Abstract

Background

Dysregulated RTK activity is closely associated with a broad range of human malignancies. Thereby, understanding the complex mechanisms by which RTKs are regulated is of clinical significance due to potential applications for therapeutic intervention against cancers [1,2]. Overexpression or constitutive activation of the Ron receptor has been demonstrated to drive strong oncogenic phenotypes in cancers derived from breast, colon, lung, and prostate [3-5]. The oncogenic potential of Ron is represented by its capacity to induce migration, invasion, growth, survival and epithelial-mesenchamal transition of epithelial tumor cells, as well as to promote pro-tumoral activities of tumor-associated macrophages [6-13]. Inhibition of Ron expression using Ron specific siRNA dramatically decreases tumorigenic and invasive activities in colorectal carcinoma cells [14]. A human neutralizing antibody against Ron exhibits partial inhibition of colon, lung and pancreatic tumor growth in xenograft models [15], and a small molecule kinase inhibitor of Ron that has antitumor activity in vivo has been described [16]. Consequently, Ron is a promising target for therapeutic intervention against tumorigenic activities and malignant phenotypes [3,17,18].

Recepteur d'originenantais (Ron) (called STK in mice and Sea in chickens) belongs to the Met proto-oncogene family [19,20]. Originally synthesized as a single chain precursor, Ron is processed into a disulfide-linked heterodimer with a transmembrane β chain and an extracellular α chain [21,22]. Binding of macrophage stimulating protein (MSP), the ligand for Ron, to the extracellular domain of Ron induces receptor dimerization, conferring catalytic activity to the receptor [23]. The intracellular region of Ron includes the juxtamembrane domain, the highly conserved kinase domain and the non-catalytic c-terminal tail. The kinase domain consists of two lobes. The N-lobe contains the αC helix and the P loop which are essential for ATP recruitment, while phosphorylation of two tyrosines (Y¹²³⁸ and Y¹²³⁹) in the C-lobe activation loop is required for receptor activation. Subsequent phosphorylation of two docking tyrosines in the c-terminal tail transduces a variety of signaling pathways mediated by Ron [22,24] (Figure 1A).

Open in a separate window

Figure 1

Schematic diagram of the Ron receptor. A) Domains of the Ron receptor. P loop (orange) and αC helix (blue) in the kinase N lobe and activation loop (yellow) in the C lobe are denoted by different colors. Critical tyrosines in the activation loop (P-2Y: Y^{1238, 1239}) and c-terminal tail (P-Y: Y^1353,1360) are indicated. B) The juxtamembrane domain of Ron is divided into four sequential regions: JM-A (residue 982–1008), the region at the N terminus; JM-B (1009–1035), the region corresponding to the sequence deleted in mRon; JM-C (1036–1047), the region with an acidic patch of residues; and JM-D (1048–1081), the region at the c-terminus of the juxtamembrane domain. Residues discussed herein are indicated.

The juxtamembrane domain, which forms an important layer of constraint, limiting promiscuous enzymatic activity of RTKs, is found frequently mutated in tumor cells [25-32]. The regulatory mechanisms mediated by the juxtamembrane domain however, varies significantly due to the varied length and lack of sequence similarity among RTKs. Murine Ron harboring a deletion of 27 residues in the juxtamembrane domain, caused by exclusion of a potential exon, exhibits significantly higher receptor activity than human Ron [33-36]. Deletion of the 27 residues in human Ron promotes receptor activation, highlighting a potential role for the juxtamembrane domain in the regulation of the Ron receptor activation [33]. However, the molecular mechanisms underlying this regulation are unknown.

Herein, we have characterized the mechanism by which the juxtamembrane domain contributes to the auto-regulation of Ron activity through structure/function analysis. While tyrosines in the Ron juxtamembrane domain are inconsequential for receptor activation, an acidic JM-C region in the juxtamembrane domain was found to be critical for Ron autoinhibition. Phosphorylation of Y¹¹⁹⁸ in the kinase domain promoted Ron activation, primarily by accelerating relief of the autoinhibition imposed by the JM-C region. Based on our experimental data, we have proposed molecular models of the inactive and the active Ron structure, which cast a new light on the mechanisms controlling Ron mediated signal transduction.

