the Gag protein of the parental retrovirus and,at its C-terminus,with a8amino acid sequence derived from a cellular locus called fox(or fau).The Fos moiety of v-FosFBR is deleted of the N-terminal24amino acids and of the C-terminal98amino acids of c-Fos.In addition,it contains two in frame deletions of13and9amino acids(triangles)andfive point mutations that are different from those in v-FosFBJ.rived from both the gag gene of the parental retrovirus and a cellular locus called fox or fau(see Figure1). This observation is interesting since it indicates that the Fos moiety of v-FosFBR either adopts a conforma-tion significantly different of that of the same region in the cellular protein or is not as accessible as the latter, which might explain the higher oncogenic activity of the viral protein.Finally,susceptibility to calpains is primarily determined by conformational determinants of the monomers and not by the quaternary structure of c-Fos and c-Jun.The multiple cleavage sites borne by both proteins can be divided into at least two classes of sensitivity.One site located at position90–91in c-Fos is extremely sensitive.However,efficient proteolysis at this site does not show any strict dependence on the nature of the amino acids on either side of the scis-sile bond in the region extending from P2to P 2.The integrity of the monomers is not crucial for recogni-tion by calpains.Rather,recognition of sensitive sites appears to be dependent on the local conformation of peptide regions that may span several tens of amino acids and,maybe,more in the case of the identified c-Fos hypersensitive site[22].Proteolysis by calpains:a possible contribution to degradation of p53Although p53is prominently degraded following the ubiquitin/proteasome pathway,several groups have shown that calpains might also contribute to its turnover[5,23,24].In vitro,wild type p53is a sensi-tive substrate for milli-and microcalpain.Degradation is dependent on p53protein conformation.Mutants of p53with altered tertiary structure display a wide range of susceptibility to calpains,some of them be-ing largely resistant to degradation and others being more sensitive.This suggests that the different mutants tested adopt slightly different conformations,to which calpains are sensitive,but that cannot be discriminated using classical sets of monoclonal antibodies(such as PAb1620and PAb240).Inhibition of calpains using the physiological inhibitor calpastatin leads to an el-evation of p53steady-state level in human MCF7cells expressing wild type p53.Conversely,activation of calpains by calcium ionophore leads to a reduction of p53in mousefibroblasts and the effect is blocked by cell-permeant calpain inhibitors.Co-transfection of p53-null cell lines with p53and calpastatin expres-sion vectors results in an increase in p53-dependent transcription activity[23].Morover,Kubbutat and V ousden[5]have shown that inhibition of calpains in MCF7cells results in longer half-life of wild type p53and Zhang et al.[24]have reported cell cycle-dependent degradation of wild type p53by calpain,or by a calpain-like protease,in a humanfibroblast cell line.Taken together,these observations support the idea that calpains may also contribute to the regulation of wild type p53protein levels in vivo although the extent of this contribution is modest.48Figure2.Potential complexity of the degradation of c-Fos,c-Jun and p53proteins.Ubiquitous milli-and micro-calpains could contribute to the degradation of c-Fos,c-Jun and p53in vivo.This pathway would only concern the degradation of the3proteins when they have not yet entered the nucleus as calpains are cytoplamic proteases.However,it has been reported that a fraction of calpains could enter the nucleus under specific conditions.c-Fos,c-Jun and p53have been described to be substrate of the ubiquitin/proteasome pathway.However the formal proof that they are degraded within the nucleus is still lacking and it cannot be excluded that they would have to return to the cytoplasm,after a possible modification,for being broken down.A fraction of cytoplasmic c-Fos can also be selectively