Research ArticleG Protein Signaling

Direct targeting of Gαq and Gα11 oncoproteins in cancer cells

See allHide authors and affiliations

Science Signaling  19 Mar 2019:
Vol. 12, Issue 573, eaau5948
DOI: 10.1126/scisignal.aau5948

Targeting uveal melanoma

Uveal melanoma (UM) is a common cancer of the eye, and about half of patients with UM develop metastatic disease. Although most cases of UM are driven by constitutively active mutants of the G protein α subunits Gαq and Gα11, therapies that target signaling pathways downstream of these oncogenic drivers have been unsuccessful. Annala et al. provide further evidence of the ability of the cyclic depsipeptide FR900359 (FR) to inhibit oncogenic Gαq/11 signaling in UM cell lines in vitro, particularly by blocking the mitogenic ERK pathway. Furthermore, FR reduced mutant Gαq–driven UM tumor growth in a mouse xenograft model, suggesting that FR or similar compounds should be further investigated for directly targeting oncogenic Gαq/11 proteins in patients with UM.


Somatic gain-of-function mutations of GNAQ and GNA11, which encode α subunits of heterotrimeric Gαq/11 proteins, occur in about 85% of cases of uveal melanoma (UM), the most common cancer of the adult eye. Molecular therapies to directly target these oncoproteins are lacking, and current treatment options rely on radiation, surgery, or inhibition of effector molecules downstream of these G proteins. A hallmark feature of oncogenic Gαq/11 proteins is their reduced intrinsic rate of hydrolysis of guanosine triphosphate (GTP), which results in their accumulation in the GTP-bound, active state. Here, we report that the cyclic depsipeptide FR900359 (FR) directly interacted with GTPase-deficient Gαq/11 proteins and preferentially inhibited mitogenic ERK signaling rather than canonical phospholipase Cβ (PLCβ) signaling driven by these oncogenes. Thereby, FR suppressed the proliferation of melanoma cells in culture and inhibited the growth of Gαq-driven UM mouse xenografts in vivo. In contrast, FR did not affect tumor growth when xenografts carried mutated B-RafV600E as the oncogenic driver. Because FR enabled suppression of malignant traits in cancer cells that are driven by activating mutations at codon 209 in Gαq/11 proteins, we envision that similar approaches could be taken to blunt the signaling of non-Gαq/11 G proteins.


Activated forms of Gαq and Gα11, two subunits of heterotrimeric G proteins of the Gq family, are oncogenic drivers in uveal melanoma (UM), an aggressive cancer of the adult eye (13). Both of these G protein α subunits harbor single amino acid substitutions at the residue glutamine-209 (Gln209, Q209), which abrogate their intrinsic guanosine triphosphatase (GTPase) activity that normally serves to inactivate the α subunits. Because of diminished regulation by GTPase activity, the nucleotide state of Gαq/11 becomes more dependent on relative nucleotide affinity and concentration. Given that guanosine triphosphate (GTP) is in molar excess over guanosine diphosphate (GDP) in living cells (4), GTPase-deficient mutants are thought to predominantly exist in the active GTP-bound state (5).

So far, most efforts to blunt signaling of mutationally activated Gαq/11 oncoproteins have focused on inhibition of their downstream effector molecules or their folding chaperones (6). Among these is protein kinase C (PKC) and mitogen-activated protein kinases (MAPKs) of the extracellular signal–regulated kinase (ERK) family, activation of which propagates mutant Gαq/11 activity, ultimately leading to cellular proliferation and malignancy (7, 8). However, clinical trials have shown limited to no activity for inhibitors of both kinases in patients with metastatic UM (9, 10). Similarly, inhibition of YAP, another relevant UM oncoprotein downstream of Gq, with verteporfin does not qualify for therapy due to high systemic toxicity (9). Once patients develop metastases, the prognosis is poor and a fatal outcome is unpreventable; median survival is 4 to 15 months after metastasis (10, 11).

Because Gq signals through multiple effectors (12, 13), and because these likely cooperate to drive malignancy on a cellular level, the most straightforward and effective therapeutic strategy should lie in targeted inhibition of the mutated G protein α subunit itself. Direct pharmacological inhibition of Gαq/11 oncoproteins has been achieved in vitro with FR900359 (FR), a highly selective Gq/11 inhibitor that was thought to preserve Gq/11-GDP in its biologically inactive form (1416). Such a mode of inhibitor action would be expected to be effective under conditions with enhanced Gq activity induced by overexpressed or mutationally activated Gq-linked G protein–coupled receptors (GPCRs) or their cognate ligands but not under conditions in which the hydrolysis of GTP is impaired, that is, when G protein α subunits are unable to revert to the inactive GDP-bound state.

Contrary to this established paradigm of FR action (5, 14, 15), yet consistent with two studies that reported FR inhibition of mutant Gαq signaling in UM cells in vitro (17, 18), we here demonstrate that direct pharmacological targeting of oncogenic Gαq/11 proteins with FR is not only feasible but also efficient, both in vitro and in vivo. In vitro, FR suppressed malignant phenotypes and prosurvival signaling of cutaneous melanoma (CM) and UM cells harboring the common oncogenic Gαq/11Q209L/P mutants. In vivo, FR inhibited tumor growth of UM cells driven by mutationally activated GαqQ209P in a mouse xenograft model. These results may be rationalized by the nature of the bound nucleotide detected within “active” GTPase-deficient Gαq: Active GαqQ209L not only is exclusively GTP-bound but also contains GDP in its nucleotide-binding pocket. Therefore, these results suggest that Gq inhibitors such as FR [and likely also YM-254890 (YM)] (19) silence not only the signaling of wild-type (WT) Gαq/11 GTPases but also that of mutationally activated Gαq/11 oncogenes.


FR suppresses malignant phenotypes in a CM line harboring an activating Gα11Q209 mutation

The current working model envisions FR as a guanine nucleotide dissociation inhibitor (GDI) for the Gq/11 family of heterotrimeric G proteins (14, 19). This mechanism of action implies prohibition of nucleotide exchange, that is, the replacement of GDP for GTP, thereby preserving Gq/11-GDP in its biologically inactive form. In apparent contrast with this notion, we found that FR suppressed the proliferation of HCmel12 cells, a mouse melanoma cell line harboring mutationally activated Gα11 (Gα11Q209L; Fig. 1A). The FR-dependent reduction in cell number was accompanied by reduced numbers of cells with active metabolism as evidenced in cell viability assays (Fig. 1B). Inhibition of proliferation and metabolic activity may result from direct cytotoxicity or disruption of the cell cycle machinery, linked to reduced energy demand that normally ensures cell cycle progression (20). To discriminate between these possibilities, we performed flow cytometric analyses. Propidium iodide (PI) staining of genomic DNA revealed that FR arrested HCmel12 cells in the G1 phase of the cell cycle (Fig. 1C). With annexin V staining, which detects phosphatidylserine exposure on the outer leaflet of the plasma membrane, we confirmed that FR did not cause substantial apoptotic cell death (Fig. 1D; see fig. S1 for the gating strategy used in the flow cytometry analysis). Instead, we observed a concentration-dependent increase of pigmentation in pellets of FR-treated HCmel12 cells (Fig. 1E). HCmel12 cells exposed to FR also showed a more dendritic morphology than that of vehicle-treated cells and were flattened and strongly pigmented, which is consistent with a more differentiated, melanocyte-like phenotype (Fig. 1F). This correlated with an increase in the abundance of the melanosome-associated glycoprotein gp100, a marker for melanocyte differentiation, as detected by Western blotting analysis (Fig. 1G). We next investigated whether FR compromised HCmel12 cell migration, a key step occurring in cancer metastasis. We found that the basal migratory capacity of HCmel12 cells in transwell migration assays was largely abolished by low nanomolar FR concentrations (Fig. 1H). Together, these data raised the possibility that FR blunts signaling of the mutationally activated Gα11 oncogene in melanoma cells.

