Somatic mutation of GRIN2A in malignant melanoma results in loss of tumor suppressor activity via aberrant NMDAR complex formation.

The ionotropic glutamate receptors (N-methyl-D-aspartate receptors (NMDARs)) are composed of large complexes of multi-protein subunits creating ion channels in the cell plasma membranes that allow for influx or efflux of mono- or divalent cations (e.g., Ca(2+)) important for synaptic transmissions, cellular migration, and survival. Recently, we discovered the high prevalence of somatic mutations within one of the ionotropic glutamate receptors, GRIN2A, in malignant melanoma. Functional characterization of a subset of GRIN2A mutants demonstrated a loss of NMDAR complex formation between GRIN1 and GRIN2A, increased anchorage-independent growth in soft agar, and increased migration. Somatic mutation of GRIN2A results in a dominant negative effect inhibiting the tumor-suppressive phenotype of wild-type (WT) GRIN2A in melanoma. Depletion of endogenous GRIN2A in melanoma cells expressing WT GRIN2A resulted in increased proliferation compared with control. In contrast, short-hairpin RNA depletion of GRIN2A in mutant cell lines slightly reduced proliferation. Our data show that somatic mutation of GRIN2A results in increased survival, and we demonstrate the functional importance of GRIN2A mutations in melanoma and the significance that ionotropic glutamate receptor signaling has in malignant melanoma.

Introduction

Glutamate receptors are involved in cell homeostasis, cell growth, neurotransmission, cell survival and programmed cell death (Kaderlik ; Komuro and Rakic, 1993; Nacher and McEwen, 2006; Schlett, 2006; Wang ). Glutamate receptors are composed of two different major types; the ionotropic and the metabotropic family (Stepulak ). Ionotropic glutamate receptors (e.g., NMDA receptors) are composed of large complexes of multi-protein subunits creating ion channels in the cell plasma membranes allowing influx or efflux of mono- or divalent cations (e.g., Ca2+) (Monyer ). Upon binding of glutamate and glycine, these ligand-gated ion channels change their conformation, resulting in ion permeability and intracellular oscillations that in many neuronal cells (cerebellar granule cells, neurons, astrocytes, and glial cells) is important for synaptic transmissions, cellular migration and survival (Komuro and Rakic, 1993; Lau and Tymianski, 2010; Traynelis ).

The NMDA receptor plays a major role in many neuronal processes and disease etiologies such as learning, memory and neurodegeneration (i.e., Schizophrenia) (Lang ). The NMDARs are heterotetrameric complexes consisting of two NR1 (GRIN1) subunits and two NR2 (GRIN2 (A-D)), or a mixture or GRIN2 and GRIN3 subunits predominately expressed in neuronal cells. Upon binding of its cognate ligands, NMDAR permits Ca2+ influx resulting in increased intracellular calcium levels leading to activation of calcium-dependent signal transduction (Traynelis ). However, hyperactivation of NMDAR by glutamate leads to calcium-dependent cell death (excitotoxicity), a phenomenon observed in neurodegenerative diseases and brain tumors (Takano ; Ye and Sontheimer, 1999). Recent studies have demonstrated the importance of functional NMDA receptors in neurological systems. GRIN2A and GRIN2B were found to be mutated in Schizophrenia patients and patients with epilepsy causing decreased channel permeation (Endele ) (Carvill ; Lemke ; Lesca ). The implications that NMDA receptor functions as a conveyor of cellular signals important for cognitive learning and apoptosis have been well established in neuronal diseases and brain tumors (Hahn ; Lau and Tymianski, 2010; Rzeski ). Interestingly, glioma cells (either in culture or in patients) secrete excess glutamate to the extracellular milieu (composed of many cell types including glial and neuronal) (Ye and Sontheimer, 1999) causing overstimulation of glutamate receptors on neurons leading to increased Ca2+-mediated apoptosis, allowing glioma cells room to grow and spread (Takano ).

Recently, our lab demonstrated through whole-exome sequencing a high prevalence of somatic mutations in GRIN2A in malignant melanoma (Wei ). Whole-exome sequencing of our matched samples revealed the unexpected discovery that GRIN2A was somatically mutated in ~25% of the melanoma cases. The mutations were distributed throughout the gene, with clustering of mutations at amino acids within important functional domains (e.g. the ligand binding domain (LBD)). We also observed three recurrent alterations at S278F, E371K, and E1175K as well as 5 nonsense mutations. Recently, another group recently published a whole-exome screen of 8 melanoma samples and found 2 additional somatic mutations in GRIN2A, suggesting that genetic alteration of this gene is important (Stark ). Furthermore, seven somatic mutations found in our screen have been discovered by others as well as some listed in the COSMIC database (D252N, S278F, W343X, G449E, M653I, R920K, S929F, and E1073K) (Berger ; Hodis ) (http://cancer.sanger.ac.uk/cancergenome/projects/cosmic/). Taken together with the recent publication of mutations found in Schizophrenic patients and our data we decided to analyze how these mutations affected GRIN2As function giving us a better understanding of how NMDA receptors work in melanomagenesis.

Results

To determine how somatic mutations in the NMDA receptor subunit GRIN2A effect the functioning of NMDARs (e.g., complex formation or cation influx), we cloned specific mutations based on location within important functional domains or if they truncated the protein product (see schematic in Supp. Fig. 1). We used wild type (WT) GRIN2A to insert mutations and first examined complex formation between GRIN1 and GRIN2A using a transient expression assay. HEK293T cells were transiently co-transfected with WT GRIN1 with GRIN2A (WT, E371K, W372X, E373K, G889E, Q891X, R920K, E1172K, or W1271X) or empty vector control and further tested for complex formation via co-immunoprecipitation using anti-GRIN1 (Fig. 1a). As can be seen, WT GRIN1 precipitated WT GRIN2A and to a lesser extent GRIN2A (W1271X). However, the rest of the mutations in GRIN2A had very little to no association with GRIN1 in vitro. To rule out a lack of protein expression in these cells we probed cell lysates from the same experiment with anti-GRIN2A and found equal expression (Fig. 1a-bottom panel). We probed for GAPDH as a loading control. Furthermore, we tested complex formation now focusing on protein-protein binding between GRIN1 and GRIN2A precipitating with GRIN2A and probing for GRIN1. HEK293T cells were transiently co-transfected with GRIN1 and GRIN2A (WT, W372X, Q891X, R920K, W1271X), or empty vector control followed by co-immunoprecipitation using anti-GRIN2A (Fig 1b). We observed that WT GRIN2A and to lesser extent GRIN2A (R920K) precipitated GRIN1. However, GRIN2A (W372X, Q891X, or W1271X) mutants showed a lack of binding to GRIN1 corroborating our previous results. The differences observed between binding of GRIN1 to GRIN2A (W1271X) may be a result of the use of antibodies directed towards GRIN1 or GRIN2A and their efficiencies or epitope-binding domains. These results suggest that GRIN2A somatic mutations resulting in truncated proteins reduce proper complex formation with GRIN1. As GRIN2A ultimately forms a cation channel, loss of such a complex formation would result in decreased influx and ultimately attenuation of the cation-dependent signaling mechanism.

Figure 1

Somatic mutations in GRIN2A have adverse effects on receptor function and formation

