Research ArticleG Protein Signaling

“Disruptor” residues in the regulator of G protein signaling (RGS) R12 subfamily attenuate the inactivation of Gα subunits

See allHide authors and affiliations

Science Signaling  12 Jun 2018:
Vol. 11, Issue 534, eaan3677
DOI: 10.1126/scisignal.aan3677

Designing specificity

Once an agonist binds to its G protein–coupled receptor, GTP replaces GDP on the α subunit of an associated G protein, which then dissociates from its βγ dimer to activate effectors. The GTPase activity of the α subunit returns the G protein to an inactive state, a process that is accelerated by interactions with members of the regulator of G protein signaling (RGS) family. Asli et al. used functional and structural analyses to identify specific amino acid residues that distinguished RGS proteins with low activity toward Gαo from those with high activity. High-activity RGS proteins that were mutated to contain these “disruptor” residues exhibited reduced inhibitory activity toward the Gα subunit. Together, these data suggest that “disruptor” residues within RGS proteins may encode specificity toward Gα subunits.

Abstract

Understanding the molecular basis of interaction specificity between RGS (regulator of G protein signaling) proteins and heterotrimeric (αβγ) G proteins would enable the manipulation of RGS-G protein interactions, explore their functions, and effectively target them therapeutically. RGS proteins are classified into four subfamilies (R4, R7, RZ, and R12) and function as negative regulators of G protein signaling by inactivating Gα subunits. We found that the R12 subfamily members RGS10 and RGS14 had lower activity than most R4 subfamily members toward the Gi subfamily member Gαo. Using structure-based energy calculations with multiple Gα-RGS complexes, we identified R12-specific residues in positions that are predicted to determine the divergent activity of this subfamily. This analysis predicted that these residues, which we call “disruptor residues,” interact with the Gα helical domain. We engineered the R12 disruptor residues into the RGS domains of the high-activity R4 subfamily and found that these altered proteins exhibited reduced activity toward Gαo. Reciprocally, replacing the putative disruptor residues in RGS18 (a member of the R4 subfamily that exhibited low activity toward Gαo) with the corresponding residues from a high-activity R4 subfamily RGS protein increased its activity toward Gαo. Furthermore, the high activity of the R4 subfamily toward Gαo was independent of the residues in the homologous positions to the R12 subfamily and RGS18 disruptor residues. Thus, our results suggest that the identified RGS disruptor residues function as negative design elements that attenuate RGS activity for specific Gα proteins.

INTRODUCTION

Heterotrimeric guanine nucleotide–binding proteins (G proteins) play a critical role in countless physiological processes, functioning as molecular switches in intracellular signal transduction pathways (1, 2). The Gα subunit determines the activation state of the G protein switch, cycling between a guanosine diphosphate–bound inactive state and a guanosine 5′-triphosphate (GTP)–bound active state that mediates downstream signaling. The duration of heterotrimeric G protein signaling is controlled by RGS (regulator of G protein signaling) proteins that inactivate Gα subunits (36). This “turn off” function of RGS proteins is achieved by allosterically accelerating GTP hydrolysis in the Gα subunits. This allosteric regulation is mediated by the ~120–amino acid “RGS domain,” which is present in all RGS proteins and is responsible for the function of RGS proteins as guanosine triphosphatase (GTPase)–activating proteins (GAPs) (4, 5, 7). As expected from their key role in G protein–coupled signaling, RGS proteins mediate numerous physiological functions, are involved in a wide range of human pathologies, and are considered promising drug targets (812).