Fig. 1 FR suppresses hallmark features downstream of Gα11oncoproteins in CM cells.

(A) Quantification of the effect of the indicated concentrations of FR on the number of adherent HCmel12 melanoma cells cultured for 72 hours. Data are means + SEM of three experiments. (B) Viable cell metabolism was analyzed by XTT assay after treatment of HCmel12 cells with vehicle or the indicated concentration of FR for 48 hours. Data are means + SEM of three experiments. Data in (A) and (B) were normalized to those of vehicle-treated control cells. (C) Flow cytometric histograms (left) and quantitative analysis (right) of cell cycle progression in HCmel12 cells treated with vehicle or the indicated concentrations of FR using the fluorescent DNA-intercalating agent PI. (D) Flow cytometric analysis (left) and quantitative analysis (right) of annexin V– and PI-stained HCmel12 cells that were left untreated or were treated with the indicated concentrations of FR for 72 hours. Plots in (C) and (D) are representative of three biological replicates. ns, not significant. (E and F) Cell pellets (E) and bright-field microscopic images (F) of HCmel12 cells after 72 hours of treatment with vehicle or FR. Scale bars, 5 μm. Four experiments are represented by the data in (E) and (F). (G) Representative Western blotting analysis (top) and quantification (bottom) of the relative abundance of the differentiation marker gp100 in HCmel12 cells after treatment with vehicle or 10 nM FR for 72 hours. Data are means ± SEM of four experiments. (H) Basal migratory capacity of HCmel12 cells after treatment with vehicle or the indicated concentrations of FR as determined in Boyden chamber transwell assays. Graph: Data are means + SEM of three experiments. Scale bar, 200 μm. Statistical significance was determined by one-sample t test for (A), (B), and (G) or by two-way analysis of variance (ANOVA) for (D).

FR preferentially silences mitogenic pERK over canonical PLCβ signaling through a direct interaction with GαqQ209L

Because previous studies indicated that FR fails to suppress oncogenic Gq signaling in canonical phospholipase Cβ (PLCβ)–dependent signaling readouts (14, 15), we hypothesized that FR may differentially affect mitogenic versus canonical PLCβ pathways downstream of Gq/11 oncoproteins. The ability to cause biased inhibition of only a subset of effectors downstream of the same Gα protein has therapeutic potential to treat G protein–mediated diseases. We found that increasing number of HCmel12 cells produced increasing amounts of inositol monophosphates (IP1), which is indicative of intrinsic Gq/11 activity (fig. S2). Consistent with previous reports (14, 15), FR was poorly effective at abrogating PLCβ-driven increases in IP1 amounts (Fig. 2A, gray squares). In apparent contrast, nanomolar FR concentrations were sufficient to completely ablate mitogenic signaling through the MAPK ERK1/2 module (Fig. 2A, blue squares). Preferential inhibition of mitogenic signaling versus canonical PLCβ signaling did not appear to be cancer cell specific but rather a mutation-specific feature because FR blunted both signaling branches comparably well in B16 melanoma cells, which harbor WT Gq but display greater Gq activity than that of other melanoma cell lines (Fig. 2B and fig. S3). In support of this assumption, we recapitulated similarly effective suppression by FR of PLCβ and phosphorylated ERK1/2 (pERK1/2) signaling in experiments with an immortalized human embryonic kidney (HEK) 293 cell line that endogenously expresses WT Gq/11 together with the Gq-coupled muscarinic M3 receptor (Fig. 2C) or that was transfected to express the Gq-coupled free fatty acid receptor FFA2 (Fig. 2D). In no instance did FR affect the abundance of total ERK protein (fig. S4).

Fig. 2 FR preferentially silences mitogenic pERK signaling over canonical PLCβ signaling in cells with mutationally activated GαqQ209L through a direct interaction with its protein target.

(A to D) Analysis of the effects of the indicated concentrations of FR on the amounts of pERK1/2 and IP1 in HCmel12 cells (A), B16 melanoma cells (B), CCh-treated HEK293 cells expressing endogenous Gq-coupled muscarinic M3 receptors (C), and propionic acid (C3)–stimulated HEK293 cells exogenously expressing free fatty acid FFA2 receptors (D). Data are from n = 7 [pERK1/2, (A)], n = 6 [IP1, (A)], n = 4 [pERK1/2, (B)], n = 3 [IP1, (B)], n = 6 [pERK1/2, (C) and (D)], and n = 4 [IP1, (C) and (D)] independent experiments. (E and F) Analysis of the effects of the indicated concentrations of FR on intrinsic IP1 accumulation (E) and SRE reporter activity (F) in Gαq/11 KO cells transiently expressing Gαqwt or GαqQ209L. Data are means ± SEM of three experiments. (G) Analysis of the effects of the indicated concentrations of FR on SRE reporter assays, with HEK293 Gαq/11 KO cells expressing WT or Q209L mutant proteins together with key FR-binding site deletions. (H) Illustration of the FR-binding site on Gαq and the sites of the individual loss-of-function mutations from (G). (I and J) Illustration of the overall structure (I) and a magnified view (J) of an AlF4-dependent complex of Gαq and PLCβ3 (PDB, 3ohm) in the absence (left) and presence (right, dark green) of FR, which was accommodated at the domain interface of the active Gαq-PLC-β3 complex. (K) Scheme for the kiss-and-run activation of PLCβ. GPCR activation promotes the exchange of GDP for GTP on the Gα subunit. Gα-GTP separates from Gβγ to bind to and activate its effector PLCβ (kiss). Deactivation of signaling occurs when PLCβ converts Gα-GTP into Gα-GDP with its intrinsic GTPase activity, causing the complex to dissociate (run) and to reset the inactive GDP-bound state. FR disrupts this cycle and attenuates PLCβ signaling by preserving Gαq in the inactive, GDP-bound state, thereby diminishing the probability of “kissing.” (L and M) Single-cell FRET imaging of HEK293T cells expressing the muscarinic M3 receptor, Gβ1γ2, YFP-PLCβ3, and CFP-Gαqwt. (L) Cells were super-fused with buffer and buffer containing 10 μM CCh consecutively as indicated, normalized to basal. Cells were (or not) additionally exposed to 1 μM FR during the period of CCh treatment (WT, n = 17 cells, four experiments; Q209L, n = 24 cells, three experiments). (M) FRET signals shown as absolute FYFP/FCFP ratio of experiments with and without 10 μM CCh (WT, n = 35 cells, four experiments; Q209L, n = 40 cells, six experiments). Data are means + SEM of the indicated number of independent experiments. Statistical significance was calculated by unpaired t test (A, B, F, L, and M) or one-way ANOVA (E). *P < 0.05, **P < 0.01, ***P < 0.001.