Mutant forms of GRIN2A binding GRIN1 with reduced affinity thus causing decreased NMDAR complex formation. a. HEK293T cells were transiently transfected with WT GRIN1 and GRIN2A (WT or mutants) or empty vectors as control and immunoprecipitated with anti-GRIN1. Immunoprecipitates were probed with anti-GRIN2A and anti-GRIN1 to confirm binding. Lysates were probed with anti-GRIN2A, and anti-GAPDH as a loading control. b. HEK293T cells were transiently transfected with WT GRIN1 and GRIN2A (WT or mutants) or empty vectors as control and immunoprecipitated with anti-GRIN2A. Immunoprecipitates were probed with anti-GRIN1 and anti-GRIN2A to confirm binding. Lysates were probed with anti-GRIN1, and anti-GAPDH as a loading control. c. Influx of calcium upon NMDA stimulation of transient transfected HEK293T cells shows decreased calcium permeability in cells expressing mutant forms of GRIN2A. Cells were pretreated with Fluo-3/AM probe prior to imaging on Ziess confocal microscope. Images were taken every 5 secs over a period of 15mins. NMDA (200μM) was added to the dishes at the 100 second mark. Graphs are fold increase in fluorescent signal vs. time (sec). Values are representative of 8–10 individual cells from the image view. d. HEK293T cells were transiently co-transfected with WT GRIN1 and WT GRIN2A and increasing amounts of empty vector or W372X and immunoprecipitated with anti-GRIN1. Immunoprecipitates were probed with anti-GRIN1 and anti-GRIN2A to confirm binding. Lysates were probed with anti-GRIN1, and anti-GAPDH as a loading control (samples are on one gel but adjusted to be next to each other in figure). e. HEK293T cells were transiently co-transfected with WT GRIN1 and GRIN2A (WT or mutants) or empty vectors as control and stimulated with 200μM NMDA in the presence of Ringer’s solution for 60 minutes. Lysates were probed with anti-P-p38 MAPK, anti-p38 MAPK, and anti-GRIN2A to confirm expression. Lysates were probed with anti-Tubulin as a loading control. Figures were adjusted to remove unwanted overall background due to western blotting and film development.

NMDA receptors require proper complex formation between GRIN1 and one of the many different GRIN2/3 subunits to allow for calcium influx into cells. We utilized a co-expression assay in HEK293T cells to test if calcium influx into cells is affected by these mutations. Transiently transfected HEK293T cells were analyzed for influx of calcium upon stimulation with NMDA, a specific ligand for the NMDA receptor. Cells expressing WT GRIN1 and WT GRIN2A, upon stimulation with 200μM of NMDA, show proper NMDA receptor function with an increase in calcium permeability in the cells as a measure of fold change over time (Fig. 1c-top left panel). However, expression of GRIN1 with GRIN2A (Q891X or R920K) results in nonfunctional NMDA receptors (Fig. 1c), confirming the previous results showing reduced binding between GRIN1 and these mutants. Interestingly, expression of GRIN1 with GRIN2A (W1271X) shows diminished calcium influx in transiently transfected cells (Fig. 1c). Our transient results demonstrate that proper complex formation is required between GRIN1 and GRIN2A for the calcium channel to function appropriately. We hypothesize that functional NMDA receptor signaling in melanoma cells containing GRIN2A truncation mutations is lost resulting in diminished activation of Ca2+-dependent pathways, potentially leading to increased survival through suppression of cell death (excitotoxicity).

To determine the nature of the loss of functioning of these mutations we examined the ability of these mutations to inhibit complex formation via a dominant negative effect. We co-expressed WT GRIN1 and WT GRIN2A in the absence or presence of mutated GRIN2A in a transient system. Figure 1d shows that increased expression of mutant GRIN2A results in loss of complex formation between WT GRIN1 and WT GRIN2A, compared to increased amounts of empty vector transfected into cells. These results suggest that somatic mutations of GRIN2A cause a loss of functioning potentially through a dominant negative effect inhibiting normal NMDAR signaling downstream to known effector molecules (e.g., p38MAPK) (Doronzo )

Hyperactivation of NMDA receptors via glutamate or NMDA stimulation leads to prolonged influx of Ca2+ ions and hyperactivation of pro-apoptotic signaling mechanisms (e.g., p38 MAPK activation) leading to reduced cell proliferation, migration, invasion and even cell death in certain cellular contexts (e.g., excitotoxicity) (Wang ). To determine how somatic mutation of GRIN2A effects its ability to activate Ca2+-mediated signaling to pro-apoptotic pathways we co-expressed GRIN2A (WT or mutants) or empty vector with GRIN1 (WT) in the presence of Ringer’s solution and NMDA. Cells expressing WT GRIN1 and WT GRIN2A had a significant increase in p38 MAPK activation when stimulated with NMDA (Fig 1e). However, co-expression of WT GRIN1 with any of the GRIN2A mutant constructs or empty vector resulted in little to no p38MAPK activation in the presence of NMDA (Fig 1e). These results further corroborate our hypothesis that GRIN2A may function as a tumor suppressor in melanoma cells whereby certain somatic mutations cause attenuation of NMDAR-dependent Ca2+-mediated pro-apoptotic signaling.

