Abstract
The expression profiles of the erythropoietin producing hepatocellular carcinoma (Eph) receptor family of tyrosine kinases have been previously shown to provide molecular signatures of normal breast cells, breast tumor cells and invasive breast carcinoma cells. In particular, the expression of EphB6 receptor is lost in invasive breast carcinoma cell line MDA-MB-231. The comparative proteomic profiles of native and EphB6-expressing MDA-MB-231 cells using difference gel electrophoresis (DIGE) and liquid chromatography–mass spectrometry of selected proteins are presented in this study. The expression of more than 70 proteins was significantly altered in EphB6-transfected MDA-MB-231 cells. These altered proteins are involved in glycolysis, cell cycle regulation, tumor suppression, cell proliferation, mitochondrial metabolism, mRNA splicing, DNA replication and repair. Although the majority of these proteins have been implicated in tumorigenesis, the impairment of energy homeostasis and altered regulation of signaling pathways appear to be noteworthy targets of EphB6. Based on the identities of altered proteins and the pathways regulated by these proteins, this study suggests that the interactions of EphB6 with a wide variety of proteins lead to altered proteomic profile of EphB6-transfected MDA-MB-231 cells.
- Mass spectrometry
- difference gel electrophoresis
- DIGE
- proteomic profile
- EPHB6
- breast cancer invasiveness
- metastasis
The erythropoietin producing hepatocellular carcinoma (Eph) receptors are the largest family of receptor tyrosine kinases (1). These receptors are activated by their cognate ephrin ligands. Eph receptors and ephrin ligands are classified into A and B classes based on their sequence similarities. There are 14 Eph receptors and 9 ephrin ligands known in mammalian cells. These receptors have been well characterized for their involvement in brain development, axon guidance and angiogenesis (2). The role of these receptors and ligands is becoming increasingly important in cancer initiation and progression. While Eph receptors and ephrin ligands are known to be up-regulated in some cancers (3-5), EphB6 is lost in more aggressive/invasive breast cancers, melanomas and neuroblastomas (4, 6, 7). It has been previously shown that down-regulation of EphB6 is related to epigenetic silencing of this gene by promoter methylation (8), as confirmed by detection of the methylated promoter using methylation sensitive PCR (8, 9). It is interesting to note that, unlike other members of the Eph receptor family, EphA10 and EphB6 are kinase-deficient. To address the biological role of EphB6, its interacting partners were characterized by using a yeast two-hybrid system. The results of that study indicated the interaction of its cytoplasmic domain with intracellular proteins such as aldolaseA, dynactin, clusterin and plekstrin homology domain-containing family B member 1 protein (10). Based on the cellular functions of these interacting partners, it was suggested that EphB6 directly or indirectly influences a variety of cellular pathways. These results have led to the speculation that proteomic profiles of mammalian cells with or without EphB6 will be significantly altered in the presence of EphB6 in mammalian cells.
Based on the unique and interesting EphB6 interactors, its knase-deficient status and involvement in cancer invasiveness, the downstream effects of EphB6 signaling are likely to illuminate the importance of proteins that are possibly influenced by EphB6 signaling. Although EphB6 interactors are largely indicated by the effect they have on other signaling proteins, the study in (10) was the first to show their direct interaction. A recent elegant study also demonstrated direct interactions of Eph receptors with other proteins (11).
The present study has focused on determining the downstream proteomic consequences of EphB6 signaling in MDA-MB-231 breast carcinoma cell line. These cells are highly invasive as evidenced by in vitro assay and in vivo animal experimentation. The invasiveness of MDA-MB-231 cells is attributed to the lack of EphB6 protein (12). Thus, the proteomic profiles of native and EphB6-transfected cells will help confirm the direct as well as the indirect involvement of EphB6 in cellular pathways.
A variety of approaches have been used to investigate the global effects of a specific protein or a specific pathway. These methods include gene profiling, micro RNA profiling, evaluation of cellular phenotype in vitro and in vivo effects of specific proteins during animal development. While the latter strategies related to in vivo phenotypic effects require several months to be performed, microarray profiling of mRNA and miRNAs are relatively straightforward. However, the profiling results do not always correspond to proteins. Such methodology-specific shortcomings can be overcome by determining the proteomic profiles of cells that are phenotypically different or made different by the presence or absence of a specific protein. Difference gel electrophoresis (DIGE) is an informative methodology for large-scale proteomic investigations (13), and it was applied in this investigation for qualitative and quantitative proteomic differences between native and EphB6-transfected MDA-MB-231 cells.