To corroborate the notion that FR preferentially inactivates only a subset of Gαq/11Q209X effectors, and to provide experimental evidence for a direct, functionally relevant interaction between FR and mutant GαqQ209L, we used CRISPR-Cas9 genome-engineered HEK293 cells lacking functional alleles for GNAQ and GNA11 [HEK-Gαq/11 KO (21)]. Gαq/11-null cells are an ideal genetic background for functional analysis of signaling phenotypes associated with WT and mutant Gαq isoforms without the confounding variable of endogenously expressed Gαq and Gα11 subunits. FR reduced tonic IP1 abundance when Gαqwt, but not GαqQ209L, was reexpressed in HEK-Gαq/11 KO cells, which is consistent with a previous study (14) and our own HCmel12 data (Fig. 2E, compare with Fig. 2A). Using a serum response element (SRE)–controlled reporter assay designed to monitor the activity of the ERK signaling cascade on the transcriptional level, we achieved sustained basal ERK signaling with forced expression of Gαqwt or GαqQ209L, enabling direct quantitative comparison of FR action (Fig. 2F). Consistent with our HCmel12 findings (Fig. 2A), we confirmed suppression of SRE activity at nanomolar FR concentrations (Fig. 2F). These data suggest that reexpression in genome-edited cells of WT Gαq and the GTPase-deficient GαqQ209L mutant accurately mimicked the effector-dependent inhibition of signaling that we observed in the CM HCmel12 cells (compare Fig. 2A with Fig. 2, E and F). The FR-mediated decrease in SRE activity was disrupted when key determinants of the FR-Gαq interaction (22) were altered by site-directed mutagenesis [Fig. 2, G (top) and H, and fig. S5]. Because a comparable loss of FR potency was observed when equivalent mutations were introduced into the GqQ209L mutant backbone (Fig. 2G, bottom), we concluded that FR not only exerted its effects by likely directly binding to and inhibiting Gqwt and GqQ209L but also shared a similar binding topology. Thus, the Q209L mutation, which inactivates the Gαq-GTPase activity but is outside the FR-binding site (Fig. 2H), likely facilitated both occupancy of GαqQ209L by FR in a manner comparable to that observed for Gαqwt and enabled potent inhibition of GαqQ209L-driven mitogenic responses by FR. This also likely provides a molecular explanation for the potent antiproliferative effects of FR in HCmel12 cells harboring Gα11Q209L.

Inefficient, and to some degree variable, inhibition of the canonical GαqQ209L-PLCβ-IP1 signaling axis by FR (Fig. 2, A and E) might then be rationalized by (i) FR accommodation within an active Gαq-PLCβ complex (Fig. 2, I and J) and (ii) disruption of the kiss-and-run mechanism that normally secures spatiotemporal control over the Gαq-PLCβ interaction (13, 23) (Fig. 2K). Because PLCβ is not only an effector of Gαq but also a GTPase-activating protein for Gαq (2325), a GTP hydrolysis–deficient Gαq can neither perform GTP hydrolysis in response to an interaction with PLCβ nor separate from PLCβ; thus, occupancy by FR of this complex may be of no functional relevance. Consistent with this hypothesis, we found rapid engagement of PLCβ3 by WT Gαq [activated here with carbachol (CCh) by stimulation of endogenous muscarinic M3 receptors], which was reversed by CCh removal or addition of FR in real-time fluorescence resonance energy transfer (FRET)–based recordings (Fig. 2, L and M). In contrast, FR was unable to separate the FRET pair formed between the phospholipase and GαqQ209L (Fig. 2L; see increased FRET between both proteins in Fig. 2M). These data suggest a potential molecular explanation for the lack of an inhibitory effect of FR on the PLCβ signaling branch.

FR effectively dampens signaling by non-PLCβ effectors downstream of Gαq/11 oncoproteins

Although canonical Gq signaling is associated with the activation of PLCβ isoforms, various noncanonical Gq interaction partners have been reported (12, 26), and for some Gq-dependent processes, the precise nature of the downstream effector has yet to be elucidated (12). Therefore, we took advantage of phenotypic, whole-cell biosensing based on dynamic mass redistribution (DMR) as a pathway-unbiased assay platform to interrogate the FR-GαqQ209L interaction hypothesis further. In cells grown on biosensor plates, DMR arises from morphology changes when cells are exposed to pharmacologically active stimuli (27). We observed DMR profiles in response to FR that were downward-deflected, concentration dependent, and strictly related to the presence of transfected GαqQ209L (Fig. 3, A and B). Instead, all cells responded with positive DMR deflection to the GPCR-independent stimuli forskolin and epidermal growth factor (EGF), which served as viability controls across all transfectants (Fig. 3A). CCh-induced DMR was also apparent in GαqQ209L-expressing cells, albeit with compromised signal amplitudes (Fig. 3A). These data suggest that GαqQ209L might not be entirely decoupled from activation by Gq-coupled GPCRs but may also be GDP bound (inactive) at steady state. Consistent with this assumption, CCh-mediated DMR was blunted by atropine, a blocker of muscarinic acetylcholine receptors (Fig. 3C), and by FR independently of the mutational state of Gαq (Fig. 3, D and E). We therefore analyzed the guanine nucleotides bound to either protein in [32P]orthophosphate-labeled cultures after immunoprecipitation. Whereas WT Gαq was immunoprecipitated bound entirely to GDP, GTPase-deficient GαqQ209L was isolated with both GTP and GDP, as determined by thin-layer chromatography (Fig. 3F). To exploit the selective interaction of Gβγ heterodimers with the GDP-bound fraction of Gα, we pulled down Gα with His-tagged Gβγ before analysis of the bound nucleotide (Fig. 3G). As expected, WT and GTPase-deficient Gα proteins formed stable complexes with Gβγ that exclusively contained GDP. Our findings may reconcile the apparent paradox that GDIs, which preserve Gαq in its inactive, GDP-bound state, also suppress signaling by GTPase-deficient Gαq.

Fig. 3 FR effectively dampens signaling of non-PLCβ effectors downstream of Gαq/11oncoproteins.

(A) Representative real-time DMR recordings of the cellular responses induced by FR and the indicated stimuli [CCh, 30 μM; forskolin (Fsk), 30 μM; EGF, 50 nM] in Gαq/11 KO cells transiently transfected with empty plasmid or with plasmid encoding Gαqwt or GαqQ209L. (B) Quantification of the effects of FR on Gαqwt and GαqQ209L from the experiments shown in (A). Data are means ± SEM of three experiments. (C and D) Representative DMR profile of CCh-induced cell responses and inhibition with 100 μM atropine (C) or FR (D) in Gαq/11 KO cells transiently expressing the indicated Gα proteins. (E) Quantification of the effects of FR on the positive DMR deflections from the experiments shown in (D). Data are means ± SEM of three experiments. (F and G) Thin-layer chromatography (TLC) of lysates of [32P]orthophosphate-labeled HEK293T cells transfected with empty vector or with plasmid encoding the indicated HA-tagged Gα subunit. Subunits were isolated either by immunoprecipitation (IP) with anti-HA antibody (F) or by their interaction with His6-Gβγ isolated by pulldown (PD:His) with TALON resin (G). The detection of Gα subunits by Western blotting analysis (G, left) confirmed their expression in total cell lysates and successful pulldown with His6-Gβγ. (H) Analysis of the effects of the indicated concentrations of FR on YAP phosphorylation (by comparing total YAP abundance to pYAP abundance) in Gαq/11 KO cells transiently transfected with empty vector or with plasmid encoding Gαqwt or GαqQ209L. (I) Analysis of the effects of the indicated concentrations of FR on pYAP abundance in Gαq/11 KO cells expressing WT Gαq or the Q209L mutant Gαq with or without the I190W mutation within the FR-binding site. For the I190W Q209L mutant, a plateau was not reached; therefore, the effect of 10 μM FR was arbitrarily set to 50%. Statistical significance was calculated using the Wilcoxon test (B) or by one-sample t test (H). *P < 0.05, **P < 0.01, ***P < 0.001.

YAP, another context-specific, PLCβ-independent effector downstream of Gαq/11 (26), is activated by dephosphorylation in the presence of high Gq activity (2830). We therefore investigated the effect of FR on YAP phosphorylation in HEK-Gαq/11 KO cells transfected to express either Gαqwt or GαqQ209L. We found that FR had no effect on cellular YAP abundance in Gαq/11-null cells transfected with empty plasmid but increased the percentage of phosphorylated YAP (pYAP) in cells transfected to express Gαqwt or GαqQ209L (Fig. 3H). The effects of FR were concentration-dependent and were diminished by about 15-fold in cells expressing WT Gαq or GαqQ209L that contained the I190W mutation, which disrupts FR binding (Fig. 3I). These data highlight the capacity of FR to dampen signaling by non-PLCβ effectors downstream of Gαq/11 oncoproteins, presumably through a direct, functionally relevant interaction with its target protein.

