Axin forms a complex with MEKK1 and activates c-Jun NH(2)-terminal kinase/stress-activated protein kinase through domains distinct from Wnt signaling.

Axin negatively regulates the Wnt pathway during axis formation and plays a central role in cell growth control and tumorigenesis. We found that Axin also serves as a scaffold protein for mitogen-activated protein kinase activation and further determined the structural requirement for this activation. Overexpression of Axin in 293T cells leads to differential activation of mitogen-activated protein kinases, with robust induction for c-Jun NH(2)-terminal kinase (JNK)/stress-activated protein kinase, moderate induction for p38, and negligible induction for extracellular signal-regulated kinase. Axin forms a complex with MEKK1 through a novel domain that we term MEKK1-interacting domain. MKK4 and MKK7, which act downstream of MEKK1, are also involved in Axin-mediated JNK activation. Domains essential in Wnt signaling, i. e. binding sites for adenomatous polyposis coli, glycogen synthase kinase-3beta, and beta-catenin, are not required for JNK activation, suggesting distinct domain utilization between the Wnt pathway and JNK signal transduction. Dimerization/oligomerization of Axin through its C terminus is required for JNK activation, although MEKK1 is capable of binding C terminus-deleted monomeric Axin. Furthermore, Axin without the MEKK1-interacting domain has a dominant-negative effect on JNK activation by wild-type Axin. Our results suggest that Axin, in addition to its function in the Wnt pathway, may play a dual role in cells through its activation of JNK/stress-activated protein kinase signaling cascade.

Axin, which was identified from analysis of the mouse Fused locus, plays a critical role in controlling axis formation during embryonic development (1) in that it is a negative regulator of the Wnt signal transduction pathway (2)(3)(4). The Fused allele harbors a mutation (Axin Fu ) that causes dominant skeletal and neurological defects characterized mainly by kinking and shortening of the tail, as well as recessive lethal embryonic defects, such as neuroectodermal abnormalities (5). Mouse embryos homozygous for either of the two spontaneous mutant alleles of Fused, i.e. Kinky (Axin Ki ) and Knobbly (Axin Kb ), or the transgenic insertional allele of Fused (Axin Tg1 ), all show recessive lethal embryonic defects and a duplication of the embry-onic axis (6 -9). The mechanism by which Axin negatively regulates the Wnt signal transduction pathway for axis formation is as yet unclear. It was recently demonstrated that Axin binds directly to adenomatous polyposis coli (APC), 1 glycogen synthase kinase-3␤ (GSK-3␤), and ␤-catenin and acts as a scaffold upon which these proteins assemble to coordinate the regulation of ␤-catenin levels for Wnt signaling (10 -15). In the absence of Wnt signals, Axin links these components together and enables the phosphorylation of ␤-catenin by GSK-3␤, leading to the degradation of ␤-catenin. In response to Wnt signals, GSK-3␤ activity is inhibited such that ␤-catenin levels become elevated. ␤-catenin is then translocated into the nucleus, where it binds to the T-cell factor or lymphoid enhancer factor transcription factors (16,17) to regulate expression of genes such as the c-MYC oncogene (18). Axin is therefore a multifunctional protein contributing to many aspects of biology ranging from the regulation of gene transcription and cell growth control to morphogenesis.