Materials and Methods
Cell culture. MDA-MB-231 cells were obtained from the American Type Culture Collection (Manassas, VA, USA) and cultured at 37°C in the presence of 7% CO2 in Dulbecco's Modified Eagle's Medium (DMEM) (Gibco Invitrogen, Carlsbad, CA, USA) containing 10% fetal bovine serum (Thermo Scientific HyClone, Logan, UT, USA), 2.0 mM L-glutamine (Gibco), 25 units/ml penicillin (Gibco) and 25 μg/ml streptomycin (Gibco).
Transfection of MDA-MB-231 with EphB6 construct. MDA-MB-231 cells were stably transfected with a full-length EphB6 cDNA cloned in pCDNA3.1+, as described previously (12).
Briefly, MDA-MB-231 cells were transfected with EphB6 cDNA construct using Lipofectamine 2000 (Invitrogen) according to the manufacturer's recommendations. The transfected cells were grown in media containing G418 (Invitrogen) at a concentration of 500 μg/ml and the surviving clones were analyzed for the levels of EphB6 transcript and protein. As a control, MDA-MB-231 cells were transfected with pCDNA3.1+ (empty vector) and G418-resistant clones were isolated.
Total RNA isolation and PCR amplification of EphB6. Approximately 2×106 native MDA-MB-231 cells or EphB6-transfected cells were lysed with Trizol reagent, and RNA was isolated as described previously (12). An aliquot of RNA was reverse transcribed and amplified by PCR using EphB6 primers (12).
Western blot analysis. The cells were grown to the logarithmic phase in appropriate media, the medium was removed and the plates were washed with ice-cold PBS. The cells were lysed by incubating with an aliquot (500 μl) of cell lysis buffer (Cell Signaling Technology, Danvers, MA, USA) containing 1 mM phenylmethylsulphonyl fluoride (Sigma, St Louis, MO, USA) for 5 minutes on ice, and the cell lysate was collected. The lysate was sonicated, the homogenate centrifuged at 16,000×g in a microfuge at 4°C for 10 minutes and the supernatant was collected.
Approximately 50 μg of protein (in a volume of 20 μl), as determined by the Bradford reagent (Sigma), were combined with 10 μl of 3× loading buffer (Cell Signaling Technology) supplemented with 125 mM dithiothreitol. The samples were loaded onto a 10% SDS-polyacrylamide gel and electrophoresed. The proteins were transferred from the gel to a PVDF membrane (Amersham Biosciences/GE Healthcare, Piscataway, NJ, USA) and processed for the detection of EphB6 or β-actin.
Two-dimensional DIGE. The conventional two-dimensional (2D) gel electrophoresis restricted the analysis of one sample at a time and thus made comparative analysis of two samples relatively labor-intensive. However, the use of fluorescent dyes and the availability of gel scanners have made simultaneous analysis of two samples in the same gel convenient by employing the DIGE strategy (13). This approach was used to distinguish between the protein complements of MDA-MB-231 cells in the presence and absence of EphB6 protein.
To isolate proteins, the culture medium containing MDA-MB-231 cells was centrifuged and the cell pellet was collected. The pellet was mixed with 0.5 ml ice-cold protein lysis solution (BioRad, Hercules, CA, USA) in the presence of a protease inhibitor cocktail (Roche, Nutley, NJ, USA), vortexed, passed repeatedly through a 20 gauge syringe needle and placed on ice for 30 min. The particulate material was removed by centrifugation at 40°C for 10 min, and the protein concentration in the supernatant was determined by using 2D Quant kit (Amersham Biosciences). The samples were cleaned using 2D Clean up kit (BioRad) before loading them on isoelectric focusing (IEF) gel strips.
