MAT2B-GIT1 interplay activates MEK1/ERK 1 and 2 to induce growth in human liver and colon cancer†
Article first published online: 14 MAY 2013
Copyright © 2013 American Association for the Study of Liver Diseases
Volume 57, Issue 6, pages 2299–2313, June 2013
How to Cite
Peng, H., Dara, L., Li, T. W.H., Zheng, Y., Yang, H., Tomasi, M. L., Tomasi, I., Giordano, P., Mato, J. M. and Lu, S. C. (2013), MAT2B-GIT1 interplay activates MEK1/ERK 1 and 2 to induce growth in human liver and colon cancer. Hepatology, 57: 2299–2313. doi: 10.1002/hep.26258
Potential conflict of interest: Nothing to report.
- Issue published online: 12 JUN 2013
- Article first published online: 14 MAY 2013
- Accepted manuscript online: 16 JAN 2013 10:56AM EST
- Manuscript Accepted: 19 DEC 2012
- Manuscript Revised: 7 DEC 2012
- Manuscript Received: 6 AUG 2012
- The National Institutes of Health. Grant Numbers: R01DK51719, R01AT004896
- J.M.M.. Grant Number: F32AA020150
- Plan Nacional of I+D. Grant Number: SAF 2011-29851
- Departamento de Educación del Gobierno Vasco (to J.M.M.)
- RKO cells were provided by the Cell Culture Core of the University of Southern California (USC) Research Center for Liver Diseases. Grant Number: P30DK48522
- Immunohistochemical staining and confocal microscopy were done at the Imaging Core of the USC Research Center for Liver Diseases. Grant Number: P30DK48522
Methionine adenosyltransferase 2B (MAT2B) encodes for two variant proteins (V1 and V2) that promote cell growth. Using in-solution proteomics, GIT1 (G Protein Coupled Receptor Kinase Interacting ArfGAP 1) was identified as a potential interacting partner of MAT2B. Here, we examined the functional significance of this interplay. Coimmunoprecipitation experiments examined protein interactions. Tissue expression levels of proteins were examined using immunohistochemistry and western blotting. Expression levels of proteins were varied using transient knockdown or overexpression to observe the effect of alterations in each protein on the entire complex. Direct interaction among individual proteins was further verified using in vitro translated and recombinant proteins. We found both MAT2B variants interact with GIT1. Overexpression of V1, V2, or GIT1 activated mitogen-activated protein kinase kinase 1 (MEK1) and extracellular signal-regulated kinase (ERK), raised cyclin D1 protein level, and increased growth, whereas the opposite occurred when V1, V2, or GIT1 was knocked down. MAT2B and GIT1 require each other to activate MEK1/ERK and increase growth. MAT2B directly interacts with MEK1, GIT1, and ERK2. Expression level of V1, V2, or GIT1 directly influenced recruitment of GIT1 or MAT2B and ERK2 to MEK1, respectively. In pull-down assays, MAT2B directly promoted binding of GIT1 and ERK2 to MEK1. MAT2B and GIT1 interact and are overexpressed in most human liver and colon cancer specimens. Increased expression of V1, V2, or GIT1 promoted growth in an orthotopic liver cancer model, whereas increased expression of either V1 or V2 with GIT1 further enhanced growth and lung metastasis. Conclusion: MAT2B and GIT1 form a scaffold, which recruits and activates MEK and ERK to promote growth and tumorigenesis. This novel MAT2B/GIT1 complex may provide a potential therapeutic gateway in human liver and colon cancer. (HEPATOLOGY 2012)
Methionine adenosyltransferase (MAT) is the sole enzyme responsible for the biosynthesis of the principal methyl donor, S-adenosylmethionine (SAMe), that is present in all mammalian cells.1 Two genes (MAT1A and MAT2A) encode for the catalytic subunit (α1 and α2) of the different MAT isoforms, whereas a third gene (MAT2B) encodes for a regulatory subunit (β) that modulates the kinetic properties of MAT2A-encoded isoenzyme.2 MAT1A is mostly expressed in healthy liver and is often silenced in hepatocellular carcinoma (HCC), whereas MAT2A is expressed in all extrahepatic tissues and induced during rapid liver growth and dedifferentiation.3-6 The β subunit regulates MAT2A-encoded isoenzyme by reducing its Km and Ki for methionine and SAMe, respectively.2 MAT2B is overexpressed in human HCC and cirrhotic livers and has been shown to confer a growth advantage.7
We previously reported that MAT2B encodes two dominant splicing variants (variant 1 [V1] and variant 2 [V2]).8 V1 is the aforementioned β subunit that differs from V2 at the N-terminal end. Although the two variants are differentially expressed, both are markedly induced in human HCC.8 V1 overexpression resulted in the activation of ERK1/2 (extracellular signal-regulated kinase), although the mechanism was unclear.8 To understand the mechanisms through which MAT2B regulates growth and the proteins the two variants interact with, we used in-solution proteomics to find binding partners for V1 and V2.9 We reported that both variants are highly expressed in the nucleus, where they interact with HuR, a messenger RNA (mRNA)-binding protein known to stabilize the mRNA of cyclins.9 In addition to HuR, we found that GIT1 (G Protein Coupled Receptor Kinase Interacting ArfGAP 1), a protein known to activate ERK1/2 by its interaction with mitogen-activated protein kinase kinase 1 (MEK1), is another potential interacting partner.9 GIT1 serves as a scaffold protein that facilitates c-SRC-dependent MAPK (mitogen-activated protein kinase) activation in response to stimulation of tyrosine kinase receptors and G-protein-coupled receptors.10
In the current article, we examine the interaction of MAT2B-splicing variants 1 and 2 with GIT1, ERK1/2, and MEK1/2. Using distinct cell lines, such as HepG2 cells that express both V1 and V2, Huh7 cells that express very little MAT2B, and RKO cells, a colon cancer cell line that express high levels of GIT1, we demonstrate the novel finding that MAT2B serves as an essential scaffold, along with GIT1, to facilitate the recruitment and activation of MEK and ERK in promoting cell growth. We also found both MAT2B and GIT1 are often overexpressed in HCC and colon cancer and that they promote tumorigenesis and metastasis in an orthotopic HCC model.
Materials and Methods
Eight HCC specimens (five paired with adjacent nontumorous tissues) were obtained from the Norris Cancer Center Tissue Repository (Keck School of Medicine, Los Angeles, CA). Sixteen colon cancers and adjacent nontumorous tissues were obtained by Prof. Giordano (Whipps Cross University Hospital, London, UK) during surgical resection for primary colon cancer. These tissues were immediately frozen in liquid nitrogen for subsequent RNA and protein extraction. For immunohistochemistry (IHC; see below), tissue array of 16 HCCs and 48 colon cancers and their adjacent nontumorous tissues were obtained from US Biomax (Rockville, MD).
Written informed consent was obtained from each patient. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki, as reflected in a previous approval by Keck School of Medicine University of Southern California's Human Research Review Committee.
HepG2 and Huh7 HCC and RKO colon carcinoma cell lines were obtained from the Cell Culture core facility at the University of Southern California Research Center for Liver Diseases. HepG2 and Huh7 cells were maintained in Dulbecco's modified Eagle's medium and RKO in modified Eagle's medium, each with 10% fetal bovine serum.
Transfection, Quantitative Polymerase Chain Reaction, and Western Blotting.
Human GIT1 as well as MAT2B V1 and V2 expression plasmids are described in the Supporting Methods. Short interfering RNAs (siRNAs) against GIT1, ERK1, and ERK2 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and siRNAs against V1 and V2 were described previously.8, 9 Gene overexpression and knockdown experiments and western blotting are described in the Supporting Methods. Total RNA isolation, complementary DNA synthesis, and quantitative polymerase chain reaction (qPCR) were done as previously described.11 Primers and TaqMan probes used are described in the Supporting Methods.
Immunoprecipitation (IP) studies were carried out as described previously, with minor modifications.12 Briefly, cells were lysed in a modified radioimmunoprecipitation assay buffer (described in Supporting Methods) with protease and phosphatase inhibitors. A total of 500 ug of whole cell lysate was used to coimmunoprecipitate with antibodies (Abs) against GIT1, MAT2B, V5, or MEK1/2. Western blottings were carried out to detect MAT2B, GIT1, V5, MEK1/2, ERK1/2, and MEK1. Clean Blot IP Detector Reagent (Thermo Scientific, Rockford, IL) was used to reduce background. Normal immunoglobulin G (IgG; Santa Cruz Biotechnology) was used as a control.
In Vitro Translation and Protein Binding Assay.
Full-length human MEK1 from pEGFP-N1-MEK1 (Addgene plasmid #14746; Addgene, Cambridge, MA) was subcloned into pCDNA3.1-V5-His vector (Life Technologies Corporation, Carlsbad, CA) to generate pCDNA3.1-V5-His-MEK1. In vitro transcription/translation of full-length human GIT1 or MEK1 proteins was performed using the TNT Quick Coupled Transcription/Translation System (Promega, Madison, WI), as per the manufacturer's protocol. In vitro pull-down assay was performed as described previously,12 with minor modifications, as described in the Supporting Methods.
Proliferation assay was measured using the bromodeoxyuridine (BrdU) Cell Proliferation Assay Kit (CalBiochem, San Diego, CA). Treatments are described in the Supporting Methods.
Immunofluorescence and Confocal Microscopy.
Immunolabeling of MAT2B and GIT1 proteins in the same cell was carried out according to a standard protocol (http://www.jacksonimmuno.com/technical/techmain.asp), with minor modifications (detailed in Supporting Methods). Images were visualized and captured by an Eclipse TE300 confocal microscope (Nikon Instruments Inc., Melville, NY).
IHC of GIT1 and MAT2B in HCC and Colon Cancer Specimens.
HCC, colon cancer, and matched adjacent tissues (catalog nos.: LVC 482 and BC05118a) were obtained from US Biomax. Slides were deparaffinized, hydrated, and stained for MAT2B or GIT1 using extended antigen retrieval (antigen unmasking solution; Vector Laboratories, Burlingame, CA). MAT2B and GIT1 Abs were diluted to 1:200 and 1:100, respectively. IHC staining of MAT2B and GIT1 were performed with the Vector ABC kit (Vector Laboratories), according to the manufacturer's method. Percent cells staining positive and intensity of staining were separately scored and totaled for each specimen as previously described, with minor modifications.13 Proportion score was as follows: 0, 0%; 1, <10%; 2, 10%-33%; 3, 34%-66%; 4, 67%-95%; and 5, >95%. Intensity score was as follows: 0, negative; 1, weak; 2, intermediate; and 3, strong.
Orthotopic Liver Cancer Model.
Effect of MAT2BV1, V2, and GIT1 overexpression on in vivo tumorigenicity was evaluated using Huh7 cells stably transfected with MAT2BV1 or V2 and treatment with injection of lentiviral vector expressing GIT1 (this was necessary to examine combined expression of V1/V2 with GIT1). Huh7 stable cell line with empty vector (vec) and MAT2BV1 or V2 were established as we have described.14 Huh7 cells stably expressing MAT2BV1, V2, or vec (1 × 106 cells/50 μL) were slowly injected into the left hepatic lobe of 6-week-old male BALB/c nude mice. Packaging was done using the pPACKH1 lentiviral vector Packaging kit (SBI System Biosciences, Mountain View, CA). Viral harvesting was done as described in the Open Biosystems protocol. A total of 1 × 105 Huh7 cells were infected at a multiplicity of 20 plaque-forming units per cell for 24 hours. Transducing units (2 × 109; final volume: 0.05 mL) were injected into the spleen (at time of injecting Huh7 cells into the left hepatic lobe) or tail vein (at 2 weeks after) of mice. Each mouse received a left hepatic lobe injection with Huh7 cells overexpressing V1/V2 or vec and injection with lentiviral vector expressing GIT1 or vec (spleen at time 0 followed by tail vein at 2 weeks). Mice were divided into six groups: group 1: vec+vec; group 2: V1 + vec; group 3: V2+vec; group 4: vec+GIT1; group 5: V1+GIT1; and group 6: V2+GIT1. Tumor size in liver tissues was measured as described above at day 21, and tumor volume was calculated as previously described.14 Lung and liver tissues were harvested for DNA, RNA, protein assays, or standard pathological studies and IHC as previously described.14 All procedure protocols, use, and care of the animals were reviewed and approved by the institutional animal care and use committee at the University of Southern California (Los Angeles, CA).
Densitometric values were measured using Quantity One Software (Bio-Rad, Hercules, CA). Data are given as mean ± standard error (SE). Statistical analysis was performed using analysis of variance, followed by Fisher's test for multiple comparisons. All P values were derived from at least three independent experiments. Statistical significance was defined by P < 0.05.
GIT1 and MAT2B Interact Endogenously.
In-solution proteomics identified GIT1 as a potential interacting protein with MAT2B in RKO cells.9 To confirm this primary interaction, a series of IP experiments were carried out in HepG2 cells. An Ab to GIT1 was able to immunoprecipitate MAT2B in HepG2 extracts, and, conversely, MAT2B Ab was able to immunoprecipitate GIT1 (Fig. 1A). Double immunofluorescence (IF) labeling of GIT1 and MAT2B also demonstrated that the two proteins colocalized to the cytoplasm, as well as the nuclei, of HepG2 cells (Fig. 1B). To determine whether the two proteins bind directly to one another, an in vitro pull-down assay was performed using human recombinant MAT2B protein (rMAT2B) immobilized to agarose beads and in vitro translated (IVT), [S35]-labeled GIT1 protein. Direct interaction between MAT2B and GIT1 was verified by both autoradiography and immunoblotting with anti-GIT1 Ab (Fig. 1C). The human MAT2B gene has two main transcripts that vary at exon 1. MAT2B V1 is a 334-amino-acid protein, whereas MAT2B V2 is shorter by 11 amino acids. To see whether both variants can interact with GIT1, we overexpressed these variants in Huh7 cells, which express very low endogenous MAT2B (Supporting Fig. 1). Using a V5 epitope tag in the overexpression vector, both MAT2B variants coprecipitated with GIT1 (Fig. 1D). Having found the association between GIT1 and MAT2B, we next set out to explore how this interaction affects ERK.
MAT2B and GIT1 Regulate Cell Growth and Increase ERK Activity.
Both MAT2B and GIT1 have been implicated to regulate ERK activation.8, 10 Individual knockdown of V1, V2, or GIT1 with siRNA in HepG2 cells decreased cell proliferation (Fig. 2A). Conversely, overexpression of V1, V2, or GIT1 alone resulted in increased cell proliferation (Fig. 2B). Similar results were found for ERK activity. When V1, V2, or GIT1 were knocked down in HepG2 cells, ERK activity decreased, as demonstrated by reduced phosphorylation of ERK1/2, resulting in a lower phosphorylated ERK1/2 (pERK1/2)/total ERK1/2 ratio (Fig. 2C). When V1, V2. or GIT1 were overexpressed in the same cell line, pERK1/2 levels and pERK1/2/total ERK1/2 ratio increased (Fig. 2D). MAT2B V1, V2, or GIT1 knockdown in HepG2 cells lowered cyclin D1 levels (Fig. 2C), whereas overexpression of these proteins led to an increase in cyclin D1 (Fig. 2D).
GIT1 and MAT2B (V1 and V2) Require One Another to Regulate Growth and Activate ERK.
To investigate whether GIT1 and MAT2B require each other to increase cell proliferation and ERK activity, we examined the effect of overexpressing GIT1 while knocking down either MAT2B V1 or V2 on proliferation and ERK phosphorylation. When GIT1 was overexpressed in HepG2 cells, BrdU incorporation increased, but this effect was abolished with knockdown of V1 or V2. The observed decrease in BrdU was similar to knocking down V1 or V2 alone (Fig. 3A). Similarly, GIT1 overexpression failed to activate ERK activity or raise cyclin D1 expression in the presence of knockdown of either V1 or V2 (Fig. 3B). To avoid nonspecific effects of knocking down V1 or V2, we examined GIT1 overexpression on growth and ERK activation in Huh7 cells, which express minimal MAT2B (Supporting Fig. 1). When GIT1 was overexpressed in these cells, no significant effect was observed with BrdU incorporation, ERK activity, or cyclin D1 expression, indicating that the effect of GIT1 on proliferation and ERK activity requires the presence of MAT2B (Fig. 3C,D). To investigate whether the reverse was true and whether MAT2B also required GIT1 to exert this effect, we used the RKO colon cancer cell line because of its abundant endogenous expression of GIT1 (Supporting Fig. 1). Knocking down GIT1 decreased RKO cell proliferation, whereas overexpression of V1 or V2 promoted cell growth and increased ERK phosphorylation. However, V1 or V2 overexpression failed to promote cell proliferation or ERK activation when GIT1 was knocked down (Fig. 4A,B).
To examine whether ERK activation is solely responsible for growth induction when V1, V2, or GIT1 was overexpressed, HepG2 cells were treated with ERK1/2 siRNA along with the overexpression vectors. In the presence of ERK1/2 knockdown, the growth-inductive effect of V1, V2, or GIT1 overexpression was significantly attenuated, but not eliminated (Fig. 4C,D).
GIT1, MAT2B, MEK1, and ERK Form a Complex.
We next investigated whether MAT2B or GIT1 interact with MEK (an upstream effector of ERK) and/or ERK endogenously through a series of IP experiments. In both HepG2 and RKO cells, MEK1/2, ERK1/2, and GIT1 all coprecipitated with MAT2B (Fig. 5A). When MEK1 was immunoprecipitated in RKO cells (i.e., expressed much higher endogenous GIT1 and MAT2B expression; Supporting Fig. 1), MAT2B, ERK1/2, and GIT1 were coprecipitated, whereas only MAT2B and GIT1 coprecipitated with MEK1 in HepG2 cells. When GIT1 was immunoprecipitated, MAT2B and MEK1 coprecipitated in RKO cells, whereas only MAT2B coprecipitated with GIT1 in HepG2 cells (Fig. 5A). To further investigate the direct interaction among the four proteins, in vitro pull-down assays were performed using rMAT2B and IVT GIT1, MEK1, or rERK2 proteins. MAT2B was able to bind to immobilized GIT1, MEK1, or ERK2, indicating that MAT2B directly binds to each individual protein in the complex (Fig. 5B). Immobilized MEK1 was able to pull down ERK2, but not GIT1, whereas immobilized GIT1 was able to pull down ERK2, but not MEK1 (Fig. 5C). This indicates that although GIT1 and MEK1/2 individually interacted with ERK2 and MAT2B, they do not directly bind to one another.
MAT2B and GIT1 Regulate MEK Activity.
Because MAT2B, MEK1/2, and GIT1 form a complex, we investigated whether MAT2B and GIT1 had any effect on MEK activity. Overexpression of V1, V2, or GIT1 in HepG2 cells had no effect on total MEK, but significantly increased phosphorylation of MEK1/2 (Fig. 5D). Knockdown of V1, V2, or GIT1 reduced phosphorylated MEK (pMEK)1/2 (Fig. 5E). This supports the notion that the MAT2B/GIT1 platform regulates MEK activity.
MAT2B Recruits ERK to the Complex.
