![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Cell Signalling Laboratory, Biochemistry and Molecular Biology, School of Biomedical, Biomolecular and Chemical Sciences, University of Western Australia, Crawley, Australia,1 School of Molecular and Microbial Sciences, Institute for Molecular Bioscience and Special Research Centre for Functional and Applied Genomics, University of Queensland, Brisbane, Australia2
SUMMARY INTRODUCTION JNK-MEDIATED PHOSPHORYLATION OF THE ARCHETYPICAL SUBSTRATE c-Jun OTHER NUCLEAR SUBSTRATES OF JNK Transcription Factors as JNK Substrates The Jun family of transcription factors. The ATF family of transcription factors. JDP2 as a JNK substrate. Elk-1 as a JNK substrate. c-Myc as a JNK substrate. p53 as a JNK substrate. The NFAT family of transcription factors. The forkhead family of transcription factors. The STAT family of transcription factors. The Pax family of transcription factors. TCFß1 as a JNK substrate. Nuclear Hormone Receptors as JNK Substrates Additional Nuclear Proteins as JNK Substrates LINKS BETWEEN JNK ACTIVATION AND PROTEIN DEGRADATION Studies of T Cells Reveal that the E3 Ligase Itch is a JNK Substrate JNK Targeting of Transcription Factors for Degradation JNK PHOSPHORYLATION OF SCAFFOLD AND ADAPTOR PROTEINS IRS-1: a JNK Substrate That Allows Signal Integration between Stress and Metabolic Events JNK-Mediated Phosphorylation of Other Adaptor and Scaffold Proteins JNK-MEDIATED PHOSPHORYLATION OF MITOCHONDRIAL PROTEINS JNK-Mediated Phosphorylation of the Bcl2 Family JNK-Mediated Phosphorylation of Sab REGULATION OF OTHER PROTEIN KINASES THROUGH JNK-MEDIATED PHOSPHORYLATION Phosphorylation of MAPK-Activated Protein Kinases JNK-Mediated Phosphorylation of the Prosurvival Protein Kinase Akt REGULATION OF CELL MOVEMENT THROUGH JNK-MEDIATED PHOSPHORYLATION JNK-Mediated Phosphorylation of the Focal Adhesion Protein Paxillin JNK Interaction with and Phosphorylation of Microtubule-Associated and Intermediate Filament Proteins GENERAL PRINCIPLES OF JNK SIGNALING VIA PROTEIN PHOSPHORYLATION JNK Peptide Specificity Substrate Specificity of JNK Isoforms Interplay of Phosphorylation by JNKs and Other Protein Kinases Complexity of JNK Signaling CONCLUSIONS ACKNOWLEDGMENTS REFERENCES
| SUMMARY |
|---|
 Top NextReferences |
|---|
| INTRODUCTION |
|---|
|
TopPreviousNextReferences |
|---|
The c-Jun N-terminal kinases (JNKs) are members of a larger group of serine/threonine (Ser/Thr) protein kinases known as the mitogen-activated protein kinase (MAPK) family. The MAPK family is one subgroup of the CMGC class of protein kinases (where CMGC is the class name derived from the major kinase members of this class, namely, cyclin-dependent kinases [CDKs], MAPKs, glycogen synthase kinase 3 [GSK3], and casein kinase 2-related protein kinases). Within the classification of all protein kinases, the CMGC class represents one of the three major protein kinase classes, in addition to classical Ser/Thr kinases and tyrosine (Tyr) kinases (122). The JNKs act within a protein kinase cascade (Fig. 1A and B). They are themselves activated by dual phosphorylation, by the MAPK kinases MKK4 and MKK7, on a specific Thr and a specific Tyr in a typical Thr-X-Tyr motif within their "activation/phosphorylation loop" sequences (for a review, see reference 73).
|
(TNF-) and interleukin-1 (for a review, see reference 81). Furthermore, the JNK pathway is activated in the innate immune response following the activation of various members of the Toll-like receptor family by invading pathogens (e.g., see references 13, 15, 178, 182, 216, 259, and 272) (Fig. 1A). The JNK pathway therefore appears to act as a critical intermediate in signaling in the immune system (81). As also shown in Fig. 1B, there are increasing links between the endoplasmic reticulum (ER) stress response and JNK activation (for a review, see reference 317). This provides at least one mechanism of activation of JNKs following the sensing of internal stress events, such as protein misfolding. There is also an increasing body of literature showing that JNK activation follows bacterial, fungal, prion, parasitic, or viral infections. Under these circumstances, JNK activation may influence important cellular consequences, such as alterations in gene expression (1, 53, 59, 162, 167, 176, 199, 294, 325, 326, 346), cell death (58, 89, 137, 139, 169, 193, 243, 293), viral replication, persistent infection or progeny release (215, 224, 251, 260), or altered cellular proliferation (178). The exact mechanism of JNK activation under each of these circumstances remains to be elucidated fully, although there may be involvement of Toll-like receptors, direct pathway modulation through interaction with upstream protein regulators, or the activation following an ER stress response (79, 87, 110, 124, 143, 191, 253, 261, 279, 294, 312).
