|Home | About | Journals | Submit | Contact Us | Français|
Raf kinases are essential for normal Ras-Raf-MEK-ERK pathway signaling, and activating mutations in components of this pathway are associated with a variety of human cancers as well as the related developmental disorders Noonan, LEOPARD, and cardiofaciocutaneous syndromes. Although the Raf kinases are known to dimerize during normal and disease-associated Raf signaling, the functional significance of Raf dimerization has not been fully elucidated. Here, using mutational analysis and a peptide inhibitor, we show that dimerization is required for normal Ras-dependent Raf activation and for the biological function of disease-associated Raf mutants with moderate, low, or impaired kinase activity. However, dimerization is not needed for the function of B-Raf mutants with high catalytic activity, such as V600E-B-Raf. Importantly, we find that a dimer interface peptide can effectively block Raf dimerization and inhibit Raf signaling when dimerization is required for Raf function, thus identifying the Raf dimer interface as a therapeutic target.
The Ras-Raf-MEK-ERK pathway controls vital cellular processes such as proliferation, differentiation and senescence, and proper regulation of the pathway is critical for normal cell function. Of the core pathway components, the mechanisms that modulate the Raf family kinases are by far the most complex, involving changes in subcellular localization, as well as protein interactions and phosphorylation/dephosphorylation events that can have either positive or negative regulatory effects (reviewed in Wellbrock et al., 2004). There are three mammalian Raf family members, A-Raf, B-Raf and C-Raf. All Raf kinases can bind activated Ras and transmit signals to the downstream kinases MEK and ERK, through the phosphorylation of MEK on activating sites. More recently, it has been appreciated that like numerous other protein kinases, the Raf kinases can form dimers. Under normal signaling conditions, Raf dimerization is Ras-dependent and occurs at the plasma membrane where the Rafs are activated (Garnett et al., 2005; Rushworth et al., 2006; Weber et al., 2001). The Raf dimers are subsequently disrupted by ERK-mediated feedback phosphorylation, which also disrupts the Ras–Raf interaction and attenuates Raf signaling (Dougherty et al., 2005; Ritt et al., 2010).
Dysregulation of the Ras-Raf-MEK-ERK pathway is a common occurrence in certain human diseases, and components of the pathway can function as disease drivers. In particular, germline-mutations in C-Raf are causative for Noonan and LEOPARD syndromes, whereas B-Raf mutations are found in Noonan, LEOPARD, and cardiofaciocutaneous (CFC) syndromes, with B-Raf mutations occurring in ~75% of CFC patients (Allanson et al., 2011; Tartaglia et al., 2011). Moreover, mutations in the Ras GTPases and B-Raf are important cancer promoters in a variety of human malignancies (Downward, 2003; Roring and Brummer, 2012). Numerous ATP-competitive Raf inhibitors have been developed, and ones such as vemurafenib are showing promise for the treatment of melanomas that express the high catalytic activity V600E-B-Raf mutant (Flaherty et al., 2010). Strikingly, however, all of the Raf inhibitors generated to date promote and stabilize Ras-dependent, Raf dimerization, thus causing paradoxical ERK activation in cells that express wild-type Raf proteins (Hatzivassiliou et al., 2010; Heidorn et al., 2010; Poulikakos et al., 2011). In addition, dimerization of the Raf kinases may contribute to several mechanisms that mediate Raf-inhibitor resistance, including mutational activation of N- and K-Ras (Nazarian et al., 2010; Su et al., 2012), upregulation of receptor tyrosine kinases (RTKs) that drive Ras activation (Nazarian et al., 2010), and expression of a V600E-B-Raf splice variant with enhanced dimerization potential (Poulikakos et al., 2011).
Taken together, the above findings have implicated Raf dimer formation as a regulatory mechanism with important consequences for disease treatment. However, key questions regarding Raf dimerization remain unanswered. In particular, it is unclear whether all Raf family members dimerize with one another and whether homo- or hetero-dimerization is most critical. Moreover, the extent to which dimerization modulates Raf kinase activity has not been fully addressed. In this study, we have used mutational analysis and a peptide inhibitor to investigate Raf dimerization and its role in Raf signaling. We find that dimerization is critical for the activation of wild-type Raf proteins and for the function of Raf mutants with all but high catalytic activity. Importantly, we show that Raf dimerization can be blocked using a peptide inhibitor and that blocking dimer formation can suppress MEK activation under conditions where Raf dimerization is required. These findings have important implications for the treatment of human disease states with elevated Ras-Raf-MEK-ERK pathway signaling.