In our current model, MAT2B, GIT1, MEK, and ERK link to form a complex (Fig. 6D). To further dissect how MAT2B affects MEK or ERK recruitment to the complex, IP experiments using MEK1/2 Ab were performed in HepG2 cells with either overexpression or knockdown of V1, V2, or GIT1. First, IP with anti-MEK1/2 Ab in HepG2 cells overexpressing V1, V2, or GIT1 resulted in increased ERK1/2 protein coprecipitation, suggesting that more ERK was recruited into the complex (Fig. 6A). Similar findings were observed when GIT1 was immunoprecipitated instead of MEK1/2 (Supporting Fig. 2). The reverse was confirmed as well, because knocking down V1, V2, or GIT1 resulted in decreased ERK1/2 binding to precipitated MEK1/2 (Fig. 6B). To confirm that up-regulation of MAT2B or GIT1 promotes recruitment of ERK to MEK1/2, we conducted an in vitro pull-down assay using recombinant and IVT proteins. Immobilized MEK1 was pulled down with ivtGIT1, along with recombinant ERK (rERK) (Fig. 6C). When rMAT2B protein was added to the mix, more GIT1 was recruited and more rERK2 protein was pulled down (by 2.25-fold) (Fig. 6C). This confirms that the addition of MAT2B recruits more ERK to the complex. In addition, under endogenous conditions in HepG2 cells and in the in vitro pull-down assay, we observed more GIT1 protein pulled down and recruited in the presence of MAT2B (Fig. 6A,C). Because we already demonstrated that overexpression or knockdown of MAT2B V1, V2, or GIT1 did not alter MEK1/2 content and alterations in V1 and V2 expression did not change GIT1 level and vice versa, we concluded that the increased presence of each protein in the precipitate was a result of increased recruitment to the complex.
GIT1 and MAT2B Interact and Are Overexpressed in HCC and Colon Cancer.
To asses the in vivo significance of our findings in the context of human cancer, we examined the tissue expression of GIT1 and MAT2B in 16 HCC and 48 colon cancer specimens, as compared to adjacent nontumorous tissues, and found that both proteins are overexpressed in most cancer specimens (Fig. 7A and Supporting Fig. 3). We also confirmed higher expression levels and interaction between MAT2B and GIT1 in HCC and colon cancer specimens (Fig. 7B-D).
Effect of MAT2BV1, V2, or GIT1 Overexpression On In Vivo Tumorigenesis.
Finally, the effect of MAT2B variants and GIT1 expression on in vivo tumorigenesis was examined using an orthotopic liver cancer model. Three weeks after injecting Huh7 cells stably expressing V1 or V2 resulted in much larger tumor volumes, as compared to vec control (Fig. 8A). Similarly, treatment with lentiviral GIT1 expression vector resulted in increased tumor volumes (Fig. 8A). Importantly, combining V1 or V2 with GIT1 had an additive effect on tumor volumes and enhanced the metastatic potential of Huh7 cells. Overexpression of MAT2B or GIT1 appears to have doubled the protein level of each other (Fig. 8B and Supporting Fig. 4). Consistent with in vitro results, overexpression of these proteins resulted in higher MEK and ERK activity (Fig. 8B).
MAT2B, which regulates MAT2A-encoded isoenzyme, is overexpressed in human cirrhosis and HCC.7 Our previous work showed that MAT2B positively regulates ERK activity, but the mechanism was unknown.8 Because GIT1 was identified as a potential binding partner of MAT2B using in-solution proteomics,9 we set out to explore this relationship further. GIT1, a scaffold protein, has been shown to activate the ERK pathway in response to mitogens, such as angiotensin II or epidermal growth factor, in vascular smooth muscle and kidney 293 cells.10 To our knowledge no data exist regarding the expression and/or role of GIT1 in any cancer. MEK and ERK are central to cell growth by increasing the synthesis of pyrimidine nucleotides and activity of transcription factors, which enhance gene expression as well as increase the activity of cyclin D1.15-19 Because both MAT2B and GIT1 have been associated with ERK signaling, we postulated that the interaction between MAT2B and GIT1 was of importance in the subsequent activation of this pathway.8, 10 Here, for the first time, we describe a complex containing GIT1 and MAT2B that serves as a scaffold to activate and recruit MEK and ERK and promote cell proliferation through this MAPK pathway. We also show, for the first time, that in both human HCC and colon cancer, MAT2B and GIT1 interact and are often overexpressed.
We began with confirming the interaction between GIT1 and MAT2B. Using co-IP in HepG2 cells, as well as direct binding assays with IVT proteins, we found that both MAT2B V1 and V2 bind to GIT1 under basal conditions (Fig. 1). Using overexpression and knockdown, we subsequently observed that this interaction promotes proliferation, ERK activation, and increases cyclin D1 levels (Fig. 2). It is important to note that knockdown in V2 expression will not alter total MAT2B levels in HepG2 cells because 75% of the protein is expressed as V1 in this cell line8 (Fig. 2C). Despite the inconsequential decrease in total MAT2B protein level with V2 knockdown, the effect on cell proliferation and decrease in ERK activity indicate the functional importance of V2. Overexpression of V1, V2, and GIT1 increased cyclin D1 expression, which is well known to reflect the cell's proliferation activity, and is a downstream target of ERK activation.15, 18 Although ERK activation is dominant, it is not the only mechanism responsible for growth induction because overexpression of V1, V2, or GIT1 still increased growth, albeit to a much lesser extent, in the presence of ERK1/2 knockdown (Fig. 4C,D). This finding is consistent with our previous report that V1 and V2 overexpression resulted in increased cytoplasmic HuR content and mRNA levels of several HuR targets, such as cyclin A and cyclin D1.9 Whether GIT1 is also involved in that process remains to be examined. Taken together, these findings support the notion that the MAT2B/GIT1 complex applies its proliferative influence and downstream effects on cell growth, in part, by ERK activation.
Next, we showed that this proliferation was the result of the interaction between GIT1 and MAT2B in complex, because knocking out V1, V2, or GIT1 abrogated this increase (Figs. 3 and 4). We went on to confirm these findings by overexpressing GIT1 in Huh7 cells, which do not express much MAT2B, and, as expected, this overexpression failed to activate ERK and promote proliferation alone (Fig. 3). The same was true in the colon cancer cell line, RKO, where V1 or V2 overexpression was unable to promote cell growth and ERK activation if GIT1 was knocked down (Fig. 4). Becase GIT1 has been shown to function as a scaffold to activate MEK1 in a yeast two-hybrid screen,10 we hypothesized that the MAT2B/GIT1 complex forms a structure necessary for MEK and subsequent ERK activation. To investigate this possible interaction, we next conducted a series of IP experiments to see whether MEK associates with, or is activated by, the MAT2B/GIT1 platform. In both liver cancer (HepG2) and colon cancer (RKO) cell lines, MEK1/2, ERK1/2, and GIT1 all coprecipitated with MAT2B (Fig. 5). To our surprise, MEK1/2 and GIT1 did not directly interact in our in vitro pull-down assay using recombinant and IVT proteins. However, all the proteins in either experiment, HepG2 cells, RKO cells, as well as in in vitro conditions, were able to coprecipitate with MAT2B. Considering the lack of direct interaction between GIT1 and MEK1, MAT2B likely serves as a central adaptor that links MEK1/2 and GIT1 together.
MEK directs the signals of growth factor or G-protein-coupled receptors to their intracellular targets, which regulate cellular processes, including proliferation, differentiation, cell morphology, and oncogenesis. Activation of MEK1/2 in mitogen-stimulated cells is directly mediated by MEK kinases, such as Raf-1 kinase, which phosphorylates two serine residues (S218 and S222) in the activation loop of MEK. MEK, in turn, then activates ERK1/2.19 Because we had observed that the MAT2B/GIT1 complex activated ERK as well as increased cyclin D1 and cell proliferation, we hypothesized that MEK phosphorylation and activation is dependent on the formation of the MAT2B/GIT1 complex. In accord with this theory, overexpression of V1, V2, or GIT1 in HepG2 cells significantly increased the phosphorylation of MEK1/2 and the opposite is true with V1, V2, or GIT1 knockdown (Fig. 6). Activation of MEK is the seminal signaling event through which the complex activates ERK and subsequently cyclin D1 and results in cell proliferation.
Overexpression of either MAT2B variants or GIT1 resulted in increased ERK recruitment to the complex without altering MEK expression. These results suggest that MAT2B V1 and V2 are the central components of a complex with the scaffold protein, GIT1, and that, together, they form a platform for MEK to recruit and activate ERK (Fig. 6). Both variants of MAT2B are necessary for the complex because knocking down either one resulted in decreased ERK activation and reduced growth. Why these two variants, which only differ at the N-terminus, cannot compensate for one another remains unclear. One possible explanation for these observations may be that V1 and V2 are both necessary and must oligomerize to form a platform with GIT1. Further research into the three-dimensional crystal structure of this multiprotein complex is necessary to explore this hypothesis. Regardless of the reason, both variants are clearly indispensable for ERK activation and cell proliferation. Interestingly, although both V1 and V2 are necessary, they are not sufficient for promotion of cell growth, because MAT2B does not induce proliferation in the absence of GIT1. To our knowledge, this is the first time that the expression of GIT1 alone has been shown to promote proliferation in HepG2 and RKO liver and colon cancer cells, respectively.
In the current body of work, we demonstrate that the MAT2B/GIT1 complex is an integral mediator of cell proliferation through recruitment and activation of the MEK/ERK MAPK pathway, which is known to be dysregulated in many cancers.20 We determined the significance of these interactions in human cancer specimens for in vivo correlation. GIT1 and MAT2B were overexpressed in most of the 16 HCC and 48 colon cancer specimens, as compared to adjacent nontumorous tissues, as assessed by IHC (Fig. 7 and Supporting Fig. 3). MAT2B-GIT1 interaction and increased expression in cancerous tissues were also confirmed using IP and western blotting. Expression of MAT2B in HCC may be underestimated because many of the nontumorous livers were cirrhotic, a condition where MAT2B is also induced.7 This is the first report of GIT1 overexpression in human cancer as well as the first documentation of MAT2B overexpression in human colon cancer specimens. We also confirmed the importance of MAT2B and GIT1 in vivo using an orthotopic HCC model. Overexpression of these proteins resulted in increased MEK and ERK activation, growth and metastasis, effects that can be, at least partly, attributed to increased ERK activity.21 Interestingly, overexpression of MAT2B or GIT1 appears to result in a higher protein level of each other, suggesting the possibility that they stabilize each other to further promote MEK/ERK activation. These in vivo results seem to be at odds with the data showing lack of growth induction when Huh7 cells were transfected with GIT1 overexpression vector (Fig. 3C). A plausible explanation lies in the difference between transient transfection (48 hours) in vitro and long term (3 weeks) of in vivo overexpression. Indeed, the ability of GIT1 and MAT2B to raise each other's protein level was only noted in the in vivo experiment, suggesting that the mechanism(s) involved required some time. The additive effect of overexpressing MAT2B and GIT1 together on growth and tumorigenesis may be the result of this feed-forward mechanism that further raised their expression.
In summary, we have provided evidence that MAT2B and GIT1, overexpressed in most HCC and colon cancers, form a scaffold that is essential to MEK/ERK activation, and that this novel MAT2B/GIT1 complex may provide a potential therapeutic gateway to inactivate ERK and halt cell proliferation. Whether they are also overexpressed in other cancers and play a similar role remain to be examined. The effect of increased MAT2B and GIT1 expression on MEK/ERK activation is summarized in Fig. 6D.
Additional Supporting Information may be found in the online version of this article.
|HEP_26258_sm_SuppFig1.tif||179K||Supporting Information Figure 1.|
|HEP_26258_sm_SuppFig2.tif||253K||Supporting Information Figure 2.|
|HEP_26258_sm_SuppFig3.tif||11864K||Supporting Information Figure 3.|
|HEP_26258_sm_SuppFig4.tif||4639K||Supporting Information Figure 4.|
|HEP_26258_sm_SuppFigLegends.doc||35K||Supporting Information Figure Legends|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.