Originally identified as stress-activated protein kinases (SAPKs) in the livers of cycloheximide-challenged rats (177), the subsequent purification, cloning, and naming of the JNKs have emphasized their ability to phosphorylate and activate the transcription factor c-Jun (77, 222, 257). The JNK-mediated phosphorylation of both Ser63 and Ser73 within the transactivation domain of c-Jun (Table 1) potentiates its transcriptional activity through the loss of repression mediated by an inhibitory complex associated with histone deacetylase 3 (316). Other events that follow c-Jun phosphorylation include its increased interaction with other binding partners, such as the transcription factor TCF4 or the E3 ubiquitin ligase Fbw7 (240, 241). The importance of c-Jun phosphorylation has been emphasized in studies of transgenic mice expressing the c-Jun mutant c-Jun Ser63
Ala Ser73Ala, which lacks the two major sites for phosphorylation by JNK (22, 23, 72, 130, 150, 311). Increasing attention has been directed towards JNK as an activator of a wide range of c-Jun-dependent events, including apoptotic cell death and oncogenic transformation (22, 23, 188).
|
Specifically, four splice forms arise from the Jnk1 gene, four arise from the Jnk2 gene, and two arise from the Jnk3 gene (Fig. 1C). While JNK1 and JNK2 are expressed in a variety of tissues, JNK3 expression is restricted primarily to the brain, heart, and testes (207, 225). This tissue-specific distribution, particularly for JNK3 expression, has led to the idea that different isoforms may perform different cellular roles. This has been explored further through studies in which the effects of deletions of the Jnk genes alone and in combination have been evaluated (for a review, see reference 30). Later sections of this review explore this isoform specificity further.
The primary structure of JNK11 is illustrated in Fig. 1D, which highlights the conserved features of the protein kinase domains of the JNKs. The subdomain numbering system shown in Fig. 1D denotes conserved sequence regions I to XI, originally defined by Hanks and colleagues for all protein kinases (123). Crystal structures were subsequently determined for JNK3 in the presence of an ATP analogue (327) and several small-molecule inhibitors (273) and for JNK1 in the presence of small-molecule ATP-competitive inhibitors (195, 290) and an inhibitory peptide from the JNK-interacting protein 1 (JIP1), with and without a small-molecule ATP-competitive inhibitor (129). Examples of these structures are shown in Fig. 2.
|
-helices (residues 113 to 337 in JNK1 and residues 150 to 374 in JNK3). The JNK C-terminal domain has an insertion typical of the MAPKs, and this is 12 residues longer in JNKs than in the related MAPKs extracellular signal-regulated kinase 2 (ERK2) and p38. These two domains are connected by two peptide segments, and based on structures of other protein kinases in complex with peptide substrates (34, 140, 171, 198, 335), the peptide substrates for JNK are expected to bind into the groove between the two lobes of JNK. The ATP molecule also binds near the domain interface. All of these structures have been determined for the inactive nonphosphorylated forms of JNK. It is expected that subtle but important changes in structure will accompany activation, as seen for the related MAPK ERK2 (183). The nonphosphorylated JNK structures are inactive due to the misalignment of the catalytic residues accompanying the relative rotation of the two domains and the obstruction of the active site by the "activation loop." In contrast, the ATP-binding site is well formed in JNK structures. In this review, an overview of the substrates of JNK is presented, beginning with a consideration of the nuclear substrates of JNK that have now been described, in addition to c-Jun. We consider the phosphorylation of 26 nuclear substrates of JNKs (Table 1) and discuss how phosphorylation alters their functions. Many of these nuclear proteins are transcription factors, and therefore their phosphorylation by the JNKs can mediate actions via a direct link to changes in gene expression following the exposure of cells to a range of cytokines and stress stimuli. However, JNK substrates in other cellular compartments have also been described, and the effects of phosphorylation of an additional 26 nonnuclear substrates of JNKs are discussed in turn (see Table 3). These substrates provide a link to a wide range of cellular functions, including cell death and cell movement, as well as allowing for modulation of other signaling events in the cell. We summarize the known effects of phosphorylation on these nuclear and nonnuclear substrates in Fig. 3. This summary shows that JNK-mediated phosphorylation may either enhance or inhibit the activities of its substrates and that, in some cases, the phosphorylation-dependent changes are more complex and involve changes in protein binding and/or localization in the cell. Therefore, the functional effects of phosphorylation following JNK activation must always be specifically tested. Lastly, the determinants of the substrate specificity of JNKs are examined in greater detail, and the effects of JNK-mediated phosphorylation are discussed within the broader context of signal transduction cross talk, integration, and diversification. This analysis reveals the complexities of signal transduction and the new challenges faced in evaluation of signal transduction pathways and their consequent effects.
|
|
| JNK-MEDIATED PHOSPHORYLATION OF THE ARCHETYPICAL SUBSTRATE c-Jun |
|---|
|
TopPreviousNextReferences |
|---|
The initial studies on JNK-mediated phosphorylation of c-Jun also revealed a requirement for amino acid sequences, known as the 
domain or the JNK-binding domain (JBD), distant from the amino acids to be phosphorylated (4, 66, 77, 152, 210). The JBD sequence within c-Jun is shown in Table 2, and a general schematic diagram illustrating the relative positions of the phosphorylated residues in relation to the JBDs of c-Jun and other substrates is shown in Fig. 4. These distant targeting domains mediate interactions of other MAPKs with their substrates, upstream activators, phosphatases, and scaffold proteins (295) and thus are more generally termed common docking (CD) domains. The use of docking domains by MAPKs can enhance the efficiency and specificity of substrate phosphorylation (19, 99, 148, 276). Furthermore, small peptides making up the JBD of c-Jun inhibit JNK activity (4). However, as we describe in this review, docking sequences for JNK have not yet been identified for all substrates of JNKs. This raises the possibility that either these substrates are recognized independently of a docking site region or their docking domains do not conform to the sequences currently recognized as forming a JBD. The docking domains have significant implications for the substrate specificity of JNKs, as discussed in later sections of this review.
