SH2-Kinase Linker Mutations Release Hck Tyrosine Kinase and Transforming Activities in Rat-2 Fibroblasts*

Biochemical and structural studies of Src and related kinases demonstrate that two intramolecular interactions suppress kinase activity. These interactions involve binding of the SH2 domain to a phosphotyrosine residue in the C-terminal tail and association of the SH3 domain with a polyproline type II helix formed by amino acids linking the SH2 and kinase domains. Recent studies have shown that high affinity interaction of the SH3 domain of Hck with the human immunodeficiency virus type I Nef protein activates Hck tyrosine kinase and biological activities, suggesting a mechanism that involves disruption of the SH3-linker interaction. To test the role of this interaction in the regulation of Hck kinase activity in living cells, we substituted alanines for prolines 225 and 228 in the linker region and observed that the resulting mutant (Hck-2PA) demonstrated strong transforming activity in a Rat-2 fibroblast focus-forming assay. Hck-2PA also exhibited elevated tyrosine kinase activity in terms of autophosphorylation, endogenous substrate phosphorylation, and in anin vitro kinase assay. The transforming and kinase activities of Hck-2PA were remarkably similar to those observed with a Hck mutant activated by Phe substitution of the conserved tail Tyr residue and with wild-type Hck following co-expression with human immunodeficiency virus Nef. Introduction of the 2PA and tail mutations into a single Hck expression construct did not increase kinase or transforming activity relative to the individual mutations. These data provide new evidence that SH3-linker interaction may represent the dominant mechanism controlling Hck tyrosine kinase activity in vivo.

Biochemical and structural studies of Src and related kinases demonstrate that two intramolecular interactions suppress kinase activity. These interactions involve binding of the SH2 domain to a phosphotyrosine residue in the C-terminal tail and association of the SH3 domain with a polyproline type II helix formed by amino acids linking the SH2 and kinase domains. Recent studies have shown that high affinity interaction of the SH3 domain of Hck with the human immunodeficiency virus type I Nef protein activates Hck tyrosine kinase and biological activities, suggesting a mechanism that involves disruption of the SH3-linker interaction. To test the role of this interaction in the regulation of Hck kinase activity in living cells, we substituted alanines for prolines 225 and 228 in the linker region and observed that the resulting mutant (Hck-2PA) demonstrated strong transforming activity in a Rat-2 fibroblast focusforming assay. Hck-2PA also exhibited elevated tyrosine kinase activity in terms of autophosphorylation, endogenous substrate phosphorylation, and in an in vitro kinase assay. The transforming and kinase activities of Hck-2PA were remarkably similar to those observed with a Hck mutant activated by Phe substitution of the conserved tail Tyr residue and with wild-type Hck following co-expression with human immunodeficiency virus Nef. Introduction of the 2PA and tail mutations into a single Hck expression construct did not increase kinase or transforming activity relative to the individual mutations. These data provide new evidence that SH3linker interaction may represent the dominant mechanism controlling Hck tyrosine kinase activity in vivo.
Hck is a member of the Src protein-tyrosine kinase family and is expressed primarily in granulocytes, monocytes, and macrophages (1)(2)(3). Several lines of evidence suggest that Hck regulates phagocyte differentiation and function. Hck expression is strongly induced by agents that promote macrophage differentiation and priming of the respiratory burst (4,5). Hck also associates with the Fc receptor and is activated following receptor engagement (6 -8). In this way, Hck may couple the Fc receptor to activation of the respiratory burst.
Other studies have implicated Hck in hematopoietic cytokine signal transduction. Interleukin-3, granulocyte-macrophage colony-stimulating factor, and leukemia inhibitory factor have all been shown to induce Hck kinase activation (9 -11). Hck has been shown to associate with the common ␤ subunit of the interleukin-3 and granulocyte-macrophage colony-stimulating factor receptors (12). Association occurs through the Hck SH3 and SH2 domains, suggesting that receptor binding may induce Hck activation (see below). Constitutive activation of Hck by gene targeting greatly reduced the leukemia inhibitory factor requirement for suppression of embryonic stem cell differentiation, suggesting a role for Hck in early development (10). This previous study demonstrated interaction of Hck with gp130, the signal transducing subunit of the leukemia inhibitory factor receptor that is also utilized by interleukin-6, ciliary neurotrophic factor, and other cytokines (13).