The protein lysates were labeled by incubating on ice for 30 min in the dark with either Cy3 or Cy5 at a concentration of 400 pmol dye/50 μg of protein. The labeling reaction was terminated by incubating with 10 mM lysine for 10 min. The labeled protein was mixed with 150 μg of unlabeled protein and used for isoelectric focusing. The isoelectric pH gradient (IPG) strips (pH 3-10, GE Healthcare, Piscataway, NJ, USA) were rehydrated in IPGphor IEF system (GE Healthcare) by incubating them with protein samples in a buffer containing 8 M urea, 4% CHAPS, bromophenol blue, 1% Phamlyte and 2 mg/ml DTT. Isoelectric focusing was performed in IPGphor II (GE Healthcare) for a period of 65.5 kVh. The strips were equilibrated for 15 min with a solution of 50 mM Tris-Cl buffer, 6 M urea, 30% glycerol, 2% SDS, bromophenol blue and 1% DTT (w/v). The equilibration was carried out for another 15 min in the mixture that contained 2.5% iodoacetamide instead of DTT. These equilibration steps led to reduction of proteins and subsequently carbamidomethylation of cysteine sulfhydryl groups. The IPG strip was carefully placed on top of a 12.5% SDS-polyacrylamide gel cast on low-fluorescence glass plates. The electrophoresis was carried out at 1W for 2 hours followed by 2W of constant power until the blue dye had run off the gel.
Visualization of fluorescent protein spots on gel and analysis of gel image. A Typhoon 9410 imager (GE Healthcare) was used to visualize protein spots. The excitation and emission wavelength combinations of 532 nm/580 nm and 633 nm/670 nm were used to capture the images of proteins labeled with Cy3 and Cy5, respectively. The spot detection, matching and normalization were carried out by using DeCyder software (GE Healthcare). The quantitative comparison of individual spots was performed by recording relative fluorescence of Cy3 and Cy5 at desired excitation/emission wavelengths. The spot intensities were compared and the fold change for each protein spot determined using PROGENESIS software (Nonlinear USA Inc., Durham, NC, USA). Protein spots showing significant changes were marked for spot picking and further analysis.
Spot excision mass spectrometry, analyses and database searches. The proteins in the gels were immobilized by incubation for 10 min in a mixture of 10% methanol and 7% acetic acid. The unlabeled portion of the proteins was visualized by incubating with deep purple stain. The gel was scanned in a Typhoon scanner and spots of interest were excised using an Ettan Spot Picker (GE Healthcare) and then digested with sequencing-grade modified trypsin (Promega Corporation, Madison, WI, USA). The digested peptides were identified based on their mass spectra using a Fourier Transform Linear Trap Quadrupole (FT-LTQ) mass spectrometer (Finnigan, San Jose, CA, USA). Briefly, the protein digests were loaded onto a trapping column for concentrating and removing salts and then separated on a reverse phase column. Subsequently peptides were infused and tandem mass spectroscopy (MS) spectra were acquired in the positive ion mode. Protein identification was carried out by searching against non-redundant NIH's NCBI databases (www.ncbi.nlm.nih.gov) using MASCOT (Matrix Science Inc., Boston, MA). The search criteria accounted for carbamidomethylation of cysteine, partial oxidation of methionine, and with one missed cleavage allowed. The mass tolerance limits for MS and MS/MS were set at 15 ppm and 0.8 Da, respectively for FT-LTQ. For LTQ, mass tolerance of 1.5 Da and 0.8 Da were set for MS and MS/MS, respectively.
Results
Stable expression of EphB6 in transfected MDA-MB-231 cells. The native and EphB6-transfected MDA-MB-231 cells were compared for the levels of EphB6 transcript and protein. As expected, the transcript and EphB6 protein were not detectable in the native cells, whereas abundant amounts of transcript and protein were observed in MDA-MB-231 cells stably transfected with an EphB6 expression construct. These observations indicate that the above cell line-pair is appropriate for investigating EphB6-dependent proteomic profiles of MDA-MB-231 cells.
Proteomic profile of native and transfected MDA-MB-231 cells. The protein lysates of native and EphB6-transfected MDA-MB-231 cells were analyzed by DIGE. Conventional methods of comparative proteomics often result in considerable variations between technical replicates. In contrast, DIGE allows simultaneous analysis of two samples in the same gel after normalization with an internal loading control (13). The fluorescence scanning of the gel indicated over 1000 spots labeled with Cy3 (Figure 1, green fluorescence of vector transfected MDA-MB-231 cell lysate) or Cy5 (Figure 1, red fluorescence of lysate from EphB6-transfected MDA-MB231 cells). The relative abundance of individual spots was assessed and quantification was performed as described in the Materials and Methods.