|
|
40 substrates. This calculation does not take into consideration the idea that some proteins will be phosphorylated at multiple sites by different protein kinases or that many different protein kinases may phosphorylate the same site on one substrate. The MAPKs are likely to be consistent with this calculation, with an initial proteomic study identifying 25 phosphoproteins following ERK activation (189). Although not all of these phosphoproteins may be direct substrates of ERK, this does confirm the complexity of signaling events downstream of MAPKs such as the ERKs. | OTHER NUCLEAR SUBSTRATES OF JNK |
|---|
|
TopPreviousNextReferences |
|---|
|
-JunD, which lacks the N-terminal menin interaction domain, suggesting that both JunD splice forms are under the control of JNK phosphorylation (339). The C-terminal part of JunD is also involved in interactions with ERKs, thus allowing JunD regulation by both the ERK and JNK pathways, and indeed, the N-terminal sites in JunD are phosphorylated by both ERK and JNK (306, 309). The sequences surrounding JunD Ser90 and Ser100 are analogous to sequences surrounding Ser63 and Ser73 of c-Jun (Table 1). Additionally, the sequence surrounding JunD Thr117 shows similarity to the sequences surrounding c-Jun Thr91, which is also phosphorylated by JNK (230). Note that JNK binding to the JunD JBD has been shown to be poor compared with JNK binding to the JBD of c-Jun or JunB (117, 152). This is reinforced by the finding that c-Jun is a more efficient substrate than JunD in vitro for four different JNK isoforms (the jnk1 splice form JNK11, the jnk2 splice forms JNK22 and JNK2ß2, and the the jnk3 splice form JNK31) (339). This study suggests that protein kinases from the JNK family exhibit considerable specificities in substrate docking and phosphorylation, even for related transcription factor substrates.
Although the related transcription factor JunB was initially not considered a JNK substrate because it lacks serine residues in the appropriate consensus sequences (Ser74 and Ser84 in murine JunB [accession number NP_032442], within the sequence G71QGS74DTGASLKLAS84TELERL90) (152), subsequent studies showed that JunB is phosphorylated by JNK at two closely spaced threonine residues, Thr102 and Thr104 (Table 1) (1). Like in the case of c-Jun, these phosphorylation sites are C-terminal to a conserved JBD sequence (Table 2 and Fig. 4). Furthermore, this phosphorylation of JunB would appear to potentiate the transcriptional activity of JunB (Fig. 3), as seen when a Thr102Glu/Thr104Glu mutant (designed to mimic the JNK-mediated phosphorylation events) showed enhanced ability to synergize with c-Maf in transcriptional activation of the interleukin-4 promoter (1). Thus, this has implicated the JunB protein in T-cell development and in directing Th2 differentiation.
The ATF family of transcription factors. Together with the transcription factors of the Jun family, the Fos and ATF2 families of bZIP transcription factors also form part of the transcription factor complexes known as the activator protein 1 (AP-1) family (188). Within this transcription factor family, there are additional substrates for JNK. While most evidence supports the presence of a Fos kinase that is not related to JNK (75, 288), ATF2 is recognized as a JNK substrate (118, 197). As seen for JNK-mediated phosphorylation of c-Jun, JNK-mediated phosphorylation of ATF2 is directed to two closely spaced residues, namely, Thr69 and Thr71, in its N-terminal transactivation domain (118, 197) (Tables 1 and 2). In a manner that therefore shows similarity to c-Jun regulation by JNK, the JNK-mediated phosphorylation of ATF2 enhances its transcriptional activity (Fig. 3) (118, 197, 304).
Knowledge of the regions of ATF2 interacting with and phosphorylated by JNK has led to the development of an ATF2-derived protein fragment (ATF250-100). This peptide, when delivered to cells, alters the balance between c-Jun and ATF2 transcriptional activities, leading to the attenuation of ATF2 activity and the induction of c-Jun activity as well as the sensitization of cultured melanoma cells to chemotherapeutic agents (26). These observations, together with recent studies on the substrate-binding characteristics of ERK2 (226), suggest that appropriate substrate-derived peptides will allow a subset of protein kinase substrates to be selectively inhibited. This is a significant advance over the ATP-competitive inhibitors of kinases currently in use. ATP-competitive inhibitors would be expected to inhibit the phosphorylation of all protein substrates for a particular protein kinase and may not be specific for a particular kinase due to the difficulty in discriminating between the conserved ATP-binding sites of various protein kinases (for a review, see reference 92). Thus, a greater understanding of the range of the possible intracellular JNK substrates is critical in the development of new approaches to achieve substrate-selective and specific inhibition of JNK.
There may be additional complex relationships between JNK and the ATF family of transcription factors. While one study has shown the importance of phosphorylation of residues Thr51 and Thr53 in the N-terminal activation of ATFa for the transcriptional activation of this specific transcription factor, it appeared that ATFa was not a direct substrate for JNK2 (74). Instead, the N-terminal domain of ATFa served as a docking site for JNK (29), allowing ATFa-associated partners, such as JunD, to then be phosphorylated by JNK (74). This relationship emphasizes the possibilities of trans-phosphorylation events in the regulation of transcription factor complexes.
JDP2 as a JNK substrate. JNK can also phosphorylate another binding partner of c-Jun, i.e., Jun dimerization protein 2 (JDP2). JDP2 is a basic leucine zipper transcription factor family member that interacts with c-Jun as well as the transcription factors ATF2 and CCAAT/enhancer-binding protein gamma (156). The site of phosphorylation of JDP2 has been mapped to Thr148 (156) and a JBD identified in subsequent studies (155) (Tables 1 and 2). It is important that although the JDP2 sequence apparently contains a classic JBD consensus sequence within its leucine zipper domain (i.e., K136NEKQHLIYMLNLH149 [residues of the consensus are shown in bold]), the site of interaction was mapped to the JDP2 C-terminal region beyond residue 153 (155). Indeed, a 14-amino-acid fragment derived from the JDP2 sequence (i.e., JDP2150-163), when added to the transcription factor ATF3, which is usually not a JNK substrate, facilitated JNK phosphorylation of ATF3; this sequence alone was therefore sufficient for JNK interaction (155). In contrast to c-Jun-binding partners such as c-Fos or ATF2, JDP2 acts as a repressor at the AP-1 site, and it also inhibits Ras-driven transformation of NIH 3T3 cells and suppresses tumor formation in vivo in a PC3 cell xenograft model (128). The functional consequences of JDP2 phosphorylation by JNK remain to be elucidated (Fig. 4).
Elk-1 as a JNK substrate. JNK also phosphorylates a number of other transcription factors that do not form part of the AP-1 complex. JNK phosphorylates the Ets domain-containing transcription factor Elk-1 on Ser383 and Ser389 in its C-terminal transactivation domain (Table 1); these same residues are also phosphorylated by the ERK MAPKs (47, 132, 206, 321). This phosphorylation of Elk-1 increases its complex formation with the serum response factor and, in this way, increases transcriptional activity (Fig. 3) (107, 108, 321). The binding of all 10 JNK isoforms (i.e., four JNK1 splice forms, four JNK2 splice forms, and two JNK3 splice forms) to Elk-1 appeared considerably weaker than the binding of these JNK proteins to either c-Jun or ATF2 (117, 337) (see Table 2 for the Elk-1 JBD sequence). Although the exact residues within the JBD required for binding either JNKs or ERKs do differ, p38 MAPK phosphorylation of Elk-1 does not appear to require an intact JBD (337). These differences in specificity determinants can thus play a pivotal role in producing unique nuclear responses following activation of the different MAPK pathways.
The requirement for specific docking domains and subsequent kinase-specific phosphorylation events is further illustrated by the observation that JNK binds to the ternary complex factor Net (also known as Elk-3) via a binding motif (Table 2) that is distinct from that bound by ERK or p38 (86). This JNK-mediated phosphorylation regulates nuclear export of Net and inhibits Net-mediated effects (Fig. 3). The mechanism of transcription factor inhibition can be explained by the actions of JNK to phosphorylate four Ser residues of Net, namely, Ser239, Ser245, Ser249, and Ser252 (Table 1), within the Net nuclear export box, enhancing nuclear export. The docking domain also shows homology to other JBD sequences (Table 2). Interestingly, the Net residues phosphorylated by JNK are distinct from those phosphorylated by ERK or p38, and the actions are distinct from those of JNK-mediated effects to increase the transcriptional activity of the related Ets family transcription factor Elk-1 (Fig. 3). Heat shock factor 1 (HSF-1) is another transcription factor that is also inactivated following its phosphorylation by JNK (65). In this case, transcriptional activity is decreased, rather than inhibition of its actions, requiring its changes in nuclear localization. The site of phosphorylation in HSF-1 appears to be Ser363 (Table 1), one of five Ser/Thr residues within S/T-P motifs. This phosphorylation also depended on an interaction motif showing similarity to the JBD of c-Jun (Table 2) (65). This inhibition by JNK-mediated phosphorylation provides a mechanism, in addition to ERK-mediated phosphorylation or interaction with a range of heat shock proteins, to downregulate the actions of this transcription factor (65).
c-Myc as a JNK substrate.
Other transcription factors have also been investigated as mediators of the nuclear actions of JNK. In c-Myc, Ser61 and Ser71 have been shown to be phosphorylated by JNK1 and, to a lesser extent, by JNK2 and JNK3 (246). This phosphorylation increases c-Myc-mediated apoptosis (Fig. 3) (246). Interestingly, the related transcription factors s-Myc and Max are not JNK substrates (246). The region of c-myc involved in the interaction with JNK was mapped to residues 1 to 262 of c-Myc (246). This was subsequently confirmed in an independent study that blocked the binding of JNK to c-Myc through the use of a peptide corresponding to residues 127 to 189 of c-Myc (6). A domain-like region of c-Myc was identified within residues 171 to 187 (6) (Table 2).
p53 as a JNK substrate. The p53 tumor suppressor protein is another transcription factor that is phosphorylated by many protein kinases, including JNK (3). The interaction between JNK and p53 has been mapped to amino acids 97 to 117 of p53 (Table 2) by the demonstration that a synthetic peptide corresponding to these p53 residues prevented phosphorylation of p53 and its interaction with JNK (3, 94). The overexpression of the JNK pathway upstream kinase MEKK1 (Fig. 1) has been shown to increase p53 stability and transcriptional activity (95), and the p53 residue phosphorylated by JNK was subsequently mapped to Thr81 (41) (Table 1). Phosphorylation of this residue appeared to be critical for p53 stabilization and conferred its transcriptional activity and ability to elicit apoptosis; in the absence of JNK expression or JNK-mediated phosphorylation, p53 was inactive (41). Thus, JNK can be considered an activator of p53 actions (Fig. 3). Interestingly, a JBD-like sequence (I80-F-K-E-Q-G-L-T-L-P-L-Y91, with clear similarity to the I33-L-K-Q-S-M-T-L-N-L-A43 sequence in c-Jun) has also been described for the tumor suppressor protein BRCA2 (221). However, the BRCA2 protein has not been shown to interact with JNK (209), and other protein kinases have been suggested to phosphorylate BRCA2 (220). This emphasizes that all potential JBD-like sequences require experimental validation before any link with JNK-dependent signaling can be suggested. Additional controversy surrounds the role of JNK-mediated phosphorylation of p53, as the sites of phosphorylation are only present in the rat sequence and not conserved in either mouse or human sequences. This raises questions on the importance of these phosphorylation sites in p53 function.
The NFAT family of transcription factors. Based on the concept that distinct docking domains mediate JNK binding to its substrates, JNK1 has been used as bait in a yeast two-hybrid screen of a mouse embryo cDNA library in a search for novel interacting partners and substrates (57). This screen revealed an interaction of JNK1 with the transcription factor nuclear factor of activated T cells c3' (NFATc3; also known as NFAT4 or NFATx); the interaction was also confirmed in mammalian cells (57). NFATs are calcium-sensitive transcription factors that have been shown to be critical regulators of T-cell development, and in addition to regulating other differentiation programs, they function in a range of tissues, being involved in skeletal muscle differentiation, cardiac valve development, and osteoclast differentiation (135, 201). The functional utility of the NFATs has been explained by their complex mechanisms of regulation and their ability to integrate calcium signaling with other signaling pathways (for a review, see reference 201). The residues mediating interaction with JNK were mapped to NFATc3 residues 162 to 207, with Ser163 and Ser165 as the sites of phosphorylation (57) (Table 1; Fig. 4). In contrast to the positive effects of JNK in enhancing transcription, as seen in many of the preceding examples of transcription factor substrates of JNK, this JNK-mediated phosphorylation of NFATc3 again resulted in the nuclear exclusion of this transcription factor (57). Thus, the activation of the JNK pathway also antagonizes the actions of NFATc3 (Fig. 3).
JNK also phosphorylates the related NFAT family member NFATc1 on Ser117 and Ser172, and this requires the presence of a JBD within residues 126 to 138 (56) (Tables 1 and 2 and Fig. 4). A comparison of the phosphorylation site sequences suggests that NFATc1 Ser172 is equivalent to NFATc3 Ser165 (Tables 1 and 2). Both NFATc1 phosphorylation sites are close to the domain that interacts with the calcium-dependent phosphatase calcineurin (56). Calcineurin can preferentially dephosphorylate Ser172 in vitro, while the phosphorylation of Ser117 was shown to be critical in regulating the targeting of calcineurin to NFATc1 (56). Thus, phosphorylation of NFATc1 by JNK inhibits the interaction with calcineurin, thus blocking its nuclear entry and providing a molecular mechanism for the observed increased nuclear localization of NFATc1 in the T cells of jnk1–/– mice (83). This antagonism of NFAT signaling by JNK activation is also seen in other systems, such as the heart, where JNK activation negatively regulates NFATc3 activation and inhibition of JNK enhances NFAT signaling, with subsequent enhanced hypertrophic growth (192). These examples again illustrate the importance of JNK in mediating the inhibition of transcriptional events, in addition to its more widely acknowledged role as a positive mediator of signaling.
In contrast to this negative regulation of NFAT signaling, the phosphorylation of a different NFAT transcription family member, NFATc2, by JNK stimulates its transcriptional activity (248) (Fig. 3). The effects of JNK required Thr116 (Table 1) within the docking site for calcineurin in the NFATc2 regulatory domain. No effect of JNK activation on the subcellular localization of NFATc2 was observed (248). Importantly, these different effects of JNK on the different NFAT isoforms highlight the danger of studying the effects of JNK on one member of a transcription factor family and then extrapolating these effects to other closely related members of the same family. Instead, functional testing appears to be required in each case.
The forkhead family of transcription factors.
The forkhead family member FOXO4 has also been shown to be phosphorylated following the exposure of cells to TNF- or oxidative stress in the form of hydrogen peroxide (90). The phosphorylation of the FOXO family has come under increasing attention as an event downstream of activation of the prosurvival protein kinase Akt (for a review, see reference 39). For example, the Akt-mediated phosphorylation of FOXO3a decreases its transcriptional activity because phosphorylated FOXO3a is bound by cytosolic 14-3-3 proteins and thus sequestered in the cytosol. This prevents upregulation of the transcription of enzymes such as catalase and Mn-dependent superoxide dismutase and thus changes the cellular levels of reactive oxygen species (for a review, see reference 40). In contrast, a role for JNKs was suggested more recently, based on the observation that FOXO4 was no longer phosphorylated in cells deficient in both JNK1 and JNK2 (i.e., jnk1–/– jnk2–/– cells) following their exposure to hydrogen peroxide (90). JNK-dependent phosphorylation enhanced FOXO4 transcriptional activity rather than changing its binding to the cytosolic 14-3-3 proteins (90) (Fig. 3). These results reveal a point of cross talk between JNK and other signal transduction pathways. Further points of cross talk will be discussed later in this review.
The positive regulation of FOXO activity in mammalian cells (90) is consistent with studies of the model organism Caenorhabditis elegans (247). In this organism, JNK interacts with and phosphorylates the FOXO homologue DAF-16. Although the phosphorylation sites appeared to be within the N-terminal region of DAF-16 (residues 83 to 307), the residues required for the interaction and for phosphorylation have not yet been identified (247). The significance of this phosphorylation lies in the consequences of DAF-16 regulation. Specifically, the negative regulation of DAF-16, as might result from enhanced signaling from the insulin-like growth factor receptor, has been associated with a shortened life span (for a review, see reference 145). Thus, JNK activation and subsequent DAF-16 phosphorylation and activation resulted in an increased life span, presumably through the upregulation of genes promoting resistance to stress (247).