Results

The acidic JM-C region in the Ron juxtamembrane domain plays a critical role in receptor autoinhibition

The juxtamembrane domain of Ron is significantly longer than that of most RTKs, therefore we divide it into four sequential regions for clarity (Figure 1B). Previous studies have mapped the functional difference between murine Ron and human Ron to the 27 residues comprising the JM-B region, which contains Y¹⁰¹² and Y¹⁰¹⁷, the only two tyrosines in the juxtamembrane domain (Additional file 1: Figure S1A). Tyrosines in the juxtamembrane domain of RTKs such as the EphR, Flt3 and Kit [29,32,37] are critical for maintenance of the inactive kinase, and phosphorylation of these tyrosines relieves receptor autoinhibition. However, consistent with previous reports [33,38], mutation of Y¹⁰¹² and Y¹⁰¹⁷ individually or in combination into phenylalanine, glutamic acid or alanine in the context of Ron did not affect receptor autophosphorylation, downstream Erk phosphorylation (Additional file 1: Figure S1B), or induction of AP1 luciferase activity (Additional file 1: Figure S1C). Thus, unlike other RTKs, tyrosines in the juxtamembrane domain of Ron do not regulate receptor activation.

While the primary sequence differs, acidic residues (glutamic acid and aspartic acid) interspersed with serines in the Ron JM-C region (Figure 2A) are relatively conserved in the closely related Met receptor, suggesting that this region might serve a conserved function. In order to determine the potential role of this region in the regulation of Ron receptor activity, we constructed a Ron mutant in which nine residues (most acidic) in the JM-C region were deleted (Ron^1039–1047) (Figure 2A). HA-tagged wild-type Ron and Ron variants were transiently transfected in HEK 293 cells. Phosphorylation of the activation loop tyrosines (Y^1238/9), a hallmark of kinase activation, and induction of a major signaling cascade downstream of Ron, the Ras-Map kinase pathway, were examined. Ron and the variants were also transiently transfected with an AP-1 luciferase reporter, and the induced AP-1 transcriptional response was measured. Equivalent levels of protein expression and cell surface expression of Ron and its variants were demonstrated by probing for HA and by the flow cytometry analysis (Additional file 2: Figure S2), respectively. Compared with the wild type receptor, Ron^1039–1047 exhibited a significant increase in receptor autophosphorylation, induced Erk phosphorylation (Figure 2B) and AP-1 transcriptional activity both in the presence and absence of MSP (Figure 2C). In order to further determine whether the negative electrostatic nature of the JM-C region is responsible for its role in the regulation of Ron activity, E¹⁰⁴⁴, D¹⁰⁴⁵ and E¹⁰⁴⁶ in the JM-C region were mutated to alanine individually and in combination. Single mutants did not show remarkable catalytic differences compared with wild type Ron (data not shown). However, the triple mutation (EDE-AAA) dramatically upregulated receptor activity and downstream signaling to a level comparable to that induced by Ron^1039–1047 (Figure 2B and C). Based on these observations, we have identified a novel segment in the juxtamembrane domain of Ron, the JM-C region, which plays a central role in kinase autoinhibition, primarily mediated by the acidic residues.

Open in a separate window

Figure 2

Acidic residues in the JM-C region of Ron promote receptor autoinhibition. A) Schematic of mutants. Mutations are highlighted in bold. Residues deleted are underlined. B) 293 cells were transfected with HA-tagged wild-type or mutant Ron as indicated in the presence and absence of MSP. Lysates were blotted for p-Ron1238/9Y, HA, p-Erk, Erk and β-tubulin. C) Relative luciferase activity in 293 cells co-transfected with wild-type or mutant Ron as indicated, and an AP-1 luciferase reporter in the presence and absence of MSP. The data were presented as the average±standard deviation from measurements collected from three independent experiments and three measurements per experiment. Wild type Ron −/+MSP was used as controls, unless denoted with ‘︹’. The significant differences of values at p<0.05, p<0.01, and p<0.001 levels are indicated by *, **and ***respectively.

Serines (S¹⁰³⁹,S¹⁰⁴¹,S¹⁰⁴³,S¹⁰⁴⁷) are major components of the acidic JM-C region, the first three of which are predicted to be potential phosphorylation sites (NetPhos 2.0 Server) (data not shown) (Figure 3A). In order to determine whether phosphorylation of these serines could play a role in the regulation of Ron receptor activity by the JM-C region, we mutated the serines to alanines individually and in combination. While individual and double mutants performed similarly as wild-type Ron (data not shown), the triple and quadruple mutations (3S-A and 4S-A) resulted in an evident increase in receptor activation and downstream signaling (Figure 3B and C). However, this activation was not as striking as that driven by mutations of the acidic ¹⁰⁴⁴EDE residues. Thus, while the acidic residues in the JM-C region maintain the primary inhibitory constraints, phosphorylation of serines in this region may also contribute to the Ron autoinhibition mediated by the JM-C region.