Although Axin is a member of the large family of regulators of G protein signaling (RGS) proteins (19,20) in that it possesses the conserved RGS domain, it does not behave as a canonical RGS protein. It does not bind to G␣ subunits and is therefore unlikely to serve as a GTPase-activating protein. 2 It is also noteworthy that Axin is ubiquitously expressed in almost all tissues, from early embryonic development through to adult stage (1,11). We suspected that Axin may play a more general role in addition to the regulation of axis formation, perhaps in cell signaling pathways. The mitogen-activated protein kinase (MAPK) signaling pathways, which are evolutionarily conserved in all eucaryotes, are instrumental in integrating and executing a diverse array of cellular programs such as cell division, cell movement, and cell differentiation (21)(22)(23)(24). We therefore investigated whether the scaffold protein Axin could regulate MAPK activities. Here we show that overexpression of Axin in human embryonic kidney 293T cells leads to differential activation of MAPK, with robust induction for c-Jun NH 2 -terminal kinase/stress-activated protein kinase (JNK/SAPK), moderate induction for p38, and very little induction for extracellular signal-regulated kinase (ERK). Two novel functional regions of Axin have been identified that are critical for JNK activation and completely distinct from domains required for Wnt signaling. We demonstrate that Axin forms a complex with MEKK1 and provide evidence that the dimerization/oligomerization of Axin is required for JNK activation. Consistent with a dual function of Dishevelled in the Wnt pathway and JNK-mediated tissue polarity signaling pathway, our studies reveal that Axin may also assume a dual role in cells through its activation of the JNK/SAPK pathway. 293T cells were transiently transfected with 1 g of HA-tagged wild-type Axin or C-terminal deletion mutant C1, plus 1 g of FLAG-JNK1, FLAG-p38, or FLAG-ERK2. For controls, cells were transfected with 0.5 g of TRAF2 or were treated for 20 min with either 400 mM sorbitol or 160 nM phorbol 12-myristate 13-acetate (PMA). Following immunoprecipitation of FLAG-JNK1, FLAG-p38, and FLAG-ERK2, their kinase activities were assayed using GST-c-Jun, GST-activating transcription factor-2 (GST-ATF2), and myelin basic protein (MBP), respectively, as substrates. The amount of the kinase in each immunoprecipitate was quantified by immunoblotting. Data are expressed as fold kinase activation compared with kinase activity produced in vector-transfected cells. The values represent the means Ϯ S.E. from three separate experiments. b, the kinase-inactive form of MEKK1, but not of ASK1 or TAK1, abolishes Axin activation of JNK. Cells were transiently transfected with 1 g of FLAG-JNK1, plus 2 g of either HA-MEKK1, HA-MEKK1-K1255M, HA-ASK1, HA-ASK1-K709M, HA-TAK1, or HA-TAK1-K63W in the presence (dark columns) or absence (light columns) of 1 g of HA-tagged Axin. Total cell lysates were probed with anti-HA to detect the expression of HA-MEKK1 (lanes 1 and 2), HA-MEKK1-K1255M (lanes 3 and 4), HA-ASK1 (lanes 5 and 6), HA-ASK1-K709M (lanes 7 and 8), HA-TAK1 (lanes 9 and 10), and HA-TAK1-K63W (lanes 11 and 12) in the presence (lanes 2, 4, 6, 8, 10, and 12) or absence (lanes 1, 3, 5, 7, 9, and 11) of Axin. Immunokinase assays were performed and are presented as described in the legend to a. c, JNK activation by Axin is inhibited by dominant-negative forms of MKK4/MKK7. Cells were transiently transfected with 1 g of FLAG-JNK1, plus 2 g of either HA-MKK4, HA-MKK4-DN, HA-MKK7, or HA-MKK7-DN in the presence (dark columns) or absence (light columns) of 1 g of HA-tagged Axin. Total cell lysates were probed with anti-HA to detect the expression of HA-MKK4 (lanes 1 and 2), HA-MKK4-DN (lanes 3 and 4), HA-MKK7 (lanes 5 and 6), and HA-MKK7-DN (lanes 7 and 8) in the presence (lanes 2, 4, 6, and 8) or absence (lanes 1, 3, 5, and 7) of Axin. Immunokinase assays were performed and are presented as described in the legend to a.

EXPERIMENTAL PROCEDURES
Construction of Plasmids-The full-length Axin cDNA was obtained by screening a mouse pituitary gt11 cDNA library using a polymerase chain reaction-generated 1-kilobase 5Ј-coding fragment as probe, tagged with either HA or Myc at the N terminus, and cloned into the ClaI and BamHI sites of the mammalian expression vector pCMV5. A series of different constructs of Axin (see Figs. 3 and 4) were created using convenient restriction enzyme sites. Briefly, to construct mutants N1 (198 -832), N2 (400 -832), and N3 (509 -832), the N terminus of wild-type Axin was deleted at the NcoI, XbaI, and SpeI sites, respectively, and fused in-frame to HA or Myc tag. To construct mutants C1(1-600), C2 (1-746), C3 (1-398), and C4 (1-230), the C terminus of wild-type Axin was deleted at SpeI, ApalI, XbaI, and PstI, respectively. For mutant M1 (1-353/509 -832), the 1.0-kilobase N-terminal fragment of wild-type Axin obtained by BamHI digestion was fused in-frame to the polymerase chain reaction-generated C-terminal fragment encoding Axin aa 509 -832. To construct mutant M2 (1-209/353-832), wild-type Axin was digested internally at AflII and BamHI sites, blunted with Klenow fragment, and religated in-frame. To construct mutant M3 (198 -353/509 -832), the N-terminal fragment of mutant N1 obtained by BamHI digestion was used to replace that of mutant M1 to generate an in-frame M3. To construct mutant M4 (198 -398), the region between NcoI and XbaI of wild-type Axin was released, fused in-frame with Myc tag at the N terminus, blunted with Klenow fragment, and ligated to pCMV5 to generate stop codon at the C terminus. Plasmids of FLAG-tagged JNK1, FLAG-tagged p38, and FLAG-tagged ERK2 were described previously (25,26).  (27,28).