Similar results with FOXO regulation have also been shown in Drosophila melanogaster, with dfoxo required for JNK-mediated life span extension (313). Thus, in this range of different systems, the FOXO forkhead transcription factors provide a point of convergence in signaling by the insulin-like growth factor and JNK signaling cascades. It will therefore be important to determine which of the mammalian FOXO family members (for a review, see reference 27) are regulated by JNK-mediated phosphorylation. Recent evidence suggests that the JNK pathway is involved in the regulation of the nuclear translocation of FOXO1, with JNK phosphorylation leading to changes in the localization of the transcription factor PDX-1, impairing PDX-1 function, as observed in pancreatic ß cells in diabetes (157). However, it is not yet possible to discount a role for JNK as a negative regulator of more membrane-proximal signaling events, such as the phosphorylation of the adaptor protein insulin receptor substrate 1 (IRS-1), as discussed below, which would alter FOXO1 subcellular distribution through the negative impact on Akt signaling. Again, this emphasizes the likely contributions of multiple signaling pathways with extensive opportunities for cross talk and control.
The STAT family of transcription factors.
Other transcription factors are subject to control by multiple phosphorylation events. One example is the signal transducer and activator of transcription (STAT) family, which has been implicated downstream of signaling by both cytokine and growth factor receptors and whose members have been considered critical growth regulators (for a review, see reference 44). The phosphorylation of Tyr705 of STAT3 is mediated by the JAK family of tyrosine kinases, whereas JNK also phosphorylates STAT3 on Ser727 (Table 1) (349). Both phosphorylation events are required for full transcriptional activation of STAT3 (Fig. 3). Similarly, the activation of the JNK pathway downstream of protein kinase C- can also result in the phosphorylation and activation of another STAT family member, STAT1 (Fig. 4). This activation requires phosphorylation of Ser727 (352) (Table 1). The involvement of both STAT1 and STAT3 as mediators in a range of diseases, including cancer, inflammatory disease, and ischemia/reperfusion injury (for reviews, see references 255, 284, and 302), warrants further evaluation of the contributions of JNKs to their initiation and development, as JNKs may therefore play critical regulatory roles.
The Pax family of transcription factors. In addition to the ability to modulate responses to stress through the phosphorylation of a range of transcription factors involved in various aspects of cell growth, as described above, JNKs may also phosphorylate additional transcription factors involved in development. The Pax family of transcription factors is required for the embryonic development of a range of tissues (for a review, see reference 55). Pax2 is required for kidney development as well as for development of the inner ear and the optic cup and has been shown to be a substrate for JNK. Pax2 can be isolated in a complex with the JNK-interacting protein JIP1 (43). Phosphorylation enhances Pax2 transcriptional activity (43) (Fig. 3). Although the phosphorylation site(s) in Pax2 was not identified, it is possible to predict possible phosphorylation sites (Table 1) by using the consensus sequence derived from other JNK substrates (329). It will now be interesting to map the JNK phosphorylation sites in Pax2 and compare these with the predicted sites. In addition, it will be critical to identify possible JNK interaction motifs and to evaluate whether other Pax family members might be regulated by the actions of JNK.
TCFß1 as a JNK substrate. Other transcription factors, such as the POU domain-containing protein T-cell factor ß1 (TCFß1), a key regulator during development and lymphocyte activation, also appear to be substrates for JNK, with TCFß1 being phosphorylated at both the Ser232 and Thr242 residues (Table 1) within its DNA-binding domain (154). This phosphorylation increases the binding of TCFß1 to DNA and thus likely mediates an increase in transcriptional actions following JNK activation in T cells (154) (Fig. 3). Therefore, in many of the examples discussed thus far, phosphorylation by JNK increases the activities of a range of transcription factor proteins (Fig. 3).
1 (PPAR-1) is a substrate for JNK (45). PPARs bind to response elements in complex with the retinoic acid receptor and activate transcription in response to a range of endogenous ligands, such as fatty acids and arachidonic acid metabolites, or foreign ligands, such as the antidiabetic drugs thiazolidinediones. JNK phosphorylates Ser82 (Table 1) in the transactivation domain of PPAR-1, and this decreases its transcriptional activity (Fig. 3) (45). This phosphorylation may contribute to the development of insulin resistance when adipose tissue releases TNF- and then signaling via the JNK pathway suppresses PPAR-1 activity in vivo.
The glucocorticoid receptor has also been shown to be negatively regulated through its phosphorylation by JNK (266) (Fig. 3). The major site of phosphorylation of the rat glucocorticoid receptor by JNK in vitro was mapped to Ser246 (Table 1). This site was confirmed within cultured cells (266) and was one of four major phosphorylation sites within the N-terminal transcriptional regulatory region (175). Thus, this JNK-mediated inhibition would decrease the actions of glucocorticoids to induce differentiation, regulate gluconeogenesis, and suppress inflammation. Further studies using a form of the human glucocorticoid receptor mutated to prevent phosphorylation by JNK (i.e., Ser226Ala mutant) suggested that JNK-mediated phosphorylation of the glucocorticoid receptor enhanced nuclear export, apparently by a leptomycin B-sensitive, exportin/CRM1-dependent mechanism (147). This provides an additional mechanism for inhibition of its effects in cells and is similar to the enhancement of nuclear export of the transcription factors Net and NFATc3 described above. The p38 MAPKs were also shown more recently to inhibit glucocorticoid receptor actions, in this case by indirectly targeting the ligand-binding domain (289). This observation demonstrates that there are multiple mechanisms of inhibition of the transcriptional activity of this nuclear hormone receptor.
Conversely, glucocorticoids also inhibit the actions of the JNK pathway (115). This has been attributed to the ability of the glucocorticoid receptor to interact directly with JNK and to inhibit its activity (37). An interaction motif showing similarities to the motif in c-Jun was seen in the glucocorticoid receptor (37) (Table 1). Furthermore, this JBD-like sequence was shown to be required for glucocorticoid-induced nuclear localization of JNK, suggesting a role for the glucocorticoid receptor in shuttling JNK to the nucleus (37). Interestingly, the nuclear translocation of JNK also increased JNK binding to the AP-1-associated response elements in the c-jun gene, and this binding of inactive JNK may maintain repression of AP-1-dependent transcription (37). Clearly, these observations show that all potential JBD-like sequences require experimental validation before any conclusions about their roles in the regulation of JNK-dependent signaling can be drawn.
The retinoic acid receptors RXR and RAR have also been shown to be substrates of JNK, providing one mechanism to explain how stress can inhibit retinoid signaling (2, 181, 283). Interestingly, RXR is a substrate for JNK as well as the dual-specificity kinase MKK4/SEK1, the latter of which is usually considered a JNK activator only (181). MKK4/SEK1-mediated phosphorylation of RXR inhibited retinoid-mediated transcriptional signaling, providing some of the first evidence that MKK4/SEK1 can initiate effects independent of its actions on JNK activation (181). The sites of phosphorylation for MKK4/SEK1 were in domains distinct from those phosphorylated by JNK (181). Although the residues phosphorylated by JNK were not identified, mutation of a single tyrosine in RXR (Tyr249) decreased phosphorylation and abrogated the ability of MKK4/SEK1 to suppress transcriptional activity (181). This is in contrast to the actions of JNKs as Ser/Thr kinases and highlights the possibility that there may be many control points for regulation of retinoid receptor activities.
Initial reports suggested that JNK mediated the phosphorylation of RXR at residues Ser61, Ser75, Thr87, and Ser265 (Table 1) (2). This phosphorylation did not appear to affect the transactivation properties of either RXR homodimers or RXR/RAR heterodimers (2). More recently, the phosphorylation of the three N-terminal residues within the transactivation domain has been shown to be required for the maximal transcriptional activity that results from the cooperation of RXR and its partner RAR (106). In addition, the importance of the phosphorylation of Ser265 was also highlighted more recently; this residue lies outside the classic transactivation domain of RXR in the omega loop of the ligand-binding domain (35). This phosphorylation enhanced the expression of some retinoic acid target genes but decreased the expression of others (35). Thus, JNK-mediated phosphorylation affects RXR function by modulating its transcriptional effects (Fig. 3). This altered regulation of retinoic acid target genes may thus have important consequences for retinoic acid actions in the cell, as seen for the cooperation of retinoic acid and arsenic trioxide in apoptosis through the JNK-mediated phosphorylation of RXR (298).
JNK has also been implicated in the inhibition of RXR transactivation when cells are exposed to stress in the form of arsenic trioxide (204). Mutational analysis has suggested the requirement for Ser32, suggesting this as the novel Ser target for JNK involved in the inhibition of nuclear receptor function (Table 1) (204). The mechanism of this inhibition requires further evaluation, as direct effects on stability, dimer formation, or interaction with DNA have not been observed (204). The JNK-mediated phosphorylation of RAR was recently mapped to residues Thr181, Ser445, and Ser461 (Table 1) (283). This phosphorylation results in the inhibition of RAR through induced proteasomal degradation of RAR (Fig. 3) (283). Specifically, when a RAR mutant lacking these JNK phosphorylation sites was expressed in cells, UV irradiation did not lead to decreases in RAR levels. Conversely, inhibition of JNK in a human lung cancer cell line by the use of the JNK inhibitor SP600125 enhanced RAR levels. This link between the JNK signaling pathway and degradation of specific proteins is explored further in the following section; for example, the E3 ligase Itch has been shown to be a specific substrate for JNK.
Other mechanisms, such as alterations in nuclear export, remain to be investigated, particularly following the observation that the orphan nuclear receptor family member nur77 is phosphorylated in its N terminus by JNK (174). This JNK-mediated phosphorylation of nur77 is involved, in conjunction with phosphorylation by Akt, in the modulation of nur77 functions through regulation of nur77 nuclear export (Fig. 3) (121). The exact residues in nur77 that are phosphorylated by JNK remain to be identified (Table 1). JNK-mediated phosphorylation of Ser650 of the androgen receptor (Table 1) was also recently shown to increase its nuclear export to decrease its transcriptional activity (Fig. 3) (109). Thus, the regulation of subcellular localization by JNK-mediated phosphorylation can be a critical control mechanism in signaling downstream of JNKs.
Gly and Leu168Ala) to allow its use of the ATP analogue N6-(2-phenythyl)-ATP. Following expression of this JNK2 mutant in 293T cells and its activation following exposure to UV irradiation, incubation of this kinase in the presence of protein extracts prepared from 293T cells and radiolabeled N6-(2-phenythyl)-ATP allowed the visualization of radiolabeled proteins separated by two-dimensional gel electrophoresis. Tandem nanoflow electrospray mass spectrometry of silver-stained spots that corresponded to phosphorylated proteins identified three peptide sequences, with each corresponding to a peptide from hnRNP-K (119).