Open in a separate window

Figure 3

Serines in the JM-C region play a subsidiary role in the regulation of Ron receptor activation. A) Schematic of mutants. Mutations are highlighted by underline. B) 293 cells were transfected with HA-tagged wild-type or mutant Ron as indicated in the presence and absence of MSP. Lysates were blotted for p-Ron1238/9Y, HA, p-Erk and Erk. C) Relative luciferase activity in 293 cells co-transfected with wild-type or mutant Ron as indicated, and an AP-1 luciferase reporter in the presence and absence of MSP.

The JM-B region regulates receptor activity indirectly through its impact on the spatial orientation of the JM-C constraint

Previous studies demonstrated that deletion of the JM-B region of human Ron (P¹⁰⁰⁹-V¹⁰³⁵), absent in murine Ron, resulted in elevated receptor activity, as observed in murine Ron [33]. However, our studies suggest that individual residues in the JM-B region including tyrosines, are dispensable for Ron activation. Therefore, we reasoned that the JM-B region might play a structural role in Ron autoinhibition. Consistent with previous studies, here we found that Ron^JM-B possessed higher intrinsic catalytic capability than the wild type receptor, as demonstrated by a dramatic increase in phosphorylation of the activation loop tyrosines and a docking tyrosine, and in expression of an AP-1 reporter in both the presence and the absence of ligand (Figure 4B and C). In order to determine the significance of the length of the JM-B region, a Ron variant with a smaller deletion of ten residues (G¹⁰²⁰-L¹⁰²⁹) in the JM-B region, Ron^JM-B-S, which did not result in deletion of the tyrosines in the JM-B region, was constructed (Figure 4A). Surprisingly, while Ron^JM-B exhibited enhanced receptor activity, Ron^JM-B-S exhibited decreased receptor activity, and the induced AP-1 transactivation was significantly reduced even in the presence of MSP (Figure 4C and D). These data indicate that the length of the JM-B region, rather than individual residues, appears to be a critical determinant for the proper regulation of Ron activity.

Open in a separate window

Figure 4

The JM-B region modulates Ron receptor activation by affecting the orientation of the JM-C region. A) Schematic of mutants. Deletions of different lengths in the JM-B region are indicated by arrows. Mutated residues are highlighted in bold. B) &D) 293 cells were transfected with HA-tagged wild-type or mutant Ron in the presence and absence of MSP. Lysates were immunoprecipitated with anti p-Y and blotted for HA, or directly blotted for p-Ron1238/9, p-Ron1360 and HA, p-Erk and Erk. C) &E) Relative luciferase activity in 293 cells co-transfected with wild-type or mutant Ron, and an AP-1 luciferase reporter in the presence and absence of MSP.

Since the acidic JM-C region lies c-terminal to the JM-B region of Ron, we hypothesized that the JM-B region might modulate receptor activity indirectly by affecting localization of the JM-C region. In order to test this hypothesis, the EDE-AAA mutation in the JM-C region was generated in the context of Ron^JM-B and Ron^JM-B-S (Figure 4A). Since both JM-B and EDE-AAA were able to relieve Ron autoinhibition, as expected, Ron^JM-B&^EDE-AAA exhibited enhanced catalytic activity both in the presence and absence of MSP (Figure 4D and E). However, the levels of AP-1 luciferase activity induced by Ron^JM-B&^EDE-AAA were similar to those induced by Ron^EDE-AAA, suggesting that the deletion of JM-B region and neutralization of the acidic JM-C region may facilitate receptor activation in an analogous manner. Alternatively, although the JM-B-S deletion prohibited receptor activation, introduction of the EDE-AAA mutation into the Ron^JM-B-S construct largely restored receptor activity (Figure 4D and E). Thus, it is likely that the JM-B region modulates activation of the Ron receptor by affecting the spatial orientation of the adjacent JM-C region, which places autoinhibitory constraints on the kinase due to its electronegative characteristics.