Proteins with RGS domains represent a large and diverse family. Of these, the 20 “classical” or “canonical” RGS proteins can recognize and inactivate Gα subunits that belong to the Gi and Gq subfamilies. Canonical RGS proteins have been further divided into four subfamilies (R4, R7, RZ, and R12) based on their sequence similarity (13). The R4 subfamily is the largest, consisting of 10 members: RGS1, RGS2, RGS3, RGS4, RGS5, RGS8, RGS13, RGS16, RGS18, and RGS21 (8, 12, 13). With the noted exception of RGS2, which has no activity toward Gi subfamily members (1416), members of the R4 subfamily are generally considered to have high GAP activity toward the Gi subfamily, with well-studied examples including RGS1 (17, 18), RGS4 (1820), and RGS16 (18, 2123). The R12 subfamily (RGS10, RGS12, and RGS14) is implicated in a range of physiological processes or pathologies. RGS10 is suggested to be involved in cardiovascular diseases (24), platelet function (25), macrophage activation (26), Parkinson’s disease (27), and chemoresistance in ovarian cancer (28). RGS12 is implicated in neuron and bone differentiation (29, 30), cardiac hypertrophy, and heart failure (31). RGS14 is abundant in the brain and is involved in spatial memory and learning (32, 33), cell division and chemotaxis (34, 35), and Parkinson’s disease (36). Therefore, the R12 subfamily members attract considerable attention both as physiological modulators of signaling and as potential drug targets. However, their cellular targets and the exact role that their GAP function plays in these cascades are unknown.

Although R12 subfamily members were suggested to act as specific GAPs toward the Gi subfamily (37), previous studies reported different relative GAP activities for this subfamily. For example, the RGS10 RGS domain has high activity toward Gαi (5), and its GAP activity is higher toward Gαi and Gαz than is the activity of other RGS proteins, such as RGS19 and RGS4 (38). Compared to high-activity RGS proteins, such as RGS4 and RGS16, RGS12 (39, 40) and RGS14 (18, 41, 42) exhibit lower GAP activity toward Gαi. Other studies measured high GAP activity of RGS14 toward both Gαo and Gαi (43) or reported that RGS14 has a higher GAP activity than that of RGS4 toward Gαo and Gαi (44, 45). Therefore, quantification of the GAP activities of this R12 subfamily in relation to those of R4 family members is lacking, as is an understanding of which amino acids determine R12 GAP activity toward the Gi subfamily.

The structural basis for the GAP activity and the selectivity of RGS domains is of great interest yet is only partially understood. To accelerate the GTPase activity of the Gα subunit, the RGS domain binds to the Gα-GTP complex and allosterically stabilizes the Gα catalytic machinery in its transition state conformation (12, 13, 21, 4649). In all of the structures of Gα-RGS complexes solved to date, the RGS domain binds mainly to the GTPase domain of the Gα subunit (16, 37, 46, 5053). However, several of these structural studies report RGS interactions with another Gα domain, the Gα helical domain (37, 5153). In the G protein superfamily, the Gα helical domain is found only in the Gα subunits of heterotrimeric G proteins, and for many years, the purpose of the Gα helical domain was unclear. Suggested functional roles included increasing the affinity of the G protein for GTP (54), acting as an intrinsic GAP (55), or participating in effector recognition (56). The Gα helical domain is implicated in binding to inhibitory proteins, such as those with the GoLoco motif (5759), and in catalyzing nucleotide exchange and Gα activation by G protein–coupled receptors (60). The previous structural studies of Gα-RGS complexes suggest that RGS contacts with the Gα helical domain are variable and heterogenic, properties consistent with this domain contributing to interaction specificity (37, 51). Structural studies of RGS2 and RGS8 bound to Gαq also indicated that interactions with its Gα helical domain might play a role in dictating affinity or GAP “potency,” the relative GTPase-accelerating activity of RGS proteins (52, 53). Together, these studies raise two interconnected questions. First, are RGS domain interactions with the Gα helical domain functionally important for RGS GAP activity? Second, how might these interactions encode specificity among members of the RGS family?