Mutational analysis of the hnRNP-K protein has suggested that JNK phosphorylates two sites, Ser216 and Ser353 (Table 1), although a JBD has not been identified (119). JNK phosphorylation of hnRNP-K did not affect inhibition of RNA translation by hnRNP-K (120), but it was shown to enhance the ability of hnRNP-K to drive AP-1-dependent reporter gene expression (Fig. 3) (119). It will be critical to identify how the JNK-mediated phosphorylation of hnRNP-K contributes to its functions in cells, particularly when there are reports that phosphorylation by other protein kinases, such as ERKs and those of the Src family, can regulate hnRNP-K function in translation (31). hnRNP-K has also been shown to function within the DNA damage response pathway, being a target of the HDM2 ubiquitin ligase that is thus stabilized in response to DNA damage stimuli, such as UV irradiation (231). The stabilized hnRNP-K protein can then act as a transcriptional coactivator of the p53 protein (231). It will thus be of considerable interest to evaluate whether phosphorylation controls hnRNP-K activity to allow fine-tuning of DNA damage-induced transcriptional events. Surprisingly, the study identifying hnRNP-K as a JNK substrate appears to be the only study to date to use a chemical genetic approach for the identification of JNK substrates (119). This may reflect the limited availability of the modified ATP analogue, although a recent report has extended this approach with the use of ATP analogue inhibitors and sensitive JNK mutants to dissect the time course of signal transduction of JNKs in primary murine embryonic fibroblasts in response to TNF- (305).
Additional substrates may also help to explain other stress-activated nuclear responses in cells. The exposure of cells to stress decreases the expression of many gene families, including the expression of genes encoding ribosomal proteins and splicing factors (232). The mechanism has been investigated, with inactivation shown to result from phosphorylation of the polymerase I (Pol I)-specific transcription factor TIF-IA by JNK at Thr200 (211) (Table 1). This phosphorylation abrogates complex formation between TIF-IA, Pol I, and the TATA-binding protein-containing factor TIF-IB/SL1 (211) (Fig. 6). The overexpression of the Thr200Val mutant of TIF-IA that cannot be phosphorylated by JNK was shown to prevent inactivation of TIF-IA and thus lead to Pol I transcription even in the presence of stress (211). Thus, JNK-dependent phosphorylation of TIF-IA following the exposure of cells to stress provides a mechanism to prevent ribosomal synthesis (Fig. 3). This mechanism represents global control of protein translation during stress and may thus act as a protective mechanism for the cell under these situations.
|
| LINKS BETWEEN JNK ACTIVATION AND PROTEIN DEGRADATION |
|---|
|
TopPreviousNextReferences |
|---|
|
KD mice) identified a role for JNK-dependent events associated with protein degradation (102). In the T cells of these animals, the increased levels of production of interleukin-2 and the downstream cytokines interleukin-5, interleukin-10, and interleukin-13 were accompanied by increased protein levels (but not mRNA levels) of the transcription factors c-Jun and JunB (102). The enhanced stability of these transcription factors was confirmed in pulse-chase experiments, suggesting that the JNK pathway may assist in accelerating protein turnover in CD4+ T cells and thus may have a role in their polarization into Th1 and Th2 effector cells.
The similarity between the phenotypes of these Mekk1KD T cells and those isolated from animals with disruption of the E3 ligase Itch prompted an evaluation of the functions of JNK (102). Specifically, the E3 ligase Itch was found to be a substrate of JNK with multiple sites of phosphorylation, including Ser199, Ser222, and Ser232 (Table 3). A specific JBD conforming to the general features observed in the JBD of c-Jun was also identified (Table 2) (100), but it should be noted that the hydrophobic residues in the Itch JBD (i.e., Val600 and Phe604) differ from those in many of the other JBDs, where these residues are usually Leu (e.g., in c-Jun [Leu40 and Leu42] or JIP-1 [Leu161 and Leu163]). It will therefore be interesting to compare the exact modes of interaction of JNK with this JBD.
The JNK-mediated phosphorylation of Itch enhances protein degradation (102) (Fig. 3), and this has been attributed to phosphorylation-dependent conformational changes in Itch (100). This mechanism differs from the E3 ligase Fbw7-containing Skp/Cullin/F-box protein complex (SCFFbw7), which is targeted to phosphorylated proteins, such as phosphorylated c-Jun (240). The regulation of Itch activity by JNK phosphorylation has the potential to allow the coordinated regulation of degradation of many different cellular proteins and thus broadens the actions of JNK phosphorylation. Within the context of TNF--induced cell death, JNK activation of Itch allows for the degradation of c-FLIP, an inhibitor of caspase-8, and thus the subsequent cleavage of Bid to form tBid (52). Thus, prolonged JNK signaling is proapoptotic. Additional known Itch targets include c-Jun (102) and JunB (91). It will be important now to identify other processes modulated by phosphorylated Itch and determine how these actions integrate with those of the known targets.
In contrast, JNK did not associate with Elk-1 or target this transcription factor for degradation (98). It remains to be tested whether some of the differences noted in the effects of JNK on protein stability might have arisen due to the different JNK isoforms tested, particularly in light of the recent report that JNK1 and JNK2 can oppositely regulate p53 levels in cells (291). It is also interesting that a recent study showed that the tyrosine kinase c-Abl was able to promote the proteolytic destruction of damaged DNA-binding proteins and that this was an action independent of the kinase activity of c-Abl (54). It therefore appears that despite the lack of consensus on the effects of JNK-mediated phosphorylation on protein degradation, there may be multiple levels of control by JNK. It will be interesting to see whether JNK regulates the stability of the other non-transcription factor substrates described in the subsequent sections of this review. Furthermore, it will be important to explore whether JNK-mediated phosphorylation can alter the activities of other enzymes involved in posttranslational modifications of proteins, such as recently shown for the Akt-mediated phosphorylation and suppression of methyltransferase activity of EZH2 (48).
| JNK PHOSPHORYLATION OF SCAFFOLD AND ADAPTOR PROTEINS |
|---|
|
TopPreviousNextReferences |
|---|