Transient Transfection and Immunokinase Assays-Human embryonic kidney 293T cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum, 100 IU of penicillin, 100 g/ml streptomycin, and 2 mM glutamine. Transfections were performed in 60-mm dishes using Superfect TM according to the manufacturer's in- Immunokinase assays were performed and are presented as described in the legend to Fig. 1. b, catalytic-inactive C-terminal kinase domain of MEKK1 inhibits Axin-induced JNK activity in a dose-dependent manner. Cells were transfected with 1 g of HA-tagged Axin and 1 g of FLAG-JNK1 together with increasing amounts of Myc-MEKK1-C-K1255M as indicated. Total cell lysates were probed with anti-HA and anti-Myc for expression of Axin and MEKK1-C-K1255M, respectively. Immunokinase assays were performed and are presented as described in the legend to Fig. 1. c, interaction of Cterminal kinase domain of MEKK1 with Axin. Cells were transfected with 2 g of either HA-MEKK1, HA-MEKK1-N3.1, HA-MEKK1-N1.9, or HA-MEKK1-C, together with 2 g of Myc-tagged Axin. Cell lysates were immunoprecipitated (IP) with anti-HA for MEKK1 proteins, anti-Myc for Axin, or control IgG. The immunoprecipitates and cell lysates were then analyzed by immunoblotting separately using anti-Myc for Axin and anti-HA for MEKK1 proteins.
structions (Qiagen). The total amount of transfected DNA was adjusted to 4 g with the empty vector pCMV5 as necessary. Cells were harvested at 40 h posttransfection and lysed in a lysis buffer (29). FLAGtagged JNK1, FLAG-tagged p38, and FLAG-tagged ERK2 were immunoprecipitated using mouse monoclonal anti-FLAG M2 beads (Sigma), and their kinase activities were determined as described previously using 1 g of GST-c-Jun (1-79), 1 g of GST-activating transcription factor-2 (1-109), and 5 g of myelin basic protein as substrates, respectively (29). Fold activation of the kinases was determined by an imaging analyzer (Molecular Dynamics model 425E) and normalized to their expression levels. Data are expressed as fold kinase activation compared with vector-transfected cells, with the values representing the means Ϯ S.E. from three separate experiments.
The MAPK cascade consists of three conserved components: a family of serine/threonine kinases called MAPK kinase kinase/MEKK, which activate MAPK kinase/MEK. MAPK kinase/MEK in turn activate MAPK by dual phosphorylation on threonine and tyrosine residues (21)(22)(23)(24). It has been shown that overexpression of MAPK kinase kinase, such as MEKK1 (31), ASK1 (32,33), and TAK1 (34), induces the activation of the JNK cascade. To explore whether these MAPK kinase kinases play a part in the induction of JNK by Axin, 293T cells were separately cotransfected with Axin, FLAG-JNK1, and wild-type HA-MEKK1, HA-ASK1, HA-TAK1, or their respective kinase-inactive mutants (Fig. 1b). Immunokinase assays showed that wild-type HA-MEKK1, HA-ASK1, and HA-TAK1 elevated the basal levels of JNK activity, as expected. Domi- nant-negative HA-ASK1-K709M and HA-TAK1-K63W did not appear to alter JNK activation mediated by Axin. In contrast, the ATP binding-deficient, dominant-negative HA-MEKK1-K1255M reduced Axin activation of JNK by 3-fold, suggesting that MEKK1 acts in the same signaling pathway in Axinmediated activation of JNK. This is consistent with the observation that coexpression of Axin with HA-MEKK1 did not cause an additive effect on JNK activation compared with JNK induction by expression of HA-MEKK1 alone (Fig. 1b).