FR blunts prosurvival signaling in UM cell lines

Because FR effectively suppressed Gαq/11Q209L-driven mitogenic signaling, we further investigated its activity in UM cell lines because UM is a cancer type in which Gαq/11Q209 mutations are frequently detected and are considered as the main oncogenic drivers (1, 3, 31, 32). Here, we used four UM cell lines with Gαq/11 mutations (Mel270, Mel202, Mel92.1, and OMM1.3) and two UM cell lines without those mutations (Mel285 and Mel290). Consistent with the presence of constitutively active Gq signaling, we found that IP1 abundance was increased in all of the UM cell lines expressing Gαq/11Q209 mutants (Fig. 4A); however, unexpectedly, IP1 abundance also increased in the cell lines expressing WT Gαq (Fig. 4A). We therefore confirmed the presence or absence of GNAQ/11 mutations across all of our UM cell lines (fig. S6) and investigated their responsiveness toward the pan–G protein activator aluminum tetrafluoride (AlF4), which mimics the presence of the terminal phosphate of GTP to generate an active transition state conformation (33, 34). Consistent with the GNAQ/11 mutational status of the cell lines, AlF4 evoked IP1 production exclusively in cells expressing WT Gαq (fig. S7). Unexpectedly, FR reduced tonic IP1 amounts in Mel285 cells (WT Gαq) and in all of the UM cell lines expressing mutant Gαq but had hardly any effect on the Mel290 cell line (WT Gαq) (Fig. 4B). These results may seem to contradict our previous findings, first, because we expected FR to be poorly effective in the UM cell lines expressing mutant Gαq and, second, because we anticipated FR to be effective in the cell line expressing WT Gαq. It is possible that the rate of GTP hydrolysis induced by the GTPase activating protein (GAP) activity of PLCβ may determine susceptibility toward inhibition by FR and may vary considerably with cellular context. However, PLCβ GAP activity in UM cells has not been previously demonstrated and cannot be easily determined or extrapolated from in vitro data. Resistance to the inhibitory effect of FR in Mel290 cells might be explained by the presence of an activating mutation in PLCβ4, a downstream effector of Gαq signaling (35). However, Sanger sequencing assigned Mel290 cells as expressing WT PLCB4 (fig. S8). Because AlF4-dependent IP1 accumulation was effectively suppressed by FR in Mel290 cells (fig. S8), FR-resistant accumulation of this second messenger is unlikely to have arisen through an activated Gq/11 pathway. Regardless, clarification of these issues requires future studies, not least because PLCβ is not a major contributing factor to propagate mutant Gq/11 activity in UM tumorigenesis (7, 28, 30, 36).

Fig. 4 FR suppresses prosurvival signaling downstream of Gαq/11 oncoproteins in UM cell lines.

(A and B) Quantification of cell-intrinsic IP1 production in the indicated UM cell lines (A) and analysis of their sensitivity to treatment with the indicated concentrations of FR (B). Gray and blue bars in (A) indicate experiments with 10,000 cells per well and 50,000 cells per well, respectively. In (B), 10,000 cells per well were used. Data are from four biological replicates. (C) Comparison of the relative amounts of pERK1/2 in the indicated UM cell lines in the absence (Basal) and presence of 1 μM trametinib (Tram). For Mel290, Mel285, and 92.1 cells, data are from four independent experiments; for Mel270, Mel202, and OMM1.3 cells, data are from three independent experiments. (D) Analysis of the effects of the indicated concentrations of FR on pERK1/2 amounts in the indicated UM cell lines. Data are means ± SEM of three experiments. (E) Comparison of the relative amounts of pAKTS473 in the indicated UM cell lines in the absence and presence of 30 μM LY294002. Data are from four independent experiments. (F) Analysis of the effects of the indicated concentrations of FR on pAKTS473 abundance in the indicated UM cell lines expressing WT (n = 3 experiments) or mutant Gαq/11 (n = 4 experiments). Statistical significance in (B) was calculated by a one-sample t test; for (C) and (E), unpaired t tests were used. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

To elaborate opportunities for targeted therapy, we next investigated the activation and sensitivity to FR of effector pathways that are relevant for survival, cell cycle progression, and cellular growth downstream of Gαq/11 oncoproteins. All of the UM cell lines showed increased activity of the MAPK cascade with varying degrees of ERK1/2 phosphorylation across cell lines, an effect that was blunted by the MAPK kinase (MEK) inhibitor trametinib irrespective of the mutational status of Gαq (Fig. 4C). In contrast, FR suppressed ERK1/2 phosphorylation selectively in those cells that carried Gαq/11 mutations (Fig. 4D; see fig. S9 for the effects of FR and trametinib on total ERK protein). A similar picture emerged when we analyzed phosphatidylinositol 3-kinase (PI3K)/AKT signaling, another survival pathway that is highly activated in UM (3739). We found increased AKT phosphorylation at Ser473 in all UM cell lines, an effect that was consistently inhibited by pretreatment with the PI3K inhibitor LY294002 (Fig. 4E). We then monitored AKT phosphorylation over time in experiments in which a single high FR concentration was used to determine the optimal time point for quantification of the effects of FR (fig. S10). Increasing concentrations of FR reduced the abundance of phosphorylated AKTS473 (pAKTS473) exclusively in UM cell lines expressing mutant Gαq, but not in cell lines expressing WT Gαq, further suggesting that the Gαq/11 mutation status was the critical determinant of whether cells responded to FR (Fig. 4F; see fig. S10 for the effects of FR on total AKT protein). Thus, despite evidence of apparent Gαq activity in UM cell lines expressing the WT protein, this pathway did not seem to promote prosurvival signaling through AKT and ERK pathways.

FR reverses YAP activity in UM cell lines

YAP, a major effector of the Hippo tumor suppressor pathway, is a key player in UM tumorigenesis (28, 30, 40). Because oncogenic Gαq/11 proteins activate YAP by dephosphorylation, which induces its nuclear translocation and transcriptional activity (30), and because YAP activation may also occur independently of mutant Gq (29, 40), we set out to use enhancement of its phosphorylated fraction by FR as a pharmacological indicator to identify those UM cell lines that might benefit from Gq-targeted therapy. We initially quantified basal amounts of pYAP across all UM cell lines using a homogeneous time-resolved fluorescence (HTRF)–based assay for the detection of total YAP and pYAP (Ser127) and noted a trend toward lower pYAP abundance in those lines carrying mutant Gq (Fig. 5A). These data suggest that YAP is activated to a larger extent in cells expressing mutant Gαq/11 as compared with UM cell lines expressing WT Gαq/11, consistent with previous reports (28, 30, 41). However, substantial increases in pYAP abundance in response to FR were detected only in the metastatic UM cell lines Mel92.1 (28, 30) and OMM1.3 (30) (Fig. 5B). Because increased pYAP abundance in response to forskolin, a Gq/11-independent YAP inhibitor, was quantitatively rather modest across all UM cell lines (Fig. 5B), we assessed YAP phosphorylation by Western blotting analysis (Fig. 5C). Treatment of cells with FR increased pYAP abundance in mutant Mel270, Mel92.1, and OMM1.3 lines but not in mutant Mel202 cells or in the cell lines expressing WT Gαq/11 (Fig. 5, C and D). Consistently, in these same three cell lines, we observed the nuclear sequestration of YAP by FR through imaging experiments with anti-YAP antibody (Fig. 5E). Thus, these data suggest that FR also inhibits Gαq/11-driven mitogenic signaling involving the transcriptional coactivator YAP in UM cell lines expressing mutant Gαq/11.

Fig. 5 FR inhibits YAP signaling in Gq mutant UM cell lines.

(A) Quantification of cytosolic pYAP amounts in the indicated UM cell lines expressing WT or mutant Gαq/11. Dots depict single experiments; horizontal lines show total averaged data ± SEM (vertical bars). (B) Effects of FR and forskolin on pYAP abundance in the indicated UM cell lines. For Mel202 and 92.1 cells, n = 7 experiments; for OMM1.3 cells, n = 5; for all others, n = 6. Data are expressed relative to vehicle-treated cells. (C and D) Western blotting analysis (C) and quantification (D) of relative pYAP abundance in the indicated cells after treatment for 1 hour with 10 μM FR or vehicle. Data are from four independent experiments. (E) Analysis of the subcellular distribution of YAP in the indicated UM cell lines after treatment with 10 μM FR or vehicle for 1 hour. Cells were analyzed by immunofluorescence microscopy after treatment with anti-YAP antibody (green) and DAPI (blue) to stain nuclei. Images are representative of three experiments. Statistical significance was calculated with an unpaired t test for (A) or a one-sample t test for (B) and (D). Data in (B) and (D) are means + SEM of the indicated number of independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

FR suppresses tumor growth in mouse xenografts driven by mutant GαqQ209P but not by mutant B-RafV600E

On the basis of our earlier findings, we investigated whether FR could inhibit proliferation selectively in UM cell lines carrying mutant Gq. Treatment with FR for 72 hours caused substantial reductions in cell number in a strictly mutation-specific manner, with effects comparable to those achieved with the MEK inhibitor trametinib (Fig. 6A). Similar results were obtained when UM cell lines in culture were treated with FR or trametinib for an equivalent period of time before being subjected to bright-field microscopic image acquisition (Fig. 6B). To investigate whether mutation-specific FR effects translated to in vivo efficacy, we subcutaneously grafted GαqQ209P Mel270 cells into severe combined immunodeficient (SCID) mice. After 14 days, tumors had grown to 415 mm3 in vehicle-treated mice but were inhibited in growth by 72% in FR-treated mice (Fig. 6C). In contrast, the growth of xenografts established from mutant B-RafV600E melanoma cells was unaffected by FR (Fig. 6D). The marked sensitivity of Gαq mutant xenografts toward treatment with FR is consistent with our in vitro data and indicates that the inhibition of oncogenic Gq activity with FR may qualify for effective on-target therapy in UM. Hence, the discovery of how molecules such as FR interact with and inhibit mutationally activated Gq changes our understanding about how G proteins can be targeted pharmacologically for therapeutic benefit.

Fig. 6 Mutation-specific inhibition of proliferation and tumor growth by FR as a strategy to treat Gαq-driven cancer.

(A) Effect of the indicated concentrations of FR on the number of the indicated UM cell lines expressing WT or mutant Gαq/11 after 72 hours of treatment. As a control, cells were treated with 1 μM trametinib. The numbers of treated cells are expressed as a percentage of the numbers of vehicle-treated controls (n = 3 biological replicates each performed in duplicate). (B) Representative bright-field images (n = 3 experiments) of UM cell lines expressing WT or mutant Gαq/11 treated with 100 nM FR or 1 μM trametinib for 72 hours. Scale bars, 100 μm. (C) Left: Effects of FR and vehicle on the growth of Mel270 (GαqQ209P) UM cells in a xenograft mouse model in vivo (n = 5 animals per control- and FR-treated group). Right: Representative histological section of a Mel270 tumor originating from xenograft experiments and stained with hematoxylin and eosin after 14 days of treatment with FR or vehicle. Tumor size was monitored every 3 to 4 days after the initiation of treatment. ***P < 0.001. (D) An experiment comparable to that shown in (C) was performed with CM A375 (B-RafV600E) xenografts (n = 6 animals per control- and FR-treated group). Data in (A), (C), and (D) are means ± SEM of the indicated numbers of experiments.


A major challenge in cancer treatment is to selectively target those gene products that initiate and maintain cancer cell proliferation and survival. In contrast to CM, in which oncogenic mutations in the genes encoding B-Raf, N-Ras, or Kit are frequently found (42, 43), UM instead harbors mutually exclusive mutations at codon 209 in GNAQ and GNA11, two closely related genes that code for the α subunits Gαq and Gα11 of the heterotrimeric Gq family of proteins (44). In UM, these GαqQ209L/P and Gα11Q209L/P GTPases are considered as the main oncogenic drivers upstream of prosurvival pathways. So far, mutationally activated Gαq and Gα11 have not been targeted therapeutically, and inhibition of downstream effectors, thus far the only rational approach for UM treatment, has been unsuccessful (11, 45, 46). Our proof-of-concept study shows that direct targeting of Gαq/11 oncoproteins is both feasible and effective. Although a priori considered mechanistically intractable [Gαq/11Q209L/P proteins were thought to be insensitive to GDIs, such as FR and YM (14, 15)], FR suppressed numerous hallmark features of malignant melanoma cells carrying oncogenic Gαq/11. With strong correlation to mutational status, FR effectively inhibited oncogenic signaling and proliferation in vitro as well as tumor growth in vivo. Mutant Gαq/11, but not intrinsic Gq activity per se, served as a biomarker in UM to predict FR responsiveness. Because more than 85% of UM tumors harbor activated forms of Gαq/11 (1, 2), we suggest that molecules such as FR or its derivatives hold promise as pharmacological agents for the direct targeting of these challenging oncoproteins in cancers with Q209X pathway mutations. This will certainly also require the development of strategies for the specific delivery of FR to target cells to avoid interference with the Gq signaling cascades of normal, nontumor cell populations.

Although it is presently unclear whether development of precision medicines only targeting mutationally activated Gq will be successful, it is clear that such inhibitors may act as GDIs even if proteins are GTPase deficient and are considered constitutively active: This is because GTPase-deficient GαqQ209L is not only exclusively GTP-bound (active) but also GDP-bound (inactive) at steady state. Hence, our findings, together with those from another study (17), clarify the enigma of how FR blunts signaling by both WT and oncogenic Gαq variants and, thereby, reconcile apparently disparate published data to settle a therapeutically relevant dispute.

While this work was under revision, two studies reported on the in vitro inhibition of mutationally activated Gαq by FR in two different UM cell lines (17, 18). The current study extends these findings by (i) analyzing six UM cell lines and two CM lines and (ii) by providing evidence for the FR-mediated inhibition of mutationally activated Gq in vivo. However, all three reports attest to the capacity of FR to blunt signaling by oncogenic Gαq/11 GTPases, despite intriguing differences in their interaction with βγ and downstream signaling partners (47). Hence, these studies offer insights into and perspectives for future treatment options of UM, a disease with high unmet medical needs.


FR extraction

FR was isolated and purified from the dried leaves of Ardisia crenata, as previously described (14). Reversed-phase high-performance liquid chromatography separation of the FR-containing fraction [column: YMC C18 Hydrosphere, 250 × 4.6 mm, 3 μm; MeOH:H2O (8:2), 0.7 ml min−1] yielded FR of 95% purity.

Cell lines and culture conditions

Cell lines were cultured at 37°C in a humidified atmosphere of 95% air and 5% CO2. All media were supplemented with penicillin (100 U/ml) and streptomycin (100 mg/ml). Fetal calf serum (FCS) (5%) was added to MCDB 153 medium (Sigma) designed for low-serum growth, whereas all other media contained 10% FCS. The UM cell lines 92.1, Mel202 (both carrying mutated GαqQ209L), Mel270 (GαqQ209P-mutated), Mel258, and Mel290 (both Gαq/11wt) were provided by M. Jager (University of Leiden, Netherlands). The metastatic UM cell line OMM1.3 (GαqQ209P) was a gift of B. R. Ksander (Harvard Medical School, USA). All UM cell lines were cultured in RPMI 1640 medium (Life Technologies). The mouse CM cell lines B16 [Gαq/11wt from the American Type Culture Collection (ATCC)] and HCmel12 [with mutated Gα11Q209L generated from a 7,12-dimethylbenz(a)anthracene–induced melanoma from HGF-CDK4(R24C) mice (48)] as well as the human CM cell lines Mamel119 (Gαq/11wt, BRAFwt) and Skmel28 (Gαq/11wt, B-RafV600E) were provided by D. Schadendorf (University of Essen, Germany). The human skin melanoma cell line A375 (carrying B-RafV600E) was obtained from Sigma. CM cell lines were cultured in RPMI 1640 medium complemented with 2 mM L-glutamine, 10 mM nonessential amino acids, 1 mM Hepes (all from Life Technologies), and 20 mM 2-mercaptoethanol (Sigma). HEK293 cells (ATCC) genome-deleted by CRISPR-Cas9 of functional alleles of GNAQ and GNA11 (HEK293 Gαq/11 KO cells) were established as previously described (14). Native and genome-edited HEK293 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Life Technologies) with supplements as described earlier.

PCR-based authentication of cell lines

The CM and UM cell lines were authenticated by polymerase chain reaction (PCR) analysis and Sanger sequencing for mutations in GNAQ/GNA11, NRAS, BRAF, and PLCB4 (Mel290). All cell lines were mycoplasma-tested by PCR on a monthly basis and were found to be mycoplasma free.

Flow cytometry

HCmel12 cells were seeded into six-well plates (1 × 105 cells per well) in complete RPMI medium. Six hours later, FR was added at various concentrations. Control cells received vehicle only. After 72 hours, cells were collected and analyzed for apoptosis induction and cell cycle arrest using standard protocols with annexin V (BD PharMingen) and PI. Data were acquired with a FACSCanto flow cytometer (BD Biosciences) and analyzed with FlowJo software (Tree Star, V7.6.5 for Windows).

XTT assays

HCmel12 cells (1 × 104 cells per well) were seeded into flat-bottom 96-well plates. Six hours later, FR was added at the concentrations indicated in the figure legends, and 1 μM trametinib (MEK inhibitor) was used as a control. XTT reagent was added 2 hours after the addition of FR. Forty-eight hours later, cell metabolic activity was measured with the XTT-based Cell Proliferation Kit II (Roche) according to the manufacturer’s protocol. The absorbance at 450 nm was recorded using a TECAN spark 10 M plate reader, and data were expressed as a percentage of the cell metabolic activity relative to that of vehicle-treated control cells.

Cell proliferation assays

For analysis of cell proliferation, HCmel12 cells (1.5 × 105 per well) and UM cell lines (2 × 105 per well) were seeded into six-well plates in complete RPMI medium. Six hours later, FR or MEK inhibitor trametinib (1 μM) was added. Reference groups were solvent-treated. Cell growth was photo-documented over time using a Zeiss Axio Vert.A1 inverted microscope. After 72 hours, cells were collected and viable cells were counted using trypan blue.

Bright-field microscopic imaging

Imaging of cell morphological changes was performed 3 days after treatment with FR, trametinib, or vehicle by bright-field microscopy. Cells were grown at a density of 15,000 cells per well for HCmel12, 92.1, and Mel285 cells or at a density of 30,000 cells per well for Mel270, Mel290, Mel202, and OMM1.3 cells in RPMI medium on 96-well ibidi plates coated with poly-d-lysine (PDL) for high-resolution microscopy. Compounds or vehicle was added 3 hours after seeding, and the plate was incubated for another 72 hours at 37°C/5% CO2. Cells were then fixed with 4% paraformaldehyde (PFA) for 20 min at room temperature and washed twice with phosphate-buffered saline (PBS). Images were taken with a Leica DFC 360FX microscope.

Transwell migration assays

Migration of HCmel12-eGFP cells was quantified using transwell assays. Briefly, 5 × 104 cells in 250 μl of RPMI 1640 with 1% FCS were placed in the upper chamber of uncoated polyethylene terephthalate (PET) filters (8 μm pore size, BD Biosciences) in a 24-well plate and subsequently incubated at 37°C/5% CO2 to adhere. Six hours later, FR was added at the concentrations indicated in Fig. 1. RPMI containing 1% FCS was added to the lower chamber. Twenty-four hours later, cells in the upper chamber were removed by a cotton swab. Transmigrated cells on the lower surface of the membrane were counted in three high-power fields (magnification, ×100) using a TE Eclipse microscope (Nikon).

HTRF-based IP1 accumulation assay

IP1 accumulation was measured using Cisbio’s HTRF technology following a suspension cell–based assay protocol in 384-well plates. Cells were washed in PBS and resuspended in assay buffer containing LiCl to prevent IP1 degradation. To quantify the effects of FR on cell-intrinsic IP1 production, cells were incubated with FR or vehicle for 1 hour at 37°C. To determine FR-mediated inhibition of ligand-induced IP1 accumulation, an additional 30-min incubation step with the following pharmacological stimuli was included: CCh, 10 μM; propionic acid C3, 1 mM; or AlF4, 300 μM. HTRF signals were measured as specified later for pERK, pYAP, and pAKT detection. Ratios were either normalized to an internal control or converted to nanomolar concentrations of IP1 using the kit’s standard curve and nonlinear regression analysis. Cell numbers were adjusted to yield IP1 amounts in the linear range of the assay kit and were as follows: HCmel12, 10,000 cells per well; B16, 50,000 cells per well; HEK293, 60,000 cells per well; all UM cell lines, 10,000 cells per well. When assays were performed with multiple cell numbers, these are indicated in the respective figures or legends.

HTRF-based pERK, pYAP, and pAKT assays

Compound-induced changes in phosphorylated fractions of prosurvival pathway proteins were quantified with the HTRF technology (Cisbio) following the two-plate assay protocol for adherent cells in PDL-coated 96-well plates as per the manufacturer’s instructions. Homogeneous time-resolved FRET signals were detected with a Mithras LB940 multimode plate reader. HTRF ratios either were normalized to internal assay controls or are shown as raw data. To optimize assay performance and window, cell density and culture conditions were modified as follows: pERK/total ERK: HEK293, 40,000 cells per well; HCmel12, 25,000 cells per well; B16, 50,000 cells per well; and UM cell lines, 25,000 cells per well. Cell-intrinsic ERK phosphorylation was detected without previous starvation, whereas GPCR agonist–mediated pERK abundance was determined after a 4-hour starvation period. For the ratio of pYAP to total YAP, to ensure activation of the Hippo pathway, cells were starved overnight and seeded at a high cell density (HEK293, 100,000 cells per well; UM cell lines, 75,000 cells per well). For the ratio of pAKT to total AKT, all cell lines were seeded at a density of 75,000 cells per well. Cells were not starved, and the culture medium was changed 1 hour before compound addition to enable proliferation and prosurvival pathway activation. Cells were seeded into assay plates 18 hours before the assay was started. The next day, culture medium was either changed for a growth-friendly environment (for ERK and AKT determination) or left unchanged for the YAP assays. FR, vehicle, or controls (trametinib, 1 μM; LY294002, 30 μM; forskolin, 10 μM) were added in medium with or without 10% FCS, consistent with the starvation conditions described earlier, and the cells were incubated for 1 hour or for different time intervals as specified for the kinetic measurements in the appropriate figure legends. Cells were then either directly lysed or, for detection of ligand-induced changes in pERK abundance, stimulated for another 3 min with CCh or propionic acid C3 before lysis. All incubation steps were performed under cell culture conditions at 37°C.

Gene expression analysis

Generation of GαqQ209L in pcDNA3.1(+) (Life Technologies) from Gαqwt was by QuikChange site-directed PCR mutagenesis as previously described (14). The same procedure was used to generate mutant expression plasmids encoding GαqV182S, GαqV184M, GαqI190N, and GαqI190W [all hemagglutinin (HA)–tagged in pcDNA3.1(+)] using the primers as specified below: V182S, 5′-GTGCTTAGAAGTCGAGTCCCCACTACAGGGATC-3′ (forward) and 5′-GATCCCTGTAGTGGGGACTCGACTTCTAAGCAC-3′ (reverse); V184 M, 5′-GCTTAGAGTTCGAATGCCCACTACAGGGATC-3′ (forward) and 5′-GTAGTGGGCATTCGAACTCTAAGCACGTCTTGTTG-3′ (reverse); I190N, 5′-CAGGATCAACGAATACCCCTTTGACTTACAAAG-3′ (forward) and 5′-CAAAGGGGTATTCGTTGATCCCTGTAGTGGG-3′ (reverse); I190W, 5′-CAGGGATCTGGGAATACCCCTTTGACTTAC-3′ (forward) and 5′-GTAAGTCAAAGGGGTATTCCCAGATCCCTG-3′ (reverse). All mutant constructs were verified by sequencing. For transient expression of the pcDNA3.1(+)-based constructs in HEK293 cells, polyethylenimine (PEI) (Polysciences) was used as transfection vehicle with a DNA/PEI ratio of 1 μg of DNA/3 μl of PEI for 1.8 × 106 cells in one well of a six-well plate. This optimal DNA/PEI ratio (quantity and volume) was scaled according to the well numbers and their surface area.

SRE assays

HEK293 Gαq/11 KO cells were transiently transfected with expression plasmids encoding Gα proteins, luciferase reporter vectors pGL4.33[luc2P/SRE/Hygro], and pCMV-Renilla (DNA ratios, 5:3:0.3). Twenty-four hours later, cells were seeded at a density of 50,000 cells per well in 96-well fibronectin-coated white opaque plates. The following day, cells were starved for at least 1 hour and incubated with FR or vehicle for 60 min before being treated with agonist or vehicle. Six hours later, the cells were treated sequentially (10 min each) with (i) Dual-Glo Luciferase reagent (Promega), which was followed by quantitation of the firefly luciferase reaction (light A), and (ii) Stop & Glo reagent (Promega) to quench firefly luciferase luminescence but activate Renilla luciferase (light B). Luminescence measurements were performed with a Mithras LB 940 multimode reader (Berthold Technologies).

FRET measurements

HEK293T cells were transfected with plasmids encoding muscarinic receptor M3, PLCß3-YFP, and the G protein subunits Gβ1-, Gɣ2-, and either cyan fluorescent protein (CFP)–Gαq or CFP-GαqQ209L and then were seeded for 24 hours in six-well plates. FRET measurements were performed 48 hours after transfection. The cells were incubated in FRET buffer [137 mM NaCl, 5.4 mM KCl, 2 mM CaCl2, 1 mM MgCl2, and 10 mM Hepes (pH 7.3)], and FRET imaging was performed with a light-emitting diode excitation system on the fluorescence microscope Eclipse Ti by Nikon. Emitted light was measured at 425 and 500 nm in an interval of 2 s. For the experimental setup, cells were exposed to a constant flow of FRET buffer in the ALA-VC3-8SP perfusion system (ALA Scientific Instruments). The buffer was replaced with buffer supplemented with 10 μM CCh for 30 s, which was followed by a washing step of 1.5 min. Thereafter, FRET buffer including CCh was re-added for another 30 s before the buffer was again replaced by a mixture of buffer with 10 μM CCh and 1 μM FR. After the cells were treated with FR, a second washing step of 1 min with buffer flow was implemented before the cells were once more stimulated with CCh.

DMR biosensing

Integrated, real-time, whole-cell activity profiles in response to FR, EGF (Sigma), CCh (Sigma), and forskolin (Tocris) were monitored with an optical biosensor on the basis of the detection of DMR (Corning Epic System). For DMR detection, CRISPR-Cas9 Gαq/11 KO cells transfected with empty vector control or plasmid encoding Gαqwt or GαqQ209L using the FuGENE transfection reagent were seeded into fibronectin-coated, 384-well biosensor microplates at a density of 15,000 cells per well 24 hours before measurements were made. On the next day (48 hours after transfection), cells were washed with Hanks’ balanced salt solution containing 20 mM Hepes followed by 1 hour of equilibration before the addition of FR and the appropriate drugs (CCh, 30 μM; forskolin, 30 μM; EGF, 50 nM). Cell responses were measured immediately after compound application. All steps, except washing, were performed at 37°C. To quantify the effects of FR, the area under the curve (AUC) of downward-deflected DMR traces within 0 and 3600 s was used and normalized to the maximum negative response, which was set at −100%.

Western blotting analysis

Cell lysates, immunoprecipitates, and protein pulldowns were loaded and separated on SDS-polyacrylamide gel electrophoresis, transferred to Immobilon-P membranes (Millipore, catalog no. IPV00010), blocked with 5% milk/1× tris-buffered saline (TBS)–0.05% Tween, and incubated overnight at 4°C on a shaker, with primary antibodies directed against gp100 (Abcam, ab137078), β-actin (Santa Cruz Biotechnology, sc-47778), 6× His-tag (Sigma, H-1029), HA-tag (Covance), Gαq (Santa Cruz Biotechnology, sc-392), pYAP Ser127 (Cell Signaling Technology, 4911), YAP (Cell Signaling Technology, 4912), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Cell Signaling Technology, D16H11). After washing three times with TBS–0.05% Tween, membranes were incubated for 1 hour at room temperature with secondary antibodies (anti-rabbit, ABIN102010, dilution 1:10,000, antibodies-online GmbH, Aachen, Germany; or anti-mouse, dilution 1:5000, KPL, catalog no. 074-1802). Last, filters were washed with TBS-Tween and revealed using Amersham ECL Prime Western Blotting Detection Reagent Kit (GE Healthcare, Chicago, USA, product no. RPN2232) as per the manufacturer’s instructions.

His6 pull-down assays and immunoprecipitations

HEK293T cells were transfected with plasmids encoding His6-Gβ1/His6-γ2 and either HA-Gαq or HA-GαqQ209L. On the next day, the cells were starved overnight. Forty-eight hours after transfection, the cells were lysed with 1.0 ml of ice-cold TBS-Triton [50 mM tris (pH 7.5), 150 mM NaCl, and 1% Triton X-100] containing 20 mM imidazole, 1 mM PMSF (phenylmethylsulfonyl fluoride), leupeptin (10 μg/ml), aprotinin (10 μg/ml), 10 mM β-glycerophosphate, 1 mM NaF, and 1 mM sodium orthovanadate. Lysates were transferred to 1.5-ml tubes and centrifuged at 4°C for 10 min at a relative centrifugal force of 15,682g. Pulldown assays to isolate His6-tagged Gβ1γ2 heterodimers were performed with TALON His-tag affinity beads (Clontech, catalog no. 8908-2). TALON beads are charged with cobalt, which is known to bind to His-tagged proteins with high specificity. Centrifuged lysates were transferred to 1.5-ml tubes and incubated with 30 μl of TALON beads in an ice bath with constant shaking for 30 min. The beads were then washed three times with lysis buffer, suspended in 1× Laemmli buffer containing β-mercaptoethanol, boiled for 5 min, and centrifuged at a relative centrifugal force of 15,682g for 5 min before being subjected to Western blotting analysis. For total cell lysates, a fraction of lysates was diluted with 4× Laemmli buffer containing β-mercaptoethanol, boiled for 5 min, and centrifuged at a relative centrifugal force of 15,682g for 5 min before Western blotting was performed. For immunoprecipitations, cell lysates obtained from HEK293T cells transfected with plasmids encoding either HA-Gαq or HA-GαqQ209L were transferred to 1.5-ml tubes and incubated with anti-HA monoclonal antibodies on a rocking platform overnight at 4°C. The next day, 30 μl of protein G Sepharose (Millipore, catalog no. 16-266) was added and the incubation was continued for 3 hours. Beads were washed three times with lysis buffer and subjected to Western blotting analysis or thin-layer chromatography.

Thin-layer chromatography analysis

HEK293T cells transfected with plasmid encoding Gαqwt or GαqQ209L, together with plasmid encoding His6-Gβγ or eGFP (enhanced green fluorescent protein), were metabolically labeled with [32P]orthophosphate (169 μCi/ml) and subjected to immunoprecipitation with anti-HA antibodies or pulldown with TALON resin. The content of either GDP or GTP in the isolated GTPase was revealed by thin-layer chromatography followed by autoradiography, as previously described (49).

Molecular modeling

FR is a close analog of YM: The atomic coordinates of YM were retrieved from the crystal structure of Gαq in the inactive state [Protein Data Bank (PDB) ID, 3ah8 (19)] and transformed to the structure of active Gαq in complex with PLC-β3 [PDB ID, 3ohm (23)] by superimposition of the backbone atoms of the two proteins using a mean square distance optimization method, as previously described (50). To model FR, the ligand was modified by adding and removing the appropriate atoms and locally minimized using MOE (Molecular Operating Environment, version 2016) using the AMBER 10 force field (51).


Cells were cultured on poly-l-lysine–coated coverslips and fixed with 4% PFA in PBS. Cells were permeabilized with 0.1% Triton X-100 for 5 min at 4°C, which was followed by blocking in PBS, 0.5% bovine serum albumin (BSA) for 30 min at room temperature. After blocking, cells were incubated in blocking solution (PBS, 0.5% BSA) containing an anti-YAP primary antibody (Santa Cruz Biotechnology) overnight at 4°C. The next day, cells on the coverslips were visualized with Alexa Fluor–labeled secondary antibodies (Thermo Fisher Scientific) and 4′,6-diamidino-2-phenylindole (DAPI) (Molecular Probes) for nuclear staining. Cells were mounted on glass slides with Shandon Immu-Mount (Thermo Fisher Scientific).

Human tumor xenografts and in vivo treatment with FR

Female NOD.Cg-Prkdcscid Il2rgtm1wjl/SzJ mice [commonly known as nonobese diabetic (NOD)/SCID gamma mice; The Jackson Laboratory, Maine], 7 to 8 weeks of age and weighing 20 to 22 g, were used in the study of Mel270 cells, whereas female athymic nude mice [NU(NCr)-Foxn1nu mice] (Charles River), 4 weeks of age and weighing 20 to 22 g, were used in the study of A375 cells. All mice were housed in appropriate sterile filter-capped cages and provided with food and water ad libitum. All procedures were essentially performed as previously described (26, 30). Briefly, exponentially growing cultures were harvested, washed, and resuspended in RPMI 1640, and 2 × 106 viable cells were transplanted subcutaneously into the flanks of the mice. For tumor growth analysis, tumor volume was assessed as (LW2/2), where L and W represent the length and the width of the tumor. The animals were monitored twice weekly for tumor development. Results of animal experiments are expressed as means ± SEM of a total of five tumors for Mel270 cells and six tumors for A375 cells. To administer FR to mice, the chemical was first dissolved in dimethyl sulfoxide (DMSO) to make a stock solution (2 mg/ml). Water was then used as the solvent to prepare the injection suspension (100 μg/ml). The final mixture was applied to the treatment group mice (100 μl per mouse, intraperitoneal injection), and vehicle was applied to the control group mice. This xenograft study was approved by the Animal Care and Use Committee, National Institute of Dental and Craniofacial Research and was in compliance with the Guide for the Care and Use of Laboratory Animals.

Tissue preparation and histology

All procedures were essentially performed as previously described (26, 30, 52). Briefly, tumor tissues were fixed in Z-fix (zinc-buffered formaldehyde, Anatech Ltd.) overnight and then transferred to 70% ethanol. Fixed tissues were embedded in paraffin, sectioned to a thickness of 4 μm, and stained with hematoxylin and eosin. Histology slides were scanned with a Scanscope Digital microscope (Aperio).

Statistical analysis

Bars and dots represent the means +/± SEM of the indicated (n number) of experiments if not stated otherwise. Statistical analysis was performed using GraphPad Prism Software. P values and the applied statistical methods including appropriate post tests are reported in the figure legends. Statistical significance is categorized as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.


Fig. S1. Exemplary gating strategy used in flow cytometry analysis to detect and quantify apoptotic and necrotic cells.

Fig. S2. HCmel12 cells display intrinsic Gq activity.

Fig. S3. Mouse B16 melanoma cells exhibit increased levels of Gq activity.

Fig. S4. FR does not affect total ERK protein abundance in HCmel12, B16, or HEK293 cells.

Fig. S5. Functional expression of WT and mutant Gαq proteins in an SRE reporter assay.

Fig. S6. Authentication of UM cell lines by Sanger sequencing at position Q209 of Gαq.

Fig. S7. Effects of AlF4 on IP1 production in UM cell lines expressing WT or mutant Gαq.

Fig. S8. FR inhibits AlF4-mediated IP1 accumulation in Mel290 cells.

Fig. S9. FR and trametinib do not alter the abundance of total ERK1/2 proteins in UM cell lines.

Fig. S10. Time course analysis of the amounts of total AKT and pAKT in UM cell lines after treatment with FR.

Fig. S11. Neither FR nor forskolin alters the abundance of total cellular YAP protein.


Acknowledgments: We thank M. Jager (University of Leiden, Netherlands) for providing the UM cell lines 92.1, Mel202, Mel270, Mel258, and Mel290; B. R. Ksander (Harvard Medical School, USA) for the metastatic OMM1.3 cell line; and D. Schadendorf (University of Essen, Germany) for the HCmel12, Mamel119, and Skmel28 cells. Funding: E.K., G.K., A.P., and R.S. were supported by the German Research Foundation (DFG)–funded Research Unit FOR2372 with grants KO 1582/10-1 and -2 (to E.K.), KO 902/17-1 and -2 (to G.K.), PF 301/19-1 and -2 (to A.P.), and MO 821/3-1 (to R.S./E.K.). S.M. was supported by DFG grant KO4095/3-1, and P.K. was supported by DFG grant KO4095/4-1. F.H. and J.Y. are members of the DFG-funded Research Training Group GRK1873 (214362475/ GRK1873/2). A.I. was funded by the Japan Agency for Medical Research and Development (AMED; grant JP17gm5910013) and the Japan Society for the Promotion of Science (JSPS) KAKENHI (grant 17 K08264). J.V.-P. was supported by CONACyT (Consejo Nacional de Ciencia y Tecnología, Mexico; grant 286274). Author contributions: E.G., T.T., and E.K. conceived the project and designed the study with support from S.A.; S.A., X.F., N.S., F.E., J.P., T.S., R.S., N.H., U.R., J.Y., E.M.P., and R.D.C.-V. designed and performed experiments. X.F. conducted the xenograft in vivo mouse study and histological analysis with supervision from J.S.G.; S.M. conducted modeling and molecular docking analyses with supervision from P.K.; R.R. and S.K. purified FR under the supervision of G.K.; J.Y. conducted YAP immunofluorescence and Western blotting under the supervision of A.P.; R.D.C.-V. and J.V.-P. designed and performed the thin-layer chromatography experiments, pull-down assays, and Western blotting; S.E. conducted FRET assays under the supervision of M.B.; F.H. provided most valuable input into figure design. All authors contributed to discussion and manuscript editing. E.K. wrote the manuscript with support from S.A. and E.G. and with comments from all authors. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials. The raw data and images associated with this manuscript are available from E.K. and E.G. upon reasonable request.

Stay Connected to Science Signaling

Navigate This Article