Hck exhibits the structural organization characteristic of all Src family kinases, including a unique N-terminal region, SH3, SH2, and kinase domains, and a negative regulatory tail (14). The N-terminal region contains sites for Hck myristylation and palmitylation, which promote plasma membrane targeting (15). The SH3 and SH2 domains bind to proline-rich and phosphotyrosine-containing peptide sequences, respectively, which drive intermolecular interactions required for substrate recruitment and subcellular localization (16). In addition, the SH3 and SH2 domains are essential for negative regulation of Src family tyrosine kinase activity. Both biochemical and structural studies support an intramolecular model of negative regulation in which the SH2 domain interacts with the conserved phosphotyrosine residue in the C-terminal tail. Mutations in the SH2 domain or conversion of the C-terminal tail Tyr residue to Phe release the kinase and transforming activities of Src family members (14). Mutations in SH3 also activate Src family tyrosine kinases (17,18). The x-ray crystal structures of both Hck (19) and Src (20,21) show that the SH3 domain makes intramolecular contacts with a polyproline type II helix formed by the amino acids linking the SH2 and kinase domains (SH2kinase linker). Linker residues outside of the polyproline helix also interact with the N-terminal lobe of the kinase domain, and this interaction may be directly responsible for suppression of kinase activity (19,22).
Recent work from our laboratory suggests that disruption of the SH3-linker interaction may be sufficient for activation of Hck in vivo (23). These studies utilized human immunodeficiency virus type I Nef as a ligand for the Hck SH3 domain, which binds to a conserved proline-rich Nef motif with the highest affinity known for an SH3-mediated interaction (24). We observed that co-expression of Hck with Nef in Rat-2 fibroblasts stimulates Hck tyrosine kinase activity, resulting in cellular transformation. Transformation correlated with Hck-Nef complex formation and required the proline-rich motif of Nef. These data suggest that activation of Hck by Nef results from displacement of the SH3 domain from the SH2-kinase linker and points to a central role for SH3-linker interaction in the negative regulation of Src family kinases in vivo. In the present study, we show that Hck mutants with proline to alanine substitutions in the SH2-kinase linker region exhibit constitutive tyrosine kinase activity and induce transformation of Rat-2 fibroblasts. The Hck linker mutant exhibited similar kinase activity and transforming potential to wild-type Hck following activation by HIV Nef. These data provide new evidence that disruption of SH3-linker interaction is sufficient for Hck kinase activation in vivo and point to SH3 engagement as a general mechanism of Src family kinase activation.

EXPERIMENTAL PROCEDURES
Retroviral Expression Constructs-The amino acid numbering of all Hck mutants is based on the p59 form of human Hck. Construction of kinase-defective (K269E; Hck-KE) and tail-activated (Y501F; Hck-YF) mutants of Hck as well as the Nef mutant lacking the proline residues essential for SH3 binding (Nef-PA) has been described elsewhere (23,25). The Hck SH2-kinase linker Pro to Ala substitution mutants (P225A single mutant and P225A/P228A double mutant, referred to hereafter as Hck-2PA) were generated using the GeneEditor in vitro site-directed mutagenesis system from Promega. The Hck linker-tail combination mutant (P225A/P228A/Y501F; Hck-2PAYF) was generated by swapping a unique restriction fragment containing the Hck-YF mutation with the corresponding fragment in Hck-2PA. All Hck proteins were expressed using the retroviral vector pSR␣MSVtkneo (26). High titer stocks of recombinant retroviruses were generated by co-transfection of 293T cells with the retroviral vectors and an ecotropic packaging vector as described elsewhere (23). A negative control retrovirus was produced using the empty pSR␣ parent vector, which confers resistance to G418 only.