To better understand how somatic mutations of GRIN2A in melanoma affect its ability to potentiate tumorigenesis we created stable pooled clones expressing a subset of the mutations initially tested. Two independent cell lines, 31T and SK-Mel-2, were used to co-express GRIN1 and GRIN2A (WT, W372X, Q891X, R920K, or W1271X) or empty vector control via sequential transduction and dual selection. We tested for expression and the ability to form proper NMDAR complexes in both cell lines (31T or SK-Mel-2) by immunoprecipitating using either anti-GRIN2A or anti-GRIN1 and analyzing via western blot analysis (Fig. 2a–b, respectively). GRIN2A expression was similar for GRIN2A W372X, Q891X, and W1271X in both cell lines, however there were some differences of expression for WT GRIN2A and GRIN2A (R920K) between 31T and SK-Mel-2 cells. GRIN1 was relatively equally expressed in both stable cell lines tested (Fig. 2a–b). We observed normal complex formation between WT GRIN1 and WT GRIN2A and to lesser extent GRIN2A (R920K) in both stable cell lines (Fig 2a–b). Conversely, expression of GRIN2A (W372X, Q891X, or W1271X) resulted in loss of complex formation to WT GRIN1, again observed in both stable cell lines. Furthermore, we tested expression of WT or mutant forms of GRIN2A on proliferation. Interestingly, little to no effect on cellular proliferation in either the 2359 (31T) or SK-Mel-2 stable cell lines was observed (Supp. Fig. 2).

Figure 2

Melanoma cells expressing mutant forms of GRIN2A have increased ability for anchorage-independent growth

Stable melanoma cell lines 31T and SK-Mel-2 expressing WT GRIN1 and GRIN2A (WT or mutants) or empty vector control and immunoprecipitated with anti-GRIN1 and anti-GRIN2A. Immunoprecipitates were probed with anti-GRIN2A and anti-GRIN1 to examine NMDAR complexes in 31T (a) or SK-Mel-2 (b). Melanoma cell line 31T (c) stably expressing GRIN1 and GRIN2A mutants or empty vector were seeded into Boyden chambers in 0.5% serum and grown for 48hrs before staining and counting. Melanoma cell line 31T (d) stably expressing GRIN1 and GRIN2A mutants or empty vector were seeded into Boyden chambers plus and minus 10μM TCN-213 for 48hrs before staining and counting. e and f are quantitative graphs of colony formation of 31T or SK-Mel-2 cells in soft agar, respectively. Error bars are representative of n=3 (s.d.) (* p<0.01 students t-test). Figures were adjusted to remove unwanted overall background due to western blotting and film development.

To determine how expression of GRIN1 and GRIN2A (WT or mutant) in melanoma affects melanoma biology we used cell migration and anchor-independent growth assays. Stable pooled clones (31T or SK-Mel-2) expressing WT GRIN1 and GRIN2A (WT or mutants) or empty vector were seeded into Boyden chamber assays and grown for 24–48hrs prior to analysis of a migratory phenotype. Expression of WT GRIN1 and WT GRIN2A in 31T caused reduced migration compared to empty vector (Fig. 2c). However, expression of mutant GRIN2A (W372X, Q891X, R920K, and W1271X) showed increased migration in 31T and SK-Mel-2 compared to both WT GRIN2A and empty vector. Migration of 2359 cells expressing WT GRIN2A was reversed in the presence of a specific NMDAR2A (GRIN2A) antagonist, TCN-213. As seen in Fig 2d, stable clones expressing WT GRIN2A migrated significantly better in the presence of 10μM TCN-213 compared to in the absence of antagonists. However, vector or mutant expressing clones were unaffected by TCN-213. SK-Mel-2 cells were tested for the ability to migrate using multiple time points and cell numbers with no migration phenotype observed at any point. Furthermore, we analyzed the tumorigenic effect of the mutants in melanoma using an anchorage-independent growth assay. Expression of WT GRIN1 and WT GRIN2A in 31T and SK-Mel-2 caused diminished colony formation compared to empty vector (Fig. 2e–f). Expression of GRIN2A mutants showed increased soft agar colony formation, again in both 31T and SK-Mel-2 stable clones. These results demonstrate that expression of somatically mutated forms of GRIN2A in melanoma cells leads to an increased migration and colony formation through the loss of tumor suppressive activity via a NMDAR-dependent Ca2+-influx.

To further delineate the role GRIN2A plays in melanoma cell biology, we stably depleted endogenous GRIN2A using short-hairpin RNA and further functionally analyzed the effects caused by knock-down. Using melanoma cell lines that express WT GRIN2A (31T and 39T) or mutant GRIN2A (E1175K) (501Mel and 125T) we stably transduced GRIN2A specific shRNAs (shRNA#1, #2, and #3) into each cell line and a vector control (pLKO.1). As seen in figure 3, GRIN2A is knocked-down at the message level by 20–70% as determined by qRT-PCR analysis in all cell lines tested. We examined growth potential of cells depleted of GRIN2A in WT or mutant backgrounds. Strikingly, shRNA knock-down of WT GRIN2A resulted in increased proliferation of cells stably transduced with shRNAs #1–3 compared to cells transduced with pLKO.1 alone (Fig. 3c–d). In contrast, depletion of mutant GRIN2A caused a slight reduction in proliferation compared to vector control cells (Fig. 3e–f), however, the level of knockdown in the two mutant lines is notably less than the WT lines, possibly indicating a higher level of cell dependency on this mutated protein. These results support our hypothesis that WT GRIN2A functions as a tumor suppressor in melanoma cells given the increase in growth potential in cells devoid of only 50–60-% of total GRIN2A mRNA. Although further analyses of the signaling mechanisms downstream of glutamate receptors are needed, our study provides an insight into the biological relevance of GRIN2A somatic mutations in melanoma and reveals a possible node of clinical intervention for the treatment of malignant melanoma.

Figure 3

GRIN2A functions as a tumor suppressor in melanoma cells

Stable depletion of endogenous GRIN2A in melanoma cells. Melanoma cells stably depleted of endogenous GRIN2A were confirmed via qRT-PCR of all clones. (a–b) Analysis of knockdown used specific GRIN2A primers and efficiency was compared to GAPDH levels in each sample thus generating a fold change. Proliferation assay of wild-type 31T (c) or 39T (d) and mutant 501Mel (e) or 125T (f) expressing GRIN2A cell lines depleted of GRIN2A. Knock-down resulted in little to no change in proliferation for 125T or 501Mel melanoma cell lines expressing mutant GRIN2A but increased proliferation for both 31T and 39T cell lines expressing WT GRIN2A. Each cell line used pLKO.1 as an empty vector control.

Discussion

Here we show for the first time in malignant melanoma the functional analysis of the ionotropic glutamate receptor subunit, GRIN2A. Based on previous whole-exome sequencing results our lab revealed GRIN2A to be highly mutated (Wei ) in melanoma demonstrating the importance that glutamate signaling plays in tumorigenesis. Recent work demonstrated GRIN2A promoter methylation present in human colorectal carcinoma (Kim ). These researchers further demonstrated that expression of WT GRIN2A with WT GRIN1 resulted in suppressed cell growth in soft agar. In our present study, we showed that somatic mutation of GRIN2A resulted in decreased complex formation of the NMDAR leading to increased migration and colony formation in soft agar, whereas expression of WT GRIN2A acted as a tumor suppressor in melanoma. Our results suggest that expression of mutant forms of GRIN2A may play a protective role in melanoma, potentially acting as a dominant negative suppressing WT GRIN2A normal activity.