MKK4 and MKK7, Which Act Downstream of MEKK1, Are Involved in Axin-mediated JNK Induction-Because both MKK4 and MKK7 are known to phosphorylate and activate JNK (28,35), we determined whether JNK activation by Axin is mediated through MKK4 and MKK7 (Fig. 1c). In agreement with the common observation that in the absence of an upstream stimulus wild-type MKK4 displays a relatively weak activation of JNK, overexpression of HA-MKK4 did not increase basal activity of JNK (Fig. 1c). However, when HA-MKK4 was coexpressed with Axin, JNK activation was more pronounced than that with expression of Axin alone, suggesting MKK4 activation by Axin. Consistently, the dominantnegative HA-MKK4-DN diminished Axin-mediated JNK activation. Overexpression of wild-type HA-MKK7 alone stimulated JNK activity, whereas coexpression with Axin further elevated JNK activity. Kinase-inactive HA-MKK7-DN partially inhibited JNK activation, despite the slightly elevated basal JNK activity. These results implicate both MKK4 and MKK7 in the Axin-mediated JNK activation, in agreement with the fact that these MAPK kinases act downstream of MEKK1 (21)(22)(23)(24).
Catalytic-inactive C-terminal Kinase Domain of MEKK1 Is Sufficient to Inhibit Axin-mediated JNK Activation-We further examined the structural requirements of MEKK1 for Axin-mediated JNK activation. It has been shown that the regulation of MEKK1 activity by proteins such as 14-3-3 (36) and Grb2 (37) is through interaction with the N-terminal region of MEKK1. To investigate whether the N terminus of MEKK1 plays a role in Axin-mediated JNK activation, we cotransfected Axin with constructs expressing either the 150-kDa (HA-MEKK1-N3.1) or the 80-kDa (HA-MEKK1-N1.9) Nterminal fragment of MEKK1 (38). In contrast to the observation that Grb2-MEKK1-mediated JNK activation by EGF is inhibited by the N-terminal regulatory domain of MEKK1 (37), we did not see a similar inhibitory effect of HA-MEKK1-N3.1 or HA-MEKK1-N1.9 on Axin-mediated JNK activation (Fig. 2a). On the contrary, we observed that the C-terminal kinase domain of MEKK1, which is mutated at its ATP-binding site (HA-MEKK1-C-K1255M), blocked Axin-induced JNK activity (Fig. 2a) in a dose-dependent manner (Fig. 2b). These results indicate that the kinase domain of MEKK1 may directly mediate Axin activation of JNK, in which case it should bind to Axin. To test this, we performed immunoprecipitation assays (Fig. 2c) and found that Myc-tagged Axin was coprecipitated with the C-terminal kinase fragment HA-MEKK1-C, as well as the full-length HA-MEKK1, but not with the N-terminal fragments HA-MEKK1-N3.1 or HA-MEKK1-N1.9. Conversely, only HA-MEKK1-C and full-length HA-MEKK1 were coprecipitated with Myc-Axin (Fig. 2c). No physical interaction between Axin and HA-ASK1 or HA-TAK1 was detected by similar immunoprecipitation assays in transfected 293T cells (data not shown). We conclude that Axin forms a complex with the Cterminal kinase domain of MEKK1 to bring about an induction in JNK activity.
A Novel Axin Domain Acts in Concert with the C Terminus to Activate JNK-The surprising finding that Axin can activate JNK activity (Fig. 1a) prompted us to examine which region(s) on the Axin protein is involved in JNK activation. Axin is known to possess multiple functional domains: an RGS homologous domain, which binds APC; distinct binding sites for GSK-3␤ and ␤-catenin; and a Dishevelled homologous (DIX) domain (1,11). We generated a series of deletion mutants of HA-tagged Axin (Fig. 3, top panel) and assayed their abilities to activate JNK (Fig. 3, bottom panels). Removal of the N-terminal region aa , which encompass the RGS homologous domain for APC binding (mutant N1), did not affect Axininduced JNK activity. Similarly, deletion of aa 354 -508, which removes the binding sites for GSK-3␤ and ␤-catenin (mutant M1), had no deleterious effect on JNK activation. On the other hand, C-terminal deletion mutants C1 and C2, in which the binding sites for APC, GSK-3␤, and ␤-catenin were retained, lost their ability to activate JNK. These results indicate that Axin activation of JNK does not require the RGS homologous domain, GSK-3␤, or ␤-catenin binding sites and that the domains on Axin responsible for Wnt signaling are distinct from those involved in JNK activation. Indeed, removal of the region of aa 210 -352 (mutant M2) completely abolished JNK activation, whereas fusion of the same region with the C-terminal region aa 509 -832 (mutant M3) fully restored JNK activation. Therefore, activation of JNK by Axin requires the region flanked by the RGS homologous domain and GSK-3␤ binding domain, and the C terminus (see below).
