A nanomolar-potency small molecule inhibitor of regulator of G-protein signaling proteins

Abstract

Regulators of G-protein signaling (RGS) proteins are potent negative modulators of signal transduction through G-protein-coupled receptors. They function by binding to activated (GTP-bound) Gα subunits and accelerating the rate of GTP hydrolysis. Modulation of RGS activity by small molecules is an attractive mechanism for fine-tuning GPCR signaling for therapeutic and research purposes. Here we describe the pharmacologic properties and mechanism of action of CCG-50014, the most potent small molecule RGS inhibitor to date. It has an IC(50) for RGS4 of 30 nM and is >20-fold selective for RGS4 over other RGS proteins. CCG-50014 binds covalently to the RGS, forming an adduct on two cysteine residues located in an allosteric regulatory site. It is not a general cysteine alkylator as it does not inhibit activity of the cysteine protease papain at concentrations >3000-fold higher than those required to inhibit RGS4 function. It is also >1000-fold more potent as an RGS4 inhibitor than are the cysteine alkylators N-ethylmaleimide and iodoacetamide. Analysis of the cysteine reactivity of the compound shows that compound binding to Cys(107) in RGS8 inhibits Gα binding in a manner that can be reversed by cleavage of the compound-RGS disulfide bond. If the compound reacts with Cys(160) in RGS8, the adduct induces RGS denaturation, and activity cannot be restored by removal of the compound. The high potency and good selectivity of CCG-50014 make it a useful tool for studying the functional roles of RGS4.

Figure 1. The chemical structure and RGS inhibitory activity of CCG-50014

A) The chemical structure of CCG-50014 (4-[(4-fluorophenyl)methyl]-2-(4-methylphenyl)-1,2,4-thiadiazolidine-3,5-dione). B) CCG-50014 concentration-dependently inhibits the binding between aluminum fluoride-activated Gα_o and RGS4 or RGS8. Data shown are an average of three independent experiments. This experiment has been independently repeated 28 times, producing average IC₅₀ values of 30±6 nM against RGS4 and 11±2 μM against RGS8. C,D) CCG-50014 also inhibits the GAP activity of RGS4 and E,F) RGS8. Using a single-turnover GAP assay, CCG-50014 concentration-dependently inhibits the GAP activity of both RGS4 and RGS8. * P <0.05, *** P <0.0001. All experiments were independently repeated a minimum of three times.

Figure 2. CCG-50014 thermally destabilizes RGS8 in a concentration-dependent manner, but has no effect on the thermal stability of Gα_o

Representative melting traces of A) RGS8 and B) Gα_o in the absence (solid trace) and presence (dashed trace) of 100 μM CCG-50014. Concentration-response curves showing the thermal destabilization effects of CCG-50014 on C) RGS4, D) RGS8 and E) Gα_o. Data are presented as the mean±SEM of three independent experiments.

Figure 3. CCG-50014 is an irreversible inhibitor of RGS4 and RGS8 and its effects are partially reversed by the thiol reductant DTT

A) RGS4 and B) RGS8 containing beads were treated for 15 minutes with 100 μM CCG-50014 prior to vigorous washing to remove any unbound compound. To determine if the compound was reacting in a thiol-sensitive manner, washing was performed in the absence or presence of 1 mM DTT. Data are presented as the mean±SEM from at least three independent experiments.

Figure 4. CCG-50014 forms a covalent adduct on RGS8

A) WT RGS8 protein (2 μM) was treated with the indicated concentrations CCG-50014 before analysis via LC-MS. After treatment with compound a predominant peak appeared with a mass shift of 317 as compared to the vehicle-treated protein, correlating with a full compound mass adduct (CCG-50014 MW: 316.4). A second minor peak with an additional mass shift of 315 was observed, which correlates to the addition of a second adduct of CCG-50014. B) No adducts are observed on RGS8Cys⁻ mutant, in which both C¹⁰⁷ and C¹⁶⁰ have been mutated to serine.

Figure 5. CCG-50014 requires at least one cysteine residue on RGS8 for full activity

WT or mutant RGS8 was biotinylated, loaded on beads, and AF532-Gα_o binding was measured by FCPIA as described in Materials and Methods. Mutating both cysteines to serine (RGS8cys⁻) produced a protein that was completely insensitive to the effect of CCG-50014. RGS8 mutants with only one cysteine, either cys¹⁰⁷ (107C) or cys¹⁶⁰ (160C), provided sensitivity to CCG-50014. The inhibition parameters (IC₅₀ (μM), Hill Coefficient) for CCG-50014 on these proteins were as follows: wildtype RGS8 (wt): 6.1 μM, −0.79; 107C: 46.5 μM, −0.54; 160C: 0.71 μM, −0.36; RGS8cys⁻: >100 μM. Data are presented as the mean±SEM of three independent experiments.