NMDAR signaling plays opposing roles depending on the partners that are found within the larger complex (Choo ; Xiao ). Neurological cells utilize different signaling mechanisms that function in a protective manner and when over stimulated via glutamate (excitotoxic) result in programmed cell death. NMDA receptor signaling uses divalent cation influx (e.g., Ca2+) via glutamate/glycine binding to mediate signaling to downstream elements involved in survival and/or cell death pathways (Choo ; Ru ; Xiao ). Recently it was demonstrated that hippocampal cells, found in the central nervous system, utilize differential signaling to either the p38 MAPK pathway or the Erk1/2 pathway depending on which GRIN2 subunit is present in the NMDAR complex with GRIN1 (Xiao ). Immature hippocampi preferentially express GRIN2B and upon stimulation of the NMDAR activate the Erk1/2. Interestingly, these authors show that mature hippocampi from rats, mainly express GRIN2A compared to GRIN2B and upon NMDAR stimulation activate p38 MAPK signaling causing a more pronounced neuronal cell death. The fact that we observed so many somatic mutations causing truncated or non-functional versions of GRIN2A may be one way that melanoma cells use to potentiate a more proliferative and pro-survival phenotype; via subunit switching to GRIN2B or a dominant negative effect.

The functional domains found in GRIN2A play major roles in Mg2+-gated calcium channel influx of NMDA receptors. Reports by Endele, Et Al., Maki et al., and others show that mutations found in both the GRIN2A or GRIN2B result in reduced Mg2+ gating of the NMDAR causing non-functional calcium signaling (Endele ; Maki ). Recent studies identified that patients with epilepsy or encephalopoathies harbor germline mutations in NMDAR2A (GRIN2A) (Carvill ; Lemke ; Lesca ). Interestingly, the spectrum of the identified mutations was very similar to the ones we observed with multiple non-synonymous mutations resulting in missense and nonsense alterations. Mutations observed in these patients also resulted in decreased or aberrant calcium (ion) influx. Another study by the Kutsche group examined patients with neurological disorders ranging from mental retardation to epilepsy. They discovered that mutations found in GRIN2A (W218X or N615K) resulted in truncated or non-functional subunits of the NMDA receptor leading ultimately to reduced calcium channel permeability. GRIN2B was also shown to carry several mutations that disrupted its functionality, resulting in reduced NMDA receptor signaling. Maki Et Al., showed that alanine screening for potentially important residues in the C-terminal domain of GRIN2A revealed two important sites of phosphorylation, Ser900 and Ser929 (Maki ). Mutation of GRIN2A at either site resulted in reduced calcium channel conductance potentially causing reduced signaling. Interestingly, we found one of the same sites to be somatically mutated (Ser929Phe) in our malignant melanoma samples (Wei ) (Supp. Fig. 1). These results not only corroborate our findings that somatic mutations in GRIN2A result in loss of function in calcium channel signaling but reveal the importance of proper NMDAR signaling is to brain (neuronal) and melanocyte development and survival. Future functional analyses are required to further examine the role GRIN2A may have in the etiology of the disease. This would allow a better understanding of the already established importance that NMDA receptor signaling plays in melanoma.

Materials and methods

Construction of wild-type and mutant GRIN2A expression vectors

Mouse GRIN1 (BC039157) and mouse GRIN2A (BC148800) were cloned by PCR as previously described (Palavalli ) using clones (MMM4769-202766680 (GRIN1) or OMM5895-202524480 (GRIN2A)) purchased from Open Biosystems with primers listed in Supplementary Table 1. The PCR products were cloned into the mammalian expression vectors pCDF-MCS2-EF1-Puro™ or pCDH1-MCS-CMV-Neo™ (Systems Biosciences, Inc., Mountain View, CA) via the XbaI and NotI restriction sites. Point mutations were introduced as previously described (Prickett ) using the primers found in Supplementary Table 1.