In a previous study, we used energy calculations to analyze structures of RGS domains with Gα subunits from the Gi subfamily (18), identifying which RGS residues contribute substantially to interactions with these Gα subunits. We further classified these Gα-interacting RGS residues into two groups. The first group contains “Significant & Conserved” (S&C) residues, which make similar and substantial energy contributions across all available structures. These residues make most of the interactions with the residues adjacent to the Gα catalytic site and presumably have a primary role in accelerating Gα GTPase activity by stabilizing Gα in a conformation optimal for GTP hydrolysis (1, 12, 49, 61). The second group contains “Modulatory” residues, which make substantial energy contributions only in some of the structures and are not conserved across all of the RGS domains. Modulatory residues are located at the periphery of the RGS domain interface with Gα subunits. We proposed that modulatory residues encode specific interactions with particular Gα subunits (18). Substitution of such modulatory residues in low-activity RGS proteins with their counterparts from high-activity RGS domains showed that these modulatory residues encode RGS selectivity in the RZ and R4 subfamilies. The hypothesis that emerged from this study is that RGS modulatory positions determine interaction specificity with Gα subunits. Yet, the mechanistic details of this putative role are lacking.

Here, we evaluated the relative GAP activities of R12 and R4 subfamily members toward the representative Gi subfamily member Gαo. We found that, compared to the high-activity R4 subfamily members RGS4 and RGS16, the R12 subfamily members RGS10 and RGS14 had lower GAP activity toward Gαo. Using structure-based computations, we identified R12-specific residues in modulatory positions that may be responsible for these differences in GAP activity. We analyzed these residues at the three-dimensional (3D) structural level and predicted that they interact with the Gα helical domain. We validated our computational predictions through mutagenesis of R4 and R12 RGS domains. Introducing the modulatory residues from R12 subfamily members into high-activity R4 subfamily members converted them into RGS proteins with low GAP activity toward Gαo, whereas replacing the modulatory residues in R12 subfamily members with low GAP activity toward Gαo resulted in higher GAP activity toward this Gα subunit. Together, our results suggest that the R12-specific disruptor residues that we identified function as negative modulatory elements that attenuate RGS GAP activity in a specific fashion.

RESULTS

R12 subfamily members show lower GAP activity toward Gαo compared to that of high-activity RGS domains

We compared the GAP activities of RGS10 and RGS14, which are members of the R12 subfamily, with the activity of RGS16 and RGS4, which are representative members of the R4 subfamily with high activity toward the Gi subfamily (18, 20, 21, 23, 62). We used single-turnover GTPase assays with Gαo and a catalytic concentration of RGS proteins to calculate the catalytic rate (kGAP) of the four RGS proteins according to previously published methodology (48, 63, 64). RGS10 and RGS14 had lower GAP activities toward Gαo (kGAP = 0.5 and 0.3 min−1, respectively), and RGS16 and RGS4 had higher GAP activities (kGAP = 1.3 and 1.3 min−1, respectively) (Fig. 1A). We also measured the GAP potency (52, 65), the concentration of RGS protein that produced a 50% maximal increase in GTP hydrolysis [half-maximal effective concentration (EC50)], of RGS10, RGS14, and RGS16 (Fig. 1B). In this in vitro assay, RGS16 had a higher GAP potency than those of RGS10 and RGS14 (RGS16 EC50 = 7 nM; RGS10 EC50 = 42 nM; RGS14 EC50 = 63 nM). We then turned to energy-based analysis to identify which R12 residues might determine the lower GAP activities of these R12 subfamily members.

Fig. 1 RGS10 and RGS14 have lower GAP activities toward Gαo compared to those of RGS16 and RGS4.

(A) kGAP constants for RGS16, RGS4, RGS10, and RGS14 for GTP hydrolyzed by Gαo (400 nM) in the presence of RGS protein (20 nM). Data are means ± SEM of at least three independent biological replicates. (B) Dose-response analysis of RGS16, RGS10, and RGS14 activity toward Gαo. The EC50 values (RGS16 = 7 ± 1 nM; RGS10 = 42 ± 2 nM; RGS14 = 63 ± 5 nM) were calculated using three-parameter sigmoidal curves. Data are means ± SEM of experiments performed in triplicate and are representative of three or more independent biological replicates for each RGS tested.