Transformation Assays-Rat-2 cells were obtained from the ATCC and grown in Dulbecco's modified Eagle's medium containing 5% fetal bovine serum and 50 g/ml gentamycin. Transformation of Rat-2 fibroblasts by Hck and Nef was assessed using a focus-forming assay as described previously (23) with the following modifications. Rat-2 fibroblasts (2 ϫ 10 4 ) were plated in each well of a six-well tissue culture plate 1 day prior to infection. The following day, cells were infected with the Nef, Nef-PA, or control retroviruses in a final volume of 5 ml. Polybrene was added to 4 g/ml, and plates were centrifuged at 1,000 ϫ g for 4 h at 18°C to enhance infection efficiency. After infection, the virus was aspirated and replaced with 5 ml of fresh medium. The following day, cells were reinfected with either the wild-type Hck or negative control virus using the same procedure. Two days later, cells were trypsinized and equally divided among four 60-mm dishes, and G418 was added to a final concentration of 800 g/ml. The selection medium was renewed every 3 days for 14 days, at which time transformed foci were visualized by Wright-Giemsa staining. Parallel cultures were lysed and tested for Nef and Hck protein expression or for Hck kinase activity as described below. To allow for a comparison of the transforming activity of the Hck-Nef combination with that of the activated Hck mutants, the same sequential infection method was used, with the control virus followed by the mutant Hck viruses.
Analysis of Hck Tyrosine Kinase Activity-Rat-2 cells (confluent 60-mm culture dish) were flash-frozen on liquid nitrogen and lysed by thawing in 1.0 ml of RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 1 mM EDTA, 1% sodium deoxycholate) or in 1.0 ml of Hck lysis buffer (50 mM Tris-HCl, pH 7.4, 50 mM NaCl, 1 mM EDTA, 10 mM MgCl 2 , 1% Triton X-100). Both buffers also contained 20 mM NaF, 1 mM Na 3 VO 4 , and 50 M NaMoO 4 . The lysates were clarified by centrifugation at 100,000 ϫ g for 15 min at 4°C, and protein concentrations were determined using the Bradford assay (Pierce). Hck was immunoprecipitated from equivalent amounts of total protein (0.5-1.0 mg) as follows: clarified lysates were incubated with 1 g of anti-Hck polyclonal antibody (Santa Cruz Biotechnology) and 20 l of protein G-Sepharose (50% slurry; Amersham Pharmacia Biotech) for 2 h at 4°C. Immunoprecipitates were washed twice with either 1.0 ml of RIPA buffer or 1.0 ml of Hck lysis buffer followed by two washes with 1.0 ml of kinase buffer (50 mM HEPES, pH 7.4, 10 mM MgCl 2 ). Kinase buffer (20 l) containing 1 g of the tyrosine kinase substrate p50 (50-kDa GST fusion protein containing residues 331-443 of the Src substrate protein Sam 68 (27,28); Santa Cruz Biotechnology) and 5 Ci of [␥-32 P]ATP (3,000 Ci/mmol; NEN Life Science Products) were added, and the reactions were incubated for 15 min at 30°C. Reactions were stopped by adding SDS-PAGE sample buffer and heating to 95°C for 5 min. Radiolabeled p50 was visualized by storage phosphor technology using a Molecular Dynamics PhosphorImager.
To detect tyrosine phosphorylation of Hck in vivo, Rat-2 cells expressing Hck proteins were flash-frozen on liquid nitrogen and lysed by thawing in 1.0 ml of RIPA buffer as described previously. Hck was immunoprecipitated from clarified cell lysates, washed three times with 1.0 ml of RIPA buffer, resolved by SDS-PAGE, transferred to polyvinylidene difluoride membranes, and probed with anti-phosphotyrosine (PY20; Transduction Laboratories) or anti-Hck monoclonal antibodies (Transduction Laboratories).

Activation of Hck in Rat-2 Cells Co-expressing Hck and Nef-
Recent work from our laboratory has shown that co-expression of Nef and Hck in Rat-2 fibroblasts induces cellular transformation (23). This effect correlated with Hck-Nef complex formation and enhanced phosphorylation of multiple cellular proteins on tyrosine, strongly suggesting that Hck is activated by Nef in vivo. To investigate whether Hck tyrosine kinase activity is elevated in the presence of Nef in the transformed cells, Hck was immunoprecipitated from populations of Rat-2 cells expressing Hck alone, Nef alone, or the two in combination. The Hck immunoprecipitates were incubated in vitro with [␥-32 P]ATP and a 50-kDa GST-Sam 68 fusion protein (p50) as substrate. As shown in Fig. 1, Hck immunoprecipitates from the Hck-Nef transformants exhibited both strong autophosphorylation and substrate phosphorylation in vitro. The extent of p50 substrate phosphorylation was consistently 8 -10-fold higher than that observed with identical immunoprecipitates from cells expressing Hck alone. The requirement for the Nef SH3-binding motif was investigated by co-expressing Hck with a Nef mutant containing alanine substitutions for the prolines, which are essential for SH3 binding (Nef-PA mutant). Previous work has shown that this Nef mutant is unable to cooperate with Hck in the Rat-2 transformation assay (23). As shown in Fig. 1, Nef-PA did not increase Hck tyrosine kinase activity 1 The abbreviations used are: HIV, human immunodeficiency virus; GST, glutathione S-transferase; PAGE, polyacrylamide gel electrophoresis. relative to Hck alone. Control blots show that equivalent amounts of Hck were present in each of the immunoprecipitates and that the Nef proteins were expressed at equivalent levels in each case. These results demonstrate that co-expression with Nef induces Hck activation in Rat-2 fibroblasts and are consistent with our previous transformation data.