Cell culture and transient expression

HEK293T cells were purchased from ATCC (Manassas, VA) and maintained in complete RPMI-1640 medium supplemented with 10% Fetal Bovine Serum (FBS). HEK293T cells were transfected with Arrest-IN reagent (Open Biosystems) at a 6:1 ratio with DNA (μl:μg) using 2–6 μg of plasmid DNA. Ringer’s solution was obtained from the Veterinary Pharmacy at the NIH. All experiments were conducted in accordance with institutional guidelines.

Immunoprecipitation and Western Blotting

Transfected cells were gently washed 2X in PBS and then lysed using 1.0 ml 1% NP-40 lysis buffer (1% NP-40, 50mM Tris-HCl pH 7.5, 150mM NaCl, Complete Protease Inhibitor tablet, EDTA-free (Roche, Indianapolis, IN), 1μM sodium orthovanadate, 1 mM sodium fluoride, and 0.1% β-mercaptoethanol) per T-75 flask for 20 minutes on ice. Lysed cells were scraped and transferred into a 1.5 mL microcentrifuge tube. Extracts were centrifuged for 10 minutes at 14,000 rpm at 4°C. 800 μl of supernatant was immunoprecipitated overnight using 20 μl of anti-GRIN1 and 30μl of 50% slurry of Protein A/G beads (Roche). The immunoprecipitates were washed and subjected to SDS-PAGE and western blotting as previously described (Palavalli ). Primary antibodies used in our signal transduction pathway analysis were anti-GRIN1 (sc162902), anti-GRIN2A (sc1468), or anti-GRIN2A (sc-136004) (Santa Cruz Biotechnology), anti-P-p38 MAPK (#9211), anti-p38 (#9212) (Cell Signaling), and anti-GAPDH (#CB1001) (Calbiochem-EMD Biosciences).

Pooled stable expression

To make lentiviruses, GRIN1 or GRIN2A constructs were co-transfected into HEK 293T cells seeded at 1.5×106 per T75 flask with pVSV-G and pFIV-34N (kind gifts from Todd Waldman, Georgetown University) helper plasmids for pCDF1 based or pPACKH1 viral production mixture from SBI for pCDH1 based using Arrest-IN as described by the manufacturer. Virus-containing media was harvested 60hr after transfection, filtered, aliquoted and stored at −80°C. 31T cells (kind gift from Dr. Rosenberg) were grown in RPMI-1640 (Lonza, Walkersville, MD) and supplemented with 10% fetal bovine serum (HyClone, Logan, UT). A375 cells were purchased from National Cancer Institute, Division of Cancer Treatment, Developmental Therapeutics Program, Frederick, MD and maintained in RPMI-1640 and supplemented with 10% FBS. 31T or SK-Mel-2 cells were seeded at 1.5 × 106 cells per T75 flask 24 hr prior to infection. Lentivirus for GRIN1 and GRIN2A (wild-type or mutants) and empty vector control were used to sequentially infect 31T or SK-Mel-2 cells as previously described (Prickett ). Stable expression of GRIN1 proteins (wild-type) was determined by immunoprecipitation and SDS-PAGE analysis followed by immunoblotting with anti-GRIN1 and anti-GAPDH to show equivalent expression among pools. Stable expression of GRIN2A (wild-type or mutants) was determined by RT-PCR analysis of mRNA from 31T or SK-Mel-2 stable pooled clones using GRIN2A specific primers and GAPDH primers as a loading control.

Proliferation assays

To examine growth potential, pooled 31T or SK-Mel-2 pooled clones were seeded into 96 well plates at 300 cells per well in either 1%, 2.5% or 10% serum-containing medium and incubated for 13–17 days. Samples were analyzed every 48 hr by lysing cells in 50 μl 0.2% SDS/well and incubating for 2 hour at 37°C prior to addition of 150 μl/well of SYBR Green I solution (1:750 SYBR Green I (Invitrogen-Molecular Probes-Carlsbad, CA) diluted in dH20). Plates were analyzed using a BMG Labtech FLOUstar Optima.