Oligomerization of Axin through the C Terminus Is Required for JNK Activation-It has recently been shown that the C terminus of Axin can bind to itself (39), suggesting that Axin may form dimers or multimers for its function. In this study, we found that the C-terminal self-binding domain is important for JNK activation (Fig. 3). We further investigated whether dimerization/oligomerization of Axin is a prerequisite for its binding to MEKK1, or whether MEKK1 binding is required for Axin oligomerization. To test these possibilities, wild-type Axin, mutant M2 (without the MEKK1-binding site), mutant C1 (without the C-terminal self-binding domain), and MEKK1-C were transfected in different combinations (Fig. 5,  a-c). Coimmunoprecipitation experiments using differently tagged Axin constructs showed that full-length Axin (Fig. 5a, i) or mutant M2 (Fig. 5a, ii) could form a complex with itself, whereas mutant C1 (Fig. 5a, iii) lost the ability to bind to itself. Whereas the mutant M2 alone did not complex with MEKK1-C (Fig. 5b), in the presence of wild-type Axin, mutant M2 was detected in the MEKK1-C immunoprecipitate (Fig. 5c), indicating that mutant M2 is capable of complexing with the wild-type Axin. Interestingly, mutants C1 and C2, without the extreme C-terminal end (which includes the DIX domain), were fully capable of binding to MEKK1-C ( Fig. 4 and data not shown), but they lost their ability to activate JNK (Fig. 3). These results suggest that Axin oligomerization is required for JNK activation through MEKK1. One would then expect Axin without the MEKK1-binding domain to display a dominant-negative effect when coexpressed with wild-type Axin. Indeed, the Axin mutant M2 exhibited an inhibitory effect on the Axin-activated JNK activity in a dose-dependent manner (Fig. 5d). DISCUSSION We have explored the possibility that the scaffold protein Axin, which plays a central role in axis formation and perhaps in tumorigenesis (40) also exerts a function in the MAPK pathway. It is demonstrated here that Axin overexpression preferentially activates JNK. We have delineated the Axin-mediated JNK activation pathway, showing that Axin forms a complex with MEKK1, but not ASK1 or TAK1. Dominant-negative forms of MEKK1 block the Axin-mediated activation of JNK. Both MKK4 and MKK7, which are downstream of MEKK1, are involved, as their dominant-negative forms can also attenuate Axin-mediated JNK activation. Interestingly, determination of the functional domains of Axin reveals that domains utilized for the Wnt pathway, i.e. the RGS homologous domain for APC binding and domains for GSK-3␤ and ␤-catenin binding (10 -14), are not required for the JNK activation. Instead, a novel domain flanked by the binding sites for APC and GSK-3␤, and the C-terminal region, are needed. We have identified MEKK1 to bind the novel domain, which we term the MEKK1-interacting domain (MID). Fusion of the MID to the C-terminal fragment of Axin fully retains the JNK activating activity, clearly demonstrating that domain utilization by the two pathways are distinct. Furthermore, C-terminal-mediated oligomerization of Axin is essential for JNK activation, in accordance with the observation that Axin deficient for the MID displays a dominant-negative effect on wild-type Axin.