Proline to Alanine Substitutions in the SH2-Kinase Linker Region Release Hck Transforming and Tyrosine Kinase Activities in Rat-2 Cells-Results presented in Fig. 1, together with previous transformation data from our laboratory, strongly suggest that Nef activates Hck by binding to its SH3 domain and displacing the negative regulatory interaction with the linker region. These data point to a central role for intramolecular SH3-linker interaction in the negative regulation of Hck and possibly other Src family members. To test this idea more directly, Hck Pro residues 225 and 228, which are critical for the formation of the linker polyproline type II helix, were changed to alanines by site-directed mutagenesis. The resulting mutant (Hck-2PA) was expressed in Rat-2 fibroblasts and assayed for transforming activity using a focus-forming assay. As shown in Fig. 2, the Hck-2PA mutant exhibited strong transforming activity, producing foci similar in number and size to those observed with a Hck mutant lacking the conserved tail tyrosine residue (Tyr-501; Hck-YF mutant). Fig. 2 also shows that Hck-2PA exhibited transforming activity close to that observed with the combination of Hck and Nef. We also investigated whether the combination of the linker and tail mutations led to any additional increase in transforming activity using a Hck mutant bearing both the linker proline to alanine substitutions and the tail Tyr to Phe mutation (Hck-2PAYF mutant). As shown in Fig. 2, the presence of both mutations did not affect overall transforming activity, suggesting that individual mutations within the linker or the tail disrupt the entire negative regulatory apparatus of Hck. Consistent with this finding is our observation that co-expression of Hck-2PA with Nef did not increase transforming activity relative to Hck-2PA alone (data not shown). Thus, the Hck-2PA mutation appears to fully disrupt negative regulation by the SH2-kinase linker region.
To determine whether the SH2-kinase linker mutants of Hck demonstrated elevated tyrosine kinase activity, Hck was immunoprecipitated from the retrovirally infected Rat-2 cell pop-ulations shown in Fig. 2 and assayed in vitro with the p50 substrate protein and [␥-32 P]ATP. As shown in Fig. 3, Hck-2PA-mediated phosphorylation of p50 was 6-fold higher than that observed with wild-type Hck. This activity level is close to that observed with Hck-YF, which exhibited 8-fold higher activity than wild-type Hck. Finally, the Hck-2PAYF mutant demonstrated kinase activity equivalent to that of Hck-YF. These results are in agreement with the transformation data shown in Fig. 2 and demonstrate that proline mutations in the Hck SH2-kinase linker region are sufficient to deregulate Hck kinase activity and induce cellular transformation. Furthermore, the SH2-kinase linker mutants showed transforming and kinase activities very similar to those observed in Rat-2 cells following co-expression of Hck and Nef. These results agree with the idea that displacement of the SH3 domain of Hck from the SH2-kinase linker is sufficient for the activation of Hck in vivo.
We also investigated whether mutagenesis of a single proline residue in the Hck SH2-kinase linker region was sufficient to release transforming activity. For these experiments, Ala was substituted for Pro-225, which is conserved among all Src kinase family members. The resulting mutant (Hck-P225A) exhibited less than 10% of the Hck-2PA transforming and tyrosine kinase activities in Rat-2 cells, suggesting that both linker proline residues contribute to negative regulation as predicted by the crystal structure (data not shown). Unlike Hck-2PA, co-expression of Hck-P225A with Nef resulted in enhanced focus-forming activity, providing additional evidence for regulation by the linker despite loss of Pro-225. The presence of two proline residues in the Hck, Lyn, Lck, and Blk linker regions suggests a more dominant role for the linker in kinase regulation relative to Src, Fyn, Yrk, Fgr, and Yes, which contain only a single Pro residue (see "Discussion").