taken up by lysosomes for degradation. The relative contribution of these pathways in the degradation of these proteins is not known,although the ubiquitin/proteasome pathway is very likely to be the principal pathway.Characterization of the thermosensitive E1 ubiquitin-activating enzyme A31N-ts20cell line According to our current knowledge,protein ubiqui-tination involves three steps:(i)activation of ubiq-uitin through formation of an energy-rich bond with an E1ubiquitin-activating enzyme,(ii)transfer of activated ubiquitin onto E2ubiquitin-conjugating en-zymes which,in turn,(iii)alone,or in combination with E3ubiquitin-protein ligase enzymes,transfer ubiquitin onto target proteins.Cell mutants harboring a thermosensitive E1 ubiquitin-activating enzyme are useful tools for study-ing the involvement of the ubiquitin pathway in the control of specific cell functions or of the degradation of specific proteins.A31N-ts20cells are thermosen-sitive mutant cells derived from Balb/c3T3mouse embryofibroblasts subjected to chemical mutagen-esis.They were initially characterized for a defect in DNA synthesis since they accumulated in early S phase upon shift from the permissive(32◦C)to the non-permissive(39◦C)temperature(for details,see [25])The ts defect was subsequently shown to be due to a deficient E1enzyme[26]and expression of a wild type E1enzyme upon cell transfection rescued the ts phenotype[27].Since it was important to determine to which degree the ubiquitin pathway is repressed upon shift to the non permissive temperature in these cells, we have undertaken a detailed biochemical character-ization of A31N-ts20cells.Our data show that:(i) A31N-ts20-E1enzymatic activity is heat-inactivatable in vitro and(ii)a major mechanism responsible for its inactivation in vivo consists of accelerated destruction. We also showed that,surprisingly,considerable reduc-tion in E1abundance little alters the formation of the bulk of protein-ubiquitin conjugates when A31N-ts20 cells are grown at the non-permissive temperature. However,specific proteins such as p53,c-Jun and his-tone H2A are no longer ubiquitinated/degraded under these conditions.Further biochemical characterization of A31N-ts20cells argues against the existence of an E1-independent ubiquitination pathway operating in this cellular context.Rather,our data support the idea that ubiquitination of the various protein substrates in A31N-ts20cells requires different amounts of a unique E1enzyme in vivo[28].Our data thus indicate that A31N-ts20cells show a leaky phenotype with regard to the inactivation of the ubiquitin pathway.In consequence,continued degradation of a specific protein in heat-treated A31N-ts20cells must be interpreted with caution and does not necessarily signify ubiquitin-independent degra-dation.However,because of the differential behavior of certain proteins in A31N-ts20cultured at the non-49Figure 3.Degradation of c-Fos and c-Jun during the g 0-to-S phase transition.E1-thermosensitive A31N-ts20mouse embryo fibrob-lasts were grown at the permissive temperature (32◦C).Cells were serum-deprived for 36hours to bring them in G 0.They were,then,stimulated for growth by addition of 20%serum.In these con-ditions,c-fos and c-jun genes are rapidly and transiently induced (A and Ba).To identify the proteolytic pathway(s)involved in the degradation of c-Fos and c-Jun,different cell-permeant protease in-hibitors were added together with the serum and possible delay in protein disappearance was assayed by immunoblotting analysis of total cell extracts using specific antibodies.A typical experiment where the proteasome is inhibited in the presence of lactacystin is presented in (Bb).Similarly,requirement upon a functional E1ubiq-uitin-activating enzyme was tested by transfering cells from 32◦C to 39◦C 16hours prior serum stimulation (A).A typical experiment,showing that c-Jun is stabilized whereas c-Fos is not,is presented in (Bc).permissive temperature,our data indicate that such a cell line might be useful for unveiling the existence of differences in the intimate mechanisms responsible for the ubiquitination of the various cell proteins in vivo .Moreover,it should constitute an invaluable tool for providing with criteria of reliability when developing in vitro ubiquitination assays for specific proteins [28].Differential directing of c-Fos and c-Jun to the proteasome during the g 0-to-S phase transition in mouse embryo fibroblasts c-Fos and c-Jun accumulation is transiently induced at the begining of the g 0-to-S phase transition in qui-escent cells stimulated for growth (Figure 3).To ad-dress the mechanisms responsible for rapid clearance of c-Fos and c-Jun proteins under these experimen-tal conditions,we used A31N-ts20mouse embryo fibroblasts.The use of cell-permeant protease in-hibitors indicates that both proteins are degraded by the proteasome and excludes any major contribution for calpains and lysosomes during the G 0-to-S phase transition.Synchronisation of A31N-ts20cells at the non permissive temperature blocks the degradation of c-Jun,indicating that this process is E1-dependent.This observation is consistent with the observation by others that the ability to be ubiquitinated corre-lates with instability in exponentially growing human HeLa cells [29].In contrast,c-Fos is broken down according to an apparently E1-independent pathway in ts20cells.Interestingly,c-Jun is highly unstable in c-Fos-null mouse embryo fibroblasts stimulated for growth.Taken together,these observations show that in vivo during a g 0-to-S phase transition (i)the precise mechanisms triggering c-Fos and c-Jun directing to the proteasome are not identical,(ii)the presence of c-Fos is not an absolute prerequisite for the degradation of c-Jun and (iii)the degradation of c-Jun is not required for that of c-Fos.In another E1-thermosensitive hamster cell line E36-ts20[30]observed a 3-fold reduction in c-Fos degradation rate in cells grown at the non-permissive temperature.Invoking a possible leakiness of the mu-tant and taking into account that c-Fos can be ubiquiti-nated in vitro in reconstituted ubiquitinylation systems [30,31],the authors reasonably concluded to an ex-clusively ubiquitin-and E1-dependent degradation for c-Fos.Although such a scenario is highly probable,we feel that additional experiments are still required be-fore a definitive conclusion can be drawn.Particularly,it will be necessary (i)to show that c-Fos is actually ubiquitinated in vivo ,(ii)to identify the critical lysines ubiquitinated in c-Fos and to analyse the stability ofcorresponding c-Fos mutants in vivo and (iii)to de-termine whether the enzymes showing ubiquitination activity on c-Fos in vitro also operate in vivo .This last issue,however,appears complex since the possibil-ity of redundancy in c-Fos ubiquitinylation enzymes has been raised by the work of two laboratories which50have developed different cell-free c-Fos ubiquitination assays[30,31].One important conclusion of our experiments is that the degradation of c-Fos is not necessary for that of c-Jun and vice-versa.Several lines of evidence, however,raise the possibility that interaction with c-Jun may modulate the degradation rate of c-Fos. Stancovski et al.[30]have observed that,in an in vitro assay,c-Jun significantly stimulates both the ubiquiti-nation and the degradation of c-Fos.In another in vitro degradation assay,c-Jun,ubiquitin,MAP kinase,ca-sein kinase II and cdc2kinase are all necessary for the rapid breakdown of c-Fos[32].In another cell-free as-say,addition of an appropriately phosphorylated c-Jun can trigger the degradation of c-Fos in the reticulocyte lysate[33].In contrast,delayed degradation of c-Fos correlates with association with Jun family members in Swiss3T3cells stimulated with serum[34,35].It is worth noting that Fos and Jun family members are constantly associating and dissociating in vivo with an interaction half-life of no more than a few minutes [34].One mechanism for c-Fos degradation taking into account all these observations would be that c-Fos is(preferentially)recognized by the proteasome in the form of a monomer and that the various kinases influencing its half-life would in fact affect the disso-ciation and/or association rates of dimers.Analysis of dimerization mutants of c-Fos might help to solve this issue.Finally,contrasting with our observation that c-Jun breakdown is both proteasome-and E1-dependent during the G0-to-S phase transition and,complicating our view of c-Jun degradation,we have previously observed that this protein can be specifically and ef-ficiently proteolyzed by the26S proteasome in an E1-and ubiquitin-independent manner in in vitro degra-dation assays using either rat liver cell extracts or purified rat liver proteasome[36].Whether ubiquitin-independent breakdown of c-Jun by the proteasome corresponds in vivo to rapid degradation under specific physiological conditions or in specific cell types or to basal degradation accelerated upon ubiquitination is however not known.Such a situation has already been described in the case of the IkBa transcription factor inhibitor,whose ubiquitin-independent basal degradation becomes ubiquitin-dependent in TNF-treated cells with a half-life shortening from30to5 minutes[37].Experiments are underway for determin-ing whether such a mechanism also applies to c-Jun in vivo.AcknowledgementsThis work was supported by grants from the Centre National de la Recherche Scientifique,the Ligue Na-tionale contre le Cancer,the Association de Recherche contre le Cancer(ARC),the Agence Nationale de Recherche contre le Sida(ANRS)and the EC Biotech Programm.References1.Angel P&Herrlich P(1994)(Angel P&Herrlich P,Eds.)CRC press,Boca Raton2.Piechaczyk M&Blanchard J-M(1994)Critic.Rev.Oncol.Hematol.17:93–1313.Salvat C,Jariel-Encontre I,Pariat M,Acquaviva C,Robbins I&Piechaczyk m(1997)Recent Res.Devel.in Molecular Biol.1:1–274.Oren M(1994)Semin.Cancer Biol.5:221–2275.Kubbutat M&V ousden K(1998)Mol.Med.Today4:250–256ner J(1994)Cancer Biol.5:211–219ner J(1995)Trends Biochem.Sci.20:49–518.Coux O,Tanaka K&Goldberg AL(1996)Ann.Rev.Biochem.65:801–8479.Kubota S,Duan L,Furuta RA,Hatanaka M&Pomerantz R(1996)J.Virol.70:1282–128710.Roth J,Dobblestein M,Freedman DA,Shenk T&Levine AJ(1998)EMBO J17:554–56411.Roux P,Blanchard J-M,Fernandez A,Lamb N,Jeanteur P&Piechaczyk M(1990)Cell63:341–35112.Roux P,Carillo S,Blanchard J-M,Jeanteur P&Piechaczyk M(1994)in:Angel P&Herrlich P(Ed)The c-fos and c-jun fam-ilies of transcription factors(pp87-93)CRL Press,Bocaraton, Floride,USA13.Croall DE&DeMartino GN(1991)Physiol.Rev.71:813–84714.Sorimachi H,Ishiura S&Suzuki K(1997)Biochem.J.328:721–73215.Carillo S,Pariat M,Steff A-M,Roux P,Etienne-Julan M,Lorca T&Piechaczyk M(1994)Oncogene9:1679–1689 16.Hirai S-I,Kawasaki H,Yaniv M&Susuki K(1991)FEBSLett.287:57–6117.Watt F&Molloy PL(1993)Nucl.Acids Res.21:5092–510018.Rogers S,Wells R&Rechsteiner M(1986)Science234:364–36819.Rechsteiner M&Rogers S(1996)Trends Biochem.Sci.21:267–27120.Carillo S,Pariat M,Jariel-Encontre I,Steff A-M,Poulat F,Bertha P&Piechaczyk M(1996)Biochem.J.313:245–251 21.Steff A-M,Carillo S,Pariat M&Piechaczyk M(1997)Biochem.J.323:685–69222.Pariat M,Salvat C,Bébien M,brockly F,Carillo S,Jariel-Encontre I&Piechaczyk M(submitted)23.Pariat M,Carillo S,Mollinari M,Salvat C,Debüssche L,Bracco L,Milner J&Piechaczyk M(1997)Molec.Cell Biol.17:2806–281524.Zhang W,Lu Q,Xie ZJ&Mellgren R(1997)Oncogene14:255–26325.Zeng G-C,Donegan HL,Ozer HL&Hand R(1984)Mol.CellBiol.4:1815–18225126.McGrath JP,Jentsch S&Varshavsky A(1991)EMBO J.10:227–23627.Chowdary DR,Bermody JJ,Jha KK&Ozer HL(1994)Mol.Cell Biol.14:1997–200328.Salvat C,Acquaviva C,Scheffner M,Robbins I,PiechaczykM&Jariel-Encontre I(submited)29.Treier M,Staszewsk LM&Bohman D(1994)Cell78:787–79830.Stancovski I,Gonen H,Orian A,Schwartz AL&CiechanoverA(1995)Mol.Cell Biol.15:7106–711631.Hermida-Matsumoto M-L,Chock PB,Curran T&Yang DCH(1996)J.Biol.Chem.271:4930–493632.Tsurumi C et al.(1995)Mol.Cell.Biol.15:5682–568733.Papavassiliou AG,Treier M,Chavrier C&Bohman D(1992)Science258:1941–194434.Kovary K&Bravo R(1992)Molec.Cell Biol.12:5015–502335.Kovary K&Bravo R(1991b)Mol.Cell Biol.11:4466–447236.Jariel-Encontre I,Pariat M,Martin F,Carillo S,Salvat C&Piechaczyk M(1995)J.Biol.Chem.270:11623–11627 37.Krappmann D,Wulczyn FG&Scheidereit C(1996)EMBOJournal15:6716–6726。