To monitor the ability of endogenous A-Raf, B-Raf and C-Raf to form heterodimers, we first conducted co-immunoprecipitation assays using antibodies specific for each Raf family member (Figure S1A,B). The human HeLa cell line was chosen for this study as these cells express significant levels of all three Raf proteins, they are responsive to growth factor stimulation, and they are of epithelial origin as are many cancers with Ras pathway mutations. As shown in Figure 1A, little to no binding between any of the Raf proteins was observed in serum-starved cells. In contrast, when the Ras pathway was activated via EGF treatment, binding between endogenous B-Raf and C-Raf was strongly induced. A low level of binding between B-Raf and A-Raf was also observed in EGF-treated cells; however, little to no interaction between C-Raf and A-Raf was detected. These findings were confirmed using an in situ proximity ligation assay (PLA) to visualize the endogenous Raf interactions (Figure 1B). In addition, similar results were obtained when Raf dimer formation was examined in PDGF-treated NIH3T3 cells (Figure S1C).
Given that dimerization occurs under conditions coincident with Raf activation, we next examined the extent to which activation of one Raf family member was dependent on the presence of another Raf member. shRNA vectors were used to stably deplete each Raf member individually, following which the intrinsic kinase activity of the remaining Raf proteins was determined (Figure 1C). A-Raf, which has very low intrinsic activity, was activated only 2-fold in EGF-treated control cells, and its activation was reduced to 1.5-fold in B-Raf-depleted cells, but not altered in C-Raf-depleted cells. In contrast, C-Raf activity increased 15-fold in response to EGF treatment, and B-Raf activity, which has the highest basal level, increased an additional 3-fold. Depletion of A-Raf had no significant effect on EGF-induced B-Raf or C-Raf activation; however, the activation of B-Raf was reduced ~50% in C-Raf-depleted cells, and C-Raf activation was reduced ~90% in B-Raf-depleted cells. Given that A-Raf showed weak dimerization and little activity change, further studies investigating Raf dimerization were focused on B-Raf and C-Raf.
Analysis of the kinase domain structure of both B-Raf and C-Raf indicates that Rafs form side-to-side dimers, with specific residues in the dimer interface being critical for dimer formation (Hatzivassiliou et al., 2010; Rajakulendran et al., 2009). Therefore, to further analyze Raf dimerization and to address the issue of homo- versus hetero-dimerization, we generated B-Raf and C-Raf proteins containing mutations in key dimer interface residues: R509H-B-Raf/R401H-C-Raf (denoted as R>H) and E586K-B-Raf/E478K-C-Raf (denoted as E>K). FLAG-tagged wild-type (WT) or mutant Raf proteins were expressed in HeLa cells and examined for their ability to interact with other endogenous Raf members (Figure 2A). Consistent with the above findings, EGF treatment strongly induced binding between FLAG-WT-B-Raf and C-Raf as well as FLAG-WT-C-Raf and B-Raf, but little to no binding with A-Raf. FLAG-R>H-B-Raf and R>H-C-Raf failed to heterodimerize following EGF treatment, whereas FLAG-E>K-B-Raf and E>K-C-Raf displayed increased heterodimerization with C-Raf and B-Raf, respectively. E>K-C-Raf also exhibited a constitutive level of B-Raf binding in serum-starved cells, and both E>K-B-Raf and E>K-C-Raf interacted with A-Raf following EGF treatment.
To assess Raf homodimerization, FLAG- and Pyo-tagged Raf proteins were co-expressed in HeLa cells, following which the Pyo-tagged Raf proteins were immunoprecipitated and examined for the presence of FLAG-tagged Raf. As shown in Figure 2B, a low level of B-Raf homodimerization was observed in serum-starved cells, and this interaction increased ~2-fold upon EGF treatment. The E>K substitution caused a slight increase in B-Raf homodimerization, whereas the R>H substitution eliminated all binding. Homodimerization of C-Raf was not detected in serum-starved cells, and EGF treatment only induced a small amount of C-Raf homodimerization, which was disrupted by the R>H substitution. Strikingly, the E>K substitution resulted in a dramatic increase in C-Raf homodimerization following EGF treatment.
Because the dimer interface substitutions strongly modulated Raf dimerization, we next examined the effects of these mutations on the specific catalytic activity of the Raf proteins (Figure 2C). The R>H mutation had no effect on the basal kinase activity of B-Raf; however, R>H-B-Raf, which cannot form dimers, failed to be activated by EGF treatment. R>H-C-Raf also showed no change in activity following EGF stimulation and even exhibited reduced basal activity. In contrast, both E>K-B-Raf and E>K-C-Raf, which have increased dimerization potential, displayed elevated basal and EGF-induced activity, with the change in activity highest for E>K-C-Raf.
Previous studies have shown that growth factor-induced Raf dimerization is dependent on the Ras–Raf interaction as well as binding of 14-3-3 to a phosphorylation site in the C-terminus of the Rafs (B-Raf pS729, C-Raf pS621; Ritt et al., 2010; Rushworth et al., 2006; Weber et al., 2001). Using phosphospecific antibodies, we were able to confirm that the dimer interface mutants were indeed phosphorylated on the C-terminal 14-3-3 binding site, indicating that the lack of dimerization for the R>H mutants was not due to a defect in the phosphorylation of this site (Figure S2A). Moreover, in co-immunoprecipitation assays, we observed no defect in the ability of the R>H-Raf mutants to bind other critical Raf interactors such as Ras, 14-3-3, Hsp90, or Cdc37 (Figure 2D), and by mass spectrometry analysis, the only difference observed for R>H-B-Raf was the lack of C-Raf binding (Figure 2E). These findings together with results from an in vitro binding assay using purified Raf proteins (Figure S1D), indicates that the R>H dimer interface mutation specifically disrupts Raf–Raf contact. Finally, we found that the increased dimerization and activation potential of the E>K mutants was still dependent on the phosphorylation of the Raf C-terminal 14-3-3 binding site, as mutation of this site (denoted as S>A) eliminated heterodimerization and activation by EGF treatment (Figure S2B,C).
Mutant Raf proteins with increased signaling potential are observed in cancer and the related developmental disorders Noonan, LEOPARD, and CFC syndromes. These disease-associated Raf mutants exhibit varying levels of intrinsic kinase activity (Pandit et al., 2007; Rodriguez-Viciana et al., 2006; Wan et al., 2004) and display constitutive B-Raf–C-Raf heterodimerization (Garnett et al., 2005; Ritt et al., 2010). To further address the importance of dimerization in disease-associated Raf signaling, we introduced the dimer interface substitutions into oncogenic B-Raf mutants with high (V600E and G469A), moderate (G464V and L597V), low (G466A) and impaired (D594G) catalytic activity, and into Noonan syndrome C-Raf mutants with moderate (L613V) and low (P261S) catalytic activity. For all Raf mutants studied, the R>H substitution prevented B-Raf–C-Raf heterodimerization, whereas the E>K substitution enhanced heterodimerization (Figure 3A). In addition, as was observed for the normal Raf proteins, the dimer interface substitutions did not alter the phosphorylation state of the disease-associated Raf mutants at the C-terminal 14-3-3 binding site (Figure S3A).
We next used a focus formation assay to measure the biological activity of the disease-associated Raf mutants. As shown in Figure 3B, the R>H and E>K dimer interface substitutions had no significant effect on focus formation induced by B-Raf mutants with high kinase activity (V600E- and G469A-B-Raf). This finding is in agreement with a recent study examining the transforming activity of V600E-B-Raf lacking dimerization capability (Roring et al., 2012). In contrast, the R>H substitution blocked focus formation induced by moderate, low, and impaired activity B-Raf and C-Raf mutants, and the E>K substitution significantly increased their focus forming ability. The increase observed with the E>K substitution was further indicated to be dependent upon 14-3-3 binding, as the S>A mutation in the C-terminal 14-3-3 binding site eliminated heterodimerization and the focus forming ability of lower activity Raf mutants containing the E>K substitution (Figure S3B,C). It should be noted that although the S>A mutation blocked heterodimerization for V600E/E>K-B-Raf, it had no effect on focus formation (Figure S3B,C), further confirming that Raf dimerization is not required for the function of high activity B-Raf mutants.
The above findings demonstrate that dimerization is critical for normal Ras-dependent Raf activation and for the function of disease-associated Raf mutants with all but high catalytic activity. Therefore, blocking Raf dimerization may be of potential therapeutic use. Given that peptides containing docking motifs for the JNK1 kinase have been successfully used to inhibit JNK1 signaling in vitro and in vivo (Gaestel and Kracht, 2009), we set out to determine if a peptide inhibitor could be used to block Raf dimerization and perhaps Raf signaling. Because B-Raf exhibited the highest degree of homo- and heterodimerization, we derived our peptides based on the sequence of the B-Raf dimer interface region (Figure 4A). Several GFP-fusion proteins containing dimer interface (DI) sequences were generated and examined for C-Raf and B-Raf binding. Although all the fusion proteins interacted with FLAG-C-Raf, the level of binding was highest for GFP-DI1 (Figure S4A). This peptide also interacted with endogenous C-Raf and B-Raf (Figure 4B), showed specificity for Raf binding (Figure S4B), and was used for all further studies. As shown in Figure 4C, GST-DI1 interacted with endogenous C-Raf in a dose-dependent manner and binding was observed in both serum-starved and EGF-treated cells. Importantly, expression of GFP-DI1 blocked the EGF-induced heterodimerization of B-Raf and C-Raf as well as signaling from Raf to MEK (as monitored by MEK phosphorylation, Figure 4C,D). In addition, the GFP-DI1 peptide suppressed MEK activation induced by disease-associated Raf mutants with moderate, low, and impaired kinase activity, but not that induced by V600E-B-Raf (Figure 4E).
To further assess the ability of this peptide to inhibit Raf signaling, the DI1 peptide was synthesized with an HIV TAT tag to allow efficient cell uptake from the media (Figure S4C). The TAT-DI1 peptide or a control TAT-peptide containing a scrambled version of the DI1 sequence (TAT-Scram) were then examined for their effects on the growth and viability of several non-small cell lung carcinoma lines (NSCLC). NSCLC lines were chosen for this analysis as oncogenic mutations causing moderate or impaired B-Raf kinase activity are observed more frequently in NSCLCs than other cancers. In addition, constitutively active EGFR and Ras proteins are prominent inducers of Raf signaling in NSCLCs. As shown in Figure 4F, treatment with TAT-DI1, but not TAT-Scram, reduced the proliferation and viability of NSCLC lines expressing activated Ras (A549, H460) and impaired activity G466V-B-Raf (Cal12T, H1666) in a dose dependent manner, with significant reductions observed at the 5 μM concentration. In contrast, the TAT-DI1 peptide had little to no effect on the A375 melanoma line expressing V600E-B-Raf and only a limited effect on the H1703 NSCLC line and HeLa cells, both of which are WT for EGFR, Ras, and Raf (http://www.sanger.ac.uk/genetics/CGP/cosmic/). Similar results were observed when an MTS assay was used to measure cell viability (Figure S4D). TAT-DI1 also reduced Raf to MEK signaling in A549 and Cal12T cells, but not in A375 cells (Figure 4G). Finally, we found that the decrease in cell viability was due, at least in part, to an induction of apoptosis, as cleavage of the caspase 3 target PARP was observed in TAT-DI1-treated A549 cells (Figure 4H).
The Raf kinases are important mediators of Ras-dependent signal transduction, and the regulation of Raf function is highly complex. Here, we have investigated the role of Raf dimerization in both normal and disease-associated Raf signaling. Using HeLa cells that have significant expression of all three mammalian Raf family members, we find that normal growth factor treatment primarily induces B-Raf–C-Raf heterodimerization, with some B-Raf and C-Raf homodimerization observed. The largest growth factor-induced change in activity occurred for C-Raf, and our findings indicate that the interaction with B-Raf is critical for C-Raf activation, as depletion of B-Raf reduced C-Raf activation ~90%. Interestingly, B-Raf, which has significantly higher basal kinase activity, was found to homodimerize constitutively. The activity of B-Raf was further increased by EGF treatment, and this increase coincided with high levels of C-Raf binding as well as some increase in homodimerization. Heterodimerization with C-Raf also appears to influence B-Raf activity in that the EGF-induced activation of B-Raf was reduced ~50% in C-Raf-depleted cells. Although A-Raf is well expressed in HeLa cells, its depletion had little effect on either B-Raf or C-Raf activation. In addition, A-Raf, which has the lowest intrinsic activity of all the Raf proteins (Marais et al., 1997; Pritchard et al., 1995), was only weakly activated by EGF treatment and exhibited low levels of dimerization. Why A-Raf appears to have a reduced dimerization potential is currently unclear, and further studies will be needed to determine if other activators or events regulate A-Raf dimer formation.
The role of dimerization in B-Raf and C-Raf activation was further confirmed by mutational analysis of the Raf dimer interface. In particular, replacing a critical arginine residue in the dimer interface with a histidine disrupted all Raf dimerization and prevented the EGF-induced activation of B-Raf and C-Raf, but had no effect on other Raf interactions, including binding to Ras, 14-3-3, MEK, Hsp90, or Cdc37. In contrast, a glutamic acid to lysine substitution was found to increase Raf dimerization with the most significant effect observed for C-Raf. E>K-C-Raf exhibited constitutive heterodimerization with B-Raf, which increased with EGF treatment, and a dramatic induction of homodimerization following growth factor treatment. Moreover, the increases in dimerization correlated with changes in both the basal and induced kinase activity of B-Raf and C-Raf. Effects of the dimer interface mutations were also observed in the context of disease-associated Raf proteins. Mutations in either B-Raf or C-Raf can function as drivers of human disease, and while previous studies have shown that these mutations can have varying effects on the intrinsic catalytic activities of the Rafs, they all appear to promote B-Raf–C-Raf heterodimerization. As expected, we observed constitutive B-Raf–C-Raf heterodimerization for all the disease-associated Raf mutants examined and found that substitutions in the dimer interface not only altered the dimerization potential of these proteins but had significant effects on the transforming potential of all but the high catalytic activity B-Raf mutants. Thus, while the high activity B-Raf mutants such as V600E-B-Raf can dimerize, it appears that their enzymatic activity is already so elevated that further activation events are unnecessary. In contrast, dimerization is critical for the mutant proteins with moderate, low, and impaired kinase activity.
The finding that the arginine to histidine substitution in the Raf dimer interface blocks dimerization as well as the function of both WT and mutant Raf proteins with all but high catalytic activity suggested that blocking Raf dimerization might have therapeutic potential. Indeed, we identified a peptide sequence corresponding to a region in the B-Raf dimer interface that when expressed as a GFP-fusion protein in cells, could interact with endogenous B-Raf and C-Raf and block EGF-induced B-Raf–C-Raf heterodimerization as well as MEK activation. Consistent with the idea that high activity B-Raf mutants do not require dimerization for function, the DI1 peptide had no effect on V600E-B-Raf-mediated MEK phosphorylation, but did inhibit MEK activation mediated by lower activity Raf mutants. Importantly, when NSCLC lines expressing either activated Ras or impaired activity B-Raf proteins were treated with a synthesized TAT-DI1 peptide, cell viability was reduced in a dose-dependent manner, indicating that the Raf dimer interface is a bona fide therapeutic target.
In summary, our findings confirm that dimerization is an important mechanism regulating Ras-dependent Raf activation and the function of disease-associated Raf mutants with all but high kinase activity. Our results also indicate that secondary mutations in the Raf dimer interface could alter the progression and treatment of disease states with elevated Ras/Raf signaling. Finally, our study provides proof of principle that an inhibitor which disrupts Raf dimerization can suppress Raf-dependent signaling and may be useful for the treatment of diseases that require dimerization for Raf function.
Antibodies to A-Raf (C-20), B-Raf (H-145, F-7), C-Raf (C-12), and 14-3-3β (K-19) were from Santa Cruz Biotechnology; antibodies to phospho-MEK, phospho-C-Raf S621, Hsp90, Cdc37, Ras and FLAG were from Cell Signaling Technology; antibodies to C-Raf, MEK1,2, and PARP were from BD Biosciences; antibody to FLAG-M2 was from Sigma-Aldrich; antibody to GFP was from Roche Diagnostics.
HeLa, NIH3T3, and Cal-12T were cultured in DMEM, and A549, H460, H1703, and H1666 were grown in RPMI. All media was supplemented with 10% fetal bovine serum (FBS) and cells were cultured at 37°C under 5% CO2. Where indicated, cells were treated with 100 ng/mL EGF for five minutes. Focus forming assays and proximity ligation assays were conducted as described in supplemental procedures.
For co-immunoprecipitation assays, cells were lysed in 1% NP-40 buffer (20 mM Tris pH 8.0, 137 mM NaCl, 10% glycerol, 1% NP-40 alternative, 0.15 U/mL aprotinin, 1 mM PMSF, 0.5 mM sodium vanadate, 20 μM leupeptin) and lysates were clarified by centrifugation. Equivalent amounts of protein lysate were incubated with the appropriate antibody and Protein G Sepharose beads for 3 hr at 4°C. Complexes were washed extensively with 1% NP-40 buffer and examined by immunoblot analysis. For in vitro kinase assays, cells were lysed under stringent conditions in RIPA buffer (1% NP-40 buffer containing 0.5% sodium deoxycholate and 0.1% SDS) to disrupt protein complexes. Raf proteins were immunopurified and their catalytic activity determined as described in supplemental procedures.
The GFP-DI1 peptide fusion protein was generated by cloning B-Raf amino acids 503-521 into the pEGFP-C1 vector via PCR. The TAT-DI1 peptide was purchased from Bachem. For viability assays, cells plated in 6-well dishes were treated with peptide in serum-free media for 1 hr before an equal volume of media was added to give a final concentration of 10% FBS. Cell counts were determined 48 hrs after plating.
This project was funded by Federal funds from the National Cancer Institute.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.