Migration assays

31T or SK-Mel-2 pooled clones were seeded into pre-conditioned migration wells (8.0 μm – BD Biocoat, BD Biosciences) at 30,000–100,000 cells per well in serum-free medium in the top chamber and incubated for 24–48 hrs with complete serum containing medium in the bottom chamber prior to harvesting. Antagonist studies used 10μM of TCN-213 (Tocris) dissolved in DMSO in the top and bottom chambers of the Boyden chamber. DMSO was used as a negative control for this assay using < 0.02% (vol/vol). Inserts were fixed and stained using Hema 3 Stat Pack as per manufacturer’s protocol. Inserts were analyzed and counted for cells migrated per field view and quantitated using ImageJ (NIH software).

Soft agar assay

31T or SK-Mel-2 pooled clones were plated in triplicate at 1000 cells/well and in top plugs consisting of sterile 0.33% Bacto-Agar (BD, Sparks, MD) and 10% fetal bovine serum (HyClone, Logan, UT) in a 24-well plate. The lower plug contained sterile 0.5% Bacto-Agar and 10% fetal bovine serum. After two weeks, the colonies were photographed and quantitated using ImageJ (NIH software).

Lentiviral shRNA

Constructs for stable depletion of GRIN2A (cat#RHS4533-EG2903) were obtained from Open Biosystems (Huntsville, AL) and were confirmed to efficiently knockdown GRIN2A at the message level. Lentiviral stocks were prepared as previously described (Prickett ). Melanoma cell lines (31T, 39T, 501Mel, or 125T) were infected with shRNA lentiviruses for each condition (vector and two-three different GRIN2A specific shRNAs). Selection of stable pooled clones was done in the presence of 3μg/ml puromycin containing normal medium for 3–5 days prior to determining knock-down efficiency. Stably infected pooled clones were tested in functional assays.

Reverse Transcription PCR

Total RNA was extracted from pooled clones of melanoma cells stably knocked down for endogenous GRIN2A following the manufacturer’s protocol for the RNeasy Mini Kit (QIAGEN #74101). Total RNA was eluted in 30 μL diethylpyrocarbonate (DEPC)-treated distilled H2O. A total of 1 μg of total RNA was used for single-strand complementary DNA (cDNA) synthesis using a SuperScript III First Strand kit (Invitrogen #18080-051). cDNA was amplified using the olido dT20 primer supplied in the kit. To test for loss of GRIN2A message, we used 1 μL of cDNA in the PCR reaction with either GRIN2A primers or GAPDH primers (Supp. Table 2).

Proliferation assays of stable knockdown cells

To examine growth potential, stable pooled knockdown clones for GRIN2A were seeded into 96 well plates at 500 cells per well in either 1%, 2.5% or 10% serum-containing medium and incubated for 6–8 days. Samples were analyzed every 2–3 days by lysing cells in 50 μl 0.2% SDS/well and incubating for 2 hour at 37°C prior to addition of 150 μl/well of SYBR Green I solution (1:750 SYBR Green I (Invitrogen-Molecular Probes-Carlsbad, CA) diluted in dH20). Plates were analyzed using a BMG Labtech FLOUstar Optima.

Calcium assays

HEK293T cells seeded in 35 mm glass-bottomed culture dishes (MatTek Corporation) were transiently transfected with GRIN1 and either GRIN2A WT or mutants (Q891X, R920K, W1271X). Thirty six hours post transfection cells were washed with Tyrode solution (5mM HEPES, 136 mM NaCl, 2.7 mM KCl, 1 mM MgCl2, 1.9 mM CaCl2, 5.6 mM glucose, buffered to pH 7.4 with Tris base). Cells were loaded with 10 μM Fluo-3AM diluted in Tyrode supplemented with 0.1% BSA for 45 min in the dark, followed by 2 washes with Tyrode-BSA and 2 washes with Tyrode alone. Loaded cells were then left for 15 min to ensure complete hydrolysis of the acetoxymethyl ester (AM) groups. Images were taken every 5 sec over a period of 15 min and collected on a Zeiss LSM-700 laser scanning microscope with a 63x oil immersion lens in a single track mode using excitation 488 nm and emission BP 505–530 filter sets.