The novel MID has been mapped to around aa 217-352 (Fig.  6). This region shares sequence similarity with the corresponding region in conductin (41) and its rat homologue Axil (42); no sequence similarity is found in other proteins. Located within the MID are multiple phosphorylation sites that have been shown to regulate Axin stability (43) so that Axin becomes dephosphorylated and destabilized upon activation of the Wnt pathway (43,44). Overexpression of Axin in this study is thus likely to mimic a situation in cells in which Axin is in high abundance and the Wnt pathway is inactive. Another region on Axin essential for JNK activation is its C-terminal region, aa 507-832 (Fig. 6), in which the extreme C-terminal segment serves as a dimerization/oligomerization site (39). It was recently demonstrated that both oligomerization of Axin and its bridging of GSK-3␤ with ␤-catenin are necessary for Axin to inhibit Lef1-dependent transcription, suggesting a role of oligomerization in ␤-catenin signaling for the Wnt pathway (45). However, evidence for the requirement of the C terminus of Axin for Wnt signaling is not definitive. On the one hand, mouse Axin mutants (Axin Fu and Axin Kb ), which produce truncated proteins at the C terminus (1,9) corresponding to the nonsubdivisible positions determined in our JNK assays (data not shown), exhibit dominant phenotypes, including a kinked tail (5,6). On the other hand, Axin with the C-terminal truncations are functionally active in the frog embryo ventralization assay (12,39,46), suggesting that the C-terminal region may not be critical for Wnt signaling. In this study, we have attributed the significance of Axin self-binding to the JNK signaling pathway. What is more provocative is that MEKK1 requires Axin-mediated dimerization to be activated, although it is capable of binding to monomeric Axin. It is possible that dimerized Axin functions as a bridge for MEKK1 to assume a dimeric configuration, which may be important for the nature of MEKK1 activation in this signaling pathway. This apparent requirement to form a dimer or a higher order of multimer is reminiscent of the activation mechanism of the ERK2 MAPK (47), extending the generality of dimerization to other kinases in the MAPK cascade. It is equally possible that Axin dimerization is required for recruitment of other factor(s) to activate MEKK1.
Our finding that Axin activates the JNK signal transduction pathway is reminiscent of the activation of JNK in Drosophila and mammalian cells by Dishevelled (48,49). As with Axin, distinct domains of Dishevelled are utilized for the Wnt pathway and JNK activation: whereas the DIX, PDZ, and DEP domains are required for Wnt signaling (48 -50), the DEP domain, perhaps with other uncharacterized domains, is essential for JNK activation (48,49). Interestingly, the JNK signaling pathways of Dishevelled in mammalian cells and Drosophila cells are different, despite their being highly conserved. Whereas RhoA is required for Dishevelled-induced activation of Drosophila JNK (51), small G proteins, including RhoA, Cdc42, and Rac1, are not involved in Dishevelled-induced JNK activation in mammalian cells (49). In our study, we found that the dominant negative mutants of Rho, Cdc42, and Rac1 could not inhibit Axin-induced JNK activation (data not shown). Despite the shared similarities, it is possible that the JNK pathways activated by Axin and Dishevelled may be distinct and have different upstream triggers, as well as downstream effector molecules. Genetic analyses of Drosophila have revealed that in addition to there being an absolute requirement of Dishevelled in reception of the Wg signal, which is then relayed to Zeste-white3 in the Wg signaling cascade, Dishevelled is also part of a tissue polarity signaling pathway, where it acts downstream of Frizzled and upstream of RhoA (2,4). Mutational studies of the Axin gene in mouse and Xenopus embryos have identified Axin to inhibit signal transduction in the Wnt pathway through its function downstream of GSK-3␤ and upstream of ␤-catenin (1,11). It is conceivable that in addition to its involvement in the Wnt pathway, Axin may also play a dual role in cells through its activation of the JNK signaling cascade, including a role in apoptosis. 3  demonstrated that aberrant positional information signals in the Drosophila wing causes JNK-dependent apoptosis of the wing cells, suggesting that the JNK apoptotic cascade is activated when normal signaling is distorted (52). Although the Dishevelled DIX domain appears to be dispensable for JNK activation (48,49), our data show that the Axin DIX domain is essential for JNK activation. It is noteworthy that the Axin DIX domain is capable of interacting with multiple proteins, which not only facilitates the formation of Axin dimers/multimers (39,45) but also enables Axin to bind Dishevelled directly (15,53). Clearly, the ability of the scaffold protein Axin to bind to a huge repertoire of proteins allows the assembly of complexes and a diverse regulatory network of molecular interactions, which can contribute to the specificity of signaling events.
In summary, our results have defined a novel functional role of Axin in the signaling pathway of MAPK activation. Structural determination of Axin in the JNK/SAPK activation pathway suggests a distinct domain utilization of Axin from the Wnt signaling pathway, pointing to the possibility that Axin may exert as yet unidentified functions. Moreover, we demonstrate that MEKK1 requires dimerized Axin to activate downstream protein kinases. Further studies will be directed to address what signals trigger Axin to activate the JNK pathway and whether Axin-mediated JNK activation may cross-talk with the Dishevelled-mediated JNK signaling pathway.