The quantification of spots revealed changes in the fluorescence of over 250 protein spots (Table I) that were up- or down-regulated in MDA-MB-231 cells transfected with EphB6. These 250 spots yielded 858 sequences with matches in the database. These protein spots were visualized with a deep purple stain and marked for extraction and sequence determination (Figure 2). Each spot marked in Figure 2 represented more than one protein. Therefore, the abundance of actual protein undergoing the change was reflected in the number of peptide fragments that matched to that protein. These proteins were classified by the degree of changes varying by at least 1.3-fold and greater (Table II). The proteins fell into four groups based on 1.3- to 1.5-fold change, 1.5- to 2.0-fold change, 2.0- to 3.0-fold change and >3.0-fold change. The majority of these alterations were 1.5- to 2.0-fold changes. Among these proteins, plastin-2 and insulin-like growth factor 2 mRNA-binding protein 2 were up-regulated by 4.3- and 3.7-fold, respectively, in MDA-MB-231 cells transfected with EphB6. In contrast, tubulin alpha-1C chain protein was down-regulated by 3.3-fold. A summary of up- and down-regulated proteins is presented in Table III.
The altered proteins were classified into categories based on their biological function (Table IV and Figure 3). These proteins were involved in processes such as cell shape and actin cytoskeleton, radiosensitivity, helicases, signal transduction, mitochondrial proteins, metabolism, transport, chaperones for protein folding, proteasome, mRNA processing, transcription factors, replication/repair, cell cycle and centrosome. The maximum numbers of genes altered in EphB6-expressing MDA-MB-231 cells belonged to signal transduction and metabolism. The miscellaneous category consisted of proteins such as galectin, transgelin, V-set and transmembrane domain containing protein, coiled-coil domain protein and some uncharacterized protein coding sequences (Table V).
Among the larger categories of signal transduction and metabolism, interestingly, all genes were altered between 1.5- and 2.0-fold. Noteworthy among the proteins affecting metabolism were enzymes of glycolysis, nucleotide synthesis, protein synthesis/turnover and estrogen synthesis. While the amounts of glyceraldehyde-3-phosphate dehydrogenase decreased significantly in EphB6-transfected cells, a notable decrease in the amounts of pyruvate kinase and phosphoglycerate kinase was also observed. Other notable proteins included plastin, cytoplasmic actin, RhoJ, nesprin 1, prefoldin subunit 5, T-complex protein 1 subunit theta, proteasome subunit beta type B, S100A3, protein assisted sugar transporter A, plasmolipin, tubulin-alpha-1 B chain and zinc finger FYVE domain containing protein-16.
Discussion
The altered transcript profiles determined by gene microarrays have frequently been implicated in cellular phenotypes. Such conclusions are based on the assumption that the abundance of transcripts relate to changes in the corresponding proteins. However, the changes in transcript levels may not always match with the amounts of proteins (Kandpal, unpublished observations). It was therefore reasoned that a direct proteomic profile may shed light on specific proteins as molecular determinants of overt phenotypes. It has been previously shown that MDA-MB-231 cells do not have any detectable expression of EphB6 (4) and forced expression of EphB6 in MDA-MB-231 cells leads to altered invasiveness and other phenotypic characteristics of these cells (12). Furthermore, EphB6 interacts with a variety of proteins such as aldolaseA, clusterin, dynactin, plekstrin homology domain-containing family B member 1 protein and transmembrane protein 25 (10). Such a diverse set of proteins interacting with EphB6 suggests that its presence may affect a wide variety of proteins. To identify the affected proteins, a comparative proteomic profile of native and EphB6-transfected MDA-MB-231 cells was performed in the present study.
The effects of EphB6 on glycolysis were predicted on the basis of EphB6 interaction with aldolase A, a glycolytic enzyme (10). It is known that energy homeostasis is altered in tumor cells (14), and if expression of a specific protein is affecting some tumorigenic processes then it is not surprising that energy balance or homeostasis in transfected cells would be altered. Such energy homeostasis was reflected in the change of enzymes such as glyceraldehyde-3-phosphate dehydrogenase (NADH production), phosphoglycerate kinase (ATP production) and pyruvate kinase (ATP production). These results indicate that EphB6 transfection is likely to decrease the availability of NADH for ATP production via oxidative phosphorylation.
