Shc (Src homology and collagen homology) proteins are considered prototypical signalling adaptors in mammalian cells. Consisting of four unique members, ShcA, B, C and D, and multiple splice isoforms, the family is represented in nearly every cell type in the body, where it engages in an array of fundamental processes to transduce environmental stimuli. Two decades of investigation have begun to illuminate the mechanisms of the flagship ShcA protein, whereas much remains to be learned about the newest discovery, ShcD. It is clear, however, that the distinctive modular architecture of Shc proteins, their promiscuous phosphotyrosine-based interactions with a multitude of membrane receptors, involvement in central cascades including MAPK (mitogen-activated protein kinase) and Akt, and unconventional contributions to oxidative stress and apoptosis all require intricate regulation, and underlie diverse physiological function. From early cardiovascular development and neuronal differentiation to lifespan determination and tumorigenesis, Shc adaptors have proven to be more ubiquitous, versatile and dynamic than their structures alone suggest.
The cell's capacity to recognize and respond to diverse stimuli is contingent upon two fundamental phenomena: inducible protein phosphorylation and molecular recognition events that enable transduction components to discern transient changes in residues and motifs. Indeed, the past three decades have witnessed both the birth and evolution of the signal transduction discipline, founded on observations that tyrosine phosphorylation could mediate oncogenic transformation of cells , and propelled by the subsequent discovery of a conserved pTyr (phosphotyrosine) recognition region, henceforth termed an SH2 (Src homology 2) domain, that was evident in several cytosolic kinases and responsible for orchestrating the cellular consequences of phosphorylation [2,3]. The concept that protein interactions are mediated by distinct, highly discerning, independently folded domains has since proven to be a unifying paradigm in cell signalling regulation [3,4]. Possession of one or more of these modules confers specific abilities to the parent protein irrespective of its other functions, thereby endowing the cell with a method of recycling and recombining successful signalling strategies .
Shortly after postulation of the domain theory, a novel protein came to light that was defined by an SH2 domain and the absence of any apparent catalytic function, thereby distinguishing it from the majority of known SH2 proteins of the day, such as Src and PLCγ (phospholipase Cγ) . Indeed, the Shc (Src homology and collagen) protein was one of the first characterized cellular adaptors, and remains a prototype for this class of molecule , that functions to recruit and juxtapose several sequential members of a pathway . Broadly, adaptors may serve to amplify or convert signals, or organize substrates in a cascade, and are defined by the unique complement of modules and amino acid motifs that equip them to participate in the transduction process .
In the ensuing twenty years since Shc was first identified, three additional related proteins have been identified in mammals, prompting the common designations ShcA, B, C and D, according to their order of discovery. The family is best characterized by its participation in the major pathways of growth factor signalling, especially the Ras/MAPK (mitogen-activated protein kinase) and PI3K (phosphatidylinositol-3 kinase)/Akt cascades proximal to various stimuli conveyed through RTKs (receptor tyrosine kinases), integrins and cytosolic kinases, as well as cytokine, antigen and G-protein-coupled receptors . Additionally, the 66 kDa isoform of ShcA performs unconventional roles in the cellular stress response . All of the homologues share a common architecture defined by an N-terminal PTB (phosphotyrosine-binding) domain and a C-terminal SH2 domain, both of which can independently co-ordinate phosphorylated tyrosine residues of active receptors. The domains are connected by a central CH (collagen homology) 1 region containing consensus tyrosine residues that are subject to phosphorylation upon engagement of the adaptor with an RTK, and subsequently serve as recognition motifs for proximal signalling effectors . Thus, by relocating to receptors in response to stimuli and recruiting downstream transduction components, Shc proteins provide a molecular platform for the development of the signalling complex. As the past decades have revealed though, the diversity of cellular processes involving these adaptors places Shc at a signalling crossroads, where its influence may extend beyond individual cascades to encompass pathway integration.
DISCOVERY AND CHARACTERIZATION
Identification of the first Shc adaptor arose from efforts to elucidate novel SH2 domain-containing proteins, in attempt to better define the mechanisms of tyrosine kinase signalling. Using a DNA probe designed against the SH2 domain of the c-fes tyrosine kinase, Pelicci et al.  screened human cDNA libraries for complementary sequences. Their search yielded a homologue sharing considerable identity with the SH2 domains of other proteins, most notably the cytosolic kinase Src (60%). Further inspection of the gene and its products revealed three protein isoforms, designated p66, p52 and p46 for their respective molecular masses in kDa (Figure 1). Structurally, the proteins are identical through the PTB–CH1–SH2 core, and are solely distinguished by length of their N-termini. p46 and p52 are generated from different start codons within the same transcript, whereas p66 is a product of alternative splicing that endows it with a unique N-terminal 110-amino-acid extension . Functionally, Shc proteins gained early notoriety as molecular mediators of pTyr-based signalling events .
Shc protein architecture
Targeted gene perturbation studies in mice have since revealed striking functional differences between the ShcA isoforms. Although concurrent disruption of p66, p52 and p46 results in embryonic lethality by E11.5 (embryonic day 11.5) , animals lacking p66 alone experience a 30% increase in lifespan, owing to the transcript's involvement in cellular oxidative stress and rendering p66 ShcA one of a small subset of human genes that has been implicated in the aging process .
Although p46 and p52 are broadly expressed across various tissues and developmental time points, p66 is notably absent from select cell types, including those of haematopoietic lineage [6,14]. It has also become apparent that all of the ShcA isoforms are markedly down-regulated in quiescent cells of the mature CNS (central nervous system) . Accordingly, several independent groups identified two novel homologues in the human and mouse brain that were believed to assume pTyr adaptor responsibilities in the absence of ShcA [16–18]. Although size estimate discrepancies exist in the literature, ShcB (Sck/Sli/Shc2) is commonly reported as a single 68 kDa protein [19–21], whereas ShcC (Rai/N-Shc/Shc3) is often cited as two isoforms of 64 kDa and 52 kDa [21–25]. In contrast with the embryonic-lethal ShcA-null animals, single and compound ShcB/C-knockout mice are viable, although subject to reductions in specific neuron populations .
The paradigm of Shc signalling within and beyond the nervous system stands to be challenged again, by the recent discovery of another mammalian homologue, ShcD [Shc4/RaLP (Rai-like protein)]. This adaptor was revealed by two separate groups as a result of BLAST queries of the mouse and human genomes [26,27]. Protein and RNA profiling in the mouse has identified ShcD in an array of embryonic and adult tissues, and has shown that it is differentially regulated throughout development . Its expression pattern is distinct from that of the other Shc proteins, and includes notable representation in the CNS, muscle, epithelia and bone precursors . Although the contributions of this protein to organismal development and maintenance have yet to be determined, incentive now exists to address this gap in adaptor knowledge.
Shc protein architecture
As discussed above, ShcA was first identified as an SH2-domain-containing protein, thereby establishing its role in pTyr signalling. However, the subsequent discovery and characterization of an additional novel domain capable of co-ordinating phosphorylated tyrosine residues [29,30] revealed this N-terminal PTB domain to be the dominant point of contact between ShcA and many of its ligands . In its unbound state prior to contacting a target ligand, the PTB domain is structurally disordered; only upon complex formation does the fully folded structure emerge . Analyses of peptide-bound domains demonstrate that, despite differences in gross architecture, the core pTyr-binding pockets of the Shc PTB and SH2 modules share a related conformation and amino acid composition, characterized by non-contiguous Arg-Arg-Lys-Ser residues . Nevertheless, the domains are selective for different amino acid motifs in their compatible ligands, with SH2 target specificity defined by residues C-terminal to the pTyr residue, and having the sequence pTyr-Φ-Xaa-Ile/Leu/Met (Φ representing a hydrophobic moiety) , whereas the PTB domain recognizes the phosphorylated residue in an N-terminal context given by Asn-Pro-Xaa-pTyr, with isoleucine or a similar hydrophobic amino acid at the −5 position relative to the pTyr . Notable exceptions to tyrosine specificity of the domain include phospho-independent interactions with cytoskeletal regulator IQGAP1 (IQ motif-containing GTPase-activating protein 1) , as well as the phosphatases PTP (protein-tyrosine phosphatase)-PEST (Pro-Glu-Ser-Thr) , PP2A (protein phosphatase 2A)  and PTPϵ .
Intriguingly, the Shc PTB domain also resembles phospholipid-binding PH (pleckstrin homology) domains. Using residues distinct from the pTyr-binding pocket, it selectively co-ordinates PtdIns(4,5)P2 and related components of the membrane bilayer, albeit with a lower affinity than the tyrosine interaction [32,39]. Although PtdIns(4,5)P2 does not affect the PTB–pTyr association , pTyr engagement with the domain does decrease its phospholipid affinity , suggesting that Shc could be thermodynamically driven from the membrane to activated receptors, but not vice versa. Indeed, lipid interaction is considered instrumental in preemptively targeting ShcA to the membrane and positioning it in proximity to certain receptors, including IL-3 (interleukin 3) . Of note, some of the residues responsible for phospholipid recognition are absent from ShcD, despite their conservation in the other Shc homologues . The consequence of this deletion on subcellular localization and signalling response of the adaptor have not directly been addressed, although ShcD is considered a physiologically relevant substrate of several RTKs [26,27].
In all of the vertebrate Shc proteins, the PTB and SH2 domains are connected by a CH1 region rich in glycine and proline residues [11,22], that can contribute to Pro-Xaa-Xaa-Pro motifs that bind SH3 (Src homology 3) domains of cellular kinases such as Src, Fyn and Lyn (v-yes-1 Yamaguchi sarcoma viral related oncogene homologue) . The best-characterized features of the region, however, are conserved tyrosine residues in positions 239/240 and 317 of p52 ShcA that become phosphorylated upon engagement of Shc proteins with active kinases, and represent canonical pTyr-Xaa-Asn-Xaa recognition motifs for the Ras/MAPK signalling adaptor, Grb2 (growth factor receptor-bound protein 2) . Additionally, ShcC and ShcD contain novel tyrosine phosphorylation sites that can co-ordinate the Crk (v-crk sarcoma virus CT10 oncogene homologue) and Grb2 adaptors respectively (Figure 1, peach residues) [26,41]. Near the C-terminal end of the CH1 region lies a binding motif for the adaptin proteins involved in clathrin-mediated internalization . This Arg-Asp-Leu-Phe-Asp-Met-Lys-Pro-Phe-Glu sequence  is notably absent from ShcD, yet conserved in ShcA, B and C (Figure 1, teal residues) [26,27].
Unique to the longer isoforms of the four Shc proteins is an N-terminal CH2 region, which is likewise enriched in glycine and proline residues . This attribute is particularly salient in ShcD, which contains several SH3-domain-recognition sequences of undetermined utility . The CH2 region has been most intensively scrutinized in p66 ShcA, as it constitutes the structural epicentre of the adaptor's oxidative stress response. As described below, at least three key sequences contribute to this function: a Ser36-Pro37 motif that undergoes phosphorylation , Cys59 that mediates tetramerization , and a CB [cyt c (cytochrome c)-binding] region that doubles as a catalytic core to generate ROS (reactive oxygen species) . Although present in p52 ShcA, the CB region has been deemed non-functional . Intriguingly, ShcD shares the Cys59 oligomerization residue , and may also possess the p66 Ser-Pro motif, as revealed by sequence alignment [26,27].
Overall, the CH2–PTB–CH1–SH2 orientation uniquely distinguishes Shc adaptors, and although PTB and SH2 domains are individually common in signalling proteins, they are rarely found within the same molecule . Amino acid sequence comparisons reveal that ShcA and D share the highest identity (Figure 1). Across the family, the PTB and SH2 domains are most conserved, whereas the CH1 and CH2 regions are more divergent . Striking similarity in exon organization between human ShcA, B and C argues that they arose from a common ancestral gene, as does the existence of apparent orthologues in Drosophila and Fugu rubripes, and putative loci in Caenorhabditis elegans . Interestingly, the nematode lacks both pTyr consensus sites of the CH1 region, whereas the fruitfly contains a Tyr239/240 sequence that becomes phosphorylated upon stimulation [22,45]. This motif, and Tyr317, are found in all of the vertebrate Shc adaptors, thus implying that expanded pTyr signalling capacity is associated with increased organismal complexity .
Fundamentally, the ability of Shc proteins to engage in controlled signal propagation in a given cell is contingent on their expression, localization, target recognition, and acquisition and removal of phosphate moieties, all of which must proceed in an orchestrated manner. Although efforts have focused on elucidating the details of ShcA dynamics, ShcB, C and D remain relatively unexplored from a regulatory perspective.
As previously noted, the isoforms of ShcA are differentially expressed . Analysis of the locus has revealed two promoters of the TATA-/Inr+ class that independently regulate transcription of p46/p52 and p66, the latter of which can be epigenetically repressed through cytosine methylation and histone deacetylation . Consistent with the role of this isoform in mediating ROS-responsive apoptosis, patients suffering ESRD (end-stage renal disease) experience considerably reduced CpG methylation in the p66 promoter .
Homologue- and isoform-specific differences also extend to subcellular localization of the proteins. In neurons, for example, ShcB is distributed throughout the soma and processes, whereas ShcC is limited to the cell body . Elsewhere, ShcA is predominantly a cytoplasmic protein that can concentrate in cellular niches. An N-terminal targeting sequence in p46 that is masked in the longer isoforms uniquely allows this adaptor to accumulate in the mitochondrial matrix . Conversely, a fraction of p66 can relocate to the mitochondrial intermembrane space in response to appropriate apoptotic stimuli [44,49,50]. Some estimates suggest that roughly 5% of p52 ShcA is already membrane-associated in unstimulated cells ; however, it is generally observed in the cytosol and rough endoplasmic reticulum, from which it translocates to the periphery upon growth factor stimulation [51,52].
Subsequently, rapid phosphorylation of Shc proteins upon binding active RTKs is a vital step in the transduction process that primes the adaptor to interact with downstream partners. Recent evidence suggests, however, that the implications of Shc post-translational modification extend beyond generating recognition motifs for effectors. George et al.  report that although the isolated SH2 domain of p52 ShcA binds representative target ligands in solution, it loses this ability in the context of the full-length Shc protein. Strikingly, phosphorylation of tyrosine residues in the CH1 region not only restores SH2-mediated binding, but also enhances the affinity of the interaction 50-fold, implying that the Shc protein employs a phospho-dependent switch that inducibly controls SH2 activation. Conversely, molecular dynamics simulations and in silico binding free energy estimates predict that the conformational change accompanying phosphorylation of Tyr317 decreases the affinity of Shc for representative ligands. This single modification was found to enhance structural rigidity of the adaptor and decouple the concerted movement of the PTB and SH2 domains [53,54], thereby disrupting the interaction of both domains with their cognate motifs on the EGF (epidermal growth factor) RTK . Considering that Shc engagement with a stimulated receptor potentiates the signal and also protects the RTK from phosphatases, this structural shift is postulated to control detachment of the adaptor following its activation, thereby facilitating eventual termination of the transduction process . However, the model does not account for additional phosphorylation of Tyr239/240, which usually occurs concomitantly with Tyr317. Whether this further alters molecular conformation remains to be seen.
Under these circumstances, biochemical and molecular simulation data appear to predict contradictory consequences of CH1 phosphorylation, although this may be the result of differences in experimental approaches. The biochemical perspective focused on the influence of phosphorylation in modifying the binding behaviour of the SH2 domain towards receptor motifs with which it failed to interact in its full-length inactive form. Unlike the in silico analysis, it did not consider the consequences of Shc phosphorylation on interactions for which the SH2 domain is a physiologically relevant principal point of contact. Thus it is possible that the models describe independent concurrent phenomena, or that the regulatory mechanisms are distinct and exclusive to certain subsets of receptor–adaptor interactions. In the former scenario, Shc phosphorylation by its partner RTK could conceivably destabilize the primary PTB-mediated interaction, while simultaneously priming a secondary association between the SH2 domain and another ligand. Indeed the utility of the double domain structure of Shc and its contributions to signalling cross-talk are presently undercharacterized. Beyond their direct implications for Shc signalling, these discoveries also challenge the assumption that independently folded domains act as fully autonomous modules within signalling proteins. Instead, they paint a more dynamic picture of contextual adaptability and regulation, hinting at subtle mechanisms by which transducers could integrate information.
The findings likewise emphasize the requirement to regulate Shc phosphorylation, and indeed studies have revealed a small repertoire of phosphatase enzymes capable of this task. The first to be identified was the ubiquitous PTP-PEST, which interacts constitutively with the PTB domains of p52 and p66 ShcA via an unconventional Asn-Pro-Leu-His recognition motif in its C-terminus [36,55]. The association is further enhanced by treatment with a DAG (diacylglycerol) second messenger analogue, which induces PKC (protein kinase C)-dependent serine phosphorylation of ShcA p52 at residue 29 (Ser138 in p66) and thereby provides an additional binding sequence for PTP-PEST [55,56]. Importantly, this association has been found to reduce Shc tyrosine phosphorylation and Shc-mediated MAPK activation following insulin signalling , and during antigen receptor-stimulated lymphocyte activation .
Tumour suppressor PTEN (phosphatase and tensin homologue deleted at chromosome 10), PP2A and PTPϵ are likewise capable of dephosphorylating Shc, impeding its interaction with Grb2 and suppressing subsequent MAPK activation [37,38,58]. Notably a role has been established for PTEN in negatively regulating Shc-mediated cell migration [59,60], thus implicating the phosphatase–adaptor interaction in the metastatic mechanism. Like PTP-PEST, PP2A and PTPϵ stably associate with the ShcA PTB domain under basal conditions; however, both phosphatases demonstrate distinct context-dependent regulation of ShcA. In the case of PP2A, growth factor stimulation and consequent phosphorylation of Shc Tyr317 inhibit the phosphatase and cause it to dissociate from the adaptor . Within 30 min of signal initiation, the interaction is restored.
Conversely, the capacities of the three ShcA isoforms to act as PTPϵ substrates depend on the kinase used initially to phosphorylate the adaptor, and whether it confers a protective effect due to complex formation. Proximal to the cytosolic kinase Src, Shc is subject to dephosphorylation by PTPϵ, whereas it is resistant to the phosphatase when downstream of RTKs including Neu and EGFR (EGF receptor) . This is postulated to reflect the difference in affinities between the kinases and adaptor, and competitive PTB domain engagement by protective RTKs that probably precludes phosphatase binding.
Although the finer details of Shc regulation are still subject to investigation, it is evident that numerous mechanisms integrate these adaptors into cellular networks that require inducible, reversible and highly specific interactions to maintain homoeostasis.
The canonical role of Shc adaptors is to mediate signal transduction events at the membrane, by providing a molecular scaffold for the recruitment and assembly of signalling components. The details of these exchanges continue to be revealed, at increasingly fine levels of resolution. Additionally, the adaptors are now appreciated to engage in a variety of homologue- and isoform-specific activities.
Growth factors comprise a spectrum of secreted compounds that mediate growth, survival and differentiation of responsive cells by selectively activating compatible transmembrane RTKs . Although numerous RTKs exist, EGFR is considered a prototype for this class of transducer and ShcA, its archetypal adaptor in mediating Ras/MAPK activation (Figure 2, A). Accordingly, we will discuss the EGFR–Shc interaction as a model of RTK signalling, with the caveat that only the general phenomenon can be extrapolated to other receptors. Nevertheless, the relevance of this system transcends vertebrate signalling, as the EGFR and Shc homologues DER (Drosophila relative of EGFR) and dShc (Drosophila Shc) are found to associate in Drosophila .
Molecular mechanisms of Shc proteins
Transduction begins when a ligand binds to the extracellular region of the RTK, stimulating receptor dimerization and autophosphorylation of conserved tyrosine residues on its cytoplasmic tail that represent recognition motifs for downstream targets. Uniquely, three such residues in EGFR seem to co-ordinate ShcA. Numbered according to classic nomenclature, pTyr992 is a minor SH2-domain-binding site, whereas pTyr1148 dominantly engages the PTB domain and pTyr1173 is considered a secondary motif that facilitates interaction with both the PTB and SH2 domains of ShcA [62–64]. Although the sites initially appeared redundant , subsequent analyses have found a more robust signalling output when the PTB domain is involved , and raise the possibility that optimal MAPK activation may be achieved through co-operation of both domains . Paradoxically, a dissenting line of evidence contends that Shc is responsive to EGF stimulation even after compound ablation of all of the aforementioned sites [67,68]; however, we offer an alternative explanation below.
Recent quantitative proteomic analyses of early signalling events have revealed that phosphorylation of EGFR Tyr1148 and Tyr1173 occurs within 5 s of ligand engagement, followed immediately by recruitment and phosphorylation of Shc . Accordingly, EGF stimulation appears to have two distinct effects on the intrinsic kinase of the RTK; in addition to increasing the general catalytic activity of the enzyme, EGF also alters substrate specificity and promotes preferential phosphorylation of Shc Tyr317 relative to EGFR autophosphorylation sites . Considering that the activation of another proximal EGFR partner, PLCγ, lags behind that of Shc , the possibility thus exists that a temporal component partly defines the consequences of transduction. Evidence further suggests that ShcA can mobilize directly from the cytoplasm to the EGFR without prior association with membrane phospholipids, although this is not true of all receptors. Interaction between ShcA and IL-3, for example, can be abrogated by mutating the three key residues in the PTB domain that constitute a phospholipid-binding pocket .
p66, p52 and p46 ShcA are all subject to phosphorylation  on two consensus binding motifs for the Grb2 adaptor, corresponding to Tyr239/240 and Tyr317 in the CH1 region of human p52 , as detected both in vitro [6,52] and in vivo in the mouse and rat [15,71]. Tyr239/240  and Tyr317  have each been proposed as the dominant Grb2-binding site that recruits the adaptor into the developing transduction complex. According to the latter scenario, Tyr317 mediates MAPK activation, whereas Tyr239/240 stimulates c-myc expression by unknown mechanisms . Notably though, the EGFR variant used in these experiments lacked the Shc-binding sites, so the outcome may not represent canonical EGFR signalling. Regardless of the motif(s) of origin, activation of the MAPK cascade, as detailed in Figure 2 (A), is the best-characterized outcome of signalling through p52 and p46 ShcA [7,11].
In contrast with these isoforms, however, p66 fails to augment MAPK phosphorylation upon ligand stimulation, despite binding both EGFR and Grb2 (Figure 2, C) [11,52,73,74]. The response is particularly perplexing considering that p66 shares identical PTB, CH1 and SH2 structures with the shorter p52/46 isoforms. Nevertheless, a model has emerged whereby the adaptor is recruited to active EGFR, phosphorylated first on tyrosine residues by the RTK, and then on serine/threonine residues of the CH2 region in a MEK [MAPK/ERK (extracellular-signal-regulated kinase) kinase]-dependent manner . Although the pTyr motifs bind and recruit Grb2, serine/threonine phosphorylation destabilizes the p66–EGFR interaction, thereby decoupling the fledgling signalling complex from Ras, and sequestering Grb2–SOS (Son of sevenless) away from the membrane and the other ShcA isoforms, which are not subject to phospho-serine/threonine regulation  (Figure 2). The consequences of this behaviour include accelerated inactivation of ERK  and suppressed transcription from the c-fos promoter, which is specifically attributed to the CH2 region .
Despite the focus on co-ordinating Grb2, phospho-independent interactions have also surfaced that diversify the downstream consequences of Shc signalling. Proximal to the EGF family receptors, ShcA and the scaffold IQGAP1 have been identified in membrane ruffles, where they associate via the Shc PTB domain . As ShcA can have a role in lamellipodia formation, the interaction is postulated to connect RTK activation to cytoskeletal reorganization through Shc .
A major component of the EGFR response, beyond signalling, involves internalizing and sorting the receptor, destining it for recycling or degradation (Figure 2, E). Within 5  to 20  min of stimulation, activated EGFR and ShcA are found to co-localize in intracellular vesicles identified as early endosomes by their immunoreactivity against EEA1 (early endosome antigen 1) . Specific tracking of phosphorylated Shc-recognition motifs on EGFR suggest that the receptor retains its signalling capacity through the initial internalization process . Conversely, only a limited subset of LAMP-1 (lysosomal-associated membrane protein 1)-positive late endosomes or lysosomes contain Tyr1173-phosphorylated EGFR, and ShcA is accordingly absent from these compartments . In addition to accompanying EGFR through part of its endocytic journey, ShcA has structural and functional precedence to participate in the internalization process. As discussed above, a motif in the CH1 region binds the α- and β-adaptin components of the AP2 (adaptor-related protein 2) complex , which is involved in clathrin-coated pit formation leading to endocytosis . The interaction between p52 ShcA and α-adaptin is markedly induced by EGF stimulation, thereby enhancing the amount of AP2 juxtaposed with EGFR. Moreover, ligand internalization is perturbed by disabling the adaptin-binding motif of ShcA . Although the AP2 complex can interact directly with the EGFR in a phospho-independent manner, it appears that Shc inducibly enhances recruitment of its components in response to signalling, therefore helping to facilitate the internalization process upon receptor activation .
By comparison, little is known about the associations between EGFR and ShcB, C and D, beyond their potential to interact [18,26,77]. Inaugural studies reported that ShcB and ShcC recruit Grb2 to the activated receptor , and that the isolated ShcC SH2 domain can act as a dominant negative to repress EGFR signalling through ERK and Elk . For ShcC, however, the phenomenon has been inconsistent and hard to recapitulate. Full-length endogenous ShcC is not phosphorylated following EGF stimulation of a neuroblastoma cell line, neither does it affect MAPK activation when overexpressed . Indeed, this adaptor appears to interact with a more limited repertoire of RTKs, and its reduced affinity for Grb2 renders it a poor transducer of the SOS–Ras cascade [21,24]. In contrast, ShcD has been found to augment the Ras/MAPK response proximal to receptors including EGFR , possibly as a result of its additional unique Grb2-recognition motif .
Overall, the ShcA cascade is thought to be a ‘redundant but dominant’ path to MAPK activation from the EGFR . Redundancy arises from the principal role of ShcA in recruiting Grb2, which can also bind directly to EGFR , and dominance is owed to the cell's predicted preference for this longer path over the more streamlined one . Insights gleaned from MEFs (mouse embryonic fibroblasts) harbouring a genetic deletion of all of the ShcA isoforms suggest that although the EGFR–MAPK pathway is functional in the absence of ShcA, the adaptor is required to sensitize cells to low concentrations of growth factor . In a different experimental system, the reduced proliferative potential of senescing hepatocytes has been attributed to diminished ERK activation, despite consistent levels of EGFR and Shc proteins and sustained phosphorylation capacity across time. Instead, the change in ERK responsiveness is postulated as a consequence of age-induced destabilization of the EGFR–Shc complex . Thus it appears that ShcA plays a key role in fine-tuning the MAPK signalling output from the EGFR.
A second major cellular hub involving Shc adaptors, particularly ShcA and ShcC, is the PI3K/Akt pathway [21,24,81], although it currently lacks the level of characterization afforded to Shc–MAPK. In particular, the consequences of ShcC proximal to the GDNF (glial-cell-line-derived neurotrophic factor) RTK Ret have received attention. In neuronal and thyroid carcinoma cells, ShcC facilitates pro-survival signalling by complexing with the adaptor Gab1 (Grb2-associated binding protein 1), which can recruit the p85 subunit of PI3K to the receptor upon stimulation, thereby activating this cascade as portrayed in Figure 2 (F) [21,24]. By contrast, ShcC appears to have little effect on PI3K signalling in neural progenitor cells , whereas it represses Akt phosphorylation in lymphocytes , thereby emphasizing the highly contextual nature of Shc-mediated transduction.
Cytosolic kinase signalling
From the infancy of Shc characterization, the role of cytosolic enzymes in regulating these adaptors has been sought. The SFKs (Src family kinases) constitute a collection of proto-oncogenic non-RTKs implicated in a variety of cellular phenomena . The founding member of the family, v-Src, was initially identified as the transforming factor in Rous sarcoma virus, and later revealed to have a cellular counterpart, c-Src, that could likewise contribute to malignancy when deregulated . SFKs are also important proximal components in RTK-mediated signalling events . In cells transformed by v-Src, all of the isoforms of ShcA are found to be highly phosphorylated . Interestingly v-Src phosphorylates ShcA at both Tyr239/240 and Tyr317, whereas c-Src targets only Tyr239/240 and depends on engagement of the Shc PTB domain with phosphatidylinositols [87–89]. Activation of these sites by SFKs likely primes Shc to engage downstream effectors in an analogous fashion to receptor-mediated signalling (Figure 2, B). Additionally, SFK regulation of Shc proximal to RTK stimulation may constitute a parallel pathway of Shc involvement in growth factor transduction. To that end, the aforementioned studies of receptor–adaptor dynamics that demonstrate ligand-induced ShcA phosphorylation in the absence of Shc-binding sites in the EGFR cytoplasmic tail have been perplexing [67,68]. However, the deletions and targeted mutations employed in these investigations failed to account for EGFR Tyr974 and Tyr845, sites of Src binding and phosphorylation respectively [90,91]. Furthermore, the EGFR–Src interaction is ligand dependent and known to enhance Shc phosphorylation . We therefore propose that EGF-mediated Shc signalling that occurs independently of a direct interaction between EGFR and the adaptor is in fact orchestrated by an SFK, which could localize to the receptor and subsequently activate ShcA.
Conversely, ShcA overexpression has been found to enhance the ligand-dependent association between EGFR and c-Src, as well as up-regulate Src enzyme activity . Both p66 and p52 directly bind and activate c-Src even in the absence of RTKs, via a non-canonical Shc interface . The consequence of the association appears to involve STAT (signal transducer and activator of transcription) phosphorylation and nuclear translocation, as well as induction of proteins including p21, WAF1 and Cip1 .
In addition to reciprocal regulatory involvement with SFKs, mounting evidence suggests that Shc is also subject to serine phosphorylation by PKC. Treatment of cells with a potent PKC activator fails to induce tyrosine phosphorylation of Shc, and instead produces a unique pattern of serine phosphorylation that can be prevented by pretreatment with a PKC inhibitor, and is not recapitulated by administering growth factors such as FGF-1 (fibroblast growth factor 1) [56,93]. Intriguingly, this stimulation induces a pTyr-independent association between Shc and Grb2 that is disrupted by the serine/threonine-specific PP1 (protein phosphatase 1) . However, the utility of this interaction is unclear, as subsequent analyses in Shc-null MEFs revealed that PKC-mediated activation of Ras/ERK proceeds independently of Shc . Intriguingly though, one of the novel phosphorylation sites corresponds to Ser29 in p52, which is involved in phosphatase regulation of ShcA, as described above .
Signalling through non-catalytic receptors
Shc adaptors are also involved in transducing signals downstream of transmembrane proteins that lack intrinsic tyrosine kinase activity, such as integrins and antigen receptors of T- and B-lymphocytes. Often the signals converge on the common Ras/MAPK pathway already discussed, although mounting evidence also highlights the existence of novel Shc functions.
Regardless of their origin, non-catalytic receptors require accessory kinases to initiate pTyr-mediated recruitment of signalling components. This is exemplified by the TCR (T-cell receptor), a multimeric complex involved in recognition of processed antigen. Upon engagement of the epitope-binding dimer, the associated CD3 signal transduction subunits are phosphorylated by the SFK Lck, on conserved ITAMs (immune-receptor tyrosine-based activation motifs). This is followed by recruitment of a second kinase, ZAP-70 [ζ-chain (TCR)-associated protein kinase 70 kDa] . As reviewed by Finetti et al. , the SH2 domain of p52 ShcA binds phospho-CD3, whereas ZAP-70 engages the PTB domain and can also phosphorylate the adaptor, in addition to Lck. In the case of p52, this can activate Ras/MAPK and Rac pathways, potentially influencing T-cell activation, survival and cytoskeletal dynamics. Similar mechanisms exist to couple p52 to the BCR (B-cell receptor). Conversely, both p66 ShcA and ShcC are considered negative regulators of TCR and BCR signalling, reducing the receptors' sensitivity to antigen and increasing levels of apoptosis [83,95]. Accordingly, loss of these antagonists triggers autoimmune phenotypes characterized by hyperactive cells and self-reactive antibodies .
In non-haematopoietic cells proteinaceous ECM (extracellular matrix) represents both a substratum for attachment as well as a source of information and stimuli for the cell. Accordingly, non-catalytic heterodimeric integrin receptors provide vital mechanical and chemical links between the ECM of the surroundings and the cytoskeletal and signalling networks of the cell interior, thereby influencing processes such as migration, proliferation and apoptosis . A subset of integrin receptors are capable of transducing signals through the ShcA adaptor protein, including those for laminin (α6β4) [97,98], collagen (α2β1) , collagen/laminin (α1β1), fibronectin (α5β1)  as well as the promiscuous receptor that can engage vitronectin (αvβ3) [99,100]. Several mechanisms have been proposed to connect Shc adaptors to active integrins and transduce ‘outside-in’ signals (Figure 2, D). In one strategy, intracellular kinases phosphorylate integrins directly to generate consensus Shc-recognition motifs. This is the case with the laminin-binding α6β4 integrin, which sustains phosphorylation on its β4 subunit upon ligation and subsequently recruits Shc, via its PTB and SH2 domains, to signal through the MAPK pathway [97,98]. β3 Integrin is also inducibly phosphorylated and becomes a direct target for Shc .
However, not all β subunits are subject to phosphorylation. In an indirect binding model, Shc recruitment occurs independently of the integrin cytoplasmic tail and any tyrosine residue motifs located therein. Instead, ligand engagement by select integrins stimulates the transmembrane region of the α subunit to interact with the membrane-spanning adaptor caveolin-1, which constitutively binds the SFK Fyn [100,102]. Alternatively, Yes (v-yes-1 Yamaguchi sarcoma viral oncogene homologue), Lck, or Lyn may substitute for Fyn . Activation of the kinase and its subsequent SH3-domain-mediated association with ShcA facilitate Shc phosphorylation of Tyr317, recruitment of Grb2 and initiation of the Ras/MAPK cascade that can lead to cell proliferation [100,102].
Owing to their promiscuity, Shc proteins can integrate simultaneous stimuli from multiple sources, such as growth factors and ECM, to tailor the cellular response. In one experimental system, ShcA was found to promote integrin-mediated haptotaxis in the absence of growth factors, by a PTB-domain-dependent mechanism. Upon the addition of EGF, migration proceeded independently of Shc, whereas proliferation required the Shc SH2 domain . Postnatal angiogenesis also appears to require the integrative powers of ShcA to assimilate signals from the VEGF (vascular endothelial growth factor) and integrin receptors, invoking a cell survival response upon engagement with a compatible ECM substrate and growth factor .
Recent developments also implicate p66 ShcA in anoikis, the apoptotic programme initiated when adherent cells detach from the matrix (Figure 2, G). Intriguingly though, this function is distinct from the role of p66 in oxidative-stress-induced apoptosis, and instead relies on recruitment of the adaptor to focal adhesions, the integrin-based mechanosensory structures that anchor the cell to the ECM . Here, p66 is speculated to infer the attachment status of the cell based upon tension in cytoskeletal stress fibres radiating from the foci. A loss of mechanical integrity triggers p66 to activate the GTPase RhoA, which initiates anoikis. Although the precise mechanism remains unclear, it appears to require the intact p66 PTB domain, and involve release of Ras-mediated repression of RhoA (Figure 2, G) [105,106]. In attached cells, or those lacking p66, the pro-apoptotic behaviours of RhoA are inhibited.
Oxidative stress and apoptosis
Arguably the most divergent of the Shc protein functions is its role in oxidative stress and PCD (programmed cell death). Although the topic has been comprehensively reviewed [107–112], the oxidoreductase activity of p66 ShcA remains a pivotal juxtaposition to canonical Shc signalling and an intensive area of investigation. The first indication of the adaptor's unprecedented function arose from a landmark study that selectively removed the p66 isoform of ShcA in mice, leaving p52 and p46 intact . Surprisingly, the genetic ablation extended rodent lifespan by 30%, and conferred resistance to oxidative stress at both the organismal and cellular level. Ser36 in the CH2 region was identified as a critical component of the p66-mediated apoptotic response to oxidative challenge, as phosphorylation of this residue is required to trigger cell death upon exposure to UV or H2O2 .
Models have since emerged that portray p66 ShcA as both a sensor and a proponent of ROS, and position it in an intricate network of cell fate decisions (Figure 2, H). It appears that the initial oxidative stimulus can be conveyed to ShcA via PKCβ, which phosphorylates Ser36 of p66 . Phospho-Ser36-Pro37 then becomes a consensus recognition site for the prolyl isomerase Pin1 (peptidylprolyl cis–trans isomerase, NIMA-interacting 1), which induces cis–trans isomerization around this bond and targets p66 for mitochondrial uptake . A further level of regulation may be achieved by context-dependent oligomerization of p66. The isolated CH1 region seems to exist as a non-covalent dimer that is incapable of eliciting apoptosis. However, under oxidizing conditions when redox scavengers are saturated, the dimer tetramerizes via disulfide bridging of the conserved Cys59, and primes the protein to initiate PCD . Strikingly, the translocated activated p66 is situated in the mitochondrial intermembrane space, from which it can interact with cyt c of the electron transport chain and engage in redox exchanges that provoke apoptosis . Contained within the CB motif of the CH1 region are three critical residues, Glu132, Glu133 and Trp134, that constitute the redox centre of p66. This core facilitates the binding and oxidation of cyt c, and serves to transfer electrons to molecular oxygen to generate H2O2. In turn, this ROS increases inner mitochondrial membrane permeability causing the organelle to swell, rupture and release pro-apoptotic factors into the cytosol .
Not surprisingly, p66 has been implicated in a number of disorders with metabolic and inflammatory characteristics that can be prevented by selective ablation of the protein in murine models. As revealed recently, however, p66 deletion is only advantageous under select, controlled circumstances. In a natural wild environment, p66-null mice fail to thrive and are found to have reduced fertility and intolerance to cold and starvation, owing to deficits in adipogenesis and glucose homoeostasis .
The collective p66 body of work briefly highlighted in the present review underscores the complexity of cellular regulatory systems and the diverse and dynamic functionality of ShcA.
The functional importance and non-redundancy of Shc proteins are best revealed by their diverse and unique contributions to animal development and maintenance. From embryonic lethality to prolonged lifespan, the consequences of Shc genetic perturbations are vast and far-reaching. In the past decade, emphasis has shifted from establishing the gross phenotypic changes in the mouse that arise from reverse-genetic analysis of Shc, to dissecting the specific intracellular requirements of signalling events that underlie physiology.
Considering the broad expression of ShcA in the adult, the discovery that it contributes most potently to heart and circulatory system morphogenesis at the early stages of mouse development came as a surprise. Global disruption of all three ShcA isoforms in a murine model causes defects in heart structure and function and reduces vascular complexity, thereby suggesting problems with angiogenic remodelling. ShcA KO (knockout) embryos succumb to gross cardiovascular dysfunction by E11.5 .
Analyses of cultured fibroblasts recovered from mutant embryos have revealed a distinct role for ShcA in facilitating ERK2 activation following stimulation with low concentrations of growth factor. In the absence of the adaptor, larger stimuli are required to elicit a comparable MAPK response. Additionally, the ShcA-null MEFs demonstrate reduced ERK phosphorylation and altered cytoskeletal organization proximal to integrin engagement with ECM. Accordingly, ShcA-deficient embryos experience considerable and specific reduction in active MAPK, corresponding with the areas involved in cardiovascular development that normally express ShcA .
Intriguingly, a similar phenotype arises from selective incapacitation of the ShcA PTB domain alone. Genetic KI (knockin) embryos expressing ShcA with a non-functional PTB domain die at approximately E11.5 and exhibit cardiac deficits reminiscent of full ShcA KO . This is not true of KI embryos with mutations to the SH2 domain or all of the CH1 region pTyr motifs (Tyr3Phe mutation) that abrogate interactions with up- or down-stream targets respectively. Instead these animals can survive to birth, but demonstrate impaired motor co-ordination due to abnormal development of muscle spindles and the monosynaptic stretch reflex . Thus it appears that mid-gestational heart morphogenesis requires ShcA with a functional PTB domain, but occurs independently of the CH1 region tyrosine residues , which have long been considered the primary effector regions of the adaptor.
Further support for this hypothesis has come from studies of the adaptor in postnatal heart function. Cardiomyocyte-specific ShcA KO produces dilated cardiomyopathy similar to cardiomyocyte-restricted expression of the same PTB mutation described above . Notably, although conditional removal of the CH1 region pTyr motifs in cardiomyocytes yields phenotypically normal offspring, these animals rapidly succumb to experimentally induced heart failure, thus implying that the Grb2 consensus sites contribute little to basal cardiac function, but are required to elicit an appropriate response to biomechanical stress .
Indeed the involvement of CH1 phosphotyrosine residues in cardiovascular homoeostasis appears to be highly contextual. In culture, cardiac fibroblasts exposed to thrombin demonstrate increased p46/52 phosphorylation on Tyr239/240 and Tyr317, concomitant with enhanced Erk activation, whereas the same treatment of cardiomyocytes fails to augment ShcA tyrosine phosphorylation .
The cardiac implications of ShcA further depend on the isoform in question. Expression of p66 has been shown to decrease during postnatal maturation of myocytes, and is re-induced upon exposure to hypertrophic agonists; the adaptor also undergoes different patterns of phosphorylation relative to p46/52 . Consistent with its role in oxidative stress and apoptosis, p66 has a deleterious role in the heart following chronic exposure to Ang II (angiotensin II), an instigating factor in cardiac senescence. Selective deletion of the p66 isoform confers resistance to Ang II-mediated hypertrophy and apoptosis of cardiomyocytes and endothelial cells in mice .
Overall, compelling evidence suggests that various tissues of the heart differentially utilize unique combinations of the signalling modalities of distinct ShcA isoforms, under the evolving demands of morphogenesis, maintenance, aging and stress, thereby underscoring the complex and dynamic contributions of this adaptor to cardiac physiology.
Genetic models and in vitro analyses have likewise fostered a vision of the dynamic interplay between ShcA, B and C proteins over the course of nervous system development and into adulthood. Contrary to its generalized expression elsewhere in the body, ShcA is highly restricted and tightly regulated in the brain. During gestation and beyond, its expression is confined to areas of active neuroblast proliferation, whereas post-mitotic cells show marked reduction of the protein coincident with cellular differentiation . Accordingly, it has been reported that the initial process of neural induction from ES (embryonic stem) cells is characterized by up-regulation of p66 ShcA . In support of this observation, experimentally induced overexpression of p66 in cultured ES cells exposed to neuralizing conditions accelerates the loss of pluripotency and acquisition of neural markers, such as nestin, as occurs very early in the commitment process . The contributions of ShcA proteins to neurogenesis from this point forward have been analysed by tissue-specific deletion or mutation in mice. McFarland et al.  reported that animals experienced considerable microencephaly when ShcA was excised under control of the nestin promoter, thereby implicating ShcA in survival mechanisms of this precursor population. Strikingly, a similar phenotype results from conditional expression of a phosphorylation-defective dominant negative adaptor (Shc Tyr3Phe) that lacks the central CH1 pTyr residues that serve as Grb2 consensus sequences. Its detrimental consequences have been attributed to neural progenitor apoptosis. Although the brain masses of nestin-induced Shc Tyr3Phe mice were reduced to 50% of control values for the lifetime of the animals, the period of enhanced apoptosis that gave rise to microencephaly was limited to a narrow developmental window .
In contrast to the CH1-region-independent contributions of ShcA to cardiovascular signalling [114,115], conserved pTyr motifs appear to be critical mediators of nervous system development . However, differences in the molecular strategies and genetic backgrounds of experimental animals used in these studies preclude direct comparison of the severity of the Tyr3Phe mutation in the heart and nervous system. Additional off-target effects of the dominant–negative protein under a non-native promoter also cannot be ruled out.
Although ShcA expression is initially high and progressively decreases with developmental stage , ShcC is absent early in embryogenesis and is strongly up-regulated prior to birth , at which point it is detected in quiescent neurons . Expression continues to escalate postnatally for 6–8 weeks in mice . Indeed ShcA and ShcC display a pattern of reciprocal expression in pre- and post-mitotic neuronal cells that has been postulated as a molecular commitment switch, wherein ShcA down-regulation and ShcC induction demarcate the transition between cellular proliferation and differentiation . This orchestration is consistent with the findings that ShcA is required to maintain neuronal precursors , whereas ShcC prevents apoptosis and promotes neuronal maturation, as gauged by morphological changes when the adaptor is overexpressed in primary neuronal cultures . The survival influence of ShcC depends on its SH2 domain, and involves sustained phosphorylation of Akt and inhibition of the pro-apoptotic protein Bad (BCL2-associated agonist of cell death), both of which are consequences of the PI3K cascade. Meanwhile, the adaptor's contribution to differentiation is linked to prolonged Erk1/2 activation . Both PI3K pathway augmentation and delayed persistent MAPK signalling are facilitated by ShcC proximal to the GDNF receptor Ret . Although a consensus has yet to be reached regarding the biochemical changes that drive and accompany the shift in Shc expression, these findings suggest that the molecular basis of the transition may lie in the divergent signalling properties of the two adaptors.
However, despite the appeal of the ShcA/ShcC developmental switch model, initial data from the genetic ShcC KO mouse did not appear to support a fundamental role for this adaptor in establishing and maintaining post-mitotic neurons. Curiously, ShcC KO animals are viable and display no overt morphological abnormalities . The possibility that ShcB, which likewise demonstrates a modest increase in CNS expression near the end of gestation, performs compensatory functions in the absence of ShcC was also tested. Double KO mice carrying global compound ShcB and ShcC ablation are viable and neuroanatomically normal in the CNS, though they exhibit reduced numbers of sympathetic neurons in the superior cervical ganglia that cannot be recapitulated by singular ablation of ShcB or ShcC . Loss occurs late in development, consistent with the hypothesis that ShcB/C support post-mitotic neurons. Both the ShcB and ShcB/ShcC KOs also lose discrete populations of sensory neurons from the dorsal root ganglia, although the animals retain sensory function. Evidently, the KO phenotypes are less severe than anticipated under a model that implicates ShcC as a critical factor in post-mitotic neuron survival. Sakai et al.  surmized that ShcB and C are not exclusive mediators of RTK signalling in neurons, but instead participate in only a subset of cellular responses. To that end, ShcD has recently been detected in post-mitotic embryonic neurons and it associates with several neuronal RTKs including TrkB (Tropomyosin-related kinase B) and MuSK (muscle, skeletal, receptor tyrosine kinase) [26,121], although its role in this context has yet to be investigated . It remains to be seen whether this uncharacterized protein will shed more light on the role of the Shc adaptor switch in neurogenesis.
Comprehensive Shc protein expression profiles have similarly been generated for the adult CNS and PNS (peripheral nervous system), complementing burgeoning functional studies to establish the roles of these adaptors after development. At maturity ShcA is present in the PNS , but is restricted to the olfactory epithelium  and SVZ (subventricular zone) of the brain , both of which are involved in adult neurogenesis. Accordingly, mice generated using McFarland's strategy to conditionally remove ShcA from nestin-expressing cells  exhibit smaller more disorganized SVZ niches postnatally . The size reduction has been attributed to decreased cell proliferation, whereas tissue architecture is altered on structural and physiological levels housing many cells of indeterminate origin . This phenotype mirrors the role of ShcA in development, bolstering the hypothesis that the adaptor is central to the maintenance and self-renewal properties of neural stem and precursor cells.
ShcB and ShcC have been widely detected in neurons of the adult CNS, where they display distinct profiles within the substructures of the brain and spinal cord. Both proteins appear to be excluded from glia [47,120] with the exception of the human gastrointestinal tract, where ShcC is evident in enteric glial cells, but absent from the surrounding neurons .
Consistent with their opposing roles in nervous system development, ShcA and ShcC are contextually regulated in motor neurons. Although ShcC predominates under homoeostasis, injury shifts the balance in favour of ShcA and suggests a related role for this protein in regeneration .
Intriguingly, further characterization of the ShcC KO mice at maturity has revealed more subtle phenotypes that emphasize the utility of this adaptor. ShcC KO animals are more susceptible to ischaemic damage and reperfusion injury following stroke-like cerebral artery occlusion, presenting larger infarct volumes, increased cellular apoptosis via the intrinsic mitochondrial pathway, more dramatic cognitive impairment and elevated incidence of mortality . Indeed cortical neurons isolated from ShcC-null animals show an increased sensitivity to hypoxic and oxidative stressors, coincident with decreased PI3K activity and Akt phosphorylation, prompting the authors to surmize that ShcC normally exerts neuroprotective effects in the wake of stress .
Moreover, a novel function in hippocampus-dependent learning and memory has been ascribed to the adaptor. ShcC-null animals demonstrate superior spatial and nonspatial learning as a result of enhanced hippocampal LTP (long-term potentiation) . Increased synaptic plasticity in the absence of ShcC coincides with elevated tyrosine phosphorylation of the postsynaptic NMDA (N-methyl-D-aspartate) receptor, which represents a physiological binding partner of ShcC and a key mediator of LTP .
Despite earlier reports to the contrary, ShcC has also recently been identified in neurogenic regions of the adult rodent brain, including the SVZ and RMS (rostral migratory stream). Evidence from the ShcC KO mice implicates this adaptor in facilitating migration of neural progenitors from the SVZ along the RMS to their final destination in the olfactory bulb, through its influence on metalloprotease expression and Wnt/β-catenin signalling .
DISEASE: CANCERS AND BEYOND
It remains to be seen whether the cardiovascular and neurological deficits apparent in Shc KO models are paralleled in naturally occurring human conditions. Yet, with such broad involvement in fundamental signalling pathways, metabolism, morphogenesis and physiology, Shc proteins are increasingly implicated in a spectrum of disorders that arise from cellular disregulation. Although ectopic or aberrant stimulation can provoke undesired cell proliferation or invasion, withdrawal of other tissues from adequate trophic support can lead to degenerative conditions. In the case of Huntington's disease, a novel link has recently been reported between p46/52 ShcA and the TrkB RTK. In vitro, the expanded mutant huntingtin protein reduces ShcA expression, effectively repressing Ras/ERK activation proximal to TrkB and its ligand, BDNF (brain-derived neurotrophic factor), and thereby rendering cells more sensitive to oxidative damage . Conversely, patients with Alzheimer's disease show enhanced glial expression of ShcA and elevated ERK activation . It has been shown that amyloid precursor protein, a neuropathological disease precursor, is cleaved to produce a membrane-bound C-terminal fragment that can become tyrosine-phosphorylated and interact with Shc and Grb2, thereby implicating the components in potentially detrimental signalling cascades [126,127].
Cancer investigations, however, have generated the most extensive insights into Shc protein contributions to pathology. From the time of its discovery, ShcA has been considered a proto-oncogene in the light of its capacity to transform cells , whereas ShcD was first characterized as a mediator of melanoma metastasis . Collectively, members of the Shc family have been associated prognostically and/or mechanistically with a variety of cancers including prostate, breast, thyroid and neuroblastoma, discussed below, in addition to malignancies of the brain , kidney  and colon .
The steroid hormones of the body influence development of neoplasms in certain tissues, including the prostate and breast. Although ShcA has been found to engage in various tumorigenic mechanisms, the prognosis depends largely on the tissue and Shc isoform in question, as they can have opposite effects in the two types of cancer .
The study of prostate cancer has focused on p66 ShcA as a driver of proliferation. This isoform in particular is up-regulated in primary prostate tumours and is induced upon steroid hormone treatment of cultured cells . Experimental p66 overexpression or knockdown can respectively increase or decrease prostate cancer cell proliferation , although the findings were initially hard to reconcile due to the protein's role in oxidative apoptosis and lack of involvement in mitogenic signalling. An alternative explanation proposes that steroid hormones act to increase p66 expression and translocation into mitochondria, where both the redox centre of the protein and the resulting ROS it generates upon interaction with cyt c are required to promote proliferation [129,132,133]. Although the participation of ROS in mitogenic signalling is well documented , prostate studies provide the first evidence implicating the oxidoreductase activity of p66 Shc in an oncogenic phenotype . However, the circumstances under which p66 ShcA stimulates cell death compared with proliferation are not yet clear, and its contributions to cellular dynamics are evidently more context-dependent than previously assumed.
Indeed breast cancer presents an opposing scenario in which p66 expression implies a favourable patient outcome, particularly in the context of the ratio of Tyr317-phosphorylated ShcA to the p66 isoform (pTyrShc/p66Shc) . Evidently, elevated levels of phosphorylated ShcA relative to the total amount of the p66 isoform are indicative of aggressive neoplasms and increased risk of relapse. Clinical predictions arising from the pTyrShc/p66Shc ratio are consistent with established prognostic indicators , and can likewise forecast disease recurrence following select pharmaceutical treatments .
To address the molecular contributions of Shc proteins to the events of mammary tumorigenesis, several transgenic mouse lines have been generated. In mammary epithelial cells, the PyV MT (polyoma virus middle T) oncogene rapidly induces metastatic tumours that develop in a manner reminiscent of human breast cancer progression . Mutation of the residue responsible for cellular transformation and Shc binding (Tyr250Phe PyV MT) has been found to reduce the oncogenic severity of the phenotype compared with that of PyV MT animals, thus implying that the Shc-recognition motif is important for promoting tumour development . An analogous study was performed to investigate the tumorigenic contributions of the Shc-binding site on the EGFR homologue Neu (ErbB-2; HER-2), a negative prognostic indicator that is up-regulated in 20–30% of human breast cancers . In the absence of five key autophosphorylation sites, the RTK loses its ability to transform cells. However, reconstituting the individual Shc site (Tyr1227) restores transformation, enhances signalling through MAPK and PI3K pathways, and promotes rapid development of multifocal tumours [139,140]. Although intriguing, the results of these transgenic studies should be interpreted with caution, as recognition sequence mutations likely affect more than one candidate binding partner. Indeed, all of the Shc family PTB domains recognize Asn-Pro-Xaa-pTyr motifs and could hypothetically co-ordinate with the PyV MT and Neu targets to mediate a response, thereby obfuscating the contributions of the individual adaptors.
The most compelling evidence of ShcA participation in breast tumorigenesis comes from mouse studies in which this locus was exclusively disrupted. In support of aforementioned findings, mammary-tissue-specific ablation of ShcA prevented Neu-mediated tumour development . Moreover, the contribution of the three ShcA CH1 phosphorylation sites to PyV MT-mediated oncogenesis was assessed using the ShcA Tyr3Phe KI strategy described above. Expression of this signalling-compromised ShcA substantially delayed tumour onset, reduced metastasis and revealed site-specific phenotypic proclivities. Although mutation of the equivalent Tyr317 site resulted in elevated tumour cell apoptosis, attenuated angiogenic signalling was observed in Tyr239/240Phe tumours, suggesting non-redundant roles for these ShcA sites in breast cancer progression .
Cancers of the nervous and neuroendocrine origin
Aberrant signalling through the GDNF receptor Ret is common to medullary and papillary thyroid carcinomas. It has been demonstrated that Tyr1062 co-ordinates both the PTB and SH2 domains of ShcA, and is important in Ret-mediated transformation . Accordingly, tyrosine-to-phenylalanine amino acid mutations of Ret Tyr1062 and Shc Tyr317 independently attenuate Ret-induced focus formation in vitro. ShcA Tyr317Phe was specifically found to reduce ERK1/2 activation, thereby suggesting the importance of the Shc adaptor and this CH1 tyrosine residue in facilitating the transformed phenotype . Another hallmark of Ret-induced oncogenesis is cellular de-differentiation, or the reduced expression of thyroid-specific proteins. This is likewise thought to occur through the Shc–Ras–MAPK pathway initiated at Ret Tyr1062, since removal of this site prevents down-regulation of thyroid markers normally observed in the presence of the oncogenic receptor . As noted previously, however, mutations rendered to the receptor do not provide definitive evidence of Shc involvement.
Notably, ShcC is also a physiological substrate of Ret that interacts through its PTB domain with Tyr1062 of the receptor . Its association with the Ret oncoprotein activates the PI3K/Akt pathway, and promotes survival in thyroid carcinoma cell lines when overexpressed . The discovery that ShcC is markedly up-regulated in human papillary thyroid cancers further suggests the relevance of this interaction , whereas observations that ShcA and C co-ordinate the same Ret receptor residue yet potentiate different signals raise the question of how their dynamics are maintained during tumorigenesis.
Perhaps nowhere is this question of Shc competition better illustrated than in the study of neuroblastoma, the most common and fatal childhood tumour that develops from sympathetic nervous system precursor cells and often presents paraspinally in the abdomen or chest . Intriguingly, it is capable of spontaneous regression in patients under 1 year of age and accumulating evidence implicates Shc proteins in cellular fate decisions. In addition to classical prognostic markers, ShcC overexpression has been found to predict poor clinical outcome [145,146]. mRNA profiling of ShcA, B and C in patient tumours revealed ubiquitous ShcA expression across all samples, although the levels were more robust in lower-grade tumours, whereas ShcB was minimal and mostly limited to earlier-stage cancers. In contrast, ShcC expression was low in early stages of the disease, and markedly elevated in advanced tumours, rendering it a strong indicator of neuroblastoma progression . This trend in ShcA/ShcC gene expression was independently confirmed by examining the respective proteins .
From the perspective of oncogenic mechanism, cultured neuroblastoma cell lines have been used to investigate the molecular determinants of the disease. Phosphorylated ShcC is evident in a number of lines, particularly those in which it associates with the constitutively active ALK (anaplastic lymphoma kinase) and becomes a substrate for this enzyme . Notably ShcA has also been detected in most of the neuroblastoma cell lines tested, although its phosphorylation pattern is reciprocal to that of ShcC . Analysis of neuroblastoma patient tissue likewise revealed that more advanced tumours were associated with elevated ShcC phosphorylation . The molecular relevance of ShcC post-translational modification was further investigated in cell lines expressing high levels of ALK, by introducing a ShcC mutant devoid of conserved CH1 region tyrosine residues. This was found to reduce activation of ERK1/2 and Akt, while impairing cellular motility, differentiation and survival . Most compelling, however, are the consequences of abrogating Shc expression entirely. Although RNAi (RNA interference)-mediated knockdown of ShcA impedes cellular growth rate, ShcC attenuation reduces in vivo tumorigenicity and promotes the appearance of neurite extensions and markers of cellular differentiation . Strikingly, neuronal maturation observed in the absence of ShcC has been attributed to ShcA through its sustained activation of ERK1/2. ShcA phosphorylation is accordingly reduced in the presence of ShcC and restored in its absence . This supports a hypothesis that the Shc adaptors have competing roles in neuroblastoma, and implies that ShcC may participate in aggressive phenotypes by maintaining cells in an undifferentiated state. The molecular processes by which it accomplishes this and impedes ShcA have not yet been elucidated. Paradoxically, this model directly opposes the putative mechanism of neuronal maturation during development, in which ShcC up-regulation is regarded as the switch from proliferation to differentiation.
It is nevertheless consistent with recent findings that ShcC is active in neurogenic regions of the adult brain where it promotes cell migration. Accordingly, ShcC is also believed to endow cancer stem cells of the high-grade CNS tumour, GBM (glioblastoma), with their infiltrative capacity . The adaptor is consistently overrepresented in the stem/progenitor subpopulation of GBM cells, regardless of its relative level in the tumour as a whole. ShcC silencing reduces the invasive properties and overall tumorigenicity of GBM-initiating cells. Intriguingly, however, the migratory influence of ShcC in GBM does not involve the Wnt/β-catenin pathway as it does in normal neuroblasts. Instead, the targets of ShcC appear more variable in the context of GBM . Collectively, these discoveries emphasize the importance of temporal and environmental factors in defining the outcome of Shc signalling. They likewise illustrate that cellular diversity in heterogeneous tumours can be a crucial determinant of disease progression.
CONCLUSIONS AND FUTURE PERSPECTIVES
In the 20 years since the first Shc protein was discovered, great strides have been made in determining the physiological roles, binding partners, molecular mechanisms, regulatory processes and oncogenic contributions of this adaptor family. Shc proteins are now appreciated to form vital links between the interstitial environment and the informatic, architectural and kinesthetic components of the cell, to integrate competing stimuli, and, in the specific case of p66 ShcA, to respond to stress and harness the redox potential of the mitochondrion to elicit apoptosis. These advances have likewise expanded the conceptual perspective of adaptor proteins in general. Although the recent discovery of ShcD offers more opportunity for this type of basic characterization, it appears overall that the reductionist chapter of the signalling discipline may be drawing to a close while larger more intractable questions remain to be addressed.
Competition or co-operation
One residual issue is the regulation of Shc protein dynamics when the adaptors are co-expressed. Owing to the considerable identity observed across the SH2 and PTB domains of this protein family, distinct members can bind the same ligands using equivalent domains that recognize the same target motifs . Yet there are well-established functional non-redundancies between different Shc adaptors and even between isoforms of the same designate. As noted by Nakamura et al. , ShcA and ShcC have equal affinity for the TrkA receptor, although the signalling output from each is distinct. These proteins co-express during development and in neuroblastoma tumours; however, they mediate entirely contradictory outcomes in the two scenarios. In a similar vein, the three ShcA isoforms bind Grb2 upon EGF stimulation, but only p66 fails to increase MAPK activation . Considering the unique opposing roles of the Shc proteins raises the question of how their interplay is determined in the cell and whether their relative levels are important for fine-tuning signals.
Given that Shc adaptors complex with a multitude of membrane receptors , they appear to represent a first point of signal convergence in the cell, wherein diverse stimuli activate unique receptors, which in turn recruit the common Shc protein. Obermeier et al.  postulated that the differences in receptor–transducer affinities could help to retain the molecular identity of distinct cascades. They found, however, that Shc exhibited similar binding characteristics with the Trk and EGF RTKs, despite their opposing roles in differentiation and proliferation respectively. It is not clear how the signal identity is retained in this context, although the concern is not unique to Shc signalling. Indeed, the ability of the ubiquitous MAPK pathway itself to transmit diverse, yet precise, signals has long baffled cell biologists. As noted by Brown and Sacks , the mechanism by which a particular stimulus provokes a targeted response through the MAPK pathway remains elusive, although contributing factors are beginning to surface.
Molecular design and function
Shc proteins are unique among adaptors in possessing both a PTB and an SH2 domain for pTyr recognition . The significance of this architecture is somewhat mysterious, although it suggests the potential for signal integration or bifurcation. Consistent with this prediction, ShcA-mediated migration and proliferation are observed to be mutually exclusive, active under different circumstances and require differential engagement of the domains .
Likewise the phenotypes exhibited by mice with ShcA mutant alleles depend on which region of the molecule is interrupted, as well as the nature of the tissue itself . Although these findings support the functional relevance of the PTB and SH2 domains, they do not provide evidence for their concomitant use. Moreover, isoforms of ShcA, B, C and D differ in the extent of their N-terminal CH2 regions, but, with the exception of p66 ShcA, the importance of this sequence is unknown.
These observations inspire an even more elemental question in the discipline: that of the overall relevance of Shc proteins to signalling cascades. Other adaptors, such as Nck and Crk, possess SH2 and SH3 domains that essentially convert a pTyr signal into a proline interaction , thereby engaging effectors that do not intrinsically contain phospho-specific domains. In contrast, Shc responds to, and perpetuates, pTyr signalling by binding a phospho motif of an activated receptor and recruiting Grb2 to its own CH1 pTyr residues. Luzi et al.  questioned the necessity of the Shc adaptor function, noting that Grb2 can localize directly to active RTKs without assistance. Although adaptor proteins can reinforce signals by expanding the repertoire of binding sites for downstream components , this remains a contentious issue for Shc as its two Grb2-recognition motifs appear to mediate distinct physiological outcomes [72,141], thereby precluding a simple additive model of amplification.
Indeed, Lai and Pawson's  discovery that ShcA sensitizes cells to low growth factor levels is among the most salient arguments that Shc adaptors constitute signal modifying components, rather than fundamental transduction architecture. As mentioned above, the MAPK specificity conundrum is concerned with the potentiation of unique signals through common cascades. The duration of ERK activation has proven important in determining outcome, as illustrated by the fate decisions occurring in PC12 phaeochromocytoma cells . Despite relying on the common MAPK pathway, EGF and nerve growth factor responses in these cells are vastly different, inducing proliferation and differentiation respectively. This effect has been attributed to temporal regulation of the MAPK cascade, wherein proliferation is a consequence of transient pathway activation, whereas differentiation is achieved only upon sustained ERK signalling . These findings raise the intriguing possibility that Shc proteins may in fact exist as signal amplitude or frequency modulators to help direct the nature of the cell response.
The emerging paradigm in the signalling discipline appears to shift emphasis from the details of molecular specificity governing interactions, to the ‘informatic’ specificity by which the meaning behind each signal is conserved, integrated and interpreted in the complex milieu of the cell. The challenge for future Shc investigations will be completing and compiling the simple characterizations of these proteins, and applying this knowledge to explore the dynamic contributions of this adaptor family to cell homoeostasis and disease.
anaplastic lymphoma kinase
adaptor-related protein 2
central nervous system
v-crk sarcoma virus CT10 oncogene homologue
c, cytochrome c
embryonic day 11.5
epidermal growth factor
glial-cell-line-derived neurotrophic factor
growth factor receptor-bound protein 2
Grb2-associated binding protein 1
IQ motif-containing GTPase-activating protein 1
v-yes-1 Yamaguchi sarcoma viral related oncogene homologue
mitogen-activated protein kinase
mouse embryonic fibroblast
programmed cell death
peptidylprolyl cis–trans isomerase, NIMA-interacting 1
protein kinase C
peripheral nervous system
protein phosphatase 2A
phosphatase and tensin homologue deleted on chromosome 10
protein tyrosine phosphatase
MT, polyoma virus middle T
rostral migratory stream
reactive oxygen species
receptor tyrosine kinase
Src family kinase
Src homology 2
Src homology 3
Src homology and collagen homology
Son of sevenless
ζ-chain (TCR)-associated protein kinase 70 kDa
This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) [grant number RG327372 (to N.J.)]. N.J. is also the recipient of an NSERC University Faculty Award and a Tier II Canada Research Chair. M.K.B.W. holds a Canadian Institutes of Health Research Vanier Canada Graduate Scholarship and the University of Guelph Brock Doctoral Scholarship.