Figure 6. CCG-50014 induced protein aggregation is dependent on the presence of 160C

A,B) Wild type, C,D) 107C, or E) 160C RGS8 was treated with a 5-fold excess of CCG-50014 before removal of the compound via gel filtration on a 20 mL S75 superdex column. A,C,E) Shown are representative UV chromatogram traces. B,D) Protein recovered from the peak was tested for Gα_o binding in an FCPIA competition assay with AF532-Ga_o binding to WT RGS8. The wild-type RGS8 chromatogram shows a slightly left shifted and suppressed peak after CCG-50014 treatment, which coincides with a 14-fold decrease in Gα_o binding. The 107C mutant protein showed no CCG-50014-induced change in migration on gel filtration and any inhibition of Gα_o binding activity was reversed by the gel filtration procedure. The 160C mutant protein completely (and visually) aggregates upon compound treatment and is removed by the prefiltration of the samples.

Figure 7. Irreversible inhibition of RGS8 is predominantly mediated by Cys¹⁶⁰

Mutant proteins A) 107C RGS8 and B) 160C RGS8 were pre-bound to beads and exposed to 20 μM CCG-50014, after which reversibility experiments were performed. C) Development of irreversible inhibition after exposure to CCG-50014 differs between the individual cysteine mutants and provides a means to understand the compound's mechanism of action. Wild-type, 160C or 107C RGS8 was treated with 20 μM CCG-50014 for the indicated amount of time before compound removal by extensive washing. The amount of irreversible inhibition was quantified by comparing the G-protein binding to CCG-50014 treated beads to DMSO treated beads. The total amount of inhibition (without a wash step) at this concentration of CCG-50014 for each protein was as follows: WT RGS8: 64±2%; 107C RGS8: 45±2%; and 160C RGS8: 56±1%. Data are presented as the mean±SEM from three independent experiments. ***P<0.0001 using an unpaired t test.

Figure 8. CCG-50014 does not inhibit the cysteine protease, papain

A) Papain (0.625 U) was mixed with self-quenching FITC-conjugated casein and the liberated fluorescence that results from casein-dependent proteolysis was observed as a function of time in the presence of different cysteine alkylators. Iodoacetamide (100 uM) markedly inhibits casein proteolysis by papain while 100 μM CCG-50014 has no effect. B) Concentration dependence of the effect of compounds on casein proteolysis (5 min). Data are presented as the mean±SEM from three independent experiments.

Figure 9. CCG-50014 is a much more potent RGS inhibitor than two general cysteine alkylators N-ethyl maleimide (NEM) and iodoacetamide (IA)

Dose response curves for NEM, IA, and CCG-50014 against A) RGS4 and B) RGS8. The only protein that displayed any sensitivity to the alkylators tested was RGS4, which was inhibited by NEM with an IC₅₀ value >3.5 Log higher than that of CCG-50014. Data are presented as the mean±SEM from three independent experiments.

Figure 10. Hypothesized binding site of CCG-50014 on RGS8

A) CCG-50014 was docked to RSG8 using Autodock software (ver 4.0) as described in Materials and Methods and one docking site showed the greatest predicted affinity (18 μM Ki). This site is far from the Gα binding interaction interface and is near the RGS4 which is important for RGS regulation by calmodulin and acidic phospholipids. Conserved residues between RGS4 and RGS8 that directly contact Gαi in the RGS4-Gαi1 structure (PDB 1AGR (24)) are shown in red. B) Assuming a static protein, this binding site places the compound close to the two cysteine residues in RGS8, but not within a distance compatible with direct covalent reaction. A conformational change must occur in the RGS to allow compound intercalation into the helix bundle. Distances are shown in angstroms. RGS8 structure from 2IHD.

Figure 11. CCG 50014 inhibits the Gα_o dependent membrane localization of RGS4

A) When overexpressed in HEK293T cells, GFP-RGS4 is localized to the cytosol. B) Co-expression with Gα_o induces a subcellular translocation of the GFP-RGS4 to the plasma membrane. C) This translocation does not occur in response to co-expression with the RGS-insensitive Gα_o mutant (G184S). D/E) Cells expressing Gαo and GFP-RGS4 show no change in the plasma membrane localization of the RGS when treated with vehicle (0.1% DMSO), however treatment with F/G) CCG-50014 (100μM) rapidly induces a loss of the plasma membrane localization of the RGS without diminishing the overall signal. Representative line scans across a cell treated with H) vehicle and I) CCG-50014 also show this effect. J) Quantification of this effect from 10 cells treated with vehicle or CCG-50014 (100 μM) shows a significant decrease in the amount of RGS-GFP located at the membrane after compound addition. A trend towards increased cytosolic localization after compound treatment is also observed, suggesting that the treatment doesn't just diminish the GFP signal. ****p<0.0001, *p<0.05.