Y¹¹⁹⁸ in the kinase domain mediates Ron autoinhibition imposed by the JM-C region

Previous studies indicate that Y¹¹⁹⁴ in Met, the equivalent residue of Y¹¹⁹⁸ in Ron, is one of the three major receptor autophosphorylation sites, and mutation of Y¹¹⁹⁴ to phenylalanine in Met results in dramatically decreased receptor activity [39,40]. Here we show that the equivalent Y1198F mutation in Ron dramatically reduced Ron autophosphorylation, as well as induction of downstream Erk phosphorylation and AP-1 transcriptional activation both in the presence and absence of MSP (Figure 5B and C). These results indicate that phosphorylation of Y¹¹⁹⁸ in the αE helix of the kinase C lobe plays a key role in the activation of wild type Ron. Although Y¹¹⁹⁸ and the juxtamembrane domain are unlikely to sterically interact with each other based on their distance, the potential functional interrelation was evaluated by introduction of the Y1198F mutation in the context of Ron^EDE-AAA and Ron^1039–1047 (Figure 5A). Alterations in the JM-C region, which were demonstrated to relieve Ron autoinhibition, successfully rescued the loss of receptor activity caused by the 1198Y-F mutation. The double mutants recovered receptor activity to a level comparable to wild type Ron, and responded to stimulation by MSP (Figure 5D and E). These studies suggest that phosphorylation of Y¹¹⁹⁸ may play an essential role in relieving the inhibitory constraints on Ron activation imposed by the JM-C region in an indirect manner.

Open in a separate window

Figure 5

Y¹¹⁹⁸ in the kinase domain of Ron regulates the autoinhibitory constraints on receptor activation imposed by the JM-C region. A) Schematic of the Ron receptor. Y¹¹⁹⁸ at the interface of the αE helix and the αC helix is highlighted. B) &D) 293 cells were transfected with HA-tagged wild-type or mutant Ron as indicated in the presence and absence of MSP. Lysates were blotted for p-Ron1238/9Y, HA, p-Erk and Erk. C) &E) Relative luciferase activity in 293 cells co-transfected with wild-type or mutant Ron as indicated, and an AP-1 luciferase reporter in the presence and absence of MSP.

Discussion

Mutations in the juxtamembrane domain are a common mechanism by which RTKs become constitutively activated and drive progression of a variety of human cancers. Internal tandem duplications in the juxtamembrane domain of Flt3 have been identified in patients with acute myeloid leukemia [41]. Deletions in the juxtamembrane domain of both KIT [42] and PDGFR [28] have been identified in human gastrointestinal stromal tumors. Germ line point mutations in the juxtamembrane domain of Met have been identified in small cell lung cancer [27] and gastric cancer [43]. An alternatively spliced form of Met, missing exon 14 which encodes part of its juxtamembrane domain [44], exhibits a highly tumorigenic phenotype [31,45]. Likewise, a splice variant of RON, missing exon 14 (BG289902) in the juxtamembrane domain, was found in a bladder papilloma cell line (human Expressed Sequence Tag data base). The juxtamembrane domain of RTKs varies in length and lacks apparent sequence similarity, suggesting that it may contribute to specific aspects of regulation within individual receptor families. Our studies demonstrate that, like other RTKs, the juxtamembrane domain of Ron plays a central role in receptor autoinhibition. However, the mode of regulation conferred by the juxtamembrane domain of Ron appears to be novel compared with other RTKs described to date.

Tyrosine phosphorylation in the juxtamembrane domain of RTKs including Kit, Flt3 and EphR, is a common mechanism by which the receptor switches from the default closed state to an active configuration [29,32,37,46]. However, phosphorylation and the aromatic side chain of tyrosines in the JM-B region are dispensable for Ron activity. Conversely, deletions of different lengths in the JM-B region were found to have significant but opposite impacts on receptor activation, probably through affecting the spatial orientation of the contiguous JM-C region.

Here, we show that the acidic cluster in the JM-C region of Ron is a crucial regulator of receptor activity, and disruption of its negative electrostatic property dramatically facilitates kinase activation. The acidic characteristics of the JM-C region are relatively conserved within the Met family members. The equivalent region in Met was found to contain a mutation (S1058P) in a sample from the non small cell lung cancer [47], suggesting the potential functional significance of this region. In order to better understand the inhibitory mechanism imposed by the JM-C region, we compared the kinase domain sequences of 150 receptors by mapping them onto the crystal structure of the Ron kinase domain (3PLS) [48] (Figure 6). Based on surface accessibility and relative location, a conserved region on the surface of the Ron kinase N lobe was identified as a putative JM-C binding region (orange in Figure 6A). It locates adjacent to the strictly conserved active site (red in Figure 6A), with a predominantly positive surface potential (blue in Figure 6B).

Open in a separate window

Figure 6

Conserved surface residues and surface potential of the Ron kinase domain (3PLS) [[48]]. The conserved region adjacent to the active site with a predominantly positive surface potential is indicated by arrows as the putative JM-C binding site. A) Conserved residues on the surface of human kinase domains are indicated, with red/orange representing highly conserved residues and blue/cyan representing highly variable residues. B) Surface potential of the Ron kinase domain, with blue representing positive potential and red representing negative potential. The units of the scale were −60.362~60.362.

Based on the crystal structure of the Ron kinase domain, we built a molecular model of the inactive Ron structure (including JM-C, JM-D region and kinase domain) [48-51] (Additional file 3: Figure S3A), which suggests that the JM-C region (magenta) promoting a helical conformation could make close contacts with the P loop (orange) and activation loop (yellow). Acidic residues in the JM-C region (E¹⁰⁴⁴-D¹⁰⁴⁵-E¹⁰⁴⁶) have the potential to interact with H¹⁰⁹² in the P-loop, Y¹²³⁸ and Q¹²⁴³ in the activation loop, and the side-chains of S¹¹¹⁷ and R¹¹¹⁸ between the β3 strand and the αC helix (lines in Additional file 3: Figure S3A). These interactions would interfere with the accommodation of ATP and substrates by the glycine rich P-loop. Spatial proximity of the JM-C region could also lock the activation loop in the inter-lobe cleft and prevent productive rotation of the N and C lobe. In addition, the rigid configuration promoted by the JM-C region, of the β3 strand and αC helix in the kinase N lobe, puts the invariant K¹¹¹⁴ and E¹¹³⁰ that participate in formation of a salt bridge required for ATP binding in catalytically disfavored positions.

Located at the interface between the αE and the αC helix, Y¹¹⁹⁸ of Ron is highly conserved among RTKs and shares structural conservation among close relatives including Met, Ret, the insulin receptor, IGF1R and FGFR [40,52-55]. Phosphorylation of the equivalent tyrosine in Met and murine Ron is closely associated with receptor activity [36,39,40]. In the crystal structure of the Ron kinase domain [48], the nonproductive orientation of the αC helix is held by two hydrogen bonds (Additional file 4: Figure S4). One is between the side chain hydroxyl of Y¹¹⁹⁸ in the αE helix and the backbone amide of N¹¹³⁹ at the C-terminus of the αC helix, and the other one is between R¹²³¹ and Q¹¹²⁴ at the N-terminus. In order to further understand the events leading to receptor activation, we modeled the activated Ron kinase domain based on the crystal structure of active Met (3Q6U) [56] (Additional file 3: Figure S3B). In this model, the hydrogen bond between Y¹¹⁹⁸ and N¹¹³⁹ is disrupted, potentially due to the phosphorylation of Y¹¹⁹⁸. Consequently, the regional stability of the inactive kinase domain collapses, and the αC helix reorients significantly compared to its position in the auto-inhibited model. This large scale movement could promote the displacement of the JM-C constraint, resulting in reorientation of the P-loop and activation loop, and realignment of the active site residues for recruitment of ATP and substrates. Therefore, when Y¹¹⁹⁸ is mutated to phenylalanine, the absence of phosphorylation at this residue would be predicted to fix the αC helix in the kinked non-productive orientation, stabilizing the kinase in the inactive state. However, when the inhibitory JM-C region is released by alterations in the juxtamembrane domain that disrupt the autoinhibition confered by the juxtamembrane domain, phosphorylation of Y¹¹⁹⁸ would not be required for reorientation of the kinase segments and receptor activation.

Conclusions

Because the mode of regulation by the juxtamembrane domain varies significantly among RTKs, targeting this tier of regulation could provide an opportunity to develop inhibitors that would exploit the unique aspects of juxtamembrane domain regulation conferred upon individual receptors. In that context, a synthesized peptide representing the juxtamembrane domain of Kit is able to impede receptor activation in vitro and delay growth of cells containing an oncogenic form of Kit [37]. Alternatively, targeting the autoinhibited form of the receptor with small molecule inhibitors could provide an additional level of specificity due to the association between the juxtamembrane domain and the active site. The work presented here has unveiled a novel regulatory mechanism governing Ron activity by its juxtamembrane domain, and provided a new perspective regarding the regulation of RTK activity through conformational plasticity and functional cooperation of different protein segments, which may provide important insights that could lead to the development of a novel class of therapeutics for the treatment against a wide range of Ron mediated carcinomas.

Methods

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

XW performed all the experiments and wrote the manuscript, NY performed all of the molecular modeling, PAH designed the experiments and helped write the manuscript. All authors read and approved the final manuscript.

Supplementary Material

References