An important enzyme related to estrogen hormone that was altered is estradiol 17 β-dehydrogenase. Interestingly, estradiol 17 β-dehydrogenase was down-regulated in EphB6 transfected cells. It warrants mention that MDA-MB-231 cells are estrogen receptor (ER) negative. It is tempting to speculate that EphB6 or its interactors/targets may influence non-classical receptor mediated non-genomic effects of estrogen (15). It has been previously reported that Eph receptors may affect the pathways leading to cell proliferation mediated by JNK proteins (16). Such changes are in conformity with the altered abundance of SH2 domain containing 3C protein (SH2DC3). Interestingly, SH2DC3, also known as SHEP1, has been shown to interact with EphB2 in yeast two-hybrid assay (17) and its binding to the conserved tyrosine-phosphorylated motif in the juxtamembrane region of EphB2 has been confirmed. These associations provide a confirmation of relevant proteins altered in EphB6-transfected cells.
Among other signaling proteins, the change in the amounts of proteins involved in signaling pathways mediated by calcium (calmodulin and annexin), wnt (APC2), ras (growth factor receptor bound protein-2), and rho proteins have been conclusively shown to participate in various tumor phenotypes (17-22). The changes in the levels of these proteins in EphB6-transfected MDA-MB-231 cells may partly be explained by some yet uncharacterized interactions between the intermediates of various pathways.
Although direct effects of EphB6 on apoptosis were not investigated in this study, the altered abundance of RUVB2, proteasome subunit b-type 8, serpin B9, endophilin B1 and DNA topoisomerase-2 binding protein 1 suggest that expression of EphB6 may influence apoptosis. Such apoptotic functionality is consistent with the proposed tumor suppressor action of EphB6 (23).
Transgelin is reported to be down-regulated in transformed cells, thus being an early indicator of tumorigenesis (24). When EphB6-transfected cells were compared to native MDA-MB-231 cells, the abundance of transgelin was higher in the transfected cells, suggesting a trend towards normal cell profile. Based on the interaction of transgelin with actin fibres (25) and the alterations in the abundance of plastin, actin, RhoJ, nesprin 1 and Rho GDP dissociation inhibitor, it was likely that the effects of EphB6 on actin cytoskeleton are responsible for the phenotypic changes observed in EphB6-transfected MDA-MB-231 cells. The alterations in proteins such as pericentrin, NPAT, DNA topoisomerase 2-β, DNA topoisomerase binding protein 1 and RUV2 indicated that the transfected cells are possibly affected at the level of cell cycle and cytokinesis. It is, therefore, reasonable to predict that EphB6 transfected cells may have altered characteristics of various phases of cell cycle.
An important set of proteins involved in transcription were also altered in EphB6-transfected MDA-MB-231 cells. These regulatory proteins are either transcription factors or participate in splicing reactions. It is interesting to note that transfection of EphB6 caused a change in the abundance of a variety of proteins. Such multitude of changes is expected due to the greater availability or lack of a general transcription factor(s). However, the altered levels of these transcription factors may not be directly attributed to EphB6.
In conclusion, EphB6-transfected MDA-MB-231 cells showed molecular profiles that were different than the native cell line. In particular, the metabolic enzymes involved in energy homeostasis, cytoskeleton regulatory proteins, a variety of signaling pathways and transcription factors were noteworthy. The observations on enzymes of glycolysis supported the reports that metabolic therapy to offset aberrant energy status may be a beneficial mode of breast cancer intervention. It warrants mention that the altered protein profiles reported here need to be validated with additional cell lines and tumor specimens that lack the expression of EphB6 receptor. The results of the present study suggest that EphB6 and its interactors influence a variety of pathways in the cell that are partly responsible for the observed cellular phenotypes.
Acknowledgments
The Author wishes to thank Drs Anand Swaroop and Rong-Fong Shen for their support, and Drs Jacob Nellissery, Brian Fox and Ganachari Nagaraja for technical assistance.
- Received May 26, 2010.
- Revision received July 17, 2010.
- Accepted July 16, 2010.
- Copyright© 2010 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved