Sandra kleiner, Joshi Venugopal, Yoshikuni Nagamine
Upcoming SlideShare
Loading in...5

Sandra kleiner, Joshi Venugopal, Yoshikuni Nagamine



First, I would like to thank my parents who ...

First, I would like to thank my parents who
supported me through all my life and are
always there for me when I need them.
I also wish to acknowledge Dr. Yoshikuni
Nagamine, who supervised me, and gave me the opportunity to develop my scientific
thinking and skills in his lab. While he offered scientific freedom, he was always available for discussions. I greatly appreciate that.

Thanks also to Prof. Gerhard Christofori and Prof. Fred Meins, the two other members of my thesis committee, for the advice they gave me during the committee meeting and for the time they will still have to invest to read and evaluate this thesis.

My special thanks go to Joshi Venugopal,
from whom I could learn a lot in many aspects of life and who became a close friend of mine. His critical and logical thinking inspired me in several things. I really appreciate the time we spent together.

I also want to acknowledge Malgorzata
Kiesielow for our fruitful collaboration and for
teaching me siRNA transfections.
In addition, I wish to thank my former and
current lab members Faisal, Hoanh, Fumiko,
Kacka, Sandra and Stephane. I always
enjoined working and spending some free time
with all of you.
Further, I want to acknowledge the technical
staff at the FMI who were always friendly and
helpful and made the scientific life at the FMI
much easier and productive. Thanks go to all
of the FMI members (especially from the
Hynes laboratory) and to all of those, who
provided me with scientific material: François
Lehembre, Kurt Ballmer, Tony Pawson, Peter
E. Shaw and Jerrold Olefsky. My thanks also
go to Pat King and Sara Oakley for critical
reading of my manuscripts.
My heartiest gratitude goes to Boris
Bartholdy who always supported me
scientifically with all of his skills and privately
with all of his love.



Total Views
Views on SlideShare
Embed Views



0 Embeds 0

No embeds


Upload Details

Uploaded via as Adobe PDF

Usage Rights

CC Attribution-ShareAlike LicenseCC Attribution-ShareAlike License

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
Post Comment
Edit your comment

Sandra kleiner, Joshi Venugopal, Yoshikuni Nagamine Sandra kleiner, Joshi Venugopal, Yoshikuni Nagamine Document Transcript

  • ISOFORM-SPECIFIC ROLES OF THE ADAPTOR PROTEIN SHCA IN CELL SIGNALING Inauguraldissertation zur Erlangung der Würde eines Doktors der Philosophie vorgelegt der Philosophisch-Naturwissenschaftlichen Fakultät der Universität Basel von SANDRA KLEINER aus Weißenborn, Deutschland Dissertationsleiter: Dr. Yoshikuni Nagamine Friedrich Miescher Institute for Biomedical Research BASEL, 2005
  • Genehmigt von der Philosophisch-Naturwissenschaftlichen Fakultätauf Antrag vonProf. Fred Meins, Dr. Yoshikuni Nagamine, Prof. Gerhard Christofori und Prof.Patrick MatthiasBasel, den 22.11.2005 Prof. Dr. Hans-Jakob Wirz (Dekan) 2
  • TABLE OF CONTENT TABLE OF CONTENTSUMMARY 51. INTRODUCTION 6 1.1 The Shc adaptor proteins 6 1.1.1. Genomic and structural organization of Shc 6 Genomic organization and regulation of Shc expression 6 Structural organization of Shc proteins 8 1.1.2 Signaling and function of ShcA 9 Role of Shc in mitogenic Ras/Erk signaling 9 Role of Shc in c-myc activation and cell survival 11 Role of Shc in cell adhesion, migration, and cytoskeletal organization 11 Role of Shc in tumorigenesis 12 In vivo function of Shc 13 Conventional Shc knockout 13 Conditional T-cell specific knockout and transgenic mice 13 Shc Role of p66 14 1.2 Signaling of the E-cadherin cell-cell adhesion protein 19 1.2.1 E-cadherin-dependent cell-cell adhesion 19 E-cadherin: a member of the classical cadherins 19 Function of catenins in the E-cadherin adhesion complex 20 Function of the E-cadherin-catenin complex 21 1.2.2 E-cadherin as a tumor suppressor 22 1.2.3 E-cadherin-mediated signaling 23 1.3 RNA interference: a new and powerful tool in molecular biology 28 1.4 Research objectives 312. RESULTS 32 2.1 Research communication 32 Isoform-specific knockdown and expression of adaptor protein ShcA using small interfering siRNA INTRODUCTION 32 EXPERIMENTAL 33 RESULTS 34 DISCUSSION 36 REFERENCES 37 2.2 Using siRNAs to study Shc function 39 3
  • TABLE OF CONTENT 2.2.1 Isoform-specific knockdown of p46/52Shc 39 2.2.2 Growth inhibition upon Shc knockdown 39 2.3 Role of Shc in EGF-induced signaling in epithelial cells 43 2.3.1 Role of Shc in EGF-induced Erk activation 43 Shc 2.3.2 Effect of p66 on EGF-driven proliferation and cell survival 44 2.4 Research Publication (under review) 46 Induction of uPA gene expression by the blockage of E-cadherin via Src- and Shc-dependent Erk signaling INTRODUCTION 46 MATERIALS AND METHODS 47 RESULTS 48 DISCUSSION 53 REFERENCES 56 2.5 Supplementary data to 2.4 59 2.5.1 Role of FAK in Decma-induced Erk activation 59 2.5.2 Disruption of cell-cell adhesion using EGTA in LLC-PK1 cells 59 2.6 Role of p66Shc in regulating cell survival in epithelial cells 613. DISCUSSION 63 3.1 Isoform-specific knockdown and knockdown-in of Shc using siRNA 63 3.2 Role of Shc in mediating Erk activation 65 3.2.1 Shc is dispensable for EGF-induced Erk activation 65 3.2.2 Shc mediates Erk activation downstream of E-cadherin 67 3.3 The role of p66Shc in stress response 70 3.4 Isoform-specific role of p46Shc 71 3.5 Conclusion 714. MATERIAL AND METHODS 735. REFERENCES 746. ACKNOWLEDGEMENTS 847. ABBREVIATIONS 858. CURRICULUM VITAE 86 4
  • SUMMARY SUMMARY ShcA is a bona fide adaptor protein without We used this technique to investigate theany enzymatic activity. Upon activation of contribution of individual ShcA isoforms toreceptor tyrosine kinases, ShcA associates EGF-induced MAPK activation in epithelialwith the receptor and becomes tyrosine cells. Knockdown of all or single ShcAphosphorylated. Phosphorylated ShcA recruits isoforms had no effect on EGF-induced Erk Shcthe Grb2/SOS complex to the membrane, activation. Moreover, overexpression of p66 Shcwhere SOS stimulates the small GTPase Ras, in non p66 -expressing MCF7 cells did notresulting in the activation of the Ras/MAPK change EGF-induced proliferation or viability.pathway. The fact that Grb2 binds directly to These data suggest that EGF-induced MAPKmost of the receptor tyrosine kinases raises activation in epithelial cells is ensured by athe question of how important is the role of Shc redundant coupling of Grb2 to the mediating MAPK activation? Moreover, In a quest for growth factor-independentbeside growth factor-induced MAPK activation, pathways involving Shc-mediated Erkare there other pathways in which ShcA- activation, we investigated signalingmediated MAPK activation is relevant? downstream of the cell-cell adhesion molecule ShcA is expressed in three different E-cadherin. We identified a previously Shc Shc Shcisoforms: p46 , p52 , and p66 . These unknown signaling pathway which is inducedisoforms are all derived from a single gene and upon disruption of E-cadherin-dependent cell-differ only in their N-terminal part. Although all cell adhesion This pathway involves Src- andisoforms are phosphorylated by receptor Shc-dependent Erk activation, which resultstyrosine kinases, and subsequently bind to subsequently in the expression of the ShcGrb2, the p66 isoform does not seem to urokinase plasminogen activator. Applying the Shcmediate MAPK activation. The individual knockdown-in technique revealed that p46contribution of p46Shc and p52Shc in mediating and p52Shc, but not p66Shc, were able toMAPK activation is also not clear. The fact that mediate MAPK activation upon disruption ofall isoforms are ubiquitously expressed, with cell-cell adhesion. This pathway directly links Shcsome restrictions for p66 , complicates the disruption of cell-cell adhesion with theexperimental investigation of each isoform. expression of proteolytic enzymes, both ShcRecently, p66 has been implicated in the processes involved in metastasis and woundregulation of apoptosis in response to oxidative healing. Shcstress. To learn more about the role of p66 in Using siRNA, we established a system which mediating oxidative stress-induced apoptosis Shcallows isoform-specific knockdown of ShcA in epithelial cells, the effect of p66 on cell Shcproteins in tissue culture. Further development viability was investigated. Although p66 hasof this technique enabled us to express a been shown to enhance stress-inducedsingle isoform in the absence of endogenous apoptosis in fibroblasts, endothelial cells, andprotein. This so-called “knockdown-in” T-cells, no effect on p66Shc expression wastechnique is applicable for most proteins which observed in two different epithelial cells,are expressed in multiple isoforms, and allows suggesting that the apoptotic response inthe investigation of specific mutations against epithelial cells is mediated in a p66Shc-a clear background without overexpression. independent manner. 5
  • INTRODUCTION 1. INTRODUCTION This chapter provides insights into three 1.1.1. Genomic and structuraldifferent topics: (i) function of Shc proteins, (ii) organization of ShcE-cadherin-mediated cell-cell adhesion and (iii)RNA interference. Genomic organization and regulation of Shc expression1.1 The Shc adaptor proteins The human shc locus maps to the Shc proteins are prototype adaptor proteins chromosome 1q21 (Huebner et al., 1994). Itwhich represent molecules that possess no contains 13 exons, which give rise to threeapparent catalytic domains or activities. different gene products: three isoforms ofAdaptor proteins contain modular protein- about 46, 52, and 66 kDa. All isoforms areprotein and protein-lipid interaction domains, generated either through RNA splicing orsuch as src-homology domain 2 (SH2) and 3 alternative translational initiation (Migliaccio et(SH3), phosphotyrosine binding domain (PTB), al., 1997; Pelicci et al., 1992) (Fig. pleckstrin homology (PH) domains, and While the p46Shc/p52Shc transcript originatesare essential in propagating signals from a from the assembly of the non-coding exon 1receptor in a coordinated fashion (Zhang et al., with the 3 portion of exon 2 (exon 2a), and2002). with exons 3−13, the p66Shc transcript is The adaptor protein ShcA was initially formed by the assembly of exons 2-13. Aidentified as an SH2-containing proto- second mechanism that regulates transcriptiononcogene involved in growth factor signaling. of the three Shc isoforms is the alternativeSince than, it has been shown to be an integral usage of in-frame translational start codons.component implicated in the action of a wide The transcript encoding p66Shc has three in-variety of receptors, including receptor tyrosine frame ATGs that are responsible for the Shckinases (RTKs), G protein-coupled receptors translation of p66 , and, to a lesser extent, Shc Shc(GPCRs), immunoglobulin receptors, and p52 , and p46 . The p52Shc/p46Shc transcriptintegrins, as well as non-receptor tyrosine contains two in-frame ATGs that are Shckinases such as Src and FAK. To date, three responsible for the translation of p52 and Shcmammalian shc genes have been identified: p46 (Migliaccio et al., 1997). The mouse shcshcA, shcB (sck), and shcC (N-shc/rai) locus is similarly organized and maps to(Nakamura et al., 1996; OBryan et al., 1996; chromosome 3 (Kojima et al., 2001; MigliaccioPelicci et al., 1996). All three shc genes et al., 1997).encode proteins that are highly related in Less is known about the moleculardomain and structure. In the following section, I mechanisms that regulate the differentialwill provide an overview of the genomic expression of the various Shc isoforms. Itorganization and structural architecture of seems that different mechanisms control theShcA, hereafter referred to as Shc, along with expression of the two main Shc transcripts in Shc Shcits known functions in signal transduction. different cell types. p46 /p52 are found 6
  • INTRODUCTIONubiquitously in every cell type, whereas p66Shc engagement of CD4 and CD3 (Pacini et al.,expression varies and is restricted to certain 2004). In vivo, p66Shc expression has beentissues and cell lines, being absent in brain, in found to be induced in circulating peripheralmost hematopoietic cell lines, in peripheral blood mononuclear cells of diabetic patientsblood lymphocytes (PBL), and in a subset of (Pagnin et al., 2005).breast cancer cell lines (Jackson et al., 2000; Overall expression analysis has shown thatPelicci et al., 1992; Stevenson and Frackelton, Shc is expressed at its highest levels in the1998; Xie and Hung, 1996). Ventura et al. placenta, adipocytes, bronchial-epithelial cells,(Ventura et al., 2002) have recently identified colorectal adenocarcinoma, cardiac myocytes,epigenetic modifications, namely histone and smooth muscle cells of humans (humandeacetylation and cytosine methylation, as GNF SymAtlas).mechanisms underlying transcriptional The family members ShcB and ShcC are shcsilencing of p66 in specific cell types. derived from different genes, and theirHistone deacetylase inhibitors, or expression is restricted to the brain anddemethylating agents, were capable of neuronal tissue (Nakamura et al., 1996; Shcrestoring p66 expression in primary, OBryan et al., 1996; Ponti et al., 2005). Unlikeimmortalized, and transformed cells. shcA, only two isoforms are encoded by the shcAdditionally, the p66 -encoding locus could shcB and shcC reactivated in human PBL and mouse T-cells by treatment with a variety of apoptogenicstimuli, such as H2O2, the calcium ionophoreA23187, Fas ligation, and sequentialFigure of humanShc locus and exonassembly of Shctranscripts. A schematicrepresentation of the exonassembly in the shc shc shcp52 /p46 and p66encoding transcripts. Shcexons are indicated byboxes (black boxes are translated exons), the exon numbers are given above, and the splicing eventsare shown by the zig-zag line. The position of the three Shc ATGs is indicated below the exons (asdescribed in (Migliaccio et al., 1997)). 7
  • INTRODUCTION1.1.1.2 Structural organization of Shc (IRS-1/2), tensin, the epidermal growth factorproteins receptor (EGFR) pathway substrate (Eps8), and the integrin cytoplasmic domain- Shc proteins are characterized by their associated protein-1 (ICAP-1) (Schlessingerspecific modular organization, consisting of an and Lemmon, 2003).amino-terminal phosphotyrosine-binding (PTB) The PTB domain shows remarkabledomain, a central proline- and glycine-rich structural similarity to pleckstrin homology (PH)collagen homology domain (CH1), and a domains, despite a very divergent primarycarboxy-terminal Src homology 2 (SH2) sequence (Zhou et al., 1995c). In a similar waydomain (Fig. The unique feature to PH domains, the Shc-PTB domain has beenthereby is the arrangement of the PTB and the shown to bind acidic phospholipids such asSH2 domain in an N to C order (Luzi et al., PI(4,5)P2 and PI(4)P (Zhou et al., 1995c), and2000). Shc proteins are evolutionarily well also PI(3,4,5)P3 (Rameh et al., 1997). Theconserved and can be found in mammals, high affinity (KD=10-50 µM) of this bindingfishes, flies and worms. suggests that the interaction of Shc with the membrane could occur independently of an interaction with tyrosine-phosphorylated receptors. Consistent with this idea was the identification of residues within the Shc-PTB domain that are critical for phospholipid binding and membrane localization and are distinct from the residues necessary for posphpo-Figure Domain structure of Shc tyrosine binding (receptor binding). Over theproteins. All Shc isoforms share the same last few years many different proteins, such asmodular organization: N-terminal PTB domain, F-actin, SHIP (SH2-containing inositolcentral collagen homology domain (CH1), and polyphosphate 5 phosphatase), IRS-1 and ShcC-terminal SH2 domain. p66 contains an PP2A (protein phosphatase type 2A), haveadditional collagen homology domain (CH2). been found to bind to the Shc-PTB domain in aAll known phosphorylation sites are indicated. phosphotyrosine-dependent or -independent manner (Kasus-Jacobi et al., 1997; Lamkin et A second phosphotyrosine-binding (PTB) al., 1997; Thomas et al., 1995; Ugi et al.,domain, distinct from the SH2 domain, was 2002).discovered in Shc proteins (Blaikie et al., 1994; On the N-terminal edge of the PTB domainKavanaugh and Williams, 1994). The unique of p52Shc and p66Shc there is a serinefeature of the Shc-PTB domain is that its phosphorylation site (Fig. (El-Shemerlybinding to target sequences is determined by et al., 1997). Further studies haveresidues N-terminal to the phosphotyrosine, demonstrated that phosphorylation of this siteand is not influenced by residues C-terminal to is necessary for Shc binding to thethe phosphotyrosine (Blaikie et al., 1997; Trub phosphatase PTP-PEST and downregulationet al., 1995; Zhou et al., 1995a). Today, more of insulin-induced Erk activation, most likelythan 160 proteins containing a PTB domain areknown, including insulin receptor substrate 1/2 8
  • INTRODUCTIONthrough dephosphorylation of Shc (Faisal et The third conserved region maps as aal., 2002). binding site for adaptins which links the The SH2 domain of Shc is located at the C- endocytic machinery of clathrin-coated pitsterminus and was thought to be the only with integral membrane proteins, suggesting adomain responsible for the recruitment of Shc potential role of Shc in endocytosis. Thisto activated growth factor receptors before the region is only weakly conserved in Drosophilaidentification of the Shc-PTB domain. It folds in (Lai et al., 1995). Shca very similar manner to other SH2 domains p66 contains an additional N-terminal CH-(Mikol et al., 1995; Zhou et al., 1995b). Unlike like domain (called CH2) (Migliaccio et al.,the Shc-PTB domain, the target binding of the 1997), which is also found in the longerShc-SH2 domain is determined by residues C- isoforms of ShcB and ShcC, but not in theterminal to the phosphotyrosine Drosophila Shc protein (Luzi et al., 2000). In(Ravichandran, 2001). contrast to the CH1 domain, the CH2 domain Between the PTB and the SH2 domain is the can be serine/threonine phosphorylated incollagen homology (CH) 1 domain. This region response to several stimuli such as oxidativeis characterized by a large number of glycine stress (Migliaccio et al., 1999), 12-O-and proline residues, but does not feature tetradecanoylphorbol-13 acetate (TPA) (El-typical collagen-like repeats. While the PTB Shemerly et al., 1997), and epidermal growthand the SH2 domains share high similarity, factor (EGF) (Okada et al., 1997). The78% and 68% respectively, the CH1 domain is phosphorylation of serine 36 (S36) has been Shcgenerally less well conserved between linked to the role of p66 in oxidative stressdifferent species. However, within the response (Migliaccio et al., 1999) and will bemammalian Shc family members, three regions discussed later. The physiological relevance ofsharing a higher degree in homology are the threonine phosphorylation site (T29) haspresent in this domain. Two of these not yet been defined.conserved regions comprise three criticaltyrosine phosphorylation sites, Y239, Y240, 1.1.2 Signaling and function of ShcAand Y317, and additional amino acidssurrounding the amino-terminal Role of Shc in mitogenic Ras/Erkphosphorylation site suggesting an important signalingrole in the recognition of effector proteins(OBryan et al., 1996). Y317 is conserved in In vivo and in vitro studies from variousmammalian Shc proteins, but not seen in those laboratories have clearly established a role forof lower organisms. Y239 and Y240 are also Shc in Ras/MAPK activation (Lai and Pawson,present in Drosophila Shc (Lai et al., 1995), but 2000; Pratt et al., 1999; Salcini et al., 1994).Shc in C. elegans does not contain any of the This is the only function of Shc of which thetyrosine residues (Luzi et al., 2000). Both molecular mechanism is understood. Activationphosphorylation sites conform to the of RTKs results in the recruitment of Shcconsensus Grb2-binding site and have been proteins and, subsequently, in Shcdemonstrated to bind Grb2 (Velazquez et al., phosphorylation. Phosphorylated, hence2000; Walk et al., 1998). activated, Shc binds to the Grb2/SOS complex. 9
  • INTRODUCTIONThe Shc/Grb2/SOS complex is then localized preferentially in complexes that also containto the membrane through the interaction of Shc Shc (Buday et al., 1995; Pronk et al., 1994;with the phosphorylated receptor via its PTB or Ravichandran et al., 1995). Still, manySH2 domain (Blaikie et al., 1994; Pelicci et al., receptors are able to directly recruit the1992; Ravichandran et al., 1993). At the Grb2/SOS complex, leading to Ras activationmembrane in vicinity to Ras, SOS stimulates without the involvement of Shc (Arvidsson etnucleotide exchange on Ras and, thereby, al., 1994; Batzer et al., 1994; Schlaepfer et al.,activation of Ras (Fig. (Ravichandran, 1998). In response to integrin ligation,2001). GPCR, integrins, and cytokine however, Shc is necessary and sufficient forreceptors without intrinsic tyrosine kinase activation of the MAP kinase pathway (Wary etactivity utilize other soluble and associated al., 1996). The ability of Shc to mediate Rastyrosine kinases to phosphorylate Shc activation is largely dependent on the three(Sayeski and Ali, 2003; Velazquez et al., 2000; tyrosine residues within its CH1 domain.Wary et al., 1996). In addition to translocating Phosphorylation-deficient mutants exertthe Grb/SOS complex to the membrane, Shc dominant-negative activity, whereby theseems to influence the extent of Ras importance of distinct Shc tyrosines differsactivation. The Shc/Grb2 interaction increases between the cell types and receptorsthe level of SOS bound to Grb2 in some (Ravichandran, 2001).systems, and SOS has been foundFigure Model for Shc-mediated Ras activation downstream of RTK. Shc binds to RTKsand recruits the Grb2/SOS complex which activates Ras. See text for details. 10
  • INTRODUCTION1.1.2.2 Role of Shc in c-myc activation and phosphorylation via the Shc/Grb2/Gab2/PI3Kcell survival pathway, and might therefore be involved in the regulation of IL-2-mediated cell survival The observation that Shc is involved in c- (Fig. (Gu et al., 2000).myc activation has led to two suggestions. The involvement of ShcB and ShcC inFirst, Shc might play a role in signaling other survival of neuronal cells has become morethan mediating Ras/MAPK activation and, evident. Whereas ShcA is only expressed insecond, the downstream signaling of proliferating neuroblasts and is downregulatedY239/Y240 and Y314 might have distinct in post-mitotic neurons, ShcB and ShcCproperties (Fig. In BaF cells, Gotoh et remain expressed (Cattaneo and Pelicci, 1998;al. (Gotoh et al., 1996) showed that Shc could Conti et al., 1997). Mice with no ShcB and/orinduce c-myc expression in response to IL-3 ShcC expression display a loss of certain typesstimulation which was dependent on of peptidergic and nociceptive neurons (SakaiY239/Y240, but not on Y137. The same et al., 2000). It appears, therefore, that ShcAsituation was demonstrated for EGF signaling plays a role in neuronal proliferation, but ShcBin NIH3T3 cells (Gotoh et al., 1997). and ShcC isoforms play a role in survival ofSubsequently, a role for Shc in c-myc gene post-mitotic neurons.activation has been shown in IL-2 signaling(Lord et al., 1998), in PDGF signaling (Blake etal., 2000), and in T-cell antigen receptor (TCR)signaling (Patrussi et al., 2005). However, itremains unclear how Shc mediates c-mycactivation and what target genes are in turnaffected by c-Myc. Induced c-myc expression downstream of IL-2/3 and TCR correlated with survival signals in Figure Distinct signaling capacitieshematopoetic cells (Gotoh et al., 1996; Lord et of the major tyrosine phosphorylation, 1998; Patrussi et al., 2005), suggesting an The three tyrosine phosphorylation sites andinvolvement of Shc in the regulation of a pro- the signaling linked to these tyrosines aresurvival pathway via c-myc. Lord et. al. (Lord et, 1998) observed Shc-dependent inductionof proliferation and expression of c-myc, bcl-2 Role of Shc in cell adhesion,and bcl-x in response to IL-2. Nevertheless, migration, and cytoskeletal organizationthe proliferative response and the expressionof bcl-family genes were not sufficient to The implication of Shc in processes such asmediate sustained cell survival and cell adhesion, migration, and cytoskeletalantiapoptotic effects associated with a organization originates from diverse reports incomplete IL-2 signal in murine T-cells. In a different contexts.different study, a Shc chimera fused to the IL-2 Embryonic fibroblasts derived from Shc-receptor β chain that lacks other cytoplasmic knockout mice have defects in spreading ontyrosines was able to evoke PKB/AKT fibronectin (Lai and Pawson, 2000). Similarly, 11
  • INTRODUCTIONthe regulation of cell adhesion and EGF- Role of Shc in tumorigenesisinduced migration on fibronectin required theinteraction of Shc and α5β1 integrin in MCF7 The ability of Shc to mediate mitogenicbreast cancer cells (Mauro et al., 1999; Nolan signaling raises the question of whether Shcet al., 1997). In addition, Shc has been shown can drive tumorigenesis. Although Shc proteinsto localize to focal adhesions and to interact do not contain any enzymatic activity, Shcwith the focal adhesion kinase (FAK) (Barberis overexpression of p46/52 was able toet al., 2000; Gu et al., 1999). Although Shc can transform mouse fibroblasts and to enablebe a substrate of FAK (Schlaepfer et al., 1998), them to form tumors in nude mice (Pelicci ettheir effects on cell migration seem to be al., 1992). In tumor cells with known tyrosinedistinct. While Shc stimulates random cell kinase gene alteration, Shc proteins weremotility through activation of the Erk signaling found to be constitutively phosphorylated andpathway, FAK regulates directional persistent complexed with Grb2 and activated tyrosinemigration via p130Cas (Gu et al., 1999). In kinases (EGFR, PDGFR, ErbB-2, Met, BCR-ErbB2-driven migration, Shc seems to be Abl, and Ret) (Pelicci et al., 1995b).required for lamellipodia formation Underscoring the role of Shc in oncogenic RTK(reorganization of the actin cytoskeleton) and signaling, dominant negative Shc has beenfor mediating the interaction between the shown to block proliferation of ErbB-2 positivereceptor and Memo, which is necessary for cell human breast cancer cell lines (Stevenson etmigration-required reorganization of the al., 1999).microtubule network (Marone et al., 2004). In More recently, an in vivo study has unveiledsupport of this report, inhibition of EGF- an unsuspected role for the Shc in RTK-induced cell migration upon downregulation of mediated vascular endothelial growth factorShc has also been observed in a different (VEGF) production and tumor angiogenesisstudy (Nolan et al., 1997). In response to HGF, (Saucier et al., 2004). Using RTK engineeredoverexpression of Shc enabled enhanced to recruit a defined signaling protein, it wasmigration and growth of melanoma cells shown that the direct recruitment of either Grb2(Pelicci et al., 1995a). Whether Shc stimulates or Shc to an RTK oncoprotein is sufficient toproliferation or migration seems, at least induce transformation and metastasis (Saucierpartially, to be determined by external stimuli. et al., 2002). The authors then extended thisIn the presence of growth factors, Shc study in order to compare and define the roleregulates DNA synthesis, but under growth of Shc and Grb2 in RTK oncoprotein-drivenfactor-limiting conditions, Shc stimulates cell tumorigenesis (Saucier et al., 2004).migration (Collins et al., 1999). To what extent Fibroblasts expressing Shc-binding RTKboth responses depend on Shc-induced MAPK oncoproteins induced tumors with short latencyactivation, or activation of and cross talk with (approximately 7 days), whereas cellsother signaling pathways, is not clear. expressing Grb2-binding RTK oncoproteinsHowever, in one case, a direct interaction induced tumors with delayed latencybetween Shc and F-actin has been observed in (approximately 24 days). The early onset ofPC12 cells in response to NGF (Thomas et al., tumor formation resulted in the ability of Shc-1995). binding RTK oncoproteins to produce (VEGF) 12
  • INTRODUCTIONin culture and an angiogenic response in vivo. growth factors. Shc-deficient mouse embryonicMoreover, the use of fibroblasts derived from fibroblasts (MEFs) also showed changes inShc-deficient mouse embryos demonstrated focal contact organization and actin stressthat Shc was essential for the induction of fibers when plated on fibronectin, underscoringVEGF by the Met/hepatocyte growth factor the role of Shc in cytoskeletal organization.RTK oncoprotein and by serum-derived growthfactors. Conditional T-cell specific knockout and transgenic mice1.1.2.5 In vivo function of Shc Efforts over the past 10 years have1. Conventional Shc knockout demonstrated that Shc plays a critical role in T- cell receptor (TCR) signaling. The earliest The conventional knockout mouse created evidence linking Shc to TCR-mediatedby Lai and Pawson (Lai and Pawson, 2000) signaling was the observation that Shcclearly established a role for Shc in vivo. becomes tyrosine phosphorylated rapidly afterAblation of exons 2 and 3, which encode the TCR/CD3 crosslinking (Ravichandran et al.,PTB domain, by gene targeting resulted in a 1993). Several studies followed showing thatloss of expression of all three Shc isoforms in expression of dominant negative mutants ofhomozygous mutants. The homozygous Shc inhibited TCR-mediated downstreammutant embryos died at day 11.5 with severe signaling (Milia et al., 1996; Pacini et al., 1998;defects in heart development and Pratt et al., 1999). To examine the relativeestablishment of mature blood vessels. The significance of Shc compared to several othercardiovascular system showed defects in adaptors in T-cells, two genetic approachesangiogenesis and cell-cell contacts. Consistent were taken in mice (Zhang et al., 2002). Thewith this, Shc was mainly expressed in the first approach involved the generation of acardiovascular system of wild-type embryos. transgenic mouse with thymocyte-specificThe Shc∆ex2/3 mutants also provided evidence expression of a dominant negative form of Shc,for Shc in MAPK signaling in vivo. There was a where all tyrosine residues were mutated toloss of MAPK activation within the phenylalanine (ShcFFF). The ShcFFF transgenic mice had a reduced thymus size,cardiovascular system of the Shc∆ex2/3 mutants, with significant reduction in thymocyteas revealed by whole mount immunostaining numbers. Further analysis revealed that T-with phospho-specific Erk antibodies, when ShcFFF cells were blocked at the double negativecompared to wild-type embryos. Studies with stage (DN) of their development, which wasShc∆ex2/3 embryonic fibroblasts have characterized by the absence of CD4 and CD8demonstrated that Shc is necessary for MAPK markers (reviewed in (Zhang et al., 2003)). Thesignaling induced by a low concentration of authors did not observe any increase in thegrowth factors, but at a high concentration of apoptotic fraction of the DN cells in ShcFFFgrowth factors (50 ng/ml EGF or 25 ng/ml transgenic mice compared to wild-type mice.PDGF) no detectable difference in MAPK More recent studies using pulse BrdU injectionactivation was observed. These data suggest have demonstrated a defect in proliferation ofthat Shc sensitizes cells to low amounts of 13
  • INTRODUCTIONthe late DN stage cells mediated by the pre- transform mouse fibroblasts (Migliaccio et al.,TCR (Fig. The same phenotype was 1997), suggesting a function distinct from the Shcalso obtained using the second approach, other two isoforms. Indeed, p66 does notconditional Shc knockout mice, with a nearly increase EGF-induced MAPK activation,complete loss of Shc protein expression in although it is tyrosine-phosphorylated uponthymocytes. Thus, both Shc expression and its EGF stimulation, binds to activated EGFRs,tyrosine phosphorylation play an essential and and forms stable complexes with Grb2non-redundant role in thymic T-cell (Migliaccio et al., 1997) (Fig. and proliferation. Furthermore, it has been shown that p66Shc expression inhibits EGF-induced c-fos promoter activation (Fig. The molecular mechanism is not understood, taken Shc into account that p66 expression did not inhibit Erk activation. However, the inhibition was attributed to the CH2 domain, since it Shc retained the inhibitory effect of p66 on the c- fos promoter (Migliaccio et al., 1997). In contrast, an independent study has shown that Shc p66 can function in a dominant-interferingFigure Role of Shc in T-cell manner and inhibits Erk activation downstreamdevelopment. Inducible expression of ShcFFF of EGFR signaling (Fig. (Okada etas a transgene or inducible loss of Shc protein al., 1997). These authors demonstrated notexpression arrests thymic development at the only tyrosine but also serine/threoninedouble negative (DN) stage. The block is seen Shc phosphorylation of p66 in response to EGF,where signaling from the pre-TCR occurs. The which impairs its ability to associate with therole of Shc during selection at the double tyrosine-phosphorylated EGFR, but not withpositive (DP) stage has not yet been Grb2. Co-immunoprecipitation of Shc and Grb2determined. SP: single positive; CD4 and CD8 from cells overexpressing the p45/52Shcare T-cell markers (adapted from (Zhang et al., Shc isoforms, versus p66 , directly demonstrated2003)). a competition of binding for a limited pool of Grb2 proteins (Fig. Inhibition of the1.1.2.6 Role of p66Shc Shc Ras/MAPK pathway by p66 in an S36 phosphorylation-dependent manner has also The cDNA encoding the largest isoform, shc been found following TCR downstreamp66 , was cloned in 1997, 5 years after the signaling (Pacini et al., 2004). Furthermore,discovery of the two smaller isoforms p66Shc-deficient T-cells have been reported to(Migliaccio et al., 1997). As already mentioned, proliferate faster than their normal counterpartsit encompasses an additional CH2 domain on in response to limiting ligand concentration,its N-terminus containing a serine (S36) and supporting an antagonistic activity of p66Shc onthreonine (T29) phosphorylation site. Unlike Shc Shc mitogenic signaling (Pacini et al., 2004). Thep46/52 , overexpression of p66 does not 14
  • INTRODUCTION mechanism whereby p66Shc-bound Grb2 becomes uncoupled from Ras remains to be Shc determined. It is possible that p66 binds Grb2 or the Grb2/SOS complex in a conformation which does not allow SOS to act as a guanine exchange factor for Ras (Fig. However, the finding that p66Shc participates in a complex which also includes RasGAP during early morphogenetic events in Xenopus gastrulation (Dupont and Blancq, 1999) suggests a different mechanism for the negative control of Ras/MAPK activation by this protein (Fig. Whatever the Shc mechanism is, p66 does not mediate growth factor-induced MAPK activation, and its expression might provide a mechanism for fine-tuning the Ras/MAPK pathway. More recently, loss-of-function studies have Shc unveiled an unexpected role of p66 in ageing and in the apoptotic response to oxidative stress (Migliaccio et al., 1999). Shc p66 -deficient mice exhibit a lifespan about 30% longer than wild-type. Moreover, they survive longer after treatment with paraquat, a drug that increases the production of reactive oxygen species (ROS) and, therefore, oxidative stress. Increased resistance to oxidative stress or oxidative stress-inducing agents such as UV and H2O2 can be correlated with a reduction in the apoptotic responses to these stimuli in p66Shc-/- fibroblasts. A Shc protective effect of p66 ablation againstFigure Possible mechanism of apoptosis in thymocyte and peripheral T- Shcp66 function in Ras/MAPK signaling. See lymphocyte has also been reported recently Shctext for details (A) p66 binds Grb2 in a (Pacini et al., 2004). Conversely, p66 Shcconformation which does not allow activation of overexpression results in enhanced stress- Shc ShcRas. (B) p66 competes with p46/52 for induced apoptosis in fibroblasts, endothelial ShcGrb2 binding. (C) p66 binds to RasGAP and cells and T-cells (Pacini et al., 2004; Trinei etnegatively influences Ras activation. al., 2002). The proapoptotic activity of p66Shc is strictly dependent on phosphorylation of S36 in the CH2 domain. S36 phosphorylation is 15