Autophosphorylation and Endogenous Substrate Phosphorylation by the Hck SH2-Kinase Linker Mutants-To determine whether the Hck linker mutants also exhibited enhanced autophosphorylation, anti-Hck immunoprecipitates were prepared from infected Rat-2 fibroblasts and probed with the anti-phosphotyrosine antibody, PY20. As shown in Fig. 4, immunoprecipitates of all three constitutively activated forms of Hck (2PA, YF, 2PAYF) showed a marked increase in phosphotyrosine content compared with wild-type Hck. A faint anti-phosphotyrosine band FIG. 2. SH2-kinase linker mutations release Hck transforming activity in Rat-2 fibroblasts. Rat-2 fibroblasts were sequentially infected with the following combinations of recombinant retroviruses as described under "Experimental Procedures": control virus carrying neomycin resistance alone followed by control again (Con); control followed by wild-type Hck (Hck-WT); control followed by kinase-defective Hck (Hck-KE); Nef followed by control (Nef); control followed by Hck mutant with Ala substitutions of linker prolines 225 and 228 (Hck-2PA); control followed by Hck tail mutant with the Phe substitution of Tyr 501 (Hck-YF); control followed by Hck linker-tail combination mutant (Hck-2PAYF); and Nef followed by Hck wild-type (HckϩNef). 48 h after infection, cells were replated and incubated under G418 selection for 14 days. Transformed foci were visualized by Wright-Giemsa staining. This experiment was repeated three times with nearly identical results. Representative plates are shown. FIG. 3. Release of Hck protein-tyrosine kinase activity by SH2kinase linker mutations. Rat-2 fibroblasts expressing Hck wild-type (WT), KE, 2PA, YF, or 2PAYF were extracted with Hck lysis buffer. Extracts from fibroblasts infected with a retrovirus carrying only the neo selection marker were included as a negative control. Clarified lysates were incubated with anti-Hck polyclonal antibodies, and Hckantibody complexes were precipitated with protein G-Sepharose. The immunoprecipitates were washed with Hck lysis buffer followed by kinase buffer and then incubated with [␥-32 P]ATP and a GST-Sam 68 fusion protein of 50 kDa (p50). Following incubation, Hck and p50 were resolved by SDS-PAGE, transferred to polyvinylidene difluoride membranes, and phosphorylated p50 was visualized by storage phosphor imaging (upper panel). The membrane was probed with anti-Hck monoclonal antibodies to ensure equivalent recovery of Hck in the immunoprecipitates (lower panel). The positions of p50 and autophosphorylated Hck are indicated by the arrows.
was observed with wild-type Hck, which is likely to correspond to the tail-phosphorylated, inactive form of the protein. The kinasedead form of Hck (Hck-KE) showed no evidence of tyrosine autophosphorylation and serves as a negative control. Aliquots of the Hck immunoprecipitates were also blotted with Hck antibodies to demonstrate that equivalent amounts of Hck were recovered in each case (Fig. 4, bottom panel). These results show that linker mutations are sufficient to promote Hck autophosphorylation and are consistent with the kinase activity assay and transformation results.
Previous work from our laboratory demonstrated the presence of many endogenous tyrosyl phosphoproteins in fibroblasts transformed by co-expression of Hck and Nef or by the tail mutant of Hck (Hck-YF) (23). Most prominent among these was a 40-kDa phosphoprotein (p40). Although the identity of this endogenous Hck substrate is unknown, it serves as a useful transformation-associated tyrosyl phosphoprotein marker. We therefore probed whole cell protein extracts of Rat-2 cells transformed by the SH2-kinase linker mutants with antiphosphotyrosine antibodies. As shown in Fig. 5, transformation by the Hck-2PA mutant is associated with tyrosine phosphorylation of a wide array of endogenous proteins, including very potent phosphorylation of p40. The extent of endogenous p40 tyrosine phosphorylation is comparable with that observed with the tail mutant of Hck as well as the linker-tail combination mutant. No detectable endogenous protein phosphorylation was observed in cells infected with a control retrovirus or expressing wild-type or kinase-dead Hck. DISCUSSION Previous work from our laboratory established that the SH3mediated interaction of Nef with Hck is sufficient to induce oncogenic transformation of Rat-2 fibroblasts (23). This result suggested that Nef constitutively stimulates Hck protein-tyrosine kinase activity in vivo, which is responsible for the transformed phenotype. Here we provide new evidence supporting this model in living cells. Hck isolated from transformed Rat-2 cells co-expressing Nef shows increased autophosphorylation and is highly active toward a substrate protein in vitro. Hck activation is dependent upon the SH3-interaction motif of Nef. Our data are in good agreement with previous work by Moarefi et al. (29), which shows that the purified inactive form of Hck is activated by SH3-mediated interaction with Nef in vitro.
A likely mechanism of Hck activation by Nef is suggested by the crystal structures of Hck and c-Src (19 -21). The structure reveals that the closed, inactive form of Hck is regulated by intramolecular interactions involving the SH2 and SH3 domains. One involves the binding of the Csk-phosphorylated tail to the SH2 domain, a finding not unexpected given the wealth of mutagenesis data demonstrating the critical role of the conserved tail tyrosine residue in the negative regulation of Src family kinases (14). A second interaction involves the binding of the SH3 domain to a polyproline type II helix formed by the linker region connecting the SH2 and kinase domains. Our results support a mechanism in which Nef binds to the SH3 domain of Hck and disrupts the interaction with the linker region, leading to kinase activation. Such a model implies that SH3-linker interaction is critical to the negative regulation of Hck. In this report, we provide further evidence for this hypothesis. Substitution of alanines for the linker proline residues within the SH2-kinase linker region releases Hck tyrosine kinase activity in vivo, leading to cellular transformation. Mutagenesis of these prolines is likely to contribute to constitutive Hck kinase activation and signaling by two mechanisms. First, loss of these prolines and attendant loss of SH3 interaction is likely to disrupt proper alignment between Trp-260 and other linker residues that contact the N-terminal lobe of the kinase domain. These interactions have been proposed to mediate suppression of kinase activity (see below). Second, loss of interaction with the linker may expose the SH3 domain, promoting recruitment of substrate proteins essential for transformation signaling. Numerous studies have suggested that Src family kinase SH3 domains have an important role in substrate recruitment (25, 30 -32). Interaction of Hck with substrate proteins via SH3 may contribute to kinase activation under physiological conditions.
Our observation that substitution of Pro residues in the Hck linker release kinase and transforming activities are in agreement with those of Gonfloni et al. (33), who demonstrated that mutations in the linker region of c-Src resulted in NIH 3T3 cell transformation and endogenous substrate phosphorylation. In contrast to our results, however, this c-Src linker mutant (K249E/P250E) showed much lower focus-forming activity when compared with c-Src activated by mutation of the conserved tail tyrosine residue, despite similar kinase activity.  This result led the authors to speculate that the Src linker region may interact with target proteins required for transformation signaling. In the case of Hck, our data suggest that the linker may play a more dominant role in kinase regulation rather than protein-protein interaction, at least in fibroblasts.
Another recent study has investigated the role of Trp-260 in the regulation of Hck kinase activity in vitro. This residue is part of the SH2-kinase linker region but lies outside of the polyproline type II helix that interacts with the SH3 domain. The crystal structure shows that Trp-260 contacts the ␣C-helix in the N-terminal lobe of the kinase domain, and this interaction has been proposed to prevent the kinase from adopting an active conformation. LaFevre-Bernt et al. (22) changed this Trp to Ala (Hck-W260A mutant) and observed higher specific activity than wild-type Hck. Interestingly, this mutant was refractory to further activation following ligand binding to either the SH2 or SH3 domain, indicating a central role for Trp-260 in coupling the regulatory domains to kinase activation. We observed that Hck-W260A exhibits focus-forming activity in the Rat-2 transformation assay (data not shown). However, the resulting foci were smaller and fewer in number than those produced by the Hck linker proline and tail mutants or the combination of Hck and Nef. This difference may result from the altered accessability of the Hck-W260A SH2 and SH3 domains to critical transformation-related signaling partners.