  • INTRODUCTIONobserved in response to many stimuli, mitochondria, and subsequent caspase 3including H2O2, UV (Migliaccio et al., 1999), activation (Fig. Again, the capacity ShcFas ligation (Pacini et al., 2004), and taxol of p66 to mediate p53-dependent apoptosis(Yang and Horwitz, 2002), but also in response requires phosphorylation of S36. The releaseto EGF (Okada et al., 1997) and insulin (Kao et of cytochrome C in oxidative stress is theal., 1997). Depending on the cellular context endpoint of the p53-dependent transcriptionaland on the identity of the stimulus, either Erk, activation of redox related genes. The resultingJNK, or p38 MAPK is responsible for S36 rise of ROS levels affects the mitochondrialphosphorylation (Le et al., 2001; Okada et al., membrane potential, leading to membrane1997; Yang and Horwitz, 2002). Taken permeability transition and cytochrome C Shctogether, these results suggest that p66 acts release (Li et al., 1999; Polyak et al., 1997).as a sensor of intracellular concentration of Cyclosporin A, an inhibitor of the mitochondrialROS (Fig. permeability transition pore which blocks Further experiments aimed at understanding oxidative stress-induced apoptosis of wild-type Shcthe mechanisms underlying the role of p66 MEFs, is able to prevent re-expressed p66Shcin regulating oxidative stress-induced from restoring apoptotic responses to oxidants Shcapoptosis have revealed that p66 is a in p66Shc-/- MEFs, suggesting that p66Shc maydownstream effector of the tumor suppressor regulate mitochondrial permeability transition,p53 (Trinei et al., 2002). It is required for p53- andinduced release of cytochrome C from ShcFigure p66 senses ROS and mediates oxidative stress-induced apoptosis. ROS Shcactivate one of the MAPKs, which in turn phosphorylates p66 on S36. S36 phosphorylation isnecessary for cytochrome C release and subsequent apoptosis. p53 acts upstream of p66Shc and Shc Shcenhances p66 protein stability, leading to p66 accumulation. p53-induced apoptosis is dependent Shcon p66 expression. 16
  • INTRODUCTIONhence cytochrome C release, by modulating apoptotic signals, suggesting that S36the production of ROS (Orsini et al., 2004). phosphorylation might serve other, ShcIndeed, intracellular ROS levels are drastically nonmitochondrial, activities of p66 which are Shc-/-reduced in p66 cells and enhanced in also needed to exert its proapoptotic function. Shc Shcp66 overexpressing cells (Nemoto and A second mechanism by which p66 couldFinkel, 2002; Orsini et al., 2004). Furthermore, influence ROS levels was suggested by Shcp66 has been found to localize to Nemoto et al. (Nemoto and Finkel, 2002) (Fig. Shcmitochondria and to be associated with Hsp70. They linked p66 expression to the(Orsini et al., 2004). The best evidence was transcriptional activity of the forkhead familyderived from a recent report by Giorgio et al. transcription factor, FKHRL1. In quiescent(Giorgio et al., 2005), which clearly established cells, FKHRL1 localizes predominantly in the Shca role for p66 in the generation of ROS. nucleus where it positively regulates Shcp66 was found to function as a redox transcription of genes such as catalase,enzyme that generates mitochondrial ROS as implicated in ROS scavenging. Oxidativesignaling molecules for apoptosis (Fig. stress most probably promotes FKHRL3). It does so by utilizing reducing equivalents phosphorylation in a PKB-dependent manner,of the mitochondrial electron transfer chain and subsequent exclusion from the nucleusthrough the oxidation of cytochrome C. results in a reduction of its transcriptionalInterestingly, S36 phosphorylation was not activity. Phosphorylation and cytoplasmic Shcobserved in the mitochondrial pool of p66 : localization of FKHRL in response to H2O2 was Shcinstead a different region was necessary for abrogated in p66 -deficient MEFs. Shcthe redox activity of p66 . It seems, therefore, Accordingly, FKHRL-dependent transcription of Shcthat p66 exists in two different pools, a the catalase gene was augmented in these Shccytoplasmic one and a mitochondrial one. cells, suggesting a pivotal role of p66 in the ShcSignificant translocation of p66 from cytosolto mitochondria does not occur followingFigure Model of p66Shc redoxactivity during mitochondrialapoptosis. Proapoptotic signals inducerelease of p66Shc from a putative Shcinhibitory complex. Active p66 thenoxidizes reduced cytochrome C (red) andcatalyzes the reduction of O2 to H2O2.Permeability transition pore opening byH2O2 then leads to swelling andapoptosis. NADH-Cyt B5 reductase isindicated as an additional putative source of reduced cytochrome C (taken from (Giorgio et al., 2005)). 17
  • INTRODUCTIONredox-dependent inactivation of FKHRL1 and, perspective, inhibition of p66Shc may bethereby, in the control of ROS. envisioned as a novel way to prevent the deleterious effects of ROS-mediated diseases in general and of Ang II on the heart in particular.Figure p66Shc regulates FKHRL1 Shctranscriptional activity. p66 expressionenhances PKB phosphorylation via anunknown mechanism. This leads to a decreasein FKHRL1 transcriptional activity due tophosphorylation by PKB which causes itsretention in the cytoplasm. Finally, ROS-detoxifying enzymes such as catalase are lessexpressed. The ability to generate ROS and to regulateexpression of scavenger proteins makes Shcp66 an attractive target for therapies againstvascular diseases, which are strongly Shcmediated by ROS. Indeed, deletion of p66reduces systemic and tissue oxidative stress,vascular cell apoptosis and earlyatherogenesis in mice fed a high-fat diet(Napoli et al., 2003). p66Shc-deficient micewere also resistant to theproapoptotic/hypertrophic action of AngiotensinII (Ang II). Consistently, in vitro experimentshave shown that Ang II causes a higher rate ofapoptotic death in cardiomyocytes isolatedfrom p66Shc(+/+) hearts than in those isolated Shc(-/-)from p66 hearts (Graiani et al., 2005). In 18
  • INTRODUCTION junctions represents a specialized form of1.2 Signaling of the E- cadherin-based adhesive contacts which helpscadherin cell-cell adhesion cells to form a tight, polarized cell layer that can perform barrier and transport functionsprotein (Gumbiner, 2005). The cadherins constitute a major class ofadhesion molecules that support calcium-dependent, homophilic cell-cell adhesion in allsolid tissues of the body. They mediate cell-cellrecognition events, bring about morphologicaltransitions that underlie tissue formation, andmaintain tissue architecture in the adultorganism. The next paragraph will give a briefintroduction of E-cadherin-dependent cell-celladhesion with major emphasis on its tumorsuppressing function and its signalingcapacities. Figure Epithelial junctional1.2.1 E-cadherin-dependent cell-cell complex. Adhesion between vertebrate cells isadhesion generally mediated by three types of adhesion junction: adherens junction (zonula adherens), E-cadherin: a member of the tight junction (zonula occludens), andclassical cadherins desmosomes. Electron micrograph of an epithelial junctional complex containing zonula Cadherins represent a large superfamily adherens (ZA), zonula occludens (O), andwhich includes classical cadherins, desmosome (D). The ZA junction completelydesmosomal cadherins, atypical cadherins, encircles the apex of the epithelial cell, but onlyproto-cadherins and cadherin-related signaling a section through the junction is shown. Themolecules (Gumbiner, 2005). E-cadherin is a membranes of the two cells align tightly at theprototype family member and belongs to the junction, with an extracellular gap of 250Å. Theclassical cadherins. Classical cadherins were cytoplasmic surface of the junction appears asoriginally named for the tissue in which they a dense plaque, presumably made up ofare most prominently expressed. Later, it cytoskeletal proteins, which associates withbecame clear that most cadherins can be actin filament (taken from (Gumbiner, 2005)).expressed in many different tissues. E-cadherin (epithelial cadherin) is expressed Classical cadherins are single-passprimarily in epithelial cells and is associated transmembrane proteins. They contain fivewith the zonula adherens (which is also known cadherin domains on their extracellular partas adherens junctions) of the epithelial which confer specific adhesive binding, andjunctional complex (Fig. Adherens homophilic protein-protein interactions 19
  • INTRODUCTION cytoplasmic proteins, the catenins, is a second characteristic which distinguishes classical cadherins from other members of the cadherin superfamily (Fig. (Takeichi, 1995). α-catenin interacts, through β-catenin, with the distal part of the cadherin cytoplasmic domain. γ-catenin (also known as plakoglobin) can bind to the same site as β-catenin in a mutually exclusive way, whereas another catenin, p120- catenin, interacts with a more proximal region of the cytoplasmic domain. Function of catenins in the E- cadherin adhesion complex The main function of catenins is the conversion of the specific homophilic bindingFigure The classical cadherin- capacity of the E-cadherin extracellular domaincatenin complex. Cadherin is a parallel, or into a stable cell-cell adhesion. Although the E-cis, homodimer. The extracellular region of cadherin extracellular domain alone possessesclassical cadherins consists of five cadherin- homophilic binding properties, stable celltype repeats (extracellular cadherin domains) adhesion requires the cadherin cytoplasmaticthat are bound together by Ca2+ ions (yellow tail and associated proteins (Yap et al., 1997).circles) to form stiff, rod-like proteins. The core α-catenin can mediate physical linksuniversal-catenin complex consists of p120- between cadherin and the actin cytoskeleton,catenin, bound to the juxtamembrane region, either by directly binding actin filaments orand β-catenin, bound to the distal region, indirectly through other actin-binding proteinswhich in turn binds α-catenin. In a less well such as vinculin and α-actinin (Fig. way, α-catenin binds to actin and Besides linking cadherins to the actinactin-binding proteins, such as vinculin, α- cytoskeleton, catenins are believed to playactinin, or formin-1 (taken from (Gumbiner, additional roles. β-catenin is a well known2005)). signaling molecule in the Wnt pathway (see below), and catenins can interact with otherbetween two cadherin molecules on two cells. signaling molecules, such as GTPasesThe exact structure of the homophilic bond is (Goodwin et al., 2003), PI3K (Woodfield et al.,still a matter of debate (Gumbiner, 2005), but 2001), and formin-1 (known to nucleate actinan intriguing possibility is that some of the polymerisation) (Kobielak et al., 2004), toexisting models represent different influence the state of the actin cytoskeletonconformational states that are important for the (see below) (Fig. of adhesion. The presence of a The core function of p120-catenin is toconserved cytoplasmic tail that associates with regulate cadherin turnover (Reynolds and 20
  • INTRODUCTIONRoczniak-Ferguson, 2004). Loss of p120- an inactive, or less adhesive, conformationcatenin leads to significantly reduced levels of (Fig. in epithelial cells (Davis et al.,2003). Thus, p120-catenin directly influences Function of the E-cadherin-cateninadhesive strength by controlling the amount of complexE-cadherin available at the cell surface for The E-cadherin-catenin complex is essentialadhesion. for the formation of epithelia in the embryo, Furthermore, the adhesive strength of and maintenance of epithelial structure in thecadherins is changed by posttranslational adult. It carries out different functions, includingmodifications of p120-catenin and β-catenin. cell-cell adhesion, cytoskeletal anchoring, andAlthough poorly understood, tyrosine signaling. The expression of different types ofphosphorylation of catenins is believed to cadherins mediates selective cell recognitionregulate the conformation or organization of events that are responsible for the sorting ofcadherins. It is thought that phosphorylation of different groups of cells in developing tissues,catenins could lead to a disruption of and the formation of selective connectionsdimerization and reduced clustering of the between neurons in the developing nervouscadherin molecules at the surface, resulting in A B CFigure Function of catenin proteins in the E-cadherin-catenin complex. There are threeways in which catenins contribute to the cadherin function. (A) α-catenin provides a direct physical linkto the actin cytoskeleton through interaction with E-cadherin-bound β-catenin and actin or actin-binding proteins such as vinculin and α-actinin. (B) Catenins bind to or influence signaling molecules(GTPases, formin-1, PI3K) known to control the actin cytoskeleton. (C) Phosphorylation of cateninsmight control the adhesive strength of the cadherin-catenin complex. Depicted is a hypotheticalexample where phosphorylation of catenins could lead to a disruption of dimerization and reducedclustering of cadherin molecules at the cell surface, resulting in an inactive or less adhesiveconformation. Ca2+ ions are indicated by yellow circles. EC: extracllular cadherin domain (taken from(Gumbiner, 2005)). 21
  • INTRODUCTIONsystem (Gumbiner, 2005). In cell culture, a 1994; Hirohashi, 1998). This observation hasmixed population of cells expressing different prompted an examination of the functional rolecadherins become sorted by adhering only to of E-cadherin in tumor progression. Behrens etthose cells expressing the same cadherin (Yap al. (Behrens et al., 1989) showed that epithelialet al., 1997). During development, segregation cells acquire invasive properties whenof cells into distinct tissues is accompanied by intercellular adhesion is specifically inhibited bychanges in the complement of cadherins the addition of E-cadherin function-blockingexpressed by the cells. The specificity of antibodies; the separated cells then invadehomophilic binding is therefore a fundamental collagen gels and embryonic heart tissue.mechanism by which cadherins influence the Subsequently, several groups haveorganization of various cell types into tissue demonstrated that re-establishing the(Yap et al., 1997). However, different functional cadherin complex by forcedcadherins can be promiscuous with regards to expression of E-cadherin results in a reversiontheir adhesive binding properties, with of an invasive, mesenchymal phenotype to aevidence for heterophilic adhesion between benign, epithelial phenotype of cultured tumordifferent classical cadherins. The level of cells (Birchmeier and Behrens, 1994; Navarrocadherin expression, and presumably therefore et al., 1991; Vleminckx et al., 1991). Based onthe overall strength of adhesion, has also been these data, it has been proposed that the lossfound to strongly influence cell-sorting of E-cadherin-mediated cell-cell adhesion is abehavior, independently of the type of cadherin prerequisite for tumor cell invasion andexpressed (Gumbiner, 2005). metastasis formation. The in vivo proof that The importance of E-cadherin-mediated cell loss of E-cadherin is not a consequence of de-adhesion is also highlighted by the fact that its differentiation, but rather the cause of tumordisturbance is causally involved in cancer progression, was made by Christofori anddevelopment. colleagues (Perl et al., 1998). Intercrossing RipTag2 mice, which provide a model of1.2.2 E-cadherin as a tumor pancreatic carcinogenesis, with transgenicsuppressor mice that maintain E-cadherin expression in β- cell-derived tumor cells resulted in the arrest of tumor development at the adenoma stage, The majority of human cancers (ca. 80-90%) whereas expression of a dominant-negativeoriginate from epithelial cells. In most, if not all, form of E-cadherin induced early invasion andof these epithelial-derived cancers, E-cadherin- metastasis. Very recently, a second study hasmediated cell-cell adhesion is lost, concomitant demonstrated causal evidence for thewith the transition from benign, non-invasive involvement of E-cadherin in tumortumor to malignant, invasive tumor. Although progression. A group from the NetherlandsE-cadherin expression is maintained in most introduced a conditional loss-of-functiondifferentiated tumors, including carcinomas of mutation in the E-cadherin gene into mice thatthe skin, head and neck, breast, lung, liver, carry p53 mutations. Although tissue-specificcolon, and prostate, there seems to be an inactivation of E-cadherin alone did not resultinverse correlation between E-cadherin levels in tumor formation, the combined inactivationand cancer grade (Birchmeier and Behrens, 22
  • INTRODUCTIONof E-cadherin and p53 led to the accelerated transcription factors, such as Snail and Slug,development of mammary gland and skin has been observed downstream of RTKtumors. Moreover, loss of E-cadherin induced signaling (Thiery, 2002). Snail, Slug, SIP1, anda phenotypic change from non-invasive to E12/47, as well as Twist, are factors whichhighly invasive mammary gland tumors, and a repress transcription from the E-cadherinconversion from ductal to lobular carcinomas promoter via the E-boxes (Cavallaro and(Birchmeier, 2005). These results show that Christofori, 2004; Yang et al., 2004).the loss of E-cadherin-mediated cell-cell β-catenin is also actively involved in EMTinvasion is one rate-limiting step in the (Fig. 1.2.2) and its role as a signaling moleculeprogression from adenoma to carcinoma and will be discussed later.subsequent formation of tumor metastases. In addition to EMT, which is a rather Downregulation of E-cadherin is often part of organized process leading to downregulationa process called epithelial-to-mesenchymal of E-cadherin expression, various othertransition (EMT), which is characterized by the mechanisms are involved in the disruption ofloss-of-expression of epithelial genes and the cell-cell adhesion during tumor progression. Again-of-expression of mesenchymal genes variety of genetic mechanisms, such as(Thiery, 2002). EMT is a crucial event during deletion or mutational inactivation of the gene,tumor metastasis but also occurs in normal or gene mutations which result in theembryonic development, for example during expression of a non-functional protein, causegastrulation (Fig. 1.2.2). Activation of RTK loss of E-cadherin expression or function,[fibroblast growth factor receptor (FGFR), especially in diffuse gastric cancer (BirchmeierEGFR family, transforming growth factor-β and Behrens, 1994; Bracke et al., 1996;(TGF-β) receptor, insulin-like growth factor Strathdee, 2002). Silencing of the E-cadherinreceptor (IGFR), hepatocyte growth factor gene by hypermethylation of promoter regionsreceptor (HGFR)] signaling is able to induce occurs frequently in carcinoma cell lines, inEMT via stimulation of PI3K, Src, Ras and thyroid carcinomas, and in several otherRac. Signaling downstream of EGFR, c-Met cancer types (Di Croce and Pelicci, 2003;and FGFR, as well as Src, results in tyrosine Hirohashi, 1998). More recently, proteolyticphosphorylation of E-cadherin, β-catenin and degradation of E-cadherin by matrix-metallop120-catenin, leading to a disassembly of the proteases (MMPs) has been described as acadherin-catenin complex, disruption of mechanism by which cell-cell adhesion can becadherin-mediated adhesion and cell disrupted. Cleavage of E-cadherin results inscattering. Tyrosine phosphorylation-mediated not only the disruption of cell-cell adhesion, butubiquitination and subsequent proteasomal also the production of a soluble 80-kDa E-degradation of E-cadherin or increased cadherin fragment that itself disrupts cell-cellendocytosis of E-cadherin seem to be adhesion in a dominant-interfering manner,mechanisms underlying this observed thereby promoting tumor progression (Noe etdisassembly (Fujita et al., 2002; Kamei et al., al., 2001; Wheelock et al., 1987).1999). Moreover, induction of expression of 23
  • INTRODUCTIONFigure 1.2.2: Epithelial-mesenchymal transition (EMT). Epithelial cells lose the expression ofepithelial-specific genes, such as E-cadherin, and acquire the expression of mesenchymal genes(vimentin, collagens, integrins). EMT causes cells to lose apical-basal polarity (shown on the left) andgain a fibroblast-like morphology, high motility and invasive properties (shown on the right). (A)Transcription factors (such as Snail and Slug) have been identified that control the expression of E-cadherin by binding directly to E-boxes in the gene promoter. Other factors, such as growth factorsand their receptors, the tyrosine kinase src, and cytoplasmic G-proteins (such as rac) can alsopromote EMT indirectly. (B) β-catenin was found to exert a dual role as an essential cytoplasmic-interaction partner of cadherins, which is essential for cell-cell adhesion, and as a nuclear partner ofthe T-cell factor (TCF)/lymphocyte-enhancer factor (LEF) family of transcription factors that regulategenes of the canonical Wnt signaling pathway. The switch of β-catenin from its action in cell adhesionto transcriptional control in the nucleus is controlled by binding to BCL9-2, which is the homologue of ahuman B-cell oncogene product, and is promoted by tyrosine phosphorylation of β-catenin (taken from(Birchmeier, 2005)). As already mentioned above, appropriate in a subset of E-cadherin-deficient tumors.cell-cell adhesion requires the cadherin-catenin However, direct evidence is lacking and itcomplex as a whole. Therefore, changes in the remains to be determined whether this wouldexpression of catenins, for example mutations represent a general process in tumorin α-catenin or expression of truncated α/β- progression.catenin, impair E-cadherin-mediated cell Proper E-cadherin function can also beadhesion and are often associated with overruled or replaced by the expression ofmalignant transformation (Hajra and Fearon, mesenchymal cadherins, such as N-cadherin,2002; Hirohashi and Kanai, 2003). Recently it which has been shown to promote cell motilityhas been shown that knockdown of p120- and migration. It becomes more and morecatenin results in the destruction of the entire evident that this “cadherin switch” is involvedcadherin complex (Reynolds and Roczniak- during the transition from a benign to anFerguson, 2004). Together with evidence of invasive tumor phenotype (Christofori, 2003).frequent p120-catenin loss in cancer, these Taken together, loss of E-cadherin-mediatedobservations suggest that p120-catenin cell-adhesion strongly contributes to tumordownregulation itself may be an initiating event progression, but it is unlikely that loss of E- 24
  • INTRODUCTIONcadherin by itself can account for the become confluent, but as cytosolic p120-metastatic phenotype, because loss of catenin becomes sequestered by the E-adhesiveness does not necessarily cause cells cadherin adhesion complex it cannot accountto become motile and/or invasive; additional for this decrease in Rho activity. Therefore,events are required. other mechanisms downstream of E-cadherin- mediated adhesion decrease Rho activity.1.2.3 E-cadherin-mediated signaling Noren et al. (Noren et al., 2003) reported that E-cadherin engagement in cell-cell adhesion An increasing body of evidence suggests suppresses Rho activity by inducingthat cadherins act at the cellular level as phosphorylation and activation ofadhesion-activated cell signaling receptors p190RhoGAP, probably through Src-family(Cavallaro and Christofori, 2004; Wheelock kinases. In other systems, E-cadherin wasand Johnson, 2003). Although signals that are found to communicate with Rho GTPases viaelicited by the formation of E-cadherin- PI3K signaling (Fig. 1.2.3). PI3K is andependent cell-cell adhesion have been upstream kinase of Rac and has previouslyextensively studied, signals that are induced by been found to interact with E-cadherin (Pece etthe loss of E-cadherin function, for example al., 1999; Woodfield et al., 2001). Yap andduring cancer progression, are only just being colleagues (Kovacs et al., 2002) showed thatelucidated. PI3K co-localized with E-cadherin at the Several studies have reported that leading edge of cadherin-based lamellipodia,establishment of E-cadherin-mediated contact and was necessary for full and sustainedinfluences the activity of Rho-family GTPases; activation of Rac. In contrast, another groupwith Rac and CDC42 being activated and Rho reported that Rac activation induced by E-being inactivated. The mechanisms underlying cadherin ligation was independent of PI3Kthis activation or inactivation vary depending activity, but dependent on EGFR signaling (seeon the model system used. One connection below) (Betson et al., 2002). Whatever thebetween cadherins and Rho GTPases is mechanisms are, E-cadherin-mediatedthrough p120-catenin. It has been shown that contacts influence the activity of Rho-familyp120-catenin activates Rac1 and CDC42, GTPases, which are believed to regulateperhaps by activating Vav2, which is a guanine dynamic organization of the actin cytoskeletonexchange factor for these GTPases (Fig. 1.2.3) and the activity of the cadherin/catenin(Grosheva et al., 2001; Noren et al., 2001). apparatus to modulate stabilization of theReynolds and colleagues showed that adhesive contact (Yap et al., 1997).cytosolic p120-catenin inhibits RhoA activity by Several studies have suggested functionalacting as guanine nucleotide dissociation interdependence of cadherins and RTK withinhibitor (Anastasiadis et al., 2000; Noren et respect to their signaling capacities. It hasal., 2000). It is worth noting that only cytosolic been demonstrated that initiation of de novo E-p120-catenin is able to modulate GTPase cadherin-mediated adhesive contacts canactivity; this function is abolished when p120- induce ligand-independent activation of thecatenin participates in the E-cadherin adhesion EGFR and subsequent activation of Erkcomplex. Rho activity decreases as cells (Munshi et al., 2002; Pece and Gutkind, 2000). 25
  • INTRODUCTIONIn contrast, it has been shown that the E- competes with α-catenin binding and cellcadherin adhesive complex can be linked to adhesion (Brembeck et al., 2004) (Fig. 1.2.2).EGFR via β-catenin (Hoschuetzky et al., 1994) Although some groups have shown thator via the extracellular domain of E-cadherin, overexpression of E-cadherin fragments ableand negatively regulate receptor tyrosine to bind β-catenin can repress TCF/LEFkinase signaling in an adhesion-dependent transcriptional activity in 293T cells (Simcha etmanner. Interaction of cadherins with al., 2001) or SW480 cells (Gottardi andrespective RTK has been observed in different Gumbiner, 2004), another group was unable tosystems (VEGFR with VE-cadherin, and FGFR find any dependence of TCF/LEF-mediatedwith N-cadherin) (Carmeliet et al., 1999; transcriptional activity on E-cadherinCavallaro et al., 2001). expression in human breast cancer cells (van It is worth noting that β-catenin, besides de Wetering et al., 2001). The effects of E-being a major component of the E-cadherin cadherin on Wnt signaling appear, therefore, toadhesion complex, is also part of the Wnt- be cell-context-dependent.mediated signaling pathway. In the absence ofWnt signaling, cytosolic β-catenin is degradedthrough a pathway that is dependent onadenomatous poliposis coli protein (APC).However, upon stimulation of the Wnt pathwaythis degradation is suppressed, resulting in theaccumulation of cytoplasmic β-catenin(Wheelock and Johnson, 2003). Subsequently,it translocates into the nucleus and acts as acoactivator of the T-cell factor(TCF)/lymphocyte-enhancer factor (LEF)transcription factors (Fig. 1.2.3). Given thatbinding to β-catenin precludes its participationin Wnt signaling, E-cadherin could potentiallyregulate Wnt signaling by sequestering β-catenin from TCF/LEF transcription factors.Gumbiner and colleagues (Gottardi andGumbiner, 2004) reported that the participationof β-catenin in adhesion and Wnt signaling isdictated by the presence of distinct molecularforms of β-catenin that have different bindingproperties. More recently it has been shownthat this switch can be regulated by the bindingof β-catenin to BCL9-2 (the homolog of thehuman B-cell oncogene product BCL-9). β-catenin/BCL9-2 binding can be promoted bytyrosine phosphorylation of β-catenin, and 26
  • INTRODUCTIONFigure 1.2.3: Signaling by the E-cadherin complex. Initiation of de novo cell-cell adhesion activatesRac and CDC42, and inhibits Rho. Soluble p120-catenin and PI3K are most likely mediating theseeffects via activation of the respective guanine exchange factor (VAV) or GTPase-activating protein(p190RhoGAP). Establishment of E-cadherin-mediated contacts can also induce ligand-independentactivation of the EGFR and, subsequently, activation of Erk and PI3K signaling. Cytosolic β-catenin isnormally degraded through the adenomatous poliposis coli (APC) complex. However, WNT signalinginhibits the APC complex, allowing β-catenin to enter the nucleus and coactivate TCF/LEFtranscription factors. By sequestering β-catenin from participation in the WNT signaling, the E-cadherinadhesion complex might also modulate WNT-induced transcription. 27
  • INTRODUCTION1.3 RNA interference: a new to trigger the non-specific dsRNA responses, but they still cause destruction ofand powerful tool in molecular complementary RNA sequences (Gitlin et al.,biology 2002). More recently, a large number of RNAi is a general term for sequence-specific endogenous microRNA (miRNAs) wasgene repression induced by double-stranded discovered. miRNAs are a specific class ofRNAs (dsRNAs) that was initially discovered in small RNAs that are encoded in gene-likeplants. It was later observed in the animal elements organized in a characteristic invertedmodel organism Caenorhabditis elegans that repeat (Grishok et al., 2001; Reinhart et al.,dsRNA triggered sequence-specific mRNA 2000). Because the active forms of miRNAscleavage (Fire et al., 1998). It soon turned out and siRNAs are sometimes biochemically orthat RNAi is not restricted to nematode and functionally indistinguishable, they arecan be induced in Drosophila melanogaster classified based on their origins (Fig. 1.3).(Kennerdell and Carthew, 1998), Trypanosoma siRNAs are derived from long dsRNAs in the(Ngo et al., 1998), and vertebrates (Elbashir et cytoplasm, whereas miRNA genes areal., 2001a; Yang et al., 2001). transcribed by RNA polymerase II to generate During RNAi, long dsRNA molecules are long primary transcripts (pri-miRNAs) (Cai etprocessed into 19- to 23-nt RNAs known as al., 2004; Lee et al., 2004). In the nucleus, pri-small-interfering RNAs (siRNAs) that serve as miRNAs are trimmed to release hairpinguides for enzymatic cleavage of intermediates (pre-miRNAs) (Lee et al., 2002)complementary RNAs (Elbashir et al., 2001b; by the RNase III type enzyme Drosha (Lee etParrish et al., 2000; Zamore et al., 2000). In al., 2003). pre-miRNAs then get exported toDrosophila and C. elegans, siRNAs can the cytoplasm (Bohnsack et al., 2004), wherefunction as primers for an RNA-dependent they are processed in a similar way toRNA polymerase that synthesizes additional dsRNAs, the precursors of siRNAs. pre-dsRNA, which in turn is processed into miRNAs and dsRNAs are processed by Dicer,siRNAs, amplifying the effects of the original the cytoplasmic RNase III type proteinsiRNAs (Lipardi et al., 2001; Sijen et al., 2001). (Bohnsack et al., 2004; Hutvagner et al., 2001; In mammalian cells, the experimental use of Ketting et al., 2001) and cleaved into the short-RNAi with dsRNA has not been successful in lived miRNA/siRNA duplexes, whose onemost cell types because of non-specific strand is degraded by an unknown nucleaseresponses elicited by dsRNA molecules longer while the other strand remains as a maturethan about 30 nt (Robertson and Mathews, miRNA/siRNA (Khvorova et al., 2003; Schwarz1996). Tuschl and coworkers (Elbashir et al., et al., 2003). The released miRNAs/siRNAs2001a) discovered that transfection of are incorporated into silencing complexes.synthetic 21-nt siRNA duplexes into Although it is difficult to assign a distinctmammalian cells effectively inhibits functional label, an siRNA-containing complexendogenous genes in a sequence-specific is commonly referred to as RNA-inducedmanner. These siRNA duplexes are too short silencing complex (RISC), whereas an miRNA- 28
  • INTRODUCTIONcontaining effector complex is referred to as a 2004). Mouse miR-196 miRNAs repress themicro ribonucleoprotein particle (miRNP) expression of the hoxb8 gene, a transcription(Meister and Tuschl, 2004). Every RISC or factor important during vertebratemiRNP contains a particular subset of developmental regulation (Yekta et al., 2004).Argonaute proteins that exert sequence- Several reports have also shown that alteredspecific gene repression by inducing cleavage expression of specific miRNA genes(‘slicing’) or, as in the case of several miRNPs, contributes to the initiation and progression ofby eliciting a block to translation (Meister and cancer (Croce and Calin, 2005; Gregory andTuschl, 2004). Shiekhattar, 2005; McManus, 2003). All these miRNAs are often only temporarily findings prove that miRNAs play importantexpressed and seem to play a role in regulatory roles in animals by targeting thedevelopmental processes. In mammals, miR- messages of protein-coding genes for181 is involved in the control of hematopoiesis translational repression or degradation.through as yet unknown target(s) (Chen et al.,Figure 1.3: Model of small-RNA-guided posttranscriptional regulation of gene expression.Primary miRNA transcripts are processed to miRNAprecursors in the nucleus by the RNase-III-likeenzyme Drosha. The miRNA precursor issubsequently exported to the cytoplasm by means ofthe export receptor exportin-5. The miRNA precursoris further processed by Dicer to siRNA-duplex-likeintermediates. The duplex is unwound whileassembling into miRNP/RISC. Mature miRNAs bindto Ago proteins, which mediate translationalrepression or cleavage of target mRNAs. Othersources of long dsRNA in the cytoplasm of a cell areviral RNAs, artificially introduced dsRNA, andgenomic sense and antisense transcripts. LikemiRNA precursors, long dsRNA is processed by theRNase III enzyme Dicer into 21-23 nucleotidedsRNA intermediates. Assisted by the RNA helicaseArmitage and R2D2, the single-stranded siRNA-containing RISC is formed. The stability of thedsRNA, and its recognition by Dicer, can beregulated by specific ADARs (deaminase) and theexonuclease ERI-1. DCR: Dicer-like protein, R2D2:dsRNA binding protein (taken from (Meister andTuschl, 2004)). 29
  • INTRODUCTION In contrast, siRNAs are produced fromdsRNAs that are synthesized from viruses,endogenously activated transposons, orrepetitive sequences introduced by geneticengineering. Thus, siRNAs have beenproposed to function in: (i) antiviral defense, (ii)silencing mRNAs that are overproduced ortranslationally aborted, and (iii) guarding thegenome from disruption by transposons(Hannon, 2002; Mello and Conte, 2004;Tabara et al., 1999). In experimental research,siRNAs can efficiently and rapidlydownregulate the level of an endogenousprotein in mammalian cells. The use of siRNAtherefore complements overexpression studiesin tissue culture by providing a powerful tool toinvestigate loss-of-function of a given protein. 30
  • RESEARCH OBJECTIVES 1.4 Research objectives chapter represents data on the role of p66Shc in regulating cell viability upon stress response in epithelial cells. Several reports have shown that Shcadaptor proteins are involved in MAPKactivation induced by several growth factors.However, the contribution of each isoform inmediating this process is still not known.Moreover, are there other signaling pathwaysin which Shc might play a role, despite growthfactor-induced signaling? The largest isoform, Shcp66 , does not seem to be involved in MAPKactivation, but whether it acts in a dominantnegative manner or in a neutral way is still a Shcmatter of debate. Gene targeting of p66 hasrevealed that this isoform is implicated in theregulation of lifespan and the response tooxidative stress. The absence of p66Shcconfers resistance to oxidative stress onmouse embryo fibroblasts, endothelial cells,and T-cells. Still, it is unclear whether this is ageneral effect or restricted to certain cell typesand what the underlying mechanisms are. This thesis aims to address some of thesequestions. To investigate isoform-specificfunctions of Shc proteins, we first generated asystem to specifically knockdown single Shcisoforms using siRNA. This method wasdeveloped further to the so-called “knockdown-in system”, where we achieved an isoform-specific expression of Shc proteins. Thesedata will be discussed in greater detail in thefirst part of the following chapter. Secondly, we used this method to explorehow necessary Shc proteins are in EGF-induced signaling. In a quest for growth factor-independent pathways in which Shc proteinsmight be involved, we investigated signalingdownstream of the cell adhesion molecule, E-cadherin. These results are described insections 2.3-2.5. The last part of the following 31
  • RESULTS 2. RESULTS2.1 RESEARCH COMMUNICATIONIsoform-specific knockdown and expression of adaptor protein ShcA using smallinterfering RNAMalgorzata KISIELOW1, Sandra KLEINER1, Michiaki NAGASAWA3, Amir FAISAL and Yoshikuni NAGAMINEFriedrich Miescher Institute for Biomedical Research, Maulbeerstrasse 66 CH-4058 Basel, Switzerland1 These authors contributed equally to the work.Many eukaryotic genes are expressed as multiple different transcripts, namely p66 and p52/p46isoforms through the differential utilization of mRNAs. An siRNA with a sequence shared by thetranscription/translation initiation sites or two transcripts suppressed all of them. However,alternative splicing. The conventional approach for another siRNA whose sequence was present only instudying individual isoforms in a clean background p66 mRNA suppressed only the p66 isoform,(i.e. without the influence of other isoforms) has suggesting that the siRNA signal did not propagatebeen to express them in cells or whole organisms in to other regions of the target mRNA. Thewhich the target gene has been deleted; this is time- expression of individual isoforms was achieved byconsuming. Recently an efficient post- first down-regulating all isoforms by the commontranscriptional gene-silencing method has been siRNA and then transfecting with an expressionreported that employs a small interfering double- vector for each isoform that harboured silentstranded RNA (siRNA). On the basis of this mutations at the site corresponding to the siRNA.method we report a rapid alternative approach for This allowed functional analysis of individual ShcAisoform-specific gene expression. We show how isoforms and may be more generally applicable forthe adaptor protein ShcA can be suppressed and studying genes encoding multiple proteins.expressed in an isoform-specific manner in ahuman cell line. ShcA exists in three isoforms, Key words: post-transcriptional gene silencing,namely p66, p52 and p46, which differ only in theirN-terminal regions and are derived from two RNAi, transient knockdown-in.INTRODUCTION ShcC are expressed specifically in the brain [8]. ShcA is recruited to, and phosphorylated by,In eukaryotes, many genes encode multiple activated receptor tyrosine kinases and, in turn,isoforms by way of differential recruits the growth-factor-receptor-bound protein 2transcription/translation initiation or alternative (Grb2)–Son-of-sevenless (Sos) complex via thesplicing, thus giving rise to related proteins with Src homology 2 (SH2) domain of Grb2, thusbiochemically as well as biologically distinct relaying growth-factor-induced signals to thefeatures [1]. Although, in many cases, multiple Ras/extracellular-signal-regulated protein kinaseisoforms are expressed in the same cell at the same (ERK) signalling pathway [9]. It is also involvedtime, the expression level and pattern of each in growth-factor-mediated activation of c-Jun N-isoform may vary with the cell type and its stage of terminal kinase (JNK) and protein kinase B, butdevelopment, making the study of each isoform the precise mechanism is unknown [9]. There areconfusing and difficult. Investigation of the three isoforms of ShcA, namely p66ShcA,function of individual isoforms ideally requires p52ShcA and p46ShcA, derived from a single geneconditions where only one isoform is expressed or through differential usage of transcription initiationeliminated. While several protocols for tissue- sites and translation start sites, which differ only inspecific expression or elimination of the gene of the N-terminal regions [7]. While all three isoformsinterest have been developed [2–4], reports of contain three tyrosine residues in theisoform-specific gene inactivation [5] or expression phosphotyrosine-binding (PTB) domain that arein a clean background are limited. Such protocols phosphorylated by activated receptor tyrosineare usually lengthy, often requiring several months kinases, they differ in the pattern ofto establish the desired conditions. During this time, serine/threonine phosphorylation induced bycells or organisms may adapt to the new conditions growth factors and tumour phorbol esters. This[6] and caution is required when interpreting the suggests a different cellular function for eachresults. isoform ([5, 10]; A.F. El-Shemerly, A. Faisal andThe signalling adaptor/scaffold protein ShcA is a Y. Nagamine, results not shown). However, amember of the Shc family, which consists of three systematic analysis of each isoform without thegenes, ShcA, ShcB/Sli/Sck and ShcC/N-Shc/Rai [7]. influence of other isoforms has not been reported.ShcA is ubiquitously expressed, whereas ShcB and 32
  • RESULTSPost-transcriptional gene silencing (PTGS) is aphenomenon originally reported in plants [11, 12], The full-length mouse p46, p52 and p66ShcAwhere introduction of the transgene causes cDNAs were isolated from NIH 3T3 cells bysilencing of the endogenous homologous gene and reverse transcriptase-PCR using the sense primersitself. The mechanism of PTGS involves enhanced 5´-CGG AAT TCA TGG GAC CTG GGG TTTmRNA degradation with double-stranded (ds)RNA CCT ACT-3´, 5´-CGG AAT TCA TGA ACA AGCas the trigger [13, 14]. A similar phenomenon TGA GTG GAG GCG-3´ and 5´-CGG AAT TCA(quelling) was observed in Neurospora [15]. In the TGG ATC TTC TAC CCC CCA AGC CGA AGTAnimal Kingdom, dsRNA-mediated gene silencing A-3´ respectively and the common antisense primerwas first described in the nematode Caenorhabditis 5´-CGG AAT TCA CAC TTT CCG ATC CACelegans [16] and was termed RNA interference GGG TTG C-3´. Full-length ShcA cDNAs were(RNAi). Subsequently, RNAi has been observed in initially cloned into pBluescriptII KS+ anda wide range of organisms, including flies, nucleotide sequences verified by thetrypanosomes, Hydra, zebrafish (Danio rerio) and dideoxynucleotide-chain-termination procedure.mice [13, 17]. The mechanism underlying RNAihas been partially elucidated, and a 21–23-nt-long Construction of expression vectorsdsRNA was found to be the intermediate/mediatorof mRNA decay [18, 19]. Elbashir et al. [20] have The haemagglutinin (HA)-tagged expression vectorshown recently that transfection of the 21-nt pcDNA3HA was constructed by inserting thedsRNA, termed small interfering RNA (siRNA), overlapping oligonucleotide pair 5´-CCC ACCcan trigger PTGS of both the co-transfected and the ATG GCT TAC CCA TAC GAT GTT CCA GATendogenous gene in cultured mammalian cells. In a TAC GCT G-3´ and 5´-AAT TCA GCG AAT TCTcell-free system of dsRNA-mediated mRNA decay GGA ACA TCG TAT GGG TAA GCC ATG GTGusing Drosophila embryonal cell extracts, the GGG TAC-3´ into the KpnI–EcoRI site of pcDNA3mRNA was shown to be cleaved only within the (Invitrogen). To construct expression vectors forregion of identity with the dsRNA [18], suggesting HA- tagged ShcA, p46HA, p52HA and p66HA, thethat endonucleolytic cleavage induced by siRNA is full-length cDNAs of p46, p52 and p66 werevery specific and that it is probably not propagated inserted into the EcoRI–EcoRV site ofto other regions of mRNA. However, the interesting pcDNA3HA. ShcA mutants in which potentialpossibility of distinguishing closely related mRNAs internal initiation methionine codons wereby siRNA has not been addressed with mammalian converted into leucine codons, thus expressing onlycells. the p66ShcA or 52ShcA forms, were created usingIn the present study, using transient transfection the QuickChange site-directed mutagenesis kitassays in HeLa cells, we established that the target (Stratagene). The overlapping oligonucleotide pairof siRNA is restricted to mRNAs containing the 5´-CTC CTC CAG GAC CTG AAC AAG CTGidentical sequence. This facilitated the isoform- AGT G-3´ and 5´-CAC TCA GCT TGT TCA GGTspecific knock-down of p66ShcA as well as CCT GGA GGA G-3´ was used to mutate Met65isoform-specific expression of ShcA isoforms. (start site for p52) to leucine in p66HA, resulting in p66HA-ml. Another overlapping oligonucleotideEXPERIMENTAL pair, 5´-CCA ACG ACA AAG TCC TGG GAC CCG GGG-3´ and 5´-CCC CGG GTC CCA GGACells and transfection CTT TGT CGT TGG- 3´, was used to mutate the initiation sites for p46 in both p66HA-ml andHeLa cells were cultured in Dulbeccos modified p52HA, resulting in p66HA-ML and p52HA-ML.Eagles medium (Gibco BRL) supplemented with Silent mutations were introduced into these vectors10% (v/v) fetal-calf serum (AMIMED; BioConcept, at the sites corresponding to h/m-shc siRNA asAllschwil, Switzerland), 0.2mg/ml streptomycin above using the overlapping oligonucleotide pairand 50 units/ml penicillin at 37°C in a humidified 5´-GGG GTT TCC TAC TTG GTC CGC TAC5% CO2 incubator. A day before transfection with ATG GGT TGT C-3´ and 5´-CAC AAC CCA TGTsiRNA, cells were plated in six-well plates in AG C GGA CCA AGT AGG AAA CCC C-3´medium without antibiotics at 1.4×10 5cells/well. (mutated nucleotides underlined) to give p46HA-The next morning, siRNAs were introduced into sm, p52HA-ML-sm and p66HA-ML-sm (h/m-shcHeLa cells using the OLI GOFECT AMINE™ means that the sequence of siRNA is common toreagent (Life Technologies) according to the both human and mouse shc sequences). Note thatmanufacturers instructions, with 10µl of 20µ M proteins expressed from these vectors are identicalsiRNA and 3µl of transfection reagent/well. with the parent proteins.Transfection with expression vectors was carriedout 2 days after the OLIG OFECTAMINE™ Oligoribonucleotidestransfection using LIPOFE CTAMINE™ 2000(Life Technologies). The following 21-mer oligoribonucleotide pairs were used: h/m-shc siRNA from nt 677–697 (in thecDNA cloning of ShcA isoforms PTB domain), 5´-CUA CUU GGU UCG GUA 33
  • RESULTSCAU GGG-3´ and 5´-CAU GUA CCG AAC CAA When HeLa cells were transfected with the h/m-shcGUA GGA-3´; and p66-shc siRNA from nt 236– siRNA, the levels of all three ShcA isoforms were256 [in the CH2 (collagen homology 2) domain], strongly decreased 24h after transfection and5´-GAA UGA GUC UCU GUC AUC GUC-3´ and reached less than 20% of the control after 48h and5´-CGA UGA CAG AGA CUC AUU CCG-3´. 4% after 60h (Figure 2A). The level of controlEntire sequences were derived from the sequence of protein (b-tubulin) was not affected under thehuman p66ShcA mRNA (accession number conditions employed, and this was also the case forHSU7377) and its complement and each pair has a Grb2, a protein that specifically interacts with ShcA3´ overhang of 2nt on each side. Designed RNA upon activation of growth-factor signalling [ 7].oligonucleotides were blasted against the Time-course analysis showed that the levels of allGenBank®/EMBL database to ensure gene ShcA isoforms remained low until the fifth dayspecificity. The RNA oligonucleotides were after transfection, but started to increase thereafter (obtained from Microsynth (Balgach, Switzerland). Figure 2B). The decrease in the three isoforms wasAnnealing was performed as described by Elbashir uniform, suggesting that both p52/p46 and p66et al. [ 20]. The complementary two strands (each at ShcA mRNAs were equally targeted by the siRNA.20µM) in 200µl of annealing buffer [100mMpotassium acetate/30mM Hepes/KOH (pH Isoform-specific ShcA knockdown7.4)/2mM magnesium acetate] were heated for1min at 90°C and then incubated for 1h at 37°C. An When cells were transfected with p66-shc siRNA,siRNA corresponding to nucleotides 753–773 of the only the p66ShcA isoform was decreased, withfirefly luciferase mRNA was used as a negative kinetics similar to that obtained with h/m-shccontrol. siRNA; the other two isoforms were not affected ( Figure 3A and 3B). In another experiment, cellsWestern-blot analysis were challenged a second time with the same p66- shc siRNA 6 days after the initial transfection,At the times indicated, cells were lysed in a buffer when the level of p66ShcA was very low, but aboutcontaining 120mM NaCl, 50mM Tris, pH 8.0, and to increase, and 10 days after the initial1% Nonidet P40 plus Complete (Roche) protein transfection, when the level of p66ShcA recoveredinhibitor tablets. The whole-cell extracts (20 µg) substantially. As shown in Figure 3 (C), the level ofwere analysed by Western blotting using a p66ShcA remained low and decreased markedlypolyclonal rabbit anti-Shc antibody (1:250; again after transfection at days 6 and 10Transduction Laboratories), mouse monoclonal respectively.anti-Grb2 (1:1000; Transduction Laboratories) or amouse monoclonal anti-b-tubulin antibody (1:1000; Isoform-specific ShcA expression: transientSigma). We used anti-rabbit or anti-mouse knockdown-in (see below)horseradish peroxidase-linked antibodies fromAmersham as secondary antibodies. An enhanced The target site of p66-shc siRNA is in the 5´ regionchemiluminescence (ECL®) detection method of p66 ShcA mRNA that is derived from the exon(Amersham) was employed, and the membrane was 1´ and is absent in p52/p46 ShcA mRNA (seeexposed to Kodak X-Omat LS film. Quantification Figure 1). The results of the above experimentsof ShcA proteins was done using ImageQuant 5.0. suggest that the effect of an siRNA is restricted to mRNAs containing a sequence identical with thatRESULTS of the siRNA used. Furthermore, there was no spreading effect of the siRNA signal, at least notEfficient down-regulation of ShcA by siRNA towards the 3´ of the target site in the mRNA. Otherwise, the p46 and p52 isoforms would alsoThree isoforms of ShcA are derived from a single have been down- regulated by p66-shc siRNA. Wegene through differential usage of transcriptional exploited this specificity to establish conditionsinitiation sites (p66 versus p52/p46) and under which ShcA would be expressed in antranslational initiation sites (p52 versus p46) ( isoform-specific manner, which we call transientFigure 1). The primary transcript of p52/p46 knockdown-in. We constructed expression vectorsmRNA contains the entire sequence of p66 mRNA; encoding mouse ShcA isoforms with or withouthowever, the very-5´ region of p66 mRNA is silent mutations at the region corresponding to h/m-present in the first intron of p52/p46 mRNA, but is shc siRNA that left the protein sequencesabsent in the latter mRNA, having been spliced out. unchanged. Two point mutations were introducedThe p46 and p52 isoforms are derived from the into each expression vector so that it was notsame mRNA using different translation initiation recognized by h/m-shc siRNA. Cells were firstsites. Target sites of two siRNAs used are shown in transfected with h/m-shc siRNA to knock-down allFigure 1 below the p66 ShcA mRNA. Note that the three isoforms and then 2 days later with ansequence of p66-shc siRNA is from a 5´ region of expression vector encoding each isoform of mousehuman p66ShcA mRNA and is absent in p52/46 ShcA. As shown in Figure 4, h/m-shc siRNAShcA mRNA. knocked down all three isoforms of endogenous 34
  • RESULTSShcA almost completely. Transfection of these cellswith mutant expression vectors for individualisoforms resulted in the expression of only thecorresponding isoforms. As expected, almost noprotein was detected in cells transfected with wild-type expression vectors. Cells which were nottransfected with h/m-shc siRNA expressed elevatedlevels of ShcA isoforms, irrespective of thepresence or absence of mutations in the expressionvectors.Figure 1 Relationship between the ShcA gene, mRNA and protein.The gene and mRNA are drawn to the same scale. Two transcriptioninitiation sites are indicated by arrows. The boxes of the gene representexons. Exon 0 that is under control of the p52/p46 promoter is ligated toexon 1 after splicing. Exons 1’ and 1 are transcribed contiguously underthe p66 promoter. Translation initiations sites are indicated both on thegene and mRNAs by triangles. The protein domains are demarcatedand indicated: CH1 and 2, collagen homology 1 and 2; PTB,phosphotyrosine binding; and SH2, Src homology 2. The two siRNAsused in this work are indicated below p66 mRNA.Figure 2 Knock-down of ShcA in HeLa cells.Cells were transfected using the oligofectAMINE™ reagent withoutsiRNA (-) or with h/m-shc siRNA (S) and firefly luciferase siRNA (L). Atdifferent times after transfection, whole-cell extracts were prepared andanalysed by Western blotting as described in the Experimental section.Membranes were probed for Shc, β-tubulin and Grb2. (A) Specificity;(B) time course.Figure 3 Isoform-specificknockdown(A) p66-selective knock-down. Cells were untreated ortransfected without siRNA (-) or with h/m-shc siRNA (S), p66-shc siRNA(66) and control firefly luciferase siRNA (L). Whole-cell extracts wereprepared 48h later and analysed for ShcA and b-tubulin levels as in Figure 2. (B) Time course. Cells were transfected with p66-shc siRNA or luc siRNA and the levels of ShcA proteins were analysed at different times as above. (C) Repeated transfection. On days 6 (white triangle) or 10 (black triangle) after the first transfection, cells were transfected again with the same p66-shc siRNA. The levels of ShcA and control b- tubulin proteins were analysed at different times as above. Abbreviation: luc siRNA, firefly luciferase mRNA. 35
  • RESULTSFigure 4 Isoform-specific expression of ShcACells were first transfected with no siRNA (mock) or with h/m-shc siRNA to down-regulate endogenous ShcA proteins. After 2 days, the cells weretransfected with an empty expression vector, pCDNA3, or an expression vector for each isoform of wild-type (-wt) and silent mutant (-sm) ShcA. A day later,whole-cell extracts were prepared and analysed by Western blotting for the ShcA expression level. Membranes were blotted with polyclonal anti-ShcA andanti-b-tubulin antibodies.DISCUSSION that long dsRNA molecules synthesized by RNA-In the present study we have shown that siRNA can dependent RNA polymerase are intermediates inefficiently, specifically and rapidly down-regulate RNAi that amplify and maintain the effect ofthe level of an endogenous protein in mammalian siRNA. This implies 5´ spreading of the silencingcells. The effect of siRNA was restricted to mRNAs signal from siRNA. Our finding in HeLa cells thatcontaining a sequence identical with that of the siRNA-mediated RNAi does not propagate tosiRNA used. That the primary action of siRNA on homologous regions of the target RNA suggestsmRNA, which is most likely an endonucleolytic that this may not be the case in mammalian cells.attack, does not propagate to other regions of the Moreover, no endogenous RNA-dependent RNAtarget mRNA was inferred from the following polymerase has been reported in mammalian cells,observations: (1) the effect of p66-shc siRNA was except after RNA virus infection [ 23]. This mayrestricted to p66ShcA ( Figure 3) and (2) h/m-ShcA also explain why there is no systemic PTGS insiRNA targeted wild-type ShcA, but not mutant, mammals, which requires the amplification ofmRNAs ( Figure 4). In the first observation, siRNA, as has been often observed in otherexpression of p52/p46 ShcA was not affected, Kingdoms [ 16, 24]. Also, if amplification isalthough p66 and p52/p46 mRNAs shared sequence involved in silencing in mammalian cells, antisenseidentity in most of the region 3´ of the siRNA site RNA oligonucleotides alone should serve as a(see Figure 1), indicating that the silencing signal primer for RNA-dependent RNA polymerase anddoes not propagate to regions of mRNA 3´ to the thus be sufficient for gene down-regulation. WesiRNA. The second observation was with found that antisense RNA did not induce silencingectopically expressed ShcA mRNAs. In this (results not shown), in agreement with the previousexperiment, sequences of wild-type ShcA mRNA report by Tuschl et al. [ 25]. Thus it may be aand mutant ShcA mRNA for each isoform were unique feature of mammalian cells that theidentical, except for two nucleotides at the siRNA silencing signal spreads neither 5´ nor 3´ of therecognition site located in the middle of the mRNA. siRNA, restricting the action of siRNA to mRNAsIf the silencing signal did spread either 5´ or 3´ of that carry a sequence identical with that of siRNA.the siRNA, ShcA expression from both wild-type An important implication of our results is that theand mutant mRNAs would have been suppressed. combination of PTGS using siRNA and theThat expression from wild-type mRNAs but not isoform-specific expression of an homologous genefrom mutant mRNAs was suppressed strongly with silent mutations, which we call transientargues for stringent specificity of siRNA-mediated knockdown-in, causes the cell to express only onemRNA decay. This possibility was already isoform, while keeping the levels of other isoformssuggested indirectly by Zamore et al. [ 18]. Using very low. Using the transient knockdown-incell-free decay reactions containing insect cell method, it should be possible to examine the effectlysates and dsRNA, they showed cleavage of target of various mutations of individual isoforms onmRNA only within the region corresponding to the cellular activity. The critical point in this method isdsRNA. However, this analysis was only on the that expression vectors should be designed so thatprimary action of dsRNA-mediated mRNA they are not recognized by siRNA. Twocleavage and did not consider the possibility of mismatches in the mutated expression vectors weresignal amplification taking place in vivo. Indeed, sufficient for them to be exempted from siRNA-results of experiments in an insect cell- free system mediated suppression ( Figure 4). We used[ 21] and C. elegans [ 22] have led to the suggestion expression vectors for mouse ShcA in the present 36
  • RESULTSstudy because there is a high degree of sequence of the manuscript before submission. The Friedrichsimilarity between mouse and human ShcA (90% Miescher Institute is a part of the Novartis ResearchmRNA sequences and 93% amino acid sequences Foundation.for p66 isoform). Thus, instead of introducingmutations into expression vectors, it would have REFERENCESbeen possible to design a different siRNA with a 1 Andreadis, A., Gallego, M.E. and Nadal-Ginard, B.sequence matching perfectly the endogenous (1987) Generation of protein isoform diversity byhuman ShcA gene, but not the transfected mouse alternative splicing: mechanistic and biologicalgene, and achieve similar results. The advantage of implications. Annu. Rev. Cell Biol. 3, 207–242our approach, however, is that we can obtain 2 Johnson, J.E., Wold, B.J. and Hauschka, S.D. (1989)isoform-specific ShcA expression in both human Muscle creatine kinase sequence elements regulating skeletal and cardiac muscle expression in transgenic mice.and mouse cells using the same set of probes. Mol. Cell. Biol. 9, 3393–3399The sequence of p66-shc siRNA is present in the 3 Tan, S.S. (1991) Liver-specific and position-effectprimary transcripts of p52/p46 ShcA mRNA, but is expression of a retinol-binding protein-lacZ fusion genespliced out of the mature mRNA (see Figure 1). (RBP-lacZ) in transgenic mice. Dev. Biol. 146, 24–37 4 Gorman, C. and Bullock, C. (2000) Site-specific geneThe fact that only the p66 ShcA isoform was down- targeting for gene expression in eukaryotes. Curr. Opin.regulated by p66-shc siRNA strongly suggests that Biotechnol. 11, 455–460the site of action of siRNA is confined to the 5 Migliaccio, E., Giorgio, M., Mele, S., Pelicci, G., Reboldi,cytoplasm. If siRNA acted in the nucleus and P., Pandolfi, P.P., Lanfrancone, L. and Pelicci, P.G. (1999) The p66shc adaptor protein controls oxidative stresstriggered the decay of transcripts containing the response and life span in mammals. Nature (London) 402,corresponding sequence, the p52ShcA and 309–313p46ShcA proteins should have been equally down- 6 Muller, U. (1999) Ten years of gene targeting: targetedregulated by the same p66-shc siRNA. Whether mouse mutants, from vector design to phenotype analysis. Mech. Dev. 82, 3–21siRNA can enter the nucleus is not known. Even if 7 Luzi, L., Confalonieri, S., Di Fiore, P.P. and Pelicci, does so, it would be able to access to mRNA only (2000) Evolution of Shc functions from nematode toafter completion of its processing to mature mRNA. human. Curr. Opin. Genet. Dev. 10, 668–674The response of cells to ShcA siRNA was rather 8 Nakamura, T., Muraoka, S., Sanokawa, R. and Mori, N. (1998) N-Shc and Sck, two neuronally expressed Shcfast and efficient; the target ShcA isoforms were adaptor homologs. Their differential regional expression inalready greatly decreased within 24h of transfection the brain and roles in neurotrophin and Src signaling. J.(>60%), and the protein levels remained low until 5 Biol. Chem. 273, 6960–6967days after transfection. After 5 days, the reduced 9 Bonfini, L., Migliaccio, E., Pelicci, G., Lanfrancone, L. and Pelicci, P.G. (1996) Not all Shcs roads lead to Ras.ShcA isoforms started to increase. There are two Trends Biochem. Sci. 21, 257–261possibilities to account for this reappearance of the 10 Migliaccio, E., Mele, S., Salcini, A.E., Pelicci, G., Lai,down-regulated isoforms: (1) adaptation of the cells K.M., Superti-Furga, G., Pawson, T., Di Fiore, P.P.,to the siRNA by establishment of a resistant Lanfrancone, L. and Pelicci, P.G. (1997) Opposite effects of the p52shc/p46shc and p66shc splicing isoforms on themechanism; or (2) dilution or degradation of siRNA EGF receptor-MAP kinase-fos signalling pathway. EMBOwith time. Because repeated transfection J. 16, 706–716maintained or re-established the low levels of target 11 van der Krol, A.R., Mur, L.A., Beld, M., Mol, J.N. andShcA isoform ( Figure 3B), the latter simple Stuitje, A.R. (1990) Flavonoid genes in petunia: addition of a limited number of gene copies may lead to a suppressiondilution/degradation mechanism seems to apply. of gene expression. Plant Cell 2, 291–299While a period of 5 days of down-regulation by a 12 Napoli, C., Lemieux, C. and Jorgensen, R.A. (1990)single transfection is long enough for some Introduction of a chimeric chalcone synthase gene inbiochemical analysis, repeated transfection should petunia results in reversible cosuppression of homologous genes in trans. Plant Cell 2, 931–943allow long-term experiments. Repeated transfection 13 Cogoni, C. and Macino, G. (2000) Post-transcriptionalwill be necessary where the target protein has a gene silencing across kingdoms. Curr. Opin. Genet. Dev.slow turnover. 10, 638–643 14 Carthew, R.W. (2001) Gene silencing by double-In summary, we have shown in mammalian cells stranded RNA. Curr. Opin. Cell Biol. 13, 244–248that the site of action of siRNA-mediated mRNA 15 Cogoni, C., Romano, N. and Macino, G. (1994)degradation is confined to the cytoplasm and that Suppression of gene expression by homologousthe target mRNA is restricted to those mRNAs transgenes. Antonie Van Leeuwenhoek 65, 205–209 16 Fire, A., Xu, S., Montgomery, M.K., Kostas, S.A.,containing a sequence identical with that of the Driver, S.E. and Mello, C.C. (1998) Potent and specificsiRNA used. These specific features of siRNA- genetic interference by double-stranded RNA inmediated gene knock-down can be employed over a Caenorhabditis elegans. Nature (London) 391, 806–811short time period in conjunction with specific 17 Bosher, J.M. and Labouesse, M. (2000) RNAexpression vectors to establish conditions for interference: genetic wand and genetic watchdog. Nat. Cell Biol. 2, E31–E36expression of the ShcA gene in an isoform-specific 18 Zamore, P.D., Tuschl, T., Sharp, P.A. and Bartel, D.P.manner. This knockdown-in method should be (2000) RNAi: double-stranded RNA directs the ATP-applicable and useful for the study of genes dependent cleavage of mRNA at 21 to 23 nucleotideexpressed as multiple isoforms. intervals. Cell 101, 25–33 19 Bernstein, E., Caudy, A.A., Hammond, S.M. and Hannon, G.J. (2001) Role for a bidentate ribonuclease inWe thank Dr Frederick Meins for very helpful discussion the initiation step of RNA interference. Nature (London)during the work and preparation of the manuscript and Dr 409, 363–366Patrick King and Dr George Thomas for critical reading 20 Elbashir, S.M., Harborth, J., Lendeckel, W., Yalcin, A., Weber, K. and Tuschl, T. (2001) Duplexes of 21-nucleotide 37
  • RESULTSRNAs mediate RNA interference in cultured mammalian On the role of RNA amplification in dsRNA-triggered genecells. Nature (London) 411, 494–498 silencing. Cell 107, 465–47621 Lipardi, C., Wei, Q. and Paterson, B.M. (2001) RNAi as 23 Cohen, S.S. (1968) Virus-induced Enzymes, Columbiarandom degradative PCR: siRNA primers convert mRNA University Press, New Yorkinto dsRNAs that are degraded to generate new siRNAs. 24 Fagard, M. and Vaucheret, H. (2000) SystemicCell 107 , 297–307 silencing signal(s). Plant Mol. Biol. 43, 285–29322 Sijen, T., Fleenor, J., Simmer, F., Thijssen, K.L., 25 Tuschl, T., Zamore, P.D., Lehmann, R., Bartel, D.P.Parrish, S., Timmons, L., Plasterk, R.H. and Fire, A. (2001) and Sharp, P.A. (1999) Targeted mRNA degradation by double-stranded RNA in vitro. Genes Dev. 13, 3191–3197 38
  • RESULTS2.2 Using siRNAs to study Shc earlier study involving siRNA-mediatedfunction knockdown of genes in human cells (Elbashir et al., 2001a). As a control for specificity, cells2.2.1 Isoform-specific knockdown of were transfected with siRNA targeting the Shc luciferase gene (si-luci), which is notp46/52 expressed in the cells, or with no siRNA As described before, siRNA transfection (buffer) (Fig 2.2.1B). During the course of theallowed the knockdown of all Shc isoforms experiments, we noticed that knockdown ofand, the specific knockdown of p66Shc. In the Shc isoforms negatively influenced the growthnext step we generated siRNA to specifically of HeLa cells. While transfection without anyknockdown p46/52Shc. The 3’ end of the siRNA (buffer) and with siRNA targeting thep46/52Shc mRNA contains exon 1 which is not luciferase gene did not impair the proliferationtranslated and, more importantly, not present of HeLa cells, transfection with Shc-targetingin the p66Shc mRNA (Fig 2.2.1A). Transfection siRNAs reduced the growth of the cells (Figof siRNA targeting this region (si-p46/52 Shc ) 2.2.2-1A). This effect was strongest afterspecifically reduced the protein levels of p46 Shc knockdown of all Shc isoforms (si-shc1), andand p52Shc, but did not change the amount of seemed to be milder after knockdown of onlyp66 Shc (Fig 2.2.1B). p66Shc or p46/p52Shc. To see whether the observed effect is due to Shc knockdown, a second siRNA (si-shc2) which targets a2.2.2 Growth inhibition upon Shc different region in the mRNA of all Shcknockdown isoforms was used. Upon transfection with si- shc2, HeLa cell proliferation was reduced to a To develop the isoform-specific knockdown similar extent to si-shc1, when measured usingand the knockdown-in system, all experiments an MTT proliferation assay (Fig 2.2.2-1B).were done using 100 nM siRNA, based on anFigure 2.2.1: Isoform-specific knockdown of Shc A. Schematic representation of p46/52Shc and Shcp66 mRNA. The triangles point to the ATG start codons, and the black boxes indicate the sitetargeted by the siRNA. B. HeLa cells were transfected with different siRNAs and collected three dayslater. Total cell lysate was used for Western blot analysis to probe for Shc and β-tubulin levels. B:buffer; L: si-luci; S: si-shc1; 46/52: si-p46/52shc; 66: si-p66shc. 39
  • RESULTSFigure 2.2.2-1: Transfection of Shc siRNA inhibits growth in HeLa cells. A. HeLa cells weretransfected with different siRNAs as indicated. Three days later pictures were made using lightmicroscopy. B. HeLa cells were transfected with si-shc1 (S1), si-shc2 (S2), or si-luci (L) andproliferation was measured after three days using an MTT proliferation assay: additionally, total celllysate was used for Western blot analysis. We next investigated whether this growth endogenous Shc proteins, and proliferationinhibition is a general effect which can also be was measured by MTT assay. Expression ofobserved in other cell types. As depicted in p52Shc did not rescue proliferation in HeLa orFig. 2.2.2.-2, knockdown of all Shc isoforms LLC-PK1 cells (Fig 2.2.2-3A). Moreover,inhibited proliferation in PNT2 cells, PC3 cells, expression of siRNA-resistant p46Shc in Shc Shcand LLC-PK1 cells that do not express p66 . addition to p52 also did not prevent the To investigate which Shc isoform was growth-inhibitory effect caused by the siRNA inresponsible for the growth inhibitory effect and LLC-PK1 cells, although the Shc expressionwould be able to rescue it, the knockdown-in level was completely rescued (Fig 2.2.2-3B).system was used. LLC-PK1 and HeLa cells These experiments demonstrated that thewere generated to stably express single Shc effect of siRNA transfection on proliferationisoforms which carry a silent mutation in the was not dependent on Shc protein levels, butregion targeted by the siRNA. These cells were presented rather an unspecific response.subsequently used to knockdown the 40
  • RESULTSFigure 2.2.2-2: Transfection of Shc siRNA inhibits growth in PNT2, PC3, and LLC-PK1 cells.Cells were transfected with si-shc1 (S) or si-luci (L), and three days later proliferation was measuredusing an MTT assay and total cell lysate was analyzed by Western blotting.Figure 2.2.2-3: Growth inhibition caused by si-shc is an unspecific effect. Stable HeLa or LLC- shcPK1 cell lines expressing an empty vector or silent mutants of HA-p52 (p52shcsm) (A) and HA- Shcp46/52 (only LLC-PK1) (B) were prepared and transfected with siRNA targeting all Shc isoforms (S)or luci siRNA (L). After three days of transfection, proliferation was measured using an MTT assay andtotal cell lysate was analyzed by Western blotting. 41
  • RESULTS To avoid unspecific events for future to buffer transfection or transfection with aexperiments, the minimum amount of siRNA scrambled control si-RNA in LLC-PK1 andsufficient for efficient knockdown was MCF7 cells (Fig. 2.2.2-4B, C).determined. Fig. 2.2.2-4A shows that In conclusion, to study the function ofknockdown using only 10 nM siRNA reduced proteins using siRNA only low concentrationsShc protein levels as effectively as knockdown of siRNA should be used to avoid unspecificusing 100 nM siRNA. Moreover, almost no effects, and rescue experiments should beeffect on proliferation was observed after performed to ensure specificity.transfection with 10 nM si-shc when comparedFigure 2.2.2-4 Knockdown of Shc using 10 nM siRNA. A. Indicated amounts of si-shc1 weretransfected into LLC-PK1 cells and, three days after transfection, total cell lysate was analyzed byWestern blotting. B, C. LLC-PK1, MCF7, and a stable MCF7 cell line expressing HA-p52Shc weretransfected with buffer, si-shc1, or si-control. Proliferation was measured by counting cells with the Vi-CELL analyzer. 42
  • RESULTS2.3 Role of Shc in EGF-induced knockdown of either all Shc isoforms or only p66Shc did not affect EGF-induced Erksignaling in epithelial cells activation at any timepoint in HeLa cells (Fig 2.3.1A right panel). In addition, knockdown of2.3.1 Role of Shc in EGF-induced any Shc isoform did not influence EGF-inducedErk activation Erk activation in the epithelial cell lines PNT2 (Fig 2.3.1B left panel), PC3, and LLC-PK1 (data not shown). Results with Shc-deficient Shc has been shown to be involved in EGF- MEFs suggest that Shc sensitizes cells to lowinduced Erk activation (1.1.2). However, the amounts of growth factors. Therefore, siRNA-contribution of each Shc isoform to this event transfected HeLa and PNT2 cells were treatedhas not been investigated. In addition, it is still with low concentrations of EGF in the nextunclear whether p66Shc influences Erk experiment. EGF-induced Erk activation wasactivation in a dominant negative manner. The reduced by low amounts of EGF; however, Shcisoform-specific knockdown and the knockdown had no influence in both HeLa andknockdown-in system provide perfect tools for PNT2 cells (Fig 2.3.1B).investigating this question. Surprisingly,Figure 2.3.1: Effect of Shc knockdown on EGF-induced Erk activation. Cells were transfectedwith siRNA and treated three days later with 50 ng/ml EGF for the indicated time (A) or for 10 min withthe indicated EGF concentration (B). Total cell lysate was subjected to Western blot analysis. U:untreated; B: buffer; L: si-luci; S: si-shc1; 66: si-p66Shc; 46/52: si-p46/52Shc. 43
  • RESULTS Shc2.3.2 Effect of p66Shc on EGF-driven p66 expression. We also tested other clones (data not shown), but the proliferation rateproliferation and cell survival seemed to be clone-dependent, rather than dependent on the expression of Shc isoforms. MCF7 epithelial breast cancer cells express Shc The same result was observed for cell viability,p66 at a very low level. Some reports Shc which started to decline 4 days after starvationsuggest that reduced p66 expression may in the absence of EGF. However, EGFplay a role in breast cancer, which is often treatment restored cell viability in all celldriven by signaling from the EGFR family Shc clones, independent of their p66 expression(Stevenson and Frackelton, 1998; Xie and level (Fig. 2.3.2-2B), suggesting that neitherHung, 1996). Since p66Shc might negatively EGF-induced proliferation nor cell survival wasinfluence Erk signaling downstream of EGFR Shc influenced by the expression of p66 infamily members, it is intriguing to speculate Shc MCF7 cells.that p66 reduces EGF-driven proliferation or Shcsurvival. To investigate p66 function inMCF7 cells, stable cell lines overexpressing ShcHA-p66 were generated. In addition, emptyvector-expressing cells or cell clones Shc Shcoverexpressing HA-p46 or HA-p52 weremade as a control (Fig 2.3.2-1).Figure 2.3.2-1: MCF7 cells stablyoverexpressing Shc isoforms. MCF7 cellswere stably transfected to overexpress HA- Shc Shc Shcp46 , HA-p52 , and HA-p66 . The clonesdepicted here were used for furtherexperiments. The responsiveness to EGF-drivenproliferation or viability of these cell clones wasthen investigated. Proliferation under starvationconditions without EGF was slowed down in allcell clones (Fig. 2.3.2-2A). The addition of 50ng/ml EFG to the starvation medium enhancedproliferation in all cell clones. Although thesecell clones exhibited different proliferationrates, there was no consistency with respect to 44
  • RESULTS ShcFigure 2.3.2-2: Effect of p66 overexpression on EGF-mediated proliferation and survival.Stably transfected MCF7 cells expressing empty vector, HA-p66Shc, or HA-p52Shc were seeded, andfrom the next day (day 0) starved in medium containing 0.1 % FCS. Alternatively, 50 ng/ml EGF wasadded to the starvation medium. At days 0, 2, and 4 after starvation, cells were counted and viabilitywas measured using trypan blue exclusion with VI-CELL analyzer. 45
  • RESULTS2.4 Research Publication (under review) INDUCTION OF UPA GENE EXPRESSION BY THE BLOCKAGE OF E-CADHERIN VIA SRC- AND SHC-DEPENDENT ERK SIGNALING Sandra Kleiner, Amir Faisal¶ and Yoshikuni Nagamine Friedrich Miescher Institute for Biomedical Research, Maulbeerstrasse 66, 4058 Basel, Switzerland Running title: Loss of E-cadherin function induces uPAAddress correspondence to: Yoshikuni Nagamine, Friedrich Miescher Institute for BiomedicalResearch, Maulbeerstrasse 66, CH-4058 Basel, Switzerland. Phone: +41 61 697 6669. Fax: +41 61697 3976. E-mail:¶ Present address: Cancer Research UK, London Research Institute, 44 Lincoln’s Inn Fields, London,WC2A 3PX, UKLoss of E-cadherin-mediated cell-cell Tumor cell invasion, a key event of metastaticadhesion and expression of proteolytic progression, requires spreading of tumor cellsenzymes characterize the transition from from the primary tumor. This is stronglybenign lesions to invasive, metastatic tumor, a dependent on the loss of homotypic cell-cellrate-limiting step in the progression from adhesion. E-cadherin is an important componentadenoma to carcinoma in vivo. A soluble E- of the cell-cell adhesion complex and requiredcadherin fragment recently found in the for the formation of epithelia in the embryo andserum and urine of cancer patients has been the maintenance of the polarized epithelialshown to disrupt cell-cell adhesion and to structure in the adult (2). As a single-spandrive cell invasion in a dominant-interfering transmembrane-domain glycoprotein, E-cadherinmanner. Physical disruption of cell-cell mediates cell-cell adhesion via calcium-adhesion can be mimicked by the function- dependent homophilic interaction of itsblocking antibody Decma. We have shown extracellular domain (3). Proteins such as p120-previously in MCF7 and T47D cells that catenin, α-catenin and β-catenin assemble theurokinase-type plasminogen activator (uPA) cytoplasmic cell adhesion complex on itsactivity is upregulated upon disruption of E- intracellular domain and link E-cadherincadherin-dependent cell-cell adhesion. We indirectly to the actin cytoskeleton (3).explored the underlying molecular E-cadherin is considered to be an importantmechanisms and found that blockage of E- tumor suppressor (2,4). In vitro studies havecadherin by Decma elicits a previously clearly established a direct correlation between aundescribed signaling pathway downstream defect in functional E-cadherin expression at theof E-cadherin that leads to Src-dependent Shc cell surface and the acquisition of an invasiveand Erk activation and results in uPA gene phenotype (3). Moreover, a partial if notactivation. siRNA-mediated knockdown of complete reversal of the invasive phenotypeendogenous Shc and subsequent expression of could be achieved by ectopic expression of E-single Shc isoforms revealed that p46Shc and cadherin (3,5). While E-cadherin expression isp52Shc but not p66Shc were able to mediate Erk maintained in most differentiated carcinomas,activation. A parallel pathway involving PI3K there is a strong correlation in several types ofcontributed partially to Decma-induced Erk cancer, including breast, gastric, prostate, andactivation. This report describes that colon carcinoma, between loss of E-cadherindisruption of E-cadherin-dependent cell-cell expression and aggravated phenotypes, e.g.adhesion induces intracellular signaling with metastasis and malignancy leading to a poorthe potential to enhance tumorigenesis and, survival rate (4). Loss of E-cadherin-mediatedthus, offers new insights into the cell-cell adhesion occurs through variouspathophysiological mechanisms of tumor mechanisms, such as downregulation of E-development. cadherin expression via promoter hypermethylation (6), transcriptional repression (7), E-cadherin gene mutation (7), modification of β-catenin (8), or the cleavage of E-cadherin by 46
  • RESULTSmatrix metalloproteases (MMPs) (9). Cleavage induce ligand-independent activation of the EGFof E-cadherin results not only in the disruption of receptor (EGFR) and subsequent activation ofcell-cell adhesion but also in a soluble 80-kDa E- (Erk) (22) (14). Moreover, it was shown that thecadherin fragment that itself disrupts cell-cell E-cadherin adhesive complex can be linkedadhesion in a dominant-interfering manner, directly to EGFR via the extracellular domain ofthereby promoting tumor progression (10). E-cadherin and negatively regulate receptorHowever, the intracellular processes subsequent tyrosine kinase signaling in an adhesion-to disruption of cell-cell adhesion remain elusive. dependent manner (24).Tumor metastasis involves, in addition to the We have shown previously that disruption of E-loss of cell-cell adhesion, degradation of the cadherin-dependent cell-cell adhesion with theextracellular matrix. MMPs and urokinase-type function-blocking antibody Decma (also termedplasminogen activator (uPA) are known to be Uvomorulin antibody or anti- Arc-1) results ininvolved in extracellular matrix degradation. disruption of cell-cell adhesion of T47D andMoreover, increased expression of uPA is MCF7 breast cancer cells (25). The loss of thedirectly related to higher tumor growth and epithelial morphology was associated with anmetastasis (1). Several analyses have already increased secretion of uPA into the extra cellularmade it clear that the expression of E-cadherin milieu. Furthermore, Decma treatment inducedand the expression of MMPs are inversely invasiveness into collagen which was inhibitedcorrelated (11,12) and that E-cadherin-dependent by the addition of uPA antibodies. The enhancedcell-cell contact regulates the expression of uPA secretion was dependent on transcriptionMMPs and uPA in vitro (13-15). However, the (25). It appears therefore that disruption E-underlying molecular mechanisms are not yet cadherin-dependent cell-cell adhesion initiatesfully understood. Expression of genes for these signaling events leading to the uPA gene.proteolytic enzymes can be induced by various However, the nature of these signaling eventsstimuli, including growth factors and integrin has remained largely unknown. Since bothligation, and has often been shown to be Erk- disruption of E-cadherin-dependent cell cell-dependent (16-18). The adaptor protein ShcA, adhesion and the expression of uPA are causallywhich is referred to here as Shc, is involved in involved in tumor progression, the understandingcoupling receptor and non-receptor tyrosine of these underlying intracellular events is ofkinases to the Ras/extracellular regulated kinase paramount importance. In the present study, we(Erk) pathway (19). Shc is expressed in three explored the signaling pathway linkingdifferent isoforms, p46Shc, p52Shc and p66Shc, but perturbation of E-cadherin-dependent cell-cellonly the two smaller isoforms seem to be adhesion to the activation of Erk and the uPAinvolved in Erk activation (20). Receptor gene expression.tyrosine kinases activated by tyrosinephosphorylation recruit and phosphorylate these MATERIALS AND METHODSShc isoforms. This creates a binding site forGrb2 and results in the recruitment of the Reagents - Decma and HA antibodies wereGrb2/Sos complex to the vicinity of Ras, where used as a hybridoma supernatant (containingSos acts as a GTP exchange factor for Ras. approximately 40-50 µg/ml antibody) dialyzedAn increasing body of evidence suggests that against DMEM (Mr cutoff 1.2×104). The Decmacadherins act at the cellular level as adhesion- hybridoma cell line was kindly provided by D.activated cell signaling receptors (3). Indeed, Vestweber and R. Kemler. The dialyzedhomophilic ligation of the E-cadherin counterpart was used for control treatments.ectodomain induces activation of several Unless indicated otherwise, cells were treated forsignaling molecules, such as Rho-family 30 min with Decma supernatant (40-50 µg/mlGTPases (3), MAPKs (22) and PI3K (3). These antibody) or control supernatant. Anti-Shcsignals are believed to regulate dynamic polyclonal, anti-Grb2, anti-E-cadherinorganization of the actin cytoskeleton and the monoclonal (used for Western analysis andactivity of the cadherin/catenin apparatus to immunostaining), and antiphospho-Src (Tyr-416)support stabilization of the adhesive contact (23). polyclonal antibodies were obtained fromSeveral studies have suggested functional Transduction Laboratories. Antiphospho-PKBinterdependence of cadherins and receptor (Ser 473) and antiphospho-Erk polyclonaltyrosine kinases with respect to their signaling antibodies were from Cell Signaling, and anti-capacities. It has been shown that initiation of de Erk polyclonal, anti E-cadherin (sc-8426) (E-novo E-cadherin-mediated adhesive contacts can 47
  • RESULTScad2) and anti-EGFR (sc-101) monoclonal AAC CAA GUA GGA-3´; control siRNA 5-antibody were obtained from Santa Cruz. Mouse GUA CCU GAC UAG UCG CAG AAG-3 andmonoclonal HA antibodies (12CA5) used for 5’-UCU GCG ACU AGU CAG GUA CGG-3’.Western blotting or immunoprecipitation were The specificities of these sequences werepurified on a protein A-Sepharose column and confirmed by blasting against themonoclonal antibodies against phosphotyrosine GenBank/EMBL database.(4G10) were used as hybridoma supernatant. Immunoprecipitation and Western blotAnti-Src mouse monoclonal antibody (clone 327) analysis - Immunoprecipitation and Westernwas a gift from K. Ballmer-Hofer (Paul Scharrer blotting were performed as described (28).Inst.). SB203580 and CGP77675 were kindly RNA isolation and Northern blot analysis -provided by E. Blum (Novartis AG), MCF7 cells (0.5×106/well) were seeded in 6-wellWortmannin, UO126 and Y27634 were obtained tissue culture plates. After 2 days, cells werefrom Calbiochem, LY294002 and Cytochalasin treated as indicated, total RNA was isolated andD (CytD) was from Sigma, SP600125 was 10µg aliquots subjected to Northern blot analysisobtained from Biomol, and TPA, horseradish as described (27).peroxidase-conjugated anti-mouse and anti- Immunofluorescence - Immunostaining andrabbit antibodies, ECL reagent, protein A and G- microscopy were carried out as describedSepharose were from Amersham Bioscience. previously (28). Briefly, cells were cultured on a Cells and transfection - MCF7 and T47D cover slip, fixed in 1 ml of pre-warmed 3%cells were cultured in DMEM/HAMs F12 paraformaldehyde in PBS for 20 min at room(Invitrogene) supplemented with 10% FCS, 0.2 temperature, permeabilized with 0.5% Triton X-mg/ml streptomycin, and 50 units/ml penicillin at 100 in PBS for 10 min, blocked with 5% normal37°C in a humidified incubator with 5% CO2. goat serum for 20 min, incubated with anti-E-For plasmid transfection, T47D cells cadherin antibody (1:200) for 2 h, washed 3(0.7×106/well) were seeded in 6-well tissue times with PBS, incubated 45 min with Alexaculture plates and incubated overnight. Plasmid 488 (1:500) and washed again 3 times. Finally,DNA (1 µg ) was transfected using 5 µl the cover slips were mounted on glass slides withLipofectamine 2000 according to the Flouromount-G (Southern Biotechnologymanufacturer’s instructions. Fugene 6 was used Associates Inc). Fluorescence was visualizedto transfect MCF7 cells (0.5×106/well). Plasmid with a Zeiss Axioplan fluorescence microscopeDNA (1 µg) was incubated with Fugene 6 (ration (63X oil objective with, numerical aperture of2:3) for 20 min and the whole mixture was added 1.4) and all images were captured usingto the cells, which were then incubated overnight Axiovision 3.0 37°C. For siRNA transfection, T47D cells Reporter gene assay (dual-luciferase-(0.18×106/well) and MCF7 cells (0.12×106/well) assay) - Cells (5x105) were plated in a 6-wellwere seeded and transfected the next day with 10 dish to be co-transfected the next day with thenM siRNA as described (26) using 5 µl and 3 µl reporter plasmid and the Renilla control plasmidOligofectamine, respectively. MCF7 cells using Fugene 6. One day after transfection, cellsexpressing siRNA against E-cadherin or mouse were pretreated for 45 min with the indicatedNCAM were generated by the stable transfection inhibitors and afterwards with Decma for 5 hof the pSuper retro vector containing the before harvesting. Luciferase expression wasrespective sequences. These cell lines were measured according to the given protocol (Dual-kindly provided by F. Lehembre and G. Luciferase Reporter Assay System, Promega)Christofori and details will be published and normalized against Renilla expression.elsewhere. Plasmids and siRNAs - Construction of RESULTSexpression vectors for HA-tagged full-lengthmouse p46shc, p52shc, and p66shc and the Decma treatment disrupts E-cadherin-introduction of silent mutations by site-directed dependent cell-cell adhesion and induces uPAmutagenesis were described previously (26).The gene expression - Under normal growthuPA-reporter plasmid pGL-2-puPA-4.6 was conditions, T47D and MCF7 breast cancerdescribed previously (27). The following 21-mer epithelial cell lines grow very compact and E-oligoribonucleotide pairs (siRNAs) were used: cadherin was concentrated at the border of theshc siRNA nt 677–697 (in the protein tyrosine cell-cell interaction, corresponding to typicalbinding domain), 5´-CUA CUU GGU UCG adhesive junction localization (Fig. 1A, a and c).GUA CAU GGG-3´ and 5´-CAU GUA CCG 48
  • RESULTSAs described previously (25), Decma treatment whether the disruption of cell-cell adhesion bydestroyed tight cell-cell interaction, resulting in Decma influenced expression of the uPA gene ,disruption of the epithelial layer (Fig 1A, b and we examined change in uPA mRNA levels.d) and acquisition of a scattered phenotype (Fig. Northern blot analysis showed only barely1B). In addition, E-cadherin disappeared from detectable levels of uPA mRNA under normalthe plasma membrane and was redistributed into growth conditions. However, an increase in uPAthe cytoplasm, suggesting its internalization (Fig. mRNA levels was observed already at 2 h after1A, b and d). The loss of functional cell-cell Decma treatment (Fig. 1D), whereas the controladhesion became more apparent when MCF7 treatment (HA supernatant) had no effect. As acells were grown in Matrigel. Decma treatment positive control, cells were treated with TPA, aled to the resolution of compact cell clusters and potent inducer of uPA gene expression (16).dissociation of the cells (Fig. 1C). To determineFig 1. Effects of Decma treatment on E-cadherin distribution, cell scattering and uPA expression. A, T47Dcells (a and b) and MCF7 cells (c and d) were treated for 4 h with control or Decma supernatant andimmunostained with anti-E-cadherin antibody recognizing the cytoplasmic part of E-cadherin. B, T47D cells (aand b) and MCF7 cells (c and d) were grown for 2 days to 60-70% confluence and then treated with control orDecma supernatant for 6 h before recording. C, MCF7 cells were grown overnight in Matrigel in the presence ofcontrol or Decma supernatant and stained with crystal violet to visualize cells before recording. D, MCF7 cellswere treated with Decma and HA supernatant or 100 ng/ml TPA as indicated and subjected to Northern blothybridization analysis for uPA and GAPDH mRNA levels. The uPA mRNA levels were normalized againstGAPDH mRNA. 49
  • RESULTSFig 2. Role of Erk in Decma-induced uPA upregulation. A,B, T47D cells were treated for 30 min withdifferent amounts of supernatant (A) or for different time periods (B) and total cell lysate was subjected toWestern blot analysis for phospho-Erk levels. C, Comparison of phospho-Erk levels induced by differentsupernatants. MCF7 cells were treated for 30 min with Decma supernatant (Decma), Decma supernatant afterDecma-depletion (Decma-depl.), anti-HA antibody-containing supernatant (HA) or control supernatant beforeanalyzing the total cell lysates by Western blotting. For Decma-depleted supernatant, Decma supernatant wasincubated with Protein A beads rotating overnight to pull down the antibody. D, MCF7 cells stably transfectedwith a pSuper retro-vector to express siRNA targeting E-cadherin or mouse-specific NCAM were treated for 30min with Decma, or 10 min with 50 ng/ml EGF, and total cell lysates were subjected to Western blot analysis forphospho-Erk status. E, MCF cells were co-transfected with a luciferase construct under the control of the uPApromoter and the Renilla plasmid overnight. Cells were then pretreated for 45 min with 10 µ M UO126 (UO) asindicated and subsequently for 5 h with Decma or control supernatant before harvesting. Luciferase activity wasmeasured and normalized against Renilla. Decma-induced uPA gene expression is medium change, which is known to activate Erk.dependent on Erk activation - We and others To test whether the observed Erkhave shown that activation of Erk plays an phosphorylation was due to the blocking activityimportant role in uPA gene expression (16,29). of Decma, we depleted Decma antibodyTo determine whether Decma treatment caused molecules with Protein A-Sepharose. Treatmentactivation of Erk, we investigated the of MCF7 cells with this Decma-depletedphosphorylation status of Erk. Western blot conditioned medium had no pronounced effectanalysis revealed a dose-dependent increase in on scattering (data not shown) or marked ErkErk phosphorylation upon Decma treatment (Fig. phosphorylation (Fig. 2C). To test whether the2A). This phosphorylation peaked 10-15 min observed Erk activation is a result of anafter Decma treatment and declined slowly, but interaction between Decma and E-cadherin, weremained at substantial levels for more than 3 h examined the effect of the Decma-conditioned(Fig. 2B). Low and transient increase in Erk medium on cells expressing low amounts of E-phosphorylation observed in control and HA- cadherin using an MCF7 cell line stablytreated cells (Fig. 2C) may be a response to transfected with a pSuper retro vector expressing 50
  • RESULTSan E-cadherin-specific siRNA. As a control, cells impact on Erk activity was not a general effect ofwere stably transfected with a pSuper vector siRNA on Erk signaling since TPA-induced Erkexpressing siRNA to target mouse-specific activation was not affected by the same siRNANCAM, an mRNA that is not expressed in this (Fig. 3B, right panel). To further test whether thecell line. Although, the knockdown of E- inhibition of Erk activation was caused by thecadherin was not complete, Decma-induced Erk reduction in Shc protein, rescue experimentsactivation was markedly lower under these were performed using the siRNA-mediatedconditions than in non-transfected or control knockdown-in approach (26). MCF7 cells werecells (Fig. 2D). Thus the effect of Decma on Erk stably transfected with plasmids encoding singleactivation depends on the presence of the E- Shc isoforms, which carry silent mutations at thecadherin protein. EGF-induced Erk activation targeting site of the siRNA. These cell lines werewas not affected in any of these cell lines. To further used for siRNA transfection toacertain that Erk activation was a result of knockdown the endogenous proteins withoutdisruption of cell-cell adhesion and not merely of affecting the silent mutant isoform. Fig. 3Cbinding of an antibody to E-cadherin or to any shows that knockdown of Shc in control cells,given surface molecule, we treated MCF7 cells transfected with the empty vector, markedlywith a second anti-E-cadherin (E-cad2) and an reduced Decma-induced Erk phosphorylation.anti-EGFR antibody. Both antibodies recognize This effect could be rescued by the expression ofthe extracellular part of their respective proteins. silent mutant p46Shc or p52Shc but not by silentIn contrast to Decma, however, none of them mutant p66Shc. To mediate Erk activation, Shcinduced disruption of cell-cell adhesion and proteins must be tyrosine phosphorylated onscattering (data not shown). Western blot either Tyr239/240 or Tyr313 (Tyr317 inanalysis revealed that in contrast to Decma humans). Accordingly, expression of p52Shc3Y3Ftreatment, neither treatment with the E-cad2 with all the three tyrosines mutated toantibody nor the EGFR antibody induced Erk phenylalanine did not rescue Decma-induced Erkactivation (SupFig). Taken together, these results activation (Fig. 3C). Moreover, Decma-inducedsuggest that the observed Erk activation was Erk activation was already reduced byspecific for the disruption of cell-cell adhesion overexpressing p52Shc3Y3F and p66Shc without theinduced by blocking of E-cadherin via Decma. knockdown of endogenous Shc, suggesting thatTo find out whether Decma-induced Erk they act in a dominant-negative manner. Theseactivation is necessary for enhanced uPA gene results indicate that Decma-induced Erkexpression, we examined the effect on uPA activation is largely dependent on the p46shc andpromoter activity of the inhibitor UO126, which p52shc proteins.blocks MEK1, the upstream kinase of Erk. Involvement of Src and PI3K in Decma-Transient transfection assays showed that Decma induced Erk activation - The E-cadherintreatment strongly enhances uPA promoter adhesion complex is linked to the actinactivity, which was efficiently suppressed by cytoskeleton via catenin proteins. We showedpretreatment of the cells with UO126 (Fig. 2E). previously that changes in the actin cytoskeletonThese results indicate that Decma treatment induce Shc-dependent Erk phosphorylation andactivates the uPA promoter through a signaling uPA upregulation in LLC-PK1 cells (28).pathway involving Erk. Therefore, we examined whether Decma-induced Shc is necessary for Decma-induced Erk Erk activation requires an intact cytoskeleton.activation - Activation of Erk by various Cytochalasin D (CytD) is a pharmacologicalextracellular signals is often preceded by Shc agent that caps actin filaments and stimulatesphosphorylation. Accordingly, we examined Shc ATP hydrolysis on G actin, leading to a veryactivation, as indicated by its tyrosine rapid dissolution of the actin cytoskeleton (30).phosphorylation, and its association with Grb2. Pretreatment with CytD as well as simultaneousBoth Shc activation and its Grb2 association treatment with Decma and CytD atincreased after Decma treatment in MCF7 cells concentrations known to disrupt the cytoskeleton(Fig. 3A) and T47D cells (data not shown). did not prevent Decma-induced ErkRNAi experiments were performed to examine phosphorylation in MCF7 cells. CytD treatmentwhether Shc activation is causally linked to Erk alone had no effect on Erk phosphorylation inactivation. Knockdown of all Shc isoforms by MCF7 cells but reduced the level in T47D cells.siRNA strongly decreased Erk phosphorylation Nevertheless, treatment with Decma resulted inin both MCF7 and T47D cell lines, while control enhanced Erk phosphorylation irrespective ofsiRNA had no effect (Fig. 3B). The observed 51
  • RESULTSFig 3. Role of Shc in Decma-induced Erk activation. A, Effect of Decma on Shc phosphorylation and itsassociation with Grb2. After treatment of cells with Decma supernatant for 30 min, 300 µg of total cell lysateswere immunoprecipitated with anti-Shc antibody and subjected to Western blot analysis. B, Effects of Shcdownregulation on Erk activation. Cells were transfected with control (C) or Shc (S) siRNA as described in“Materials and Methods” and treated 3 days later with Decma supernatant or 100 ng/ml TPA as indicated,followed by Western blotting for Shc, phospho-Erk and total Erk levels. C, Rescue by ectopic Shc isoformexpression of Erk activation that was suppressed by downregulation of endogenous Shc. Stable cell linesexpressing empty vector or silent mutants of HA-p46shc, HA-p52shc, HA-p52shc3Y3F, and HA-p66shc were preparedand transfected with siRNA targeting all Shc isoforms (S) or control siRNA (C). After 3 days transfection, cellswere treated with Decma supernatant and total cell lysates were analyzed by Western blotting.CytD treatment (Fig. 4A). These results suggest as by the MEK1 inhibitor UO126 and partiallythat the actin cytoskeleton is not required for attenuated by the PI3K inhibitor Wortmannin.Decma-induced Erk activation. No effect was observed with inhibitors of p38Some reports show a functional cross talk MAPK, Rho kinase or JNK, although theirbetween E-cadherin and the EGFR (14,22,31). activities were confirmed by different controlTo determine whether Decma-induced Erk experiments (data not shown). These resultsactivation is a result of cross talk between E- show that not only MEK1 and Shc but also Srccadherin and the EGFR, which might then and PI3K are upstream of Erk in Decma-inducedactivate the Shc/Erk pathway, we examined signaling.whether EGFR activity was required for Erk Role of Src in Decma-induced Erkactivation. As shown in Fig. 4B, Decma-induced activation - Since Src kinase activity was foundErk phosphorylation was not affected by the to be necessary for Decma-induced ErkEGFR-specific inhibitor PKI166, while EGF- phosphorylation, we examined the activation ofinduced Erk activation was completely Src. Western blot analysis showed that Decmasuppressed, indicating that Decma-induced Erk treatment enhanced Src phosphorylation ofactivation does not rely on transactivation of the Tyr416, an indicator of Src activation (Fig. 5A).EGFR. To assess whether Src is upstream of Shc,In a search for molecules other than Shc lying Decma-induced Shc tyrosine phosphorylation inbetween E-cadherin and Erk in Decma-induced the presence of the Src inhibitor CGP77675 wassignaling, we made use of specific inhibitors of examined. Decma-induced Shc tyrosinevarious kinases potentially involved in this phosphorylation and its association with Grb2signaling. Fig. 4C shows that Decma-induced were suppressed by the inhibitor, suggesting thatErk phosphorylation was completely suppressed Src is located upstream of Shc in this signalingby the Src-specific inhibitor CGP77675 as well cascade (Fig. 5B). Again, the Rho kinase 52
  • RESULTS also blocked PKB phosphorylation, suggesting that Src acts upstream of PI3K in Decma- induced signaling. Neither Shc phosphorylation nor its association with Grb2 were affected by Wortmannin (Fig. 6B). Shc knockdown resulted in attenuation of basal PI3K signaling as measured by PKB phosphorylation, but Decma treatment still enhanced PKB phosphorylation (Fig. 6C). Erk phosphorylation was completely suppressed when Shc knockdown and Wortmannin treatment were combined (Fig. 6C). Taken together, these results imply the presence of two parallel pathways downstream of Src leading to Erk activation, one mediated by Shc with a major contribution to Erk activation and the other mediated by PI3K with a minor contribution to Erk activation. Decma-induced uPA expression is dependent on Src and PI3K in addition to Erk - We show here that Src activation is necessary for Erk activation and that PI3K contributes partially. Also, Erk activation is necessary for Decma-induced uPA gene expression (Fig. 2E). As expected, we found that pretreatment withFig 4. Effect of CytD and several kinase inhibitors UO126 but also with CGP77675 abolishedon Decma-induced Erk activation. A, MCF7 and Decma-induced uPA activation (Fig. 7).T47D cells were treated separately with Decma Wortmannin, which only partially inhibited Erksupernatant (Decma) for 30 min, with 3 µM CytD for activation (Fig. 6A), also reduced uPA gene30 min, with 3 µM CytD for 45 min followed by expression to some extent. These results indicateDecma for 30 min, or simultaneously with Decma and that blockage of E-cadherin function inducesCytD for 30 min (boxed), and total cell lysates were uPA gene expression through signaling pathwaysanalyzed by Western blotting for total andphosphorylated Erk levels. B, C, MCF7 cells were involving these kinases .pretreated for 45 min with 5 µM PKI166 (B), 10 µMUO126 (UO), 5 µM CGP077675 (CGP), 1 µM DISCUSSIONSB263580 (SB), 100 nM Wortmannin (W), 10 µMY27632 (Y27) or 20 µM SP600125 (SP) (C) and then Using the function-blocking antibody Decma, wetreated with EGF for 10 min (B) or Decma for 30 min(C). Total cell lysates were analyzed as above. showed previously that blockage of E-cadherin- mediated cell adhesion results in theinhibitor Y27634 affected neither Decma- upregulation of uPA activation in MCF7 andinduced Erk activation nor Shc phosphorylation T47D breast cancer cell lines (25). In this presentand its association with Grb2 (Fig. 5B). study, we investigated the underlying molecularInterestingly, Src inhibition also suppressed the mechanisms and showed that disruption of cell-disruption of cell-cell adhesion, the scattered cell adhesion induces Erk signaling downstreamphenotype of the cells and the redistribution of of E-cadherin. This Erk activation was Src- andE-cadherin into the cytoplasm (Fig. 5C). Shc-dependent and resulted in enhanced Role of PI3K in Decma-induced Erk expression of the uPA gene and, to a lesseractivation - The partial suppression of the extent, of the MMP-9 gene (data not shown).Decma-induced Erk phosphorylation by Disruption of cell-cell adhesion by calciumWortmannin suggests the involvement of PI3K chelation using EGTA has been reported toin this signaling (Fig. 4C). Both Wortmannin and increase Erk activity (13). Conversely, it wasLY29400, two structurally distinct PI3K shown that E-cadherin adhesion suppresses basalinhibitors, partially attenuated Erk Erk activity and concomitantly MMP-9phosphorylation but completely blocked Decma- expression (32). It may seem contradictory thatinduced PKB phosphorylation (Fig. 6A). Erk is also activated upon re-establishment ofInterestingly, the Src kinase inhibitor CGP77675 53
  • RESULTSFig 5. Involvement of Src in Decma-induced Erk activation. A, Activation of Src by Decma. T47D cellswere pretreated with 5 µM CGP077675 for 45 min (CGP) as indicated and then treated with Decma supernatant.Total cell lysates (400 µg protein) were immunoprecipitated with a polyclonal anti-Src antibody and thensubjected to Western blot analysis for Src and phospho-Src (Y416). To discriminate between Src and the heavychain antibody, the antibody was incubated with only protein A beads and lysis buffer (C1) or the cell lysate wasincubated only with protein A beads (C2). B, Effect of Src inhibitor on Shc and Erk phosphorylation. T47D cellswere pretreated for 45 min with 5 µM CGP077675 (CGP) or 10 µM Y27632 (Y27) and then treated with Decmasupernatant for 30 min. Total lysates (250 µg total protein) were immunoprecipitated with anti-Shc antibody andthen subjected to Western blot analysis (upper panel). In parallel, the total cell lysates (CL) were examined forErk and phospho-Erk levels by Western blotting (lower panel). C, T47D and MCF7 cells were grown for 2 daysto ca. 60% confluence. The cells were then treated for 45 min with 5 µM CGP077675 (CGP) (c, f, i, l) andsubsequently for 4 h with Decma supernatant (b-c, e-f, h-i, k-l) before recording (a-f) or before immunostainingwith the anti-E-cadherin antibody (g-l).cell-cell adhesion. However, the duration of Src has been implicated previously in thethis activation is much shorter and the control of cell adhesion. Inhibition of Srcunderlying molecular mechanisms of the two catalytic activity by overexpression ofsystems are different. While Erk activation by dominant inhibitory c-Src or by specificthe establishment of new cell-cell adhesion is inhibitors stabilizes E-cadherin-dependent cell-transient (5-60 min) and dependent on EGFR cell adhesion (33). Conversely, elevated Src(14) (22), Erk activation by the blockage of E- activity leads to disorganization of E-cadherin-cadherin was sustained (>3 h) and EGFR dependent cell-cell adhesion and cell scatteringindependent but Src- and Shc-dependent. (34). Fujita et al. (35) showed that E-cadherinInterestingly, RNAi-mediated downregulation and β-catenin become ubiquitylated by Hakaiof E-cadherin reduced Decma-induced Erk upon Src activation, ultimately leading toactivation (Fig. 2D) but did not elevate basal endocytosis of the E-cadherin complex.Erk phosphorylation. These results suggest that Accordingly, in the course of Decma-inducedit is not the absence per se of E-cadherin- disruption of cell-cell adhesion, Src activationdependent cell-cell interaction that induces the cell scattering and the redistributionsignaling pathway. 54
  • RESULTS Fig 7. Role of Src, PI3K and Erk in Decma- induced uPA upregulation. MCF7 cells were grown to 60-70% confluency, treated for 45 min with 10 µM UO126 (UO), 5 µM CGP077675 (CGP) or 100 nM Wortmannin (W), and then with Decma supernatant as indicated. Total RNA (10 µg) was subjected to Northern blot analysis (lower panel). The uPA mRNA levels were normalized against GAPDH and presented graphically (upper panel). cytoskeleton seems not to be necessary for Src activation since CytD failed to prevent Decma- induced Erk activation (Fig. 4A). One possible mechanism of Src activation is through the interaction with p120-catenin. p120-catenin is a Src substrate and has been shown to interact with Src kinase family members (36). ThisFig 6. Role of PI3K in Decma-induced Erk interaction is thought to keep Src kinases in anactivation. A, MCF7 cells were pretreated for 45min with 100 nM Wortmannin (W), 5 µM inactive state. Disruption of E-cadherin-LY294002 (LY), or 5 µM CGP077675 (CGP) and dependent cell-cell adhesion might change thethen treated with Decma supernatant as indicated. interaction between E-cadherin/p120/SrcTotal cell lysates were subjected to Western blot family members which could allow Srcanalysis for phospho-PKB (Ser473), phospho-Erk activation. Alternatively, Src activation couldand total Erk levels. B, MCF7 cells were treated be a result of a functional crosstalk between E-with 100 nM Wortmannin for 45 min and then with cadherin and integrins. IntercommunicationDecma supernatant as indicated. Total cell lysates between integrins and cadherins has been(400 µ g protein) were immunoprecipitated with observed several times (37,38) andanti-Shc antibody and then analyzed for phospho- Chattopadhyay et al. (39) recently reported aShc, total Shc and Grb2 by Western blotting. C,MCF7 cells were transfected with Shc-specific or complex containing α3β1-integrins and E-control siRNA as described in Materials and cadherin besides other proteins. In a indirectMethods. After 3 days, transfected cells were way loss of cell-cell adhesion might generatepretreated with 100 nM Wortmannin for 45 min and forces on focal adhesions which could producethen with Decma supernatant as indicated for 30 integrin-dependent signals. All thesemin. Levels of phospho-PKB (Ser473), Shc, possibilities are currently under investigation.phospho-Erk and Erk in total cell lysates were Preliminary experiments suggest a functionaldetermined by Western blotting. cross talk between E-cadherin and integrins, given that siRNA-induced knockdown of β1-was necessary for the initiation of the signaling integrin reduced Decma-induced Erkof E-cadherin into the cytoplasm. However, the phosphorylation (unpublished data).mechanism of Src activation by Decma Using siRNA against all isoforms of Shc, wetreatment has not been elucidated. The actin found this adaptor protein to be essential for Decma-induced Erk activation. Moreover, 55
  • RESULTSexpression of silent mutants of Shc isoforms contribute to Erk activation by acting at any ofshowed that only p46Shc and p52Shc rescued the these sites, except upstream of Shc.effect of the siRNA. Overexpression of p66Shc We disrupted cell-cell adhesion by physicalnot only failed to rescue Erk activation, but in means using the function-blocking antibodyfact had a negative effect comparable to the Decma in order to reproduce a processeffect of overexpressed dominant negative observed in some types of tumorigenesis.p52Shc3Y3F. These results support previous During the course of tumor progression, theobservations that p66Shc is a negative regulator ectodomain of E-cadherin can be detached byof EGF-induced Erk activation and c-fos matrilysin and stomilysin-1, releasing an 80-promoter activation (21). In agreement with kDA soluble E-cadherin fragment (sE-this, we showed recently that p66Shc is unable cad)(10). sE-cad has been found in urine andto rescue cytoskeleton reorganization-induced serum of cancer patients and correlates with aErk activation after siRNA mediated poor prognosis (46-48). In tissue culture, itknockdown of all isoforms of endogenous Shc induces scattering of epithelial cells (49),(28). inhibition of E-cadherin-dependent cellDecma-induced Shc tyrosine phosphorylation aggregation and invasion of cells into type Iand its binding to Grb2 were completely collagen (10). Furthermore, sE-cad stimulatesrepressed by pretreatment with a Src inhibitor, the upregulation of MMP-2, MMP-9 and MTI-suggesting that Src acts upstream of Shc in MMP expression in human lung tumor cells, asDecma-induced signaling. In accordance with reported by Noe and colleagues (10). Tothis observation, in vitro kinase assays have explain all these effects, the authors suggesteddemonstrated that Src is able to phosphorylate the presence of a signal transduction pathwayall three tyrosine residues of Shc proteins induced either directly by sE-cad or indirectlydirectly (40) and is responsible for Shc by the disruption of cell-cell contact (50). Itphosphorylation upon fibronectin (41) and may be argued that the signaling described inPDGF stimulation (42). FAK has been this report is a consequence of Decma acting asreported to form a complex with Shc and Grb2 a ligand for E-cadherin. However, several linesupon CytD treatment in LLC-PK1 cells (28) of evidence suggest it is the disruption of cell-and upon fibronectin stimulation in NIH3T3 cell adhesion that is attributable for Decma-fibroblasts (41). However, it is unlikely that induced signaling activation. First, a secondFAK plays a role in Decma-induced Erk antibody against the extracellular domain of E-activation. No interaction of Shc and FAK was cadherin which did not disrupt cell-celldetected upon Decma treatment and adhesion did not induce Erk activation.overexpression of dominant-negative FRNK Second, inhibition of Src blocked thefailed to abrogate Erk activation (data not disruption of cell-cell adhesion and at the sameshown). time prevented Erk activation. Third, DecmaWhile the Src/Shc/Erk pathway plays a major recognizes an eptiope located close to therole in Decma-induced Erk activation, Decma membrane proximal part of the extracellularalso induced the Src/PI3K/Erk pathway. domain of E-cadherin (51). Structural changesTreatment of MCF7 or T47D cells with in this membrane proximal region has beenWortmannin partially reduced Erk activation shown to change the adhesive properties ofwithout affecting Shc phosphorylation. RNAi- cells. Finally, Decma and sE-cad sharemediated knockdown of Shc reduced the basal common features that they disrupt cell-cellactivity of the PI3K pathway as measured by adhesion and induce signaling despite the factPKB phosphorylation. Nevertheless, Decma that they interact with E-cadherin in differenttreatment still enhanced PKB phosphorylation, manners. Therefore, it is most likely that it isindicating a Shc-independent pathway for the disruption of E-cadherin-mediated cell-cellDecma-induced PKB activation. Effects of adhesion that triggers a signal transductionWortmannin on Erk have been reported in pathway leading to Erk activation and uPAseveral cell systems: in T lymphocytes (43), gene expression.Cos 7 cells (44), and a CHO-derived cellline(45). However, the site at which the PI3K REFERENCESsignaling feeds into the Erk activating pathwayvaries in these systems: at the Ras, Raf or 1. Rabbani, S. A., and Xing, R. H. (1998) IntMEK activation level. In the Decma-induced J Oncol 12(4), 911-920pathway, the signal from PI3K could 56
  • RESULTS2. Takeichi, M. (1991) Science 251(5000), 20. Luzi, L., Confalonieri, S., Di Fiore, P. P., 1451-1455 and Pelicci, P. G. (2000) Curr Opin Genet3. Cavallaro, U., and Christofori, G. (2004) Dev 10(6), 668-674 Nat Rev Cancer 4(2), 118-132 21. Pellegrini, M., Pacini, S., and Baldari, C.4. Wijnhoven, B. P., Dinjens, W. N., and T. (2005) Apoptosis 10(1), 13-18 Pignatelli, M. (2000) Br J Surg 87(8), 992- 22. Pece, S., and Gutkind, J. S. (2000) J Biol 1005 Chem 275(52), 41227-412335. Chen, W. C., and Obrink, B. (1991) Cell- 23. Adams, C. L., and Nelson, W. J. (1998) cell contacts mediated by E-cadherin Curr Opin Cell Biol 10(5), 572-577 (uvomorulin) restrict invasive behavior of 24. Qian, X., Karpova, T., Sheppard, A. M., L-cells. In. J Cell Biol McNally, J., and Lowy, D. R. (2004)6. Chang, H. W., Chow, V., Lam, K. Y., Embo J 23(8), 1739-1784 Wei, W. I., and Yuen, A. (2002) Cancer 25. Frixen, U. H., and Nagamine, Y. (1993) 94(2), 386-392 Cancer Res 53(15), 3618-36237. Thiery, J. P. (2002) Nat Rev Cancer 2(6), 26. Kisielow, M., Kleiner, S., Nagasawa, M., 442-454 Faisal, A., and Nagamine, Y. (2002)8. Kawanishi, J., Kato, J., Sasaki, K., Fujii, Biochem J 363(Pt 1), 1-5 S., Watanabe, N., and Niitsu, Y. (1995) 27. Irigoyen, J. P., Besser, D., and Nagamine, Mol Cell Biol 15(3), 1175-1181 Y. (1997) J Biol Chem 272(3), 1904-19099. Lochter, A., Galosy, S., Muschler, J., 28. Faisal, A., Kleiner, S., and Nagamine, Y. Freedman, N., Werb, Z., and Bissell, M. J. (2004) J Biol Chem 279(5), 3202-3211 (1997) J Cell Biol 139(7), 1861-1872 29. Nagamine, Y., Medcalf, R. L., and Munoz-10. Noe, V., Fingleton, B., Jacobs, K., Canoves, P. (2005) Thromb Haemost Crawford, H. C., Vermeulen, S., Steelant, 93(4), 661-675 W., Bruyneel, E., Matrisian, L. M., and 30. Sampath, P., and Pollard, T. D. (1991) Mareel, M. (2001) J Cell Sci 114(Pt 1), Biochemistry 30(7), 1973-1980 111-118 31. Betson, M., Lozano, E., Zhang, J., and11. Polette, M., Gilles, C., de Bentzmann, S., Braga, V. M. (2002) J Biol Chem 277(40), Gruenert, D., Tournier, J. M., and 36962-36969 Birembaut, P. (1998) Clin Exp Metastasis 32. Laprise, P., Langlois, M. J., Boucher, M. 16(2), 105-112 J., Jobin, C., and Rivard, N. (2004) J Cell12. Kitadai, Y., Ellis, L. M., Tucker, S. L., Physiol 199(1), 32-39 Greene, G. F., Bucana, C. D., Cleary, K. 33. Owens, D. W., McLean, G. W., Wyke, A. R., Takahashi, Y., Tahara, E., and Fidler, I. W., Paraskeva, C., Parkinson, E. K., J. (1996) Am J Pathol 149(5), 1541-1551 Frame, M. C., and Brunton, V. G. (2000)13. Ara, T., Deyama, Y., Yoshimura, Y., Mol Biol Cell 11(1), 51-64 Higashino, F., Shindoh, M., Matsumoto, 34. Avizienyte, E., Wyke, A. W., Jones, R. J., A., and Fukuda, H. (2000) Cancer Lett McLean, G. W., Westhoff, M. A., 157(2), 115-121 Brunton, V. G., and Frame, M. C. (2002)14. Munshi, H. G., Ghosh, S., Mukhopadhyay, Nat Cell Biol 4(8), 632-638 S., Wu, Y. I., Sen, R., Green, K. J., and 35. Fujita, Y., Krause, G., Scheffner, M., Stack, M. S. (2002) J Biol Chem 277(41), Zechner, D., Leddy, H. E., Behrens, J., 38159-38167 Sommer, T., and Birchmeier, W. (2002)15. Nawrocki-Raby, B., Gilles, C., Polette, M., Nat Cell Biol 4(3), 222-231 Martinella-Catusse, C., Bonnet, N., 36. Reynolds, A. B., and Carnahan, R. H. Puchelle, E., Foidart, J. M., Van Roy, F., (2004) Semin Cell Dev Biol 15(6), 657-663 and Birembaut, P. (2003) Am J Pathol 37. Zhang, F., Tom, C. C., Kugler, M. C., 163(2), 653-661 Ching, T. T., Kreidberg, J. A., Wei, Y.,16. Irigoyen, J. P., Munoz-Canoves, P., and Chapman, H. A. (2003) J Cell Biol Montero, L., Koziczak, M., and Nagamine, 163(1), 177-188 Y. (1999) Cell Mol Life Sci 56(1-2), 104- 38. Yano, H., Mazaki, Y., Kurokawa, K., 132 Hanks, S. K., Matsuda, M., and Sabe, H.17. Liu, J. F., Crepin, M., Liu, J. M., (2004) J Cell Biol 166(2), 283-295 Barritault, D., and Ledoux, D. (2002) 39. Chattopadhyay, N., Wang, Z., Ashman, L. Biochem Biophys Res Commun 293(4), K., Brady-Kalnay, S. M., and Kreidberg, J. 1174-1182 A. (2003) J Cell Biol 163(6), 1351-136218. Ghosh, S., Munshi, H. G., Sen, R., Linz- 40. van der Geer, P., Wiley, S., Gish, G. D., McGillem, L. A., Goldman, R. D., Lorch, and Pawson, T. (1996) Curr Biol 6(11), J., Green, K. J., Jones, J. C., and Stack, M. 1435-1444 S. (2002) Cancer 95(12), 2524-2533 41. Schlaepfer, D. D., Jones, K. C., and19. Ravichandran, K. S. (2001) Oncogene Hunter, T. (1998) Mol Cell Biol 18(5), 20(44), 6322-6330 2571-2585 57
  • RESULTS42. Blake, R. A., Broome, M. A., Liu, X., Wu, C. H., Shenton, B. K., Neal, D. E., and J., Gishizky, M., Sun, L., and Courtneidge, Mellon, J. K. (1996) Br J Cancer 74(4), S. A. (2000) Mol Cell Biol 20(23), 9018- 579-584 9027 48. Banks, R. E., Porter, W. H., Whelan, P.,43. Karnitz, L. M., Burns, L. A., Sutor, S. L., Smith, P. H., and Selby, P. J. (1995) J Clin Blenis, J., and Abraham, R. T. (1995) Mol Pathol 48(2), 179-180 Cell Biol 15(6), 3049-3057 49. Wheelock, M. J., Buck, C. A., Bechtol, K.44. King, W. G., Mattaliano, M. D., Chan, T. B., and Damsky, C. H. (1987) J Cell O., Tsichlis, P. N., and Brugge, J. S. Biochem 34(3), 187-202 (1997) Mol Cell Biol 17(8), 4406-4418 50. Chunthapong, J., Seftor, E. A., Khalkhali-45. McIlroy, J., Chen, D., Wjasow, C., Ellis, Z., Seftor, R. E., Amir, S., Lubaroff, Michaeli, T., and Backer, J. M. (1997) Mol D. M., Heidger, P. M., Jr., and Hendrix, Cell Biol 17(1), 248-255 M. J. (2004) J Cell Biochem 91(4), 649-46. Katayama, M., Hirai, S., Kamihagi, K., 661 Nakagawa, K., Yasumoto, M., and Kato, I. 51. Ozawa, M., Hoschutzky, H., (1994) Br J Cancer 69(3), 580-585 Herrenknecht, K., and Kemler, R. (1990)47. Griffiths, T. R., Brotherick, I., Bishop, R. Mech Dev 33(1), 49-56 I., White, M. D., McKenna, D. M., Horne,Supplementary Fig. MCF7 cells were treated with 50 µg/ml of the indicated antibody for the indicated time.Total cell lysates were subjected to Western blot analysis for phospho-Erk and total Erk levels.11 We are grateful to François Lehembre and Gerhard Christofori (University Basel) for providing uswith MCF7 cell lines expressing siRNA against E-cadherin and NCAM. We thank Stéphane Thiry fortechnical assistance and Joshi Venugopal for stimulating discussions. Boris Bartholdy (HarvardInstitutes of Medicine) and Pat King are acknowledged for critical reading of the manuscript. FriedrichMiescher Institute is part of the Novartis Research Foundation. This work was partly supported bySwiss Cancer league. 58
  • RESULTS2.5 Supplementary data to 2.4 a role for FAK in coupling Src activation to the Erk signaling pathway in this system unlikely (Fig. 2.5.1B). The functionality of the construct2.5.1 Role of FAK in Decma-induced has been shown in another study from ourErk activation laboratory (Irigoyen and Nagamine, 1999). We have shown that Src activity is 2.5.2 Disruption of cell-cell adhesionnecessary for Decma-induced Erk activation. using EGTA in LLC-PK1 cellsThe Focal Adhesion Kinase (FAK) is not onlyone of the major interacting partners of Src, In the non-transformed pig epithelial cell linebut it can also activate Erk signaling through its LLC-PK1, the E-cadherin function-blockinginteraction with Shc and Grb2 (Schlaepfer et antibody Decma does not disrupt cell-cellal., 1999). We, therefore, asked whether FAK adhesion. However, E-cadherin-dependentis activated upon Decma treatment, and found cell-cell adhesion is calcium-dependent andthat its tyrosine phosphorylation is enhanced can therefore be disrupted using calcium-compared with the basal level (Fig. 2.5.1A). To chelating agents such as EGTA. As illustratedfurther investigate whether FAK participates in in Fig. 2.5.2A, almost all E-cadherin proteinDecma-induced Erk activation, the effect of that is normally localized at sites of cell-celloverexpressed FRNK, a dominant-negative adhesion was internalized into the cytoplasmFAK mutant (Richardson and Parsons, 1996), 30 min after EGTA treatment. Similar towas examined. FRNK expression had no effect Decma treatment in MCF7 and T47D cells,on Erk phosphorylation, making EGTA treatment caused Erk activation, Shc phosphorylation, and Shc association with Grb2 (Fig. 2.5.2B, C). Knockdown of Shc using siRNA prevented EGTA-induced Erk phosphorylation, suggesting that Erk activation is dependent on Shc. Moreover, expression of Shcsm Shcsm HA-p46 and HA-p52 , which excaped targeting of the siRNA (knockdown-in), rescued sustained Erk phosphorylationFigure 2.5.1: Role of FAK in Decma-induced induced by EGTA (Fig. 2.5.2D). In conclusion,Erk activation. T47D cells were transiently EGTA disrupts cell-cell adhesion and inducestransfected with pcDNA, HA-FAK, or HA-FRNK Shc-dependent sustained Erk activation in aand then treated with Decma supernatant. manner, which is reminiscent of the signalingTotal cell lysates were immunoprecipitated induced by disruption of cell-cell adhesionwith HA antibody (A) or directly subjected to using Decma in MCF7 and T47D cells.Western blot analysis (B). 59
  • RESULTSFigure 2.5.2: A. Effect of EGTA on LLC-PK1 cells. LLC-PK1 cells were treated for 30 min with 4 mMEGTA and 1 mM MgCl2, and then immunostained with anti-E-cadherin antibody. B. LLC-PK1 cellswere treated with 4 mM EGTA/1 mM MgCl2 or 50 ng/ml EGF for the indicated time, and total celllysate was subjected to Western blot analysis. C. After treatment of cells with 4 mM EGTA/1 mMMgCl2 for 15 min, 300 µg of total cell lysates were immunoprecipitated with anti-Shc antibody and Shcsmsubjected to Western blot analysis. D. Cell lines stably expressing empty vector or HA-p46/52were transfected with control or all-shc siRNA. Three days later, cells were treated with 4 mM EGTA/1mM MgCl2 for the indicated time and total cell lysate was analyzed by Western blotting. 60
  • RESULTS2.6 Role of p66Shc in regulating MCF7 cells (Fig. 2.3-2) were analyzed. Starvation, UV and H2O2 treatment of MCF7cell survival in epithelial cells cells resulted in loss of cell viability as measured by trypan blue exclusion. However, Shc p66 is implicated in the regulation of cell no p66 Shc -dependent change in cell viabilitysurvival in response to oxidative stress, and was observed after any of these treatmentsthe absence of p66Shc renders cells more (Fig 2.6-1). The same experiments wereresistant to UV-, H2O2-, or starvation-induced repeated with LLC-PK1 epithelial cells.cell death ( (Migliaccio et al., 1999; Similarly to MCF7 cells, wild-type LLC-PK1Nemoto and Finkel, 2002). This function of cells do not express p66Shc. Again, p66Shc- Shcp66 has been attributed to the overexpressing cell lines were generated.phosphorylation of the serine residue S36. Because the effect of p66Shc on stress-inducedStill, the precise mechanism underlying this cell death was reported to be dependent onrole is not fully understood. In addition, it is not serine 36, cells overexpressing p66 Shc carryingclear whether this is only restricted to certain a serine to alanine mutation (S36A) at this sitecell types, such as fibroblasts, endothelial were made. Since p66 Shc contains ancells, and T-cells, or whether this represents a additional threonine phosphorylation site withgeneral mechanism. To investigate the role of as yet unknown function, a mutation was Shcp66 in regulating the survival of epithelial introduced at this site too (T29A). Finally, cell Shccells, p66 -expressing and non-expressing lines overexpressing p66Shc containing serine ShcFigure 2.6-1: Effect of p66 overexpressionon stress-induced cell death in MCF7 cells.Stable MCF7 cells expressing empty vector, HA- Shc Shc Shcp46 , HA-p52 , or HA-p66 were treated 2with 100 J/m UV or 1 mM H2O2, or were starvedfor 4 days. Cell viability was measured by trypanblue exclusion using the Vi-CELL analyzer at theindicated time point. 61
  • RESULTS36/threonine 29 double mutation (TSA) were LLC-PK1 cells were more resistant to allgenerated and, as control, cell clones stress-inducing agents, and higher doses were Shcexpressing p66 mutated at serine 138 necessary to induce cell death. Again, Shc(S138A), which is implicated in PTP-PEST overexpression of p66 did not render thesebinding (Fig 2.6-2). cells more sensitive to UV- or H2O2-induced cell death (Fig 2.6-3A). In addition, no changes in cell viability were observed when either of Shc the p66 phosphorylation mutants was overexpressed in these cells (Fig 2.6-3B). Various different experimental conditionsFigure 2.6-2: LLC-PK1 cells stably (dosage of stress, duration after stress,overexpressing p66Shc. LLC-PK1 cells were starvation) were used, but we were unable tostably transfected to overexpress wild-type HA- see p66 Shc -dependent effects on cell viability in Shc Shcp66 and HA-p66 threonine or serine to neither MCF7 nor LLC-PK1 cells, suggestingalanine mutants. The depicted clones were that p66Shc might not be involved in regulatingused for further experiments. cell survival upon stress in epithelial cells. ShcFigure 2.6-3: Effect of p66 overexpression on stress-induced cell death in LLC-PK1 cells. A.Two clones of stable LLC-PK1 cells expressing empty vector or HA-p66Shc were treated with 200 J/m2 ShcUV or 1 mM H2O2. B. Stable LLC-PK1 cells expressing empty vector, HA-p66 wild-type, HA- ShcTSA ShcT29A ShcS36A ShcS138A 2p66 , HA-p66 , HA-p66 , or HA-p66 were treated with 200 J/m UV or 1.5 mMH2O2. Cell viability was measured by trypan blue exclusion using the Vi-CELL analyzer at the indicatedtimepoints. 62
  • DISCUSSION 3. Discussion3.1 Isoform-specific knockdown and revealed siRNA-specific, rather than target- specific, signatures including silencing of non-knockdown-in of Shc using siRNA targeted genes (Jackson et al., 2003). Another We have generated a system which allows study has shown that transfection of siRNA atdownregulation and expression of single Shc concentrations of 100 nM induced a significantisoforms in a short time period in tissue number of genes, many of which are known toculture. This system should be applicable and be involved in apoptosis and stress responsefunctional for the study of almost all genes (Semizarov et al., 2003). Reduction of thewhich are expressed as multiple isoforms. siRNA concentration to 20 nM eliminated thisMoreover, the knockdown-in system can be nonspecific response. Taking our observationuseful for validating specificity or mutational and these reports into account, a number ofanalysis in follow-up experiments in which the different controls for siRNA experiments aretarget gene is restored by vector-based recommended. In line with this, a recentexpression of a wild-type or mutated form of editorial in Nature Cell Biology (2003)the gene. In our studies, coding sequences published a list of control experiments requiredwere targeted by siRNA. Therefore, the for siRNA experiments. These include: (i)expression of siRNA-resistant proteins minimization of siRNA concentration, (ii) use ofrequired point mutations in the region scrambled siRNA control, (iii) use of multiplecorresponding to the siRNA. Alternatively, 3’ or siRNAs for a single target, (iv) examination of5’ UTR regions could be targeted downregulation of mRNA and protein levels,( and (v) rescue experiments as ultimate and0302.html) (Elbashir et al., 2002; Tsuda et al., best control. If a careful design of siRNAs,2005). Targeting UTR sequences facilitates simlar to those published by Elbashir et al.knockdown-in experiments because under (Elbashir et al., 2002) and Semizarov et al.these conditions it is unnecessary to introduce (Semizarov et al., 2003), is combined with thesilent mutations in the vector-encoded genes. suggested controls, siRNAs are a highlyHowever, since UTR sequences are known specific tool for targeted gene knockdown.sites of mRNA-binding proteins, some In contrast to fungi, plants, and worms,researches prefer to avoid these regions which can replicate siRNA, there is no(Dykxhoorn et al., 2003; Elbashir et al., 2002). indication of siRNA amplification in mammals Although siRNAs are thought to act in a very (Dykxhoorn et al., 2003). Likewise, we andspecific manner, we observed side effects others have shown the transient nature ofwhen siRNAs were used at high siRNA-directed silencing by transfection intoconcentrations (100 nM). Nonspecific events mammalian cells (Dykxhoorn et al., 2003). Towere also noticed by other laboratories when prolong siRNA-induced knockdown, wesiRNA concentrations of 100 nM or higher suggested repeated transfection of siRNAswere used in mammalian cells. A comparison into cells. Later, several reports appearedof expression profiles resulting from silencing which described vector-based siRNAof the same target gene by different siRNAs expression systems (Amarzguioui et al., 2005; 63
  • DISCUSSIONBrummelkamp et al., 2002; Miyagishi and silencing were not fully understood. In thisTaira, 2002; Yu et al., 2002). These allow regard, our report not only offered a usefulstable expression of siRNA and can be application of siRNA for the investigation ofdesigned in an inducible manner. Today, proteins expressed in multiple isoforms, butseveral strategies have been developed for also gave some insights into the mechanismsRNAi in mammalian cells: they are of RNAi. These are discussed in section 2.1.summarized in Fig 3.1 (Kim, 2005). Today, the mechanisms underlying RNAi have The study describing the isoform-specific been well investigated (section 1.3). The useknockdown and the knockdown-in was of RNAi has been further developed and it haspublished at the beginning of 2002, when the become an effective, widely used method formechanisms underlying siRNA-directed the analysis of gene function.Figure 3.1: Various strategies for RNAi in mammalian cells. A. Long dsRNAs can induce specificRNAi in oocytes, early embryos, and undifferentiated embryonic stem cells. B. Chemically synthesizedsiRNA duplexes can be efficiently transfected into cultured cells. C. siRNA can be prepared in vitrofrom dsRNAs by incubating with recombinant Dicer protein. The diced products are purified based ontheir size (~21 nt) and transfected into cells. D. Short hairpin RNAs (shRNAs) are expressed in thenucleus from expression plasmids: the RNA polymerase (pol) III-derived expression system is shownhere as an example. Upon export by Exportin 5 (Exp5), shRNAs are processed by Dicer-releasingsiRNAs. E. ShRNA expression cassette can be delivered by viral vectors such as retroviral vector,lentiviral vector, and adenoviral vector (taken from (Dykxhoorn et al., 2003)). 64
  • DISCUSSION3.2 Role of Shc in mediating activation which would still lead to the activation of the Raf/Erk pathway but wouldErk activation reduce Ras-mediated activation of PI3K. However, it is not clear whether this would3.2.1 Shc is dispensable for EGF- have any physiological role, because PI3K is directly activated by EGFR. In a differentinduced Erk activation cellular system, Hashimoto et al. (Hashimoto et al., 1999) have shown that Shc is not Shc has been found as an adaptor protein necessary for Ras activation by EGFR, but isthat recruits the Grb2/SOS complex to the important for JNK activation. It remains to bemembrane in order to activate Ras, thereby investigated whether JNK activation iscoupling activated RTKs, such as EGFR, to changed upon Shc knockdown in epithelialErk activation (Ravichandran, 2001). Through Shc cells. p66 has been proposed as a negativethis action, Shc is involved in regulating the regulator of EGF-induced Erk activation.proliferation of mammalian cells (Pelicci et al., Shc Isoform-specific knockdown of p66 did not1992). In an attempt to study the isoform- change EGF-induced Erk activation in HeLa,specific contribution of Shc proteins to EGF- PNT2, or PC3 epithelial cells. Conversely, noinduced signaling, we applied the siRNA- change in EGF-mediated cell viability ordirected knockdown technique to several Shc proliferation was observed when p66 wasepithelial cell lines. Isoform-specific overexpressed. However, the experimentsknockdown and knockdown of all Shc proteins were done using 50 ng/ml EGF. Furtherdid not influence EGF-induced Erk activation Shc experiments will show whether p66or response. Therefore, the results argue that influences these parameters at low EGFEGF signaling is not dependent on Shc concentrations.proteins in epithelial cells. Interestingly, Shc Ignoring these cautions for a moment, theknockdown had also no effect on NGF-induced data represented here seem contradictory toErk activation in PC12 cells (data not shown). the current established role of Shc proteins inHowever, further experiments are required to cell signaling. However, reports whichinvestigate whether the absence of Shc has demonstrate that Shc is essential in mediatingmore subtle effects on EGF signaling. Shc acts Erk activation downstream of EGFR are basedupstream of Ras in EGFR signaling. We on studies in which dominant negative Shclimited our investigation on Erk mutants were overexpressed (Gotoh et al.,phosphorylation to the classical output of Ras 1995; Gotoh et al., 1997). It must be noted thatactivation. However, there are several lines of the binding of Grb2 to EGFR, similar to most ofevidence suggesting that PI3K is another Ras- the RTK, is redundant. Grb2 can be associatedeffector molecule (Rommel and Hafen, 1998). with the receptor either directly via its SH2Whereas stimulation of the Ras/Raf/Erk domain or indirectly via binding to EGFR-pathway is seen at very low concentrations of associated Shc and, most likely, Shp1 (Srcgrowth factors, PI3K activation is only homology phosphatase) (Chen et al., 1996;observed at higher concentrations (Pawson Hynes and Lane, 2005; Minoo et al., 2004)and Saxton, 1999). Therefore, it can be argued (Fig 3.2.1-1). Domain-mediated proteinthat Shc knockdown reduces the extent of Ras 65
  • DISCUSSIONinteractions are dependent on both the affinityand the relative concentration of the bindingpartner: high affinity interaction would befavored at low concentrations of the targetmolecule, but could be displaced by low affinityinteractions driven by high concentrations ofaffinity partners. Therefore, it is possible thatoverexpression of dominant negative Shcprevents the association of the receptor notonly with endogenous Shc, but also with otherproteins, such as Grb2, resulting in theinhibition of Erk activation. In contrast, the Figure 3.2.1-1: Schematic representation ofremoval of Shc by siRNA would not interfere the main autophosphorylation sites inwith the association of Grb2 or Shp1 with the EGFR and of signaling moleculesreceptor. Redundancy of adaptor proteins was associated with these sites. Each receptoralso found in Drosophila. DShc, DRK (Grb2 in chain becomes phosphorylated on multiplemammals) and DOS (daughter of sevenless) sites and binds specific SH2-containinghave been shown to act in parallel to proteins. The receptor has redundanttransduce signals from the RTK torso interactions, for example with Grb2. This figure(Luschnig et al., 2000). Therefore, loss-of- is illustrative and not comprehensive: somefunction mutations in dshc affected RTK binding partners and pathways, including PI3K,function only partially. Moreover, dShc seems Src, and STATS, are not act in signaling of only a subset of RTKs,indicating that dShc confers specificity to As already mentioned above, proteinreceptor signaling. In mammals, Shc is interactions are influenced by their bindingphosphorylated downstream of all RTKs affinity to each other and by their relativeknown to date (Luzi et al., 2000). However, the concentrations. It is also important tophysiological relevance of Shc phosphorylation recognize that the “local” abundance of amight depend on the repertoire of other protein may determine its availability foradaptor proteins expressed at the same time. binding: co-localization will favor interactionThe relative contribution of Shc to signaling of even when affinity is low. Protein associationeach of the receptors should be addressed patterns can vary within the same cellusing siRNA-mediated knockdown, and not depending on the stimulus and timing of theoverexpression dominant negative Shc. The stimulation. Preliminary data suggest that Shcexistence of interactions between Shc and proteins might play a role in stress-inducedGrb2 with EGFR, with each other, and with a activation of MAPKs. In LLC-PK1 cells,subset of cellular proteins raises the question knockdown of Shc did not alter Erk activationof how interactions are controlled. Do all induced by EGF and TPA (Fig 3.2.1-2 lanes 1-possible interactions occur in a single cell and, 6). However, Erk activation was slightlyif so, does activation of one pathway influence reduced after H2O2 and UV treatment in theactivation of alternative pathways? absence of Shc (Fig 3.2.1-2 lanes 7-10). 66
  • DISCUSSIONInterestingly, only expression of p66Shc mechanisms of Shc should target these(knockdown-in) enhanced the reduction of tissues.H2O2- and UV-induced Erk activation seenbefore (Fig 3.2.1-2 lanes 17-20). Similar 3.2.2 Shc mediates Erk activationeffects were seen for JNK activation (Fig 3.2.1- downstream of E-cadherin2 lane 13-18). These results indicate that Erkand JNK activation stimulated by stress, such Loss of Shc affects fibrocectin-induced Erkas UV and H2O2, are mediated, at least activity, focal complex distribution, the actinpartially, by Shc proteins and that p66Shc is cytoskeleton, and cell-cell contacts (Lai andable to modulate this MAPK activation. Pawson, 2000). In addition, Wary et al (Wary Interactions between various proteins are et al., 1996) demonstrated that Shc isalso controlled in a tissue-dependent manner. necessary for integrin-induced Erk activation,Different cell types may favor different protein and we have shown that Shc plays anassociations. Shc is most highly expressed in essential role in CSR-induced Erk activationadipocytes, smooth muscle cells, and cardiac (Faisal et al., 2004). Taken together, thesemyocytes. The Shc knockout embryo died results suggest that Shc might be important forbecause of defects in heart development and Erk activation mediated by non-receptorthe cardiovascular system. Therefore, Shc is tyrosine kinases. We have recently shown thatexpected to play a more important role in these uPA secretion is induced upon disruption oftissues, and future experiments to deepen ourunderstanding of the underlying molecular ShcFigure 3.2.1-2: Effect of Shc knockdown and p66 overexpression on MAPK activation. LLC- ShcPK1 cells expressing either empty vector or p66 were transfected with buffer (B) or si-shc1 (S) andthree days later treated with 50 ng/ml EGF, 100 nM TPA, 1 mM H2O2, or 200 J/m2 UV. Total cell lysatewas analyzed by Western blotting. Boxes indicate changes in the phosphorylation level of Erk or JNKin dependence on Shc proteins. 67
  • DISCUSSION cell-cell adhesion (Frixen and Nagamine, E-cadherin-dependent cell-cell adhesion to the1993). Knowing that uPA gene expression is induction of the uPA gene, two events whichoften regulated through stimulation of the are strongly involved in metastasis. UponMAPK pathway (Besser et al., 1995a; Besser disruption of cell-cell adhesion through theet al., 1995b; Irigoyen and Nagamine, 1999) function-blocking antibody Decma, Srcthe question was raised of whether disruption becomes activated and induces Shc/Grb2of cell-cell adhesion induces MAPK activation, association leading to the activation of Erk andand whether Shc would play a role in finally to uPA gene expression. On the othermediating this MAPK activation. hand, Src mediates PI3K and PKB activation. In section 2.3 we describe and discuss the PI3K activation contributes partially to Erkfact that disruption of E-cadherin-dependent activation.cell-cell adhesion does indeed stimulate apreviously unknown signaling pathway (Fig3.4.2). This pathway directly links disruption ofFigure 3.2.2: Schematic representation of the signaling pathway induced by disruption of E-cadherin-dependent cell-cell adhesion. Perturbation of E-cadherin-dependent cell-cell adhesion Shcleads to Src activation, followed by p46/p52 phosphorylation, which mediates Erk activation throughRas/Raf/MEK1. This results in enhanced uPA gene expression. Active Src also mediates PI3K/PKBactivation. PI3K signaling contributes to Erk activation. 68
  • DISCUSSION Proteolytic cleavage of E-cadherin by MMPs, induced Erk activation is dependent on E-such as ADAM10, has been shown to induce cadherin endocytosis. However, CytDtranslocation of β-catenin into the nucleus and treatment, which is believed to block E-increase the expression of the β-catenin cadherin internalization through disruption ofdownstream gene cyclin D1 (Maretzky et al., the actin cytoskeleton, did not inhibit Erk2005). uPA has also been described as a activity, arguing against this possibility (sectiontarget of β-catenin (Hiendlmeyer et al., 2004). 2.3 Fig 4). To find a conclusive answer,Therefore, it would be interesting to see experiments using specific pharmacologicalwhether Decma-induced uPA gene expression inhibitors of endocytosis have to be partially mediated by β-catenin. It is worth In this study, we clearly showed that Decma-noting that the same amount of β-catenin co- induced Erk activation can be mediated byimmunoprecipitated with E-cadherin at least 30 p46/p52Shc but not by p66Shc. Moreover, Shcmin after Decma treatment (data not shown). overexpression of p66 led to a dominant ShcSimilarly, using immunostaining we could not negative pattern. The reason for p66 being unable to rescue Decma-induced Erkdetect nuclear β-catenin localization upon activation is not understood. In a differentDecma treatment (data not shown). However, study, we have shown that p46/p52Shc, but notwe did not perform a promoter activation assay p66Shc, rescued CSR-induced Erk activation into see whether TCF/LEF-dependent LLC-PK1 cells (Faisal et al., 2004).transcription becomes activated upon Shc Investigation of the p66 phosphorylationdisruption of cell-cell adhesion through Decma. Shc status revealed that p66 was not Disruption of cell-cell adhesion using EGTA phosphorylated at its tyrosine residues uponalso induced Shc-dependent Erk activation in CSR. On the other hand, overexpression ofLLC-PK1 cells (section 2.5.2 Fig. 2.5.2), Shc p66 had no dominant negative effect onindicating that the signaling pathway described CSR-induced Erk activation. We have notearlier is not restricted to Decma-induced investigated yet whether p66Shc becomesdisruption of cell-cell adhesion. Interestingly, a tyrosine-phosphorylated upon disruption ofrecent report shows that disruption of cell-cell cell-cell adhesion. The fact that itsadhesion using EGTA leads to the activation of overexpression interfered with Decma-inducedthe small GTPase Rap1, a crucial regulator of Erk activation argues for an active role ofinside-out activation of integrins (Balzac et al., p66Shc in this signaling. A mechanism where2005). The authors also observed an increase Shc p66 competes with the other isoforms forin Src activity, which was required for Rap1 Grb2 binding, or where p66Shc binds toactivation. Further investigation revealed that RasGAP to downregulate Ras activation asE-cadherin endocytosis is necessary for Rap1 described in section (Fig, canactivation. Src is a major player in regulating be envisioned. More experiments will beendocytosis of E-cadherin (Frame, 2002; Fujita needed to test this hypothesis. However, theet al., 2002; Palovuori et al., 2003). Treatment Shc action of p66 seems to be cell type- and/orof the Src inhibitor prevented disruption of cell- stimulus-dependent.cell adhesion and internalization of E-cadherinupon treatment with Decma. Therefore, itwould be interesting to see whether Decma- 69
  • DISCUSSION Taken together, we found that Shc is not experimental setup was changed, no Shcessential for growth-factor-induced Erk conditions were found in which p66activation. In contrast, we found a novel expression decreased the viability of these cellpathway downstream of E-cadherin in which lines upon stress induction. The functionality of Shcp46/p52 are required to mediate Erk the HA-p66shc constructs have been provenactivation. In context with other studies before in other studies (Faisal et al., 2002;showing that the absence of Shc impairs focal Faisal et al., 2004).complex distribution, the actin cytoskeleton, The simplest explanation for theseand cell-cell contacts (Lai and Pawson, 2000), contradictory results would be that we did notas well as that Shc is essential for CSR- and find the right experimental conditions, or thatintegrin-induced Erk activation (Faisal et al., we chose the wrong cell lines. LLC-PK1 is a2004; Wary et al., 1996), we propose that Shc non-transformed proximal tubular epithelial cellproteins play a more important role in line which has been used to study apoptosis inmediating growth factor-independent Erk response to oxidative stress (Al-Ghamdi et al.,activation which involves the action of soluble 2004; Allen et al., 2003; Rustom et al., 2003).tyrosine kinases. This hypothesis is also However, LLC-PK1 cells turned out to besupported by in vivo studies demonstrating that highly resistant to any stress-inducing agents.Shc is essential for signaling downstream of MCF7 cells are breast cancer cells. It is notthe pre-TCR that uses non-receptor tyrosine clear whether the apoptotic response inkinases to phosphorylate Shc proteins (Pacini transformed cells is regulated by al., 1998; Zhang et al., 2002). However, both cell lines express p53, which becomes stabilized upon treatment with UV or H2O2 (data not shown).3.3 The role of p66Shc in stress These results could also imply that, inresponse epithelial cells, p66Shc is not involved in the regulation of cell viability in response to Shc The absence of p66 confers resistance to oxidative stress. To date there has been nooxidative stress in mouse embryo fibroblasts, report demonstrating a p66Shc-dependentendothelial cells, and T-cells (section regulation of apoptosis in epithelial cells. OtherWe wanted to use two epithelial cell lines to factors specific for epithelial cells could take Shcstudy p66 function and the role of its serine over the role of p66Shc in this regard. Oneand threonine phosphorylation sites in possible candidate is the newly identifiedmediating the oxidative stress response. It was protein REDD1 which is a downstream targetexpected that sensibility to oxidative stress- of p63, a p53 family member (Ellisen et al.,induced apoptosis would be enhanced upon 2002). REDD1 was identified as a hypoxia- Shc Shcintroduction of p66 in non-p66 expressing inducible gene involved in the regulation ofcells, similar to what has been reported for cellular ROS (Shoshani et al., 2002). Similar to Shc Shcp66 -deficient MEFs (Migliaccio et al., 1999). p66 , overexpression of REDD1 desensitizes ShcSurprisingly, p66 expression did not change cells to apoptotic stimuli, and loss of REDD1the viability of these cells upon treatment with expression results in reduced intracellular ROSvarious stress-inducing agents. Although the levels and enhanced resistance to oxidative 70
  • DISCUSSIONstress-induced apoptosis (Ellisen et al., 2002; more internally located, display a remarkablySchwarzer et al., 2005). Remarkably, REDD1 different subcellular localization. These Shcis involved in epithelial development and is findings indicate that p46 may exert a non-specifically expressed in tissues derived from redundant biological function in signalthe ectoderm (Ellisen et al., 2002). Therefore, it transduction pathways involving mitochondria. Shcis possible that REDD1 represents an Another group has reported that p46 isepithelial-specific factor involved in the specifically and heavily phosphorylated inregulation of ROS and cell viability. However, proliferating hepatocytes and cancer cells Shcto exclude a role for p66 in this process, derived from liver, and localizes in both themore epithelial cell lines should be examined. nuclei and the cytoplasm of these cells (Yoshida et al., 2004; Yuji et al., 2004). The3.4 Isoform-specific role of authors suggest that p46Shc localization in the nuclei may be closely related top46Shc hepatocarcinogenesis and represents a useful marker for the detection of hepatocytes with Beside the isoform-specific role of p66Shc in high proliferative activity. The same laboratoryregulating apoptosis, it remains unclear reported p46Shc nuclear localization in gastric Shc Shcwhether p46 and p52 display different normal mucosa and cancer, also suggesting afunctions. Both seem to be equally involved in role for p46 Shc in gastric carcinogenesismediating MAPK activation, as shown here (Yukimasa et al., 2005). However, the questionand in other reports (Faisal et al., 2004; of what is the function of p46 Shc in the nucleusRavichandran, 2001). On the other hand, there remains unanswered.are some indications for variation in their In contrast, Murayama et al. (Murayama et Shcfunction. p52 can be phosphorylated on al., 2004) demonstrated that antibody ligationserine 29 and it has been shown that this of CD9 induced apoptosis of gastrointestinalphosphorylation is necessary for its binding to cancer cell lines in a p46Shc-dependentthe phosphatase PTB-PEST in order to manner.downregulate insulin-induced Erk activation Taken together, these results indicate that in(Faisal et al., 2002). An investigation of the some tissues p46 Shc seems to play a moresubcellular localization of Shc isoforms important role than p52 Shc . Specific conditionalrevealed a specific and selective localization of knockout studies would help to further ourp46Shc to the mitochondrial matrix (Ventura et understanding of the physiological role of Shcal., 2004). Deletion mapping experiments have isoforms.demonstrated that targeting of p46Shc tomitochondria is mediated by its first 32 amino 3.5 Conclusionacids, which behave as a bona fidemitochondrial targeting sequence. Further, it This thesis introduces a tool for the analysishas been shown that the N-terminal location of of single Shc isoforms in tissue culture. Usingthe signal peptide is critical for its function. This Shc this tool, we found that p46/p52 , but notaccounts for the observation that p52Shc and Shc Shc p66 , play an essential role in a previouslyp66 , containing the same sequence but unknown signaling pathway downstream of E- 71
  • DISCUSSIONcadherin. This pathway directly links disruption the MAPK pathway and the individualof E-cadherin-dependent cell-cell adhesion to contribution of various adaptor proteins is notexpression of the uPA gene and might known. Data provided in this thesis suggesttherefore play a role in tumor progression. Shc that p46/p52Shc are exchangeable for growthproteins play an important role in mediating factor-induced MAPK activation but play aMAPK activation. However, receptor tyrosine more important role in growth factor-kinases have redundant ways of coupling to independent MAPK activation. 72
  • MATERIAL AND METHODS 4. MATERIAL AND METHODS The following chapter adds additional time 10 µl of a 5 mg/ml stock solution MTT (3-material and methods used in chapter 2 which (4,-dimethylthiazol-2-yl)-2,5-diphenyl-have not been described before. tetrazolium) (Boehringer Mannheim) was added to each well and incubated for an Cell lines. The renal proximal tubular cell additional 4 h. The MTT is converted to a colorline LLC-PK1 was cultured in Dulbeccos crystal product by mitochondrial enzymes,modified Eagles medium (Invitrogen), which are then dissolved by adding 100 µl of asupplemented with 10% (v/v) fetal calf serum 0.01 M HCl solution containing 10% SDS for 8(AMIMED, Allschwil, Switzerland), 0.2 mg/ml h. Color development was measuredstreptomycin, and 50 units/ml penicillin. For spectrophotometrically on an ELISA reader atserum starvation, cells were incubated in a wavelength of 590 nm. Proliferation of controlDulbeccos modified Eagles medium cells was set 100%.containing 0.1% fetal calf serum. The humanprostate epithelial cell line PNT2, and the p46/52Shc siRNA. The following 21-merhuman prostate adenomcarcinoma cell line oligoribonucleotide pair was used as si-PC3, were cultured in RPMI medium p46/p52Shc: 5’-GUG CGG AGA CUC CAUsupplemented with 10% (v/v) fetal calf serum, GAG GCC-3’ and 5’-CCU CAU GGA GUC0.2 mg/ml streptomycin, and 50 units/ml UCC GCA CGC-3’. The specificities of thesepenicillin. All cells were grown at 37° in a C sequences were confirmed by blasting againsthumidified incubator with 5% CO2. For serum the GenBank/EMBL database.starvation, cells were incubated in theirmedium containing 0.1% fetal calf serum. Viability assay. 1.5x106 cells were seeded in 6 cm plates. The next day, first samples EGTA treatment. LLC-PK1 cells were were collected (day 0) and the others werestarved for 16 h and then treated with 4 mM starved or treated with EGF, H2O2, or UVEGTA and 1 mM MgCl2 for an additional 30 according to the experiment. At the indicatedmin if not indicated differently. timepoint, cells were trypsinized and diluted in 1 ml of medium, and cell number as well as MTT proliferation assay. HeLa, PNT2, viability was measured in the Vi-CELL analyzerPC3, and LLC-PK1 cells were transfected with using the trypan blue exclusion. All samplessiRNA. The next day, 100-500 cells/well were done in duplets.(depending on the cell line) were seeded intriplicate in 96-well plates. The cells weregrown in 100 µl medium for 3 days, at which 73
  • REFERENCES 5. REFERENCES(2003). Whither RNAi? Nat Cell Biol 5, 489-490. Besser, D., Presta, M., and Nagamine, Y. (1995a). Elucidation of a signaling pathway induced byAl-Ghamdi, S. S., Chatterjee, P. K., Raftery, M. J., FGF-2 leading to uPA gene expression in NIH Thiemermann, C., and Yaqoob, M. M. (2004). 3T3 fibroblasts. Cell Growth Differ 6, 1009- Role of cytochrome P4502E1 activation in 1017. proximal tubular cell injury induced by hydrogen peroxide. Ren Fail 26, 103-110. Besser, D., Urich, M., Sakaue, M., Messerschmitt, A., Ballmer-Hofer, K., and Nagamine, Y.Allen, D. A., Harwood, S., Varagunam, M., (1995b). Urokinase-type plasminogen activator Raftery, M. J., and Yaqoob, M. M. (2003). High gene regulation by polyomavirus middle-T glucose-induced oxidative stress causes apoptosis antigen. Oncogene 11, 2383-2391. in proximal tubular epithelial cells and is mediated by multiple caspases. Faseb J 17, 908- Betson, M., Lozano, E., Zhang, J., and Braga, V. 910. M. (2002). Rac activation upon cell-cell contact formation is dependent on signaling from theAmarzguioui, M., Rossi, J. J., and Kim, D. (2005). epidermal growth factor receptor. J Biol Chem Approaches for chemically synthesized siRNA 277, 36962-36969. and vector-mediated RNAi. FEBS Lett. Birchmeier, W. (2005). Cell adhesion and signalAnastasiadis, P. Z., Moon, S. Y., Thoreson, M. A., transduction in cancer. Conference on cadherins, Mariner, D. J., Crawford, H. C., Zheng, Y., and catenins and cancer. EMBO Rep 6, 413-417. Reynolds, A. B. (2000). Inhibition of RhoA by p120 catenin. Nat Cell Biol 2, 637-644. Birchmeier, W., and Behrens, J. (1994). Cadherin expression in carcinomas: role in the formation ofArvidsson, A. K., Rupp, E., Nanberg, E., cell junctions and the prevention of invasiveness. Downward, J., Ronnstrand, L., Wennstrom, S., Biochim Biophys Acta 1198, 11-26. Schlessinger, J., Heldin, C. H., and Claesson- Welsh, L. (1994). Tyr-716 in the platelet-derived Blaikie, P., Immanuel, D., Wu, J., Li, N., Yajnik, growth factor beta-receptor kinase insert is V., and Margolis, B. (1994). A region in Shc involved in GRB2 binding and Ras activation. distinct from the SH2 domain can bind tyrosine- Mol Cell Biol 14, 6715-6726. phosphorylated growth factor receptors. J Biol Chem 269, 32031-32034.Balzac, F., Avolio, M., Degani, S., Kaverina, I., Torti, M., Silengo, L., Small, J. V., and Retta, S. Blaikie, P. A., Fournier, E., Dilworth, S. M., F. (2005). E-cadherin endocytosis regulates the Birnbaum, D., Borg, J. P., and Margolis, B. activity of Rap1: a traffic light GTPase at the (1997). The role of the Shc phosphotyrosine crossroads between cadherin and integrin interaction/phosphotyrosine binding domain and function. J Cell Sci 118, 4765-4783. tyrosine phosphorylation sites in polyoma middle T antigen-mediated cell transformation. J BiolBarberis, L., Wary, K. K., Fiucci, G., Liu, F., Chem 272, 20671-20677. Hirsch, E., Brancaccio, M., Altruda, F., Tarone, G., and Giancotti, F. G. (2000). Distinct roles of Blake, R. A., Broome, M. A., Liu, X., Wu, J., the adaptor protein Shc and focal adhesion kinase Gishizky, M., Sun, L., and Courtneidge, S. A. in integrin signaling to ERK. J Biol Chem 275, (2000). SU6656, a selective src family kinase 36532-36540. inhibitor, used to probe growth factor signaling. Mol Cell Biol 20, 9018-9027.Batzer, A. G., Rotin, D., Urena, J. M., Skolnik, E. Y., and Schlessinger, J. (1994). Hierarchy of Bohnsack, M. T., Czaplinski, K., and Gorlich, D. binding sites for Grb2 and Shc on the epidermal (2004). Exportin 5 is a RanGTP-dependent growth factor receptor. Mol Cell Biol 14, 5192- dsRNA-binding protein that mediates nuclear 5201. export of pre-miRNAs. Rna 10, 185-191.Behrens, J., Mareel, M. M., Van Roy, F. M., and Bracke, M. E., Van Roy, F. M., and Mareel, M. M. Birchmeier, W. (1989). Dissecting tumor cell (1996). The E-cadherin/catenin complex in invasion: epithelial cells acquire invasive invasion and metastasis. Curr Top Microbiol properties after the loss of uvomorulin-mediated Immunol 213 (Pt 1), 123-161. cell-cell adhesion. J Cell Biol 108, 2435-2447. 74
  • REFERENCESBrembeck, F. H., Schwarz-Romond, T., Bakkers, J., and proliferative signaling by the adaptor protein Wilhelm, S., Hammerschmidt, M., and Shc. J Cell Biol 147, 1561-1568. Birchmeier, W. (2004). Essential role of BCL9-2 in the switch between beta-catenins adhesive and Conti, L., De Fraja, C., Gulisano, M., Migliaccio, transcriptional functions. Genes Dev 18, 2225- E., Govoni, S., and Cattaneo, E. (1997). 2230. Expression and activation of SH2/PTB- containing ShcA adaptor protein reflects theBrummelkamp, T. R., Bernards, R., and Agami, R. pattern of neurogenesis in the mammalian brain. (2002). A system for stable expression of short Proc Natl Acad Sci U S A 94, 8185-8190. interfering RNAs in mammalian cells. Science 296, 550-553. Croce, C. M., and Calin, G. A. (2005). miRNAs, cancer, and stem cell division. Cell 122, 6-7.Buday, L., Warne, P. H., and Downward, J. (1995). Downregulation of the Ras activation pathway by Davis, M. A., Ireton, R. C., and Reynolds, A. B. MAP kinase phosphorylation of Sos. Oncogene (2003). A core function for p120-catenin in 11, 1327-1331. cadherin turnover. J Cell Biol 163, 525-534.Cai, X., Hagedorn, C. H., and Cullen, B. R. (2004). Di Croce, L., and Pelicci, P. G. (2003). Tumour- Human microRNAs are processed from capped, associated hypermethylation: silencing E- polyadenylated transcripts that can also function cadherin expression enhances invasion and as mRNAs. Rna 10, 1957-1966. metastasis. Eur J Cancer 39, 413-414.Carmeliet, P., Lampugnani, M. G., Moons, L., Dupont, H., and Blancq, M. (1999). Formation of Breviario, F., Compernolle, V., Bono, F., complexes involving RasGAP and p190 RhoGAP Balconi, G., Spagnuolo, R., Oostuyse, B., during morphogenetic events of the gastrulation Dewerchin, M., et al. (1999). Targeted deficiency in xenopus. Eur J Biochem 265, 530-538. or cytosolic truncation of the VE-cadherin gene in mice impairs VEGF-mediated endothelial Dykxhoorn, D. M., Novina, C. D., and Sharp, P. A. survival and angiogenesis. Cell 98, 147-157. (2003). Killing the messenger: short RNAs that silence gene expression. Nat Rev Mol Cell BiolCattaneo, E., and Pelicci, P. G. (1998). Emerging 4, 457-467. roles for SH2/PTB-containing Shc adaptor proteins in the developing mammalian brain. El-Shemerly, M. Y., Besser, D., Nagasawa, M., and Trends Neurosci 21, 476-481. Nagamine, Y. (1997). 12-O- Tetradecanoylphorbol-13-acetate activates theCavallaro, U., and Christofori, G. (2004). Cell Ras/extracellular signal-regulated kinase (ERK) adhesion and signalling by cadherins and Ig- signaling pathway upstream of SOS involving CAMs in cancer. Nat Rev Cancer 4, 118-132. serine phosphorylation of Shc in NIH3T3 cells. J Biol Chem 272, 30599-30602.Cavallaro, U., Niedermeyer, J., Fuxa, M., and Christofori, G. (2001). N-CAM modulates Elbashir, S. M., Harborth, J., Lendeckel, W., tumour-cell adhesion to matrix by inducing FGF- Yalcin, A., Weber, K., and Tuschl, T. (2001a). receptor signalling. Nat Cell Biol 3, 650-657. Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells. NatureChen, C. Z., Li, L., Lodish, H. F., and Bartel, D. P. 411, 494-498. (2004). MicroRNAs modulate hematopoietic lineage differentiation. Science 303, 83-86. Elbashir, S. M., Harborth, J., Weber, K., and Tuschl, T. (2002). Analysis of gene function inChen, H. E., Chang, S., Trub, T., and Neel, B. G. somatic mammalian cells using small interfering (1996). Regulation of colony-stimulating factor 1 RNAs. Methods 26, 199-213. receptor signaling by the SH2 domain-containing tyrosine phosphatase SHPTP1. Mol Cell Biol 16, Elbashir, S. M., Lendeckel, W., and Tuschl, T. 3685-3697. (2001b). RNA interference is mediated by 21- and 22-nucleotide RNAs. Genes Dev 15, 188-Christofori, G. (2003). Changing neighbours, 200. changing behaviour: cell adhesion molecule- mediated signalling during tumour progression. Ellisen, L. W., Ramsayer, K. D., Johannessen, C. Embo J 22, 2318-2323. M., Yang, A., Beppu, H., Minda, K., Oliner, J. D., McKeon, F., and Haber, D. A. (2002).Collins, L. R., Ricketts, W. A., Yeh, L., and REDD1, a developmentally regulated Cheresh, D. (1999). Bifurcation of cell migratory transcriptional target of p63 and p53, links p63 to 75
  • REFERENCES regulation of reactive oxygen species. Mol Cell domain of Shc suppresses EGF-induced 10, 995-1005. mitogenesis in a dominant negative manner. Oncogene 11, 2525-2533.Faisal, A., el-Shemerly, M., Hess, D., and Nagamine, Y. (2002). Serine/threonine Gotoh, N., Tojo, A., and Shibuya, M. (1996). A phosphorylation of ShcA. Regulation of protein- novel pathway from phosphorylation of tyrosine tyrosine phosphatase-pest binding and residues 239/240 of Shc, contributing to suppress involvement in insulin signaling. J Biol Chem apoptosis by IL-3. Embo J 15, 6197-6204. 277, 30144-30152. Gotoh, N., Toyoda, M., and Shibuya, M. (1997).Faisal, A., Kleiner, S., and Nagamine, Y. (2004). Tyrosine phosphorylation sites at amino acids Non-redundant role of Shc in Erk activation by 239 and 240 of Shc are involved in epidermal cytoskeletal reorganization. J Biol Chem 279, growth factor-induced mitogenic signaling that is 3202-3211. distinct from Ras/mitogen-activated protein kinase activation. Mol Cell Biol 17, 1824-1831.Fire, A., Xu, S., Montgomery, M. K., Kostas, S. A., Driver, S. E., and Mello, C. C. (1998). Potent and Gottardi, C. J., and Gumbiner, B. M. (2004). specific genetic interference by double-stranded Distinct molecular forms of beta-catenin are RNA in Caenorhabditis elegans. Nature 391, 806- targeted to adhesive or transcriptional complexes. 811. J Cell Biol 167, 339-349.Frame, M. C. (2002). Src in cancer: deregulation Graiani, G., Lagrasta, C., Migliaccio, E., Spillmann, and consequences for cell behaviour. Biochim F., Meloni, M., Madeddu, P., Quaini, F., Padura, Biophys Acta 1602, 114-130. I. M., Lanfrancone, L., Pelicci, P., and Emanueli, C. (2005). Genetic deletion of the p66Shc adaptorFrixen, U. H., and Nagamine, Y. (1993). protein protects from angiotensin II-induced Stimulation of urokinase-type plasminogen myocardial damage. Hypertension 46, 433-440. activator expression by blockage of E-cadherin- dependent cell-cell adhesion. Cancer Res 53, Gregory, R. I., and Shiekhattar, R. (2005). 3618-3623. MicroRNA biogenesis and cancer. Cancer Res 65, 3509-3512.Fujita, Y., Krause, G., Scheffner, M., Zechner, D., Leddy, H. E., Behrens, J., Sommer, T., and Grishok, A., Pasquinelli, A. E., Conte, D., Li, N., Birchmeier, W. (2002). Hakai, a c-Cbl-like Parrish, S., Ha, I., Baillie, D. L., Fire, A., protein, ubiquitinates and induces endocytosis of Ruvkun, G., and Mello, C. C. (2001). Genes and the E-cadherin complex. Nat Cell Biol 4, 222- mechanisms related to RNA interference regulate 231. expression of the small temporal RNAs that control C. elegans developmental timing. CellGiorgio, M., Migliaccio, E., Orsini, F., Paolucci, 106, 23-34. D., Moroni, M., Contursi, C., Pelliccia, G., Luzi, L., Minucci, S., Marcaccio, M., et al. (2005). Grosheva, I., Shtutman, M., Elbaum, M., and Electron transfer between cytochrome c and Bershadsky, A. D. (2001). p120 catenin affects p66Shc generates reactive oxygen species that cell motility via modulation of activity of Rho- trigger mitochondrial apoptosis. Cell 122, 221- family GTPases: a link between cell-cell contact 233. formation and regulation of cell locomotion. J Cell Sci 114, 695-707.Gitlin, L., Karelsky, S., and Andino, R. (2002). Short interfering RNA confers intracellular Gu, H., Maeda, H., Moon, J. J., Lord, J. D., antiviral immunity in human cells. Nature 418, Yoakim, M., Nelson, B. H., and Neel, B. G. 430-434. (2000). New role for Shc in activation of the phosphatidylinositol 3-kinase/Akt pathway. MolGoodwin, M., Kovacs, E. M., Thoreson, M. A., Cell Biol 20, 7109-7120. Reynolds, A. B., and Yap, A. S. (2003). Minimal mutation of the cytoplasmic tail inhibits the Gu, J., Tamura, M., Pankov, R., Danen, E. H., ability of E-cadherin to activate Rac but not Takino, T., Matsumoto, K., and Yamada, K. M. phosphatidylinositol 3-kinase: direct evidence of (1999). Shc and FAK differentially regulate cell a role for cadherin-activated Rac signaling in motility and directionality modulated by PTEN. J adhesion and contact formation. J Biol Chem Cell Biol 146, 389-403. 278, 20533-20539. Gumbiner, B. M. (2005). Regulation of cadherin-Gotoh, N., Muroya, K., Hattori, S., Nakamura, S., mediated adhesion in morphogenesis. Nat Rev Chida, K., and Shibuya, M. (1995). The SH2 Mol Cell Biol 6, 622-634. 76
  • REFERENCES and Linsley, P. S. (2003). Expression profilingHajra, K. M., and Fearon, E. R. (2002). Cadherin reveals off-target gene regulation by RNAi. Nat and catenin alterations in human cancer. Genes Biotechnol 21, 635-637. Chromosomes Cancer 34, 255-268. Jackson, J. G., Yoneda, T., Clark, G. M., and Yee,Hannon, G. J. (2002). RNA interference. Nature D. (2000). Elevated levels of p66 Shc are found 418, 244-251. in breast cancer cell lines and primary tumors with high metastatic potential. Clin Cancer Res 6,Hashimoto, A., Kurosaki, M., Gotoh, N., Shibuya, 1135-1139. M., and Kurosaki, T. (1999). Shc regulates epidermal growth factor-induced activation of the Kamei, T., Matozaki, T., Sakisaka, T., Kodama, A., JNK signaling pathway. J Biol Chem 274, 20139- Yokoyama, S., Peng, Y. F., Nakano, K., Takaishi, 20143. K., and Takai, Y. (1999). Coendocytosis of cadherin and c-Met coupled to disruption of cell-Hiendlmeyer, E., Regus, S., Wassermann, S., cell adhesion in MDCK cells--regulation by Rho, Hlubek, F., Haynl, A., Dimmler, A., Koch, C., Rac and Rab small G proteins. Oncogene 18, Knoll, C., van Beest, M., Reuning, U., et al. 6776-6784. (2004). Beta-catenin up-regulates the expression of the urokinase plasminogen activator in human Kao, A. W., Waters, S. B., Okada, S., and Pessin, J. colorectal tumors. Cancer Res 64, 1209-1214. E. (1997). Insulin stimulates the phosphorylation of the 66- and 52-kilodalton Shc isoforms byHirohashi, S. (1998). Inactivation of the E- distinct pathways. Endocrinology 138, 2474- cadherin-mediated cell adhesion system in human 2480. cancers. Am J Pathol 153, 333-339. Kasus-Jacobi, A., Perdereau, D., Tartare-Deckert,Hirohashi, S., and Kanai, Y. (2003). Cell adhesion S., Van Obberghen, E., Girard, J., and Burnol, A. system and human cancer morphogenesis. Cancer F. (1997). Evidence for a direct interaction Sci 94, 575-581. between insulin receptor substrate-1 and Shc. J Biol Chem 272, 17166-17170.Hoschuetzky, H., Aberle, H., and Kemler, R. (1994). Beta-catenin mediates the interaction of Kavanaugh, W. M., and Williams, L. T. (1994). An the cadherin-catenin complex with epidermal alternative to SH2 domains for binding tyrosine- growth factor receptor. J Cell Biol 127, 1375- phosphorylated proteins. Science 266, 1862- 1380. 1865.Huebner, K., Kastury, K., Druck, T., Salcini, A. E., Kennerdell, J. R., and Carthew, R. W. (1998). Use Lanfrancone, L., Pelicci, G., Lowenstein, E., Li, of dsRNA-mediated genetic interference to W., Park, S. H., Cannizzaro, L., and et al. (1994). demonstrate that frizzled and frizzled 2 act in the Chromosome locations of genes encoding human wingless pathway. Cell 95, 1017-1026. signal transduction adaptor proteins, Nck (NCK), Shc (SHC1), and Grb2 (GRB2). Genomics 22, Ketting, R. F., Fischer, S. E., Bernstein, E., Sijen, 281-287. T., Hannon, G. J., and Plasterk, R. H. (2001). Dicer functions in RNA interference and inHutvagner, G., McLachlan, J., Pasquinelli, A. E., synthesis of small RNA involved in Balint, E., Tuschl, T., and Zamore, P. D. (2001). developmental timing in C. elegans. Genes Dev A cellular function for the RNA-interference 15, 2654-2659. enzyme Dicer in the maturation of the let-7 small temporal RNA. Science 293, 834-838. Khvorova, A., Reynolds, A., and Jayasena, S. D. (2003). Functional siRNAs and miRNAs exhibitHynes, N. E., and Lane, H. A. (2005). ERBB strand bias. Cell 115, 209-216. receptors and cancer: the complexity of targeted inhibitors. Nat Rev Cancer 5, 341-354. Kim, V. N. (2005). Small RNAs: classification, biogenesis, and function. Mol Cells 19, 1-15.Irigoyen, J. P., and Nagamine, Y. (1999). Cytoskeletal reorganization leads to induction of Kobielak, A., Pasolli, H. A., and Fuchs, E. (2004). the urokinase-type plasminogen activator gene by Mammalian formin-1 participates in adherens activating FAK and Src and subsequently the junctions and polymerization of linear actin Ras/Erk signaling pathway. Biochem Biophys cables. Nat Cell Biol 6, 21-30. Res Commun 262, 666-670. Kojima, T., Yoshikawa, Y., Takada, S., Sato, M.,Jackson, A. L., Bartz, S. R., Schelter, J., Kobayashi, Nakamura, T., Takahashi, N., Copeland, N. G., S. V., Burchard, J., Mao, M., Li, B., Cavet, G., Gilbert, D. J., Jenkins, N. A., and Mori, N. 77
  • REFERENCES (2001). Genomic organization of the Shc-related degraded to generate new siRNAs. Cell 107, 297- phosphotyrosine adaptors and characterization of 307. the full-length Sck/ShcB: specific association of p68-Sck/ShcB with pp135. Biochem Biophys Res Lord, J. D., McIntosh, B. C., Greenberg, P. D., and Commun 284, 1039-1047. Nelson, B. H. (1998). The IL-2 receptor promotes proliferation, bcl-2 and bcl-x induction, but notKovacs, E. M., Ali, R. G., McCormack, A. J., and cell viability through the adapter molecule Shc. J Yap, A. S. (2002). E-cadherin homophilic Immunol 161, 4627-4633. ligation directly signals through Rac and phosphatidylinositol 3-kinase to regulate Luschnig, S., Krauss, J., Bohmann, K., Desjeux, I., adhesive contacts. J Biol Chem 277, 6708-6718. and Nusslein-Volhard, C. (2000). The Drosophila SHC adaptor protein is required for signaling byLai, K. M., Olivier, J. P., Gish, G. D., Henkemeyer, a subset of receptor tyrosine kinases. Mol Cell 5, M., McGlade, J., and Pawson, T. (1995). A 231-241. Drosophila shc gene product is implicated in signaling by the DER receptor tyrosine kinase. Luzi, L., Confalonieri, S., Di Fiore, P. P., and Mol Cell Biol 15, 4810-4818. Pelicci, P. G. (2000). Evolution of Shc functions from nematode to human. Curr Opin Genet DevLai, K. M., and Pawson, T. (2000). The ShcA 10, 668-674. phosphotyrosine docking protein sensitizes cardiovascular signaling in the mouse embryo. Maretzky, T., Reiss, K., Ludwig, A., Buchholz, J., Genes Dev 14, 1132-1145. Scholz, F., Proksch, E., de Strooper, B., Hartmann, D., and Saftig, P. (2005). ADAM10Lamkin, T. D., Walk, S. F., Liu, L., Damen, J. E., mediates E-cadherin shedding and regulates Krystal, G., and Ravichandran, K. S. (1997). Shc epithelial cell-cell adhesion, migration, and beta- interaction with Src homology 2 domain catenin translocation. Proc Natl Acad Sci U S A containing inositol phosphatase (SHIP) in vivo 102, 9182-9187. requires the Shc-phosphotyrosine binding domain and two specific phosphotyrosines on SHIP. J Marone, R., Hess, D., Dankort, D., Muller, W. J., Biol Chem 272, 10396-10401. Hynes, N. E., and Badache, A. (2004). Memo mediates ErbB2-driven cell motility. Nat CellLe, S., Connors, T. J., and Maroney, A. C. (2001). Biol 6, 515-522. c-Jun N-terminal kinase specifically phosphorylates p66ShcA at serine 36 in response Mauro, L., Sisci, D., Bartucci, M., Salerno, M., to ultraviolet irradiation. J Biol Chem 276, Kim, J., Tam, T., Guvakova, M. A., Ando, S., 48332-48336. and Surmacz, E. (1999). SHC-alpha5beta1 integrin interactions regulate breast cancer cellLee, Y., Ahn, C., Han, J., Choi, H., Kim, J., Yim, adhesion and motility. Exp Cell Res 252, 439- J., Lee, J., Provost, P., Radmark, O., Kim, S., and 448. Kim, V. N. (2003). The nuclear RNase III Drosha initiates microRNA processing. Nature 425, 415- McManus, M. T. (2003). MicroRNAs and cancer. 419. Semin Cancer Biol 13, 253-258.Lee, Y., Jeon, K., Lee, J. T., Kim, S., and Kim, V. Meister, G., and Tuschl, T. (2004). Mechanisms of N. (2002). MicroRNA maturation: stepwise gene silencing by double-stranded RNA. Nature processing and subcellular localization. Embo J 431, 343-349. 21, 4663-4670. Mello, C. C., and Conte, D., Jr. (2004). RevealingLee, Y., Kim, M., Han, J., Yeom, K. H., Lee, S., the world of RNA interference. Nature 431, 338- Baek, S. H., and Kim, V. N. (2004). MicroRNA 342. genes are transcribed by RNA polymerase II. Embo J 23, 4051-4060. Migliaccio, E., Giorgio, M., Mele, S., Pelicci, G., Reboldi, P., Pandolfi, P. P., Lanfrancone, L., andLi, P. F., Dietz, R., and von Harsdorf, R. (1999). Pelicci, P. G. (1999). The p66shc adaptor protein p53 regulates mitochondrial membrane potential controls oxidative stress response and life span in through reactive oxygen species and induces mammals. Nature 402, 309-313. cytochrome c-independent apoptosis blocked by Bcl-2. Embo J 18, 6027-6036. Migliaccio, E., Mele, S., Salcini, A. E., Pelicci, G., Lai, K. M., Superti-Furga, G., Pawson, T., DiLipardi, C., Wei, Q., and Paterson, B. M. (2001). Fiore, P. P., Lanfrancone, L., and Pelicci, P. G. RNAi as random degradative PCR: siRNA (1997). Opposite effects of the p52shc/p46shc primers convert mRNA into dsRNAs that are and p66shc splicing isoforms on the EGF 78
  • REFERENCES receptor-MAP kinase-fos signalling pathway. tumor progression of mouse epidermal Embo J 16, 706-716. carcinogenesis. J Cell Biol 115, 517-533.Mikol, V., Baumann, G., Zurini, M. G., and Nemoto, S., and Finkel, T. (2002). Redox Hommel, U. (1995). Crystal structure of the SH2 regulation of forkhead proteins through a p66shc- domain from the adaptor protein SHC: a model dependent signaling pathway. Science 295, 2450- for peptide binding based on X-ray and NMR 2452. data. J Mol Biol 254, 86-95. Ngo, H., Tschudi, C., Gull, K., and Ullu, E. (1998).Milia, E., Di Somma, M. M., Baldoni, F., Chiari, Double-stranded RNA induces mRNA R., Lanfrancone, L., Pelicci, P. G., Telford, J. L., degradation in Trypanosoma brucei. Proc Natl and Baldari, C. T. (1996). The aminoterminal Acad Sci U S A 95, 14687-14692. phosphotyrosine binding domain of Shc associates with ZAP-70 and mediates TCR Noe, V., Fingleton, B., Jacobs, K., Crawford, H. C., dependent gene activation. Oncogene 13, 767- Vermeulen, S., Steelant, W., Bruyneel, E., 775. Matrisian, L. M., and Mareel, M. (2001). Release of an invasion promoter E-cadherin fragment byMinoo, P., Zadeh, M. M., Rottapel, R., Lebrun, J. matrilysin and stromelysin-1. J Cell Sci 114, 111- J., and Ali, S. (2004). A novel SHP-1/Grb2- 118. dependent mechanism of negative regulation of cytokine-receptor signaling: contribution of SHP- Nolan, M. K., Jankowska, L., Prisco, M., Xu, S., 1 C-terminal tyrosines in cytokine signaling. Guvakova, M. A., and Surmacz, E. (1997). Blood 103, 1398-1407. Differential roles of IRS-1 and SHC signaling pathways in breast cancer cells. Int J Cancer 72,Miyagishi, M., and Taira, K. (2002). U6 promoter- 828-834. driven siRNAs with four uridine 3 overhangs efficiently suppress targeted gene expression in Noren, N. K., Arthur, W. T., and Burridge, K. mammalian cells. Nat Biotechnol 20, 497-500. (2003). Cadherin engagement inhibits RhoA via p190RhoGAP. J Biol Chem 278, 13615-13618.Munshi, H. G., Ghosh, S., Mukhopadhyay, S., Wu, Y. I., Sen, R., Green, K. J., and Stack, M. S. Noren, N. K., Liu, B. P., Burridge, K., and Kreft, B. (2002). Proteinase suppression by E-cadherin- (2000). p120 catenin regulates the actin mediated cell-cell attachment in premalignant cytoskeleton via Rho family GTPases. J Cell Biol oral keratinocytes. J Biol Chem 277, 38159- 150, 567-580. 38167. Noren, N. K., Niessen, C. M., Gumbiner, B. M.,Murayama, Y., Miyagawa, J., Oritani, K., Yoshida, and Burridge, K. (2001). Cadherin engagement H., Yamamoto, K., Kishida, O., Miyazaki, T., regulates Rho family GTPases. J Biol Chem 276, Tsutsui, S., Kiyohara, T., Miyazaki, Y., et al. 33305-33308. (2004). CD9-mediated activation of the p46 Shc isoform leads to apoptosis in cancer cells. J Cell OBryan, J. P., Songyang, Z., Cantley, L., Der, C. Sci 117, 3379-3388. J., and Pawson, T. (1996). A mammalian adaptor protein with conserved Src homology 2 andNakamura, T., Sanokawa, R., Sasaki, Y., Ayusawa, phosphotyrosine-binding domains is related to D., Oishi, M., and Mori, N. (1996). N-Shc: a Shc and is specifically expressed in the brain. neural-specific adapter molecule that mediates Proc Natl Acad Sci U S A 93, 2729-2734. signaling from neurotrophin/Trk to Ras/MAPK pathway. Oncogene 13, 1111-1121. Okada, S., Kao, A. W., Ceresa, B. P., Blaikie, P., Margolis, B., and Pessin, J. E. (1997). The 66-Napoli, C., Martin-Padura, I., de Nigris, F., kDa Shc isoform is a negative regulator of the Giorgio, M., Mansueto, G., Somma, P., epidermal growth factor-stimulated mitogen- Condorelli, M., Sica, G., De Rosa, G., and activated protein kinase pathway. J Biol Chem Pelicci, P. (2003). Deletion of the p66Shc 272, 28042-28049. longevity gene reduces systemic and tissue oxidative stress, vascular cell apoptosis, and early Orsini, F., Migliaccio, E., Moroni, M., Contursi, C., atherogenesis in mice fed a high-fat diet. Proc Raker, V. A., Piccini, D., Martin-Padura, I., Natl Acad Sci U S A 100, 2112-2116. Pelliccia, G., Trinei, M., Bono, M., et al. (2004). The life span determinant p66Shc localizes toNavarro, P., Gomez, M., Pizarro, A., Gamallo, C., mitochondria where it associates with Quintanilla, M., and Cano, A. (1991). A role for mitochondrial heat shock protein 70 and regulates the E-cadherin cell-cell adhesion molecule during trans-membrane potential. J Biol Chem 279, 25689-25695. 79
  • REFERENCES Pelicci, G., Dente, L., De Giuseppe, A., Verducci-Pacini, S., Pellegrini, M., Migliaccio, E., Patrussi, Galletti, B., Giuli, S., Mele, S., Vetriani, C., L., Ulivieri, C., Ventura, A., Carraro, F., Naldini, Giorgio, M., Pandolfi, P. P., Cesareni, G., and A., Lanfrancone, L., Pelicci, P., and Baldari, C. Pelicci, P. G. (1996). A family of Shc related T. (2004). p66SHC promotes apoptosis and proteins with conserved PTB, CH1 and SH2 antagonizes mitogenic signaling in T cells. Mol regions. Oncogene 13, 633-641. Cell Biol 24, 1747-1757. Pelicci, G., Giordano, S., Zhen, Z., Salcini, A. E.,Pacini, S., Ulivieri, C., Di Somma, M. M., Isacchi, Lanfrancone, L., Bardelli, A., Panayotou, G., A., Lanfrancone, L., Pelicci, P. G., Telford, J. L., Waterfield, M. D., Ponzetto, C., Pelicci, P. G., and Baldari, C. T. (1998). Tyrosine 474 of ZAP- and et al. (1995a). The motogenic and mitogenic 70 is required for association with the Shc responses to HGF are amplified by the Shc adaptor and for T-cell antigen receptor-dependent adaptor protein. Oncogene 10, 1631-1638. gene activation. J Biol Chem 273, 20487-20493. Pelicci, G., Lanfrancone, L., Grignani, F.,Pagnin, E., Fadini, G., de Toni, R., Tiengo, A., McGlade, J., Cavallo, F., Forni, G., Nicoletti, I., Calo, L., and Avogaro, A. (2005). Diabetes Grignani, F., Pawson, T., and Pelicci, P. G. induces p66shc gene expression in human (1992). A novel transforming protein (SHC) with peripheral blood mononuclear cells: relationship an SH2 domain is implicated in mitogenic signal to oxidative stress. J Clin Endocrinol Metab 90, transduction. Cell 70, 93-104. 1130-1136. Pelicci, G., Lanfrancone, L., Salcini, A. E.,Palovuori, R., Sormunen, R., and Eskelinen, S. Romano, A., Mele, S., Grazia Borrello, M., (2003). SRC-induced disintegration of adherens Segatto, O., Di Fiore, P. P., and Pelicci, P. G. junctions of madin-darby canine kidney cells is (1995b). Constitutive phosphorylation of Shc dependent on endocytosis of cadherin and proteins in human tumors. Oncogene 11, 899- antagonized by Tiam-1. Lab Invest 83, 1901- 907. 1915. Perl, A. K., Wilgenbus, P., Dahl, U., Semb, H., andParrish, S., Fleenor, J., Xu, S., Mello, C., and Fire, Christofori, G. (1998). A causal role for E- A. (2000). Functional anatomy of a dsRNA cadherin in the transition from adenoma to trigger: differential requirement for the two carcinoma. Nature 392, 190-193. trigger strands in RNA interference. Mol Cell 6, 1077-1087. Polyak, K., Xia, Y., Zweier, J. L., Kinzler, K. W., and Vogelstein, B. (1997). A model for p53-Patrussi, L., Savino, M. T., Pellegrini, M., Paccani, induced apoptosis. Nature 389, 300-305. S. R., Migliaccio, E., Plyte, S., Lanfrancone, L., Pelicci, P. G., and Baldari, C. T. (2005). Ponti, G., Conti, L., Cataudella, T., Zuccato, C., Cooperation and selectivity of the two Grb2 Magrassi, L., Rossi, F., Bonfanti, L., and binding sites of p52Shc in T-cell antigen receptor Cattaneo, E. (2005). Comparative expression signaling to Ras family GTPases and Myc- profiles of ShcB and ShcC phosphotyrosine dependent survival. Oncogene 24, 2218-2228. adapter molecules in the adult brain. Neuroscience 133, 105-115.Pawson, T., and Saxton, T. M. (1999). Signaling networks--do all roads lead to the same genes? Pratt, J. C., van den Brink, M. R., Igras, V. E., Cell 97, 675-678. Walk, S. F., Ravichandran, K. S., and Burakoff, S. J. (1999). Requirement for Shc in TCR-Pece, S., Chiariello, M., Murga, C., and Gutkind, J. mediated activation of a T cell hybridoma. J S. (1999). Activation of the protein kinase Immunol 163, 2586-2591. Akt/PKB by the formation of E-cadherin- mediated cell-cell junctions. Evidence for the Pronk, G. J., de Vries-Smits, A. M., Buday, L., association of phosphatidylinositol 3-kinase with Downward, J., Maassen, J. A., Medema, R. H., the E-cadherin adhesion complex. J Biol Chem and Bos, J. L. (1994). Involvement of Shc in 274, 19347-19351. insulin- and epidermal growth factor-induced activation of p21ras. Mol Cell Biol 14, 1575-Pece, S., and Gutkind, J. S. (2000). Signaling from 1581. E-cadherins to the MAPK pathway by the recruitment and activation of epidermal growth Rameh, L. E., Arvidsson, A., Carraway, K. L., 3rd, factor receptors upon cell-cell contact formation. Couvillon, A. D., Rathbun, G., Crompton, A., J Biol Chem 275, 41227-41233. VanRenterghem, B., Czech, M. P., Ravichandran, K. S., Burakoff, S. J., et al. (1997). A comparative analysis of the phosphoinositide 80
  • REFERENCES binding specificity of pleckstrin homology Yancopoulos, G. D., Muller, W. J., Pawson, T., domains. J Biol Chem 272, 22059-22066. and Park, M. (2004). The Shc adaptor protein is critical for VEGF induction by Met/HGF andRavichandran, K. S. (2001). Signaling via Shc ErbB2 receptors and for early onset of tumor family adapter proteins. Oncogene 20, 6322- angiogenesis. Proc Natl Acad Sci U S A 101, 6330. 2345-2350.Ravichandran, K. S., Lee, K. K., Songyang, Z., Saucier, C., Papavasiliou, V., Palazzo, A., Cantley, L. C., Burn, P., and Burakoff, S. J. Naujokas, M. A., Kremer, R., and Park, M. (1993). Interaction of Shc with the zeta chain of (2002). Use of signal specific receptor tyrosine the T cell receptor upon T cell activation. Science kinase oncoproteins reveals that pathways 262, 902-905. downstream from Grb2 or Shc are sufficient for cell transformation and metastasis. Oncogene 21,Ravichandran, K. S., Lorenz, U., Shoelson, S. E., 1800-1811. and Burakoff, S. J. (1995). Interaction of Shc with Grb2 regulates the Grb2 association with Sayeski, P. P., and Ali, M. S. (2003). The critical mSOS. Ann N Y Acad Sci 766, 202-203. role of c-Src and the Shc/Grb2/ERK2 signaling pathway in angiotensin II-dependent VSMCReinhart, B. J., Slack, F. J., Basson, M., Pasquinelli, proliferation. Exp Cell Res 287, 339-349. A. E., Bettinger, J. C., Rougvie, A. E., Horvitz, H. R., and Ruvkun, G. (2000). The 21-nucleotide Schlaepfer, D. D., Hauck, C. R., and Sieg, D. J. let-7 RNA regulates developmental timing in (1999). Signaling through focal adhesion kinase. Caenorhabditis elegans. Nature 403, 901-906. Prog Biophys Mol Biol 71, 435-478.Reynolds, A. B., and Roczniak-Ferguson, A. Schlaepfer, D. D., Jones, K. C., and Hunter, T. (2004). Emerging roles for p120-catenin in cell (1998). Multiple Grb2-mediated integrin- adhesion and cancer. Oncogene 23, 7947-7956. stimulated signaling pathways to ERK2/mitogen- activated protein kinase: summation of both c-Richardson, A., and Parsons, T. (1996). A Src- and focal adhesion kinase-initiated tyrosine mechanism for regulation of the adhesion- phosphorylation events. Mol Cell Biol 18, 2571- associated proteintyrosine kinase pp125FAK. 2585. Nature 380, 538-540. Schlessinger, J., and Lemmon, M. A. (2003). SH2Robertson, H. D., and Mathews, M. B. (1996). The and PTB domains in tyrosine kinase signaling. regulation of the protein kinase PKR by RNA. Sci STKE 2003, RE12. Biochimie 78, 909-914. Schwarz, D. S., Hutvagner, G., Du, T., Xu, Z.,Rommel, C., and Hafen, E. (1998). Ras--a versatile Aronin, N., and Zamore, P. D. (2003). cellular switch. Curr Opin Genet Dev 8, 412-418. Asymmetry in the assembly of the RNAi enzyme complex. Cell 115, 199-208.Rustom, R., Wang, B., McArdle, F., Shalamanova, L., Alexander, J., McArdle, A., Thomas, C. E., Schwarzer, R., Tondera, D., Arnold, W., Giese, K., Bone, J. M., Shenkin, A., and Jackson, M. J. Klippel, A., and Kaufmann, J. (2005). REDD1 (2003). Oxidative stress in a novel model of integrates hypoxia-mediated survival signaling chronic acidosis in LLC-PK1 cells. Nephron Exp downstream of phosphatidylinositol 3-kinase. Nephrol 95, e13-23. Oncogene 24, 1138-1149.Sakai, R., Henderson, J. T., OBryan, J. P., Elia, A. Semizarov, D., Frost, L., Sarthy, A., Kroeger, P., J., Saxton, T. M., and Pawson, T. (2000). The Halbert, D. N., and Fesik, S. W. (2003). mammalian ShcB and ShcC phosphotyrosine Specificity of short interfering RNA determined docking proteins function in the maturation of through gene expression signatures. Proc Natl sensory and sympathetic neurons. Neuron 28, Acad Sci U S A 100, 6347-6352. 819-833. Shoshani, T., Faerman, A., Mett, I., Zelin, E.,Salcini, A. E., McGlade, J., Pelicci, G., Nicoletti, I., Tenne, T., Gorodin, S., Moshel, Y., Elbaz, S., Pawson, T., and Pelicci, P. G. (1994). Formation Budanov, A., Chajut, A., et al. (2002). of Shc-Grb2 complexes is necessary to induce Identification of a novel hypoxia-inducible factor neoplastic transformation by overexpression of 1-responsive gene, RTP801, involved in Shc proteins. Oncogene 9, 2827-2836. apoptosis. Mol Cell Biol 22, 2283-2293.Saucier, C., Khoury, H., Lai, K. M., Peschard, P., Sijen, T., Fleenor, J., Simmer, F., Thijssen, K. L., Dankort, D., Naujokas, M. A., Holash, J., Parrish, S., Timmons, L., Plasterk, R. H., and 81
  • REFERENCES Fire, A. (2001). On the role of RNA Tsuda, N., Kawano, K., Efferson, C. L., and amplification in dsRNA-triggered gene silencing. Ioannides, C. G. (2005). Synthetic microRNA Cell 107, 465-476. and double-stranded RNA targeting the 3- untranslated region of HER-2/neu mRNA inhibitSimcha, I., Kirkpatrick, C., Sadot, E., Shtutman, HER-2 protein expression in ovarian cancer cells. M., Polevoy, G., Geiger, B., Peifer, M., and Ben- Int J Oncol 27, 1299-1306. Zeev, A. (2001). Cadherin sequences that inhibit beta-catenin signaling: a study in yeast and Ugi, S., Imamura, T., Ricketts, W., and Olefsky, J. mammalian cells. Mol Biol Cell 12, 1177-1188. M. (2002). Protein phosphatase 2A forms a molecular complex with Shc and regulates ShcStevenson, L. E., and Frackelton, A. R., Jr. (1998). tyrosine phosphorylation and downstream Constitutively tyrosine phosphorylated p52 Shc mitogenic signaling. Mol Cell Biol 22, 2375- in breast cancer cells: correlation with ErbB2 and 2387. p66 Shc expression. Breast Cancer Res Treat 49, 119-128. van de Wetering, M., Barker, N., Harkes, I. C., van der Heyden, M., Dijk, N. J., Hollestelle, A.,Stevenson, L. E., Ravichandran, K. S., and Klijn, J. G., Clevers, H., and Schutte, M. (2001). Frackelton, A. R., Jr. (1999). Shc dominant Mutant E-cadherin breast cancer cells do not negative disrupts cell cycle progression in both display constitutive Wnt signaling. Cancer Res G0-G1 and G2-M of ErbB2-positive breast 61, 278-284. cancer cells. Cell Growth Differ 10, 61-71. Velazquez, L., Gish, G. D., van Der Geer, P.,Strathdee, G. (2002). Epigenetic versus genetic Taylor, L., Shulman, J., and Pawson, T. (2000). alterations in the inactivation of E-cadherin. The shc adaptor protein forms interdependent Semin Cancer Biol 12, 373-379. phosphotyrosine-mediated protein complexes in mast cells stimulated with interleukin 3. BloodTabara, H., Sarkissian, M., Kelly, W. G., Fleenor, 96, 132-138. J., Grishok, A., Timmons, L., Fire, A., and Mello, C. C. (1999). The rde-1 gene, RNA interference, Ventura, A., Luzi, L., Pacini, S., Baldari, C. T., and and transposon silencing in C. elegans. Cell 99, Pelicci, P. G. (2002). The p66Shc longevity gene 123-132. is silenced through epigenetic modifications of an alternative promoter. J Biol Chem 277, 22370-Takeichi, M. (1995). Morphogenetic roles of classic 22376. cadherins. Curr Opin Cell Biol 7, 619-627. Ventura, A., Maccarana, M., Raker, V. A., andThiery, J. P. (2002). Epithelial-mesenchymal Pelicci, P. G. (2004). A cryptic targeting signal transitions in tumour progression. Nat Rev induces isoform-specific localization of p46Shc Cancer 2, 442-454. to mitochondria. J Biol Chem 279, 2299-2306.Thomas, D., Patterson, S. D., and Bradshaw, R. A. Vleminckx, K., Vakaet, L., Jr., Mareel, M., Fiers, (1995). Src homologous and collagen (Shc) W., and van Roy, F. (1991). Genetic protein binds to F-actin and translocates to the manipulation of E-cadherin expression by cytoskeleton upon nerve growth factor epithelial tumor cells reveals an invasion stimulation in PC12 cells. J Biol Chem 270, suppressor role. Cell 66, 107-119. 28924-28931. Walk, S. F., March, M. E., and Ravichandran, K. S.Trinei, M., Giorgio, M., Cicalese, A., Barozzi, S., (1998). Roles of Lck, Syk and ZAP-70 tyrosine Ventura, A., Migliaccio, E., Milia, E., Padura, I. kinases in TCR-mediated phosphorylation of the M., Raker, V. A., Maccarana, M., et al. (2002). A adapter protein Shc. Eur J Immunol 28, 2265- p53-p66Shc signalling pathway controls 2275. intracellular redox status, levels of oxidation- damaged DNA and oxidative stress-induced Wary, K. K., Mainiero, F., Isakoff, S. J., apoptosis. Oncogene 21, 3872-3878. Marcantonio, E. E., and Giancotti, F. G. (1996). The adaptor protein Shc couples a class ofTrub, T., Choi, W. E., Wolf, G., Ottinger, E., Chen, integrins to the control of cell cycle progression. Y., Weiss, M., and Shoelson, S. E. (1995). Cell 87, 733-743. Specificity of the PTB domain of Shc for beta turn-forming pentapeptide motifs amino-terminal Wheelock, M. J., Buck, C. A., Bechtol, K. B., and to phosphotyrosine. J Biol Chem 270, 18205- Damsky, C. H. (1987). Soluble 80-kd fragment of 18208. cell-CAM 120/80 disrupts cell-cell adhesion. J Cell Biochem 34, 187-202. 82
  • REFERENCESWheelock, M. J., and Johnson, K. R. (2003). Identification of p46 Shc expressed in the nuclei Cadherin-mediated cellular signaling. Curr Opin of hepatocytes with high proliferating activity: Cell Biol 15, 509-514. Study of regenerating rat liver. Int J Mol Med 13, 721-728.Woodfield, R. J., Hodgkin, M. N., Akhtar, N., Morse, M. A., Fuller, K. J., Saqib, K., Thompson, Yukimasa, S., Masaki, T., Yoshida, S., Uchida, N., N. T., and Wakelam, M. J. (2001). The p85 Watanabe, S., Usuki, H., Yoshiji, H., Maeta, T., subunit of phosphoinositide 3-kinase is Ebara, K., Nakatsu, T., et al. (2005). Enhanced associated with beta-catenin in the cadherin- expression of p46 Shc in the nucleus and p52 Shc based adhesion complex. Biochem J 360, 335- in the cytoplasm of human gastric cancer. Int J 344. Oncol 26, 905-911.Xie, Y., and Hung, M. C. (1996). p66Shc isoform Zamore, P. D., Tuschl, T., Sharp, P. A., and Bartel, down-regulated and not required for HER-2/neu D. P. (2000). RNAi: double-stranded RNA signaling pathway in human breast cancer cell directs the ATP-dependent cleavage of mRNA at lines with HER-2/neu overexpression. Biochem 21 to 23 nucleotide intervals. Cell 101, 25-33. Biophys Res Commun 221, 140-145. Zhang, L., Camerini, V., Bender, T. P., andYang, C. P., and Horwitz, S. B. (2002). Distinct Ravichandran, K. S. (2002). A nonredundant role mechanisms of taxol-induced serine for the adapter protein Shc in thymic T cell phosphorylation of the 66-kDa Shc isoform in development. Nat Immunol 3, 749-755. A549 and RAW 264.7 cells. Biochim Biophys Acta 1590, 76-83. Zhang, L., Lorenz, U., and Ravichandran, K. S. (2003). Role of Shc in T-cell development andYang, J., Mani, S. A., Donaher, J. L., Ramaswamy, function. Immunol Rev 191, 183-195. S., Itzykson, R. A., Come, C., Savagner, P., Gitelman, I., Richardson, A., and Weinberg, R. Zhou, M. M., Harlan, J. E., Wade, W. S., Crosby, A. (2004). Twist, a master regulator of S., Ravichandran, K. S., Burakoff, S. J., and morphogenesis, plays an essential role in tumor Fesik, S. W. (1995a). Binding affinities of metastasis. Cell 117, 927-939. tyrosine-phosphorylated peptides to the COOH- terminal SH2 and NH2-terminal phosphotyrosineYang, S., Tutton, S., Pierce, E., and Yoon, K. binding domains of Shc. J Biol Chem 270, (2001). Specific double-stranded RNA 31119-31123. interference in undifferentiated mouse embryonic stem cells. Mol Cell Biol 21, 7807-7816. Zhou, M. M., Meadows, R. P., Logan, T. M., Yoon, H. S., Wade, W. S., Ravichandran, K. S.,Yap, A. S., Brieher, W. M., and Gumbiner, B. M. Burakoff, S. J., and Fesik, S. W. (1995b). (1997). Molecular and functional analysis of Solution structure of the Shc SH2 domain cadherin-based adherens junctions. Annu Rev complexed with a tyrosine-phosphorylated Cell Dev Biol 13, 119-146. peptide from the T-cell receptor. Proc Natl Acad Sci U S A 92, 7784-7788.Yekta, S., Shih, I. H., and Bartel, D. P. (2004). MicroRNA-directed cleavage of HOXB8 mRNA. Zhou, M. M., Ravichandran, K. S., Olejniczak, E. Science 304, 594-596. F., Petros, A. M., Meadows, R. P., Sattler, M., Harlan, J. E., Wade, W. S., Burakoff, S. J., andYoshida, S., Masaki, T., Feng, H., Yuji, J., Fesik, S. W. (1995c). Structure and ligand Miyauchi, Y., Funaki, T., Yoshiji, H., recognition of the phosphotyrosine binding Matsumoto, K., Uchida, N., Watanabe, S., et al. domain of Shc. Nature 378, 584-592. (2004). Enhanced expression of adaptor molecule p46 Shc in nuclei of hepatocellular carcinoma cells: study of LEC rats. Int J Oncol 25, 1089- 1096.Yu, J. Y., DeRuiter, S. L., and Turner, D. L. (2002). RNA interference by expression of short- interfering RNAs and hairpin RNAs in mammalian cells. Proc Natl Acad Sci U S A 99, 6047-6052.Yuji, J., Masaki, T., Yoshida, S., Kita, Y., Feng, H., Uchida, N., Yoshiji, H., Kitanaka, A., Watanabe, S., Kurokohchi, K., and Kuriyama, S. (2004). 83
  • ACKNOWLEDGEMENTS 6. ACKNOWLEDGEMENTS First, I would like to thank my parents who go to Pat King and Sara Oakley for criticalsupported me through all my life and are reading of my manuscripts.always there for me when I need them. My heartiest gratitude goes to Boris I also wish to acknowledge Dr. Yoshikuni Bartholdy who always supported meNagamine, who supervised me, and gave me scientifically with all of his skills and privatelythe opportunity to develop my scientific with all of his love.thinking and skills in his lab. While he offeredscientific freedom, he was always available fordiscussions. I greatly appreciate that. Thanks also to Prof. Gerhard Christofori andProf. Fred Meins, the two other members of mythesis committee, for the advice they gave meduring the committee meeting and for the timethey will still have to invest to read andevaluate this thesis. My special thanks go to Joshi Venugopal,from whom I could learn a lot in many aspectsof life and who became a close friend of mine.His critical and logical thinking inspired me inseveral things. I really appreciate the time wespent together. I also want to acknowledge MalgorzataKiesielow for our fruitful collaboration and forteaching me siRNA transfections. In addition, I wish to thank my former andcurrent lab members Faisal, Hoanh, Fumiko,Kacka, Sandra and Stephane. I alwaysenjoined working and spending some free timewith all of you. Further, I want to acknowledge the technicalstaff at the FMI who were always friendly andhelpful and made the scientific life at the FMImuch easier and productive. Thanks go to allof the FMI members (especially from theHynes laboratory) and to all of those, whoprovided me with scientific material: FrançoisLehembre, Kurt Ballmer, Tony Pawson, PeterE. Shaw and Jerrold Olefsky. My thanks also 84
  • ABBREVIATIONS 7. ABBREVIATIONSAng II Angiotensin II MMP matrix-metallo proteaseAPC adenomatous poliposis coli NGF neuronal growth factorCH collagen homology domain PBL peripheral blood lymphocytesCSR cytoskeletal reorganization PDGF platelet-derived growthDCR dicer-like protein factorDN double negative stage PH pleckstrin homology domainDP double positive stage PI3K phosphatitylinositol-3 kinasedsRNA double-stranded RNA PI(4)P phosphoinositol 4-phosphateEGF epithelial growth factor PI(4,5)P2 phosphoinositol 4,5-diphosphateEGFR EGF receptor PI(3,4,5)P3 phosphoinositol 3,4,,5-phosphateEMT epithelia-to-mesenchymal PKB protein kinase B (Akt) transition PP2A protein phosphatase type 2AErk extracellular activated kinase PTB phosphotyrosine binding domainFAK focal adhesion kinase R2D2 dsRNA binding proteinFKHRL Forkhead family transcription RISC RNA-induced silencing factor complexGab Grb2-associated binding protein RNAi RNA interferenceGAP GTPase activating protein ROS reactive oxygen speciesGPCR G protein-coupled receptors Shc Src-homology and collagen-Grb2 growth-factor-receptor binding like protein protein-2 Shp2 Src homology phosphatase-GTP guanonsine triphosphate 1HGF hepatocyte growth factor SHIP SH2-containing inositolICAP-1 integrin cytoplasmic domain- polyphosphate 5-phosphatase associated protein-1 SOS son of sevenlessIL interleukin SP single positive stageIRS insulin receptor substrate siRNA small interfering RNAIGFR insulin-like growth factor SH src-homology domain receptor TCR T-cell receptorJNK jun N-terminal kinase TCF T-cell factorLEF lymphocyte-enhancer factor TGF-β transforming growth factor-βMAPK mitogen activated protein uPA urokinase plasminogen activator kinase UV ultraviolet lightMEFs mouse embryonic fibroblasts TPA 12-O-tetradecanoylphorbol-13Memo mediator of ErbB2-driven cell acetate motility VEGF vascular endothelial growth factormiRNA microRNAmiRNP micro ribonucleoprotein particle 85
  • CURRICULUM VITAE 8. CURRICULUM VITAE Sandra KleinerPERSONAL DETAILSDate and Place of birth: 10.06.1976, Gera, GermanyMarital status: single, no childrenNationality: GermanPrivate Address: Oetlingerstrasse 150, CH-4057 Basel, Switzerland Phone: 061 681 8324Office Address: Friedrich Miescher Institute, Maulbeerstrasse 66, CH-4058 Basel, Switzerland, Phone 061 697 9944Email: skleiner@fmi.chEDUCATIONAL QUALIFICATIONDate Institute Qualification08.2001-to Friedrich Miescher Institute, Ph.D. in Molecular Biology, Thesis: Isoformpresent University of Basel, Switzerland specific-roles of the ShcA adaptor protein in Supervisor: Dr. Yoshikuni cellular signaling, in progress Nagamine (Prof. Fred Meins)02.1999-07.1999 University degli studi di L’Aquila, Study abroad Italy cell biology course, parasitology course, Italian language course08.1995-06.2001 Friedrich Schiller University, Jena, German Diploma (best mark: 1), Diploma Germany thesis: Development of a model system for the investigation of the c-Met signaling pathway in murine fibroblasts06.1991-07.1995 High School at the Holzland- German Abitur (best mark: 1) general Gymnasium, Hermsdorf, Germany qualification for university entranceGRANTS AND FELLOWSHIPS“” internet fellowship from McKinsey and German Telekom, 06.2001–to presentResearch Grant from Swiss cancer league, 01.2002-12.2002International Ph.D. Program Scholarship from Friedrich Miescher Institute, 08.2001–to presentErasmus fellowship to study abroad, 02.1999-08.1999 86
  • CURRICULUM VITAEPROFESSIONAL ACTIVITIESTechnical expertise Northern/Western blotting, transfection methods, proliferation assays, apoptosis assays, FACS analysis, immunostaining, GST protein preparation, kinase assays, cloningAcademic/Teaching experience Supervision of two trainees in the laboratory of Dr. Y. Nagamine Maya Zimmermann 08.2003-12.2003 Lauren Smith 01.2005-to present Tutorial for biology students of the first semester at University of Basel, Switzerland, 11.2002- 01.2003 Organized and taught a 3 week biology course to train especially talented pupils in a summer school academy (Association of Education and Giftedness), Rostock, Germany 07.2001 Teaching assistant in a practical course for students at the Friedrich Schiller University, Jena, Germany, 09.1999-01.2000 Service on Academic Appointment Committees: Committee to appoint a “Titularprofessor”, University of Basel, Switzerland, 06.2005 Committee to appoint a professor of Genetics, Friedrich Schiller University, Jena, Germany, 10.1999 Research assistant in the Faculty of Microbiology (Prof. Dr. J. Wöstemeyer) at the Friedrich Schiller University, Jena, Germany, 04.1998-01.1999 “Cloning of the laccase gene and its expression in Saccharomyces cerevisiae”Conferences attended Gordon Conference: Cell Contact & Adhesion, Andover, USA, 06.2005 Novartis – FMI joint symposium: Signal transduction pathways - past, present and future, Füringen, Switzerland, 03.2004 Novartis Corporate Research Conference, Boston, USA, 10.2004 European Life Science Organization Meeting, Nice, France, 07.2002 Signal Transduction Society Meeting: Signal Transduction: Receptors, Mediators and Genes, Berlin, Germany, 11.2000 3rd World Congress of Cellular and Molecular Biology – Modern Microscopy in Biology, Biotechnology and Medicine, Jena, Germany, 10.2000Event organization Member of the organizing committee of the Career Guidance Conference in Life science under the hospices of Novartis, Basel, Switzerland, 05.2005 th Co-organization of the traditional Biology Ball, an event organized by biology students of the 4 semester, Jena, Germany, 05.1998 87
  • CURRICULUM VITAEPUBLICATIONSKleiner, S., Faisal, A., and Nagamine, Y. 2005, J Biol Chem, Induction of uPA gene expression uponblockage of E-cadherin via Src- and Shc-dependent Erk. (submitted)Cramer, A., Kleiner, S., Westermann, M., Meissner, A., Lange, A., and Friedrich KH. 2005, J CellBiochem, Activation of the c-Met receptor complex in fibroblasts drives invasive cell behavior bysignaling through transcription factor STAT3.Faisal, A., Kleiner, S., and Nagamine, Y. 2004, J Biol Chem, Non-redundant role of Shc in Erkactivation by cytoskeletal reorganization.Kisielow, M*., Kleiner, S*., Nagasawa, M., Faisal, A., and Nagamine, Y. 2002, Biochem J., Isoform-specific knockdown and expression of adaptor protein ShcA using small interfering RNA. *both authorscontributed equallyPATENTKisielow, M,. Kleiner, S., Nagamine, Y. Methods of obtaining isoform-specific expression inmammalian cells. 1-32330A/FMI, 01.2002, Pending EP, JP and US.INTERESTS AND ACTIVITIESOutside of laboratory and office work I like physical exercise which includes regularly dancing in ournewly built up dancing group “Blickfang” but also outdoor sports like biking, jogging and inline skating.I am interested in foreign languages (Italian, French) and I enjoy learning and speaking them. 88