Energy calculations predict that R12-specific residues in modulatory positions perturb productive interactions with the Gα helical domain

On the basis of the methodology that we developed previously (18), we can use energy-based calculations to identify which RGS residues contribute substantially to interactions with Gα subunits. We analyzed four x-ray complexes of Gα subunits bound to three different RGS domains from R4 subfamily members with high GAP activity toward the Gi subfamily: RGS1, RGS4, and RGS16 (see Materials and Methods). Two of the complexes are of RGS16, which was solved in complexes with Gαi and Gαo (37, 51). We used the finite-difference Poisson-Boltzmann (FDPB) method to calculate the net electrostatic and polar contributions (ΔΔGelec) of each RGS residue that is within 15 Å of the RGS-Gα interface in these complexes. Nonpolar energy contributions (ΔΔGnp) were calculated as a surface-area proportional term by multiplying the per-residue surface area buried upon complex formation by a surface tension constant of 0.05 kcal/mol per Å2. Our calculations indicated that electrostatic interactions represent most of the interactions between R4 RGS residues and Gαi/o (fig. S1). We classified RGS S&C and modulatory residues (18) and visualized them on the superimposed structures of these four complexes (Fig. 2A). Our analysis showed that RGS S&C residues interact only with the GTPase domain in all four structures. In contrast, we found that RGS modulatory residues interact both with the GTPase domain and with the Gα helical domain. Using the structures of Gαi with RGS16 (Fig. 2B) and RGS4 (Fig. 2C), we identified four modulatory residues in each RGS protein that interacted with the Gα helical domain. In RGS16, the modulatory residues Glu135, Glu164, Lys165, and Lys173 contributed substantially to the interaction with the helical domain of the Gα subunit (Fig. 2B). In RGS4, Glu161, Lys162, Arg166, and Lys170 contributed substantially to the interaction with the Gα helical domain (Fig. 2C). The RGS4 residues Glu161, Lys162, and Lys170 are equivalent to the RGS16 residues Glu164, Lys165, and Lys173, respectively. Although there is no contributing RGS4 residue that corresponds to RGS16 Glu135, Arg166, which is unique to RGS4 among all classical RGS proteins (and corresponds to Pro169 in RGS16), contributed substantially to the interaction with the Gα helical domain.

Fig. 2 RGS S&C residues interact with the Gα GTPase domain and RGS modulatory residues interact with the GTPase and helical domains of Gα.

(A) Superimposition of four crystal structures of Gα-RGS complexes (with PDB codes): RGS1-Gαi1 (PDB ID: 2GTP), RGS4-Gαi1 (PDB ID: 1AGR), human RGS16–Gαi1 (PDB ID: 2IK8), and mouse RGS16–Gαo (PDB ID: 3C7K). RGS domains (cyan) and Gα subunits (green and orange) are shown as backbone traces. Gαi in the RGS4-Gαi1 structure is also visualized as a light gray transparent molecular surface. RGS S&C residues are shown as red spheres, and modulatory residues are shown as purple spheres. (B) Four RGS16 modulatory residues (Glu135, Glu164, Lys165, and Lys173, shown as sticks) interact with the helical domain of the Gα subunit. RGS and Gα subunits are shown as ribbon diagrams colored as in (A), rotated 20° about the y axis relative to (A). (C) Four RGS4 modulatory residues (Glu161, Lys162, Arg166, and Lys170; shown as sticks) interact with the helical domain of the Gα subunit, shown in the same orientation as in (B).

To identify which R12 subfamily residues are responsible for the low GAP activity of these RGS proteins toward Gαo, we mapped the results of our energy analysis onto the sequences of the representative R4 subfamily members and compared them to the sequences of the R12 RGS domains (Fig. 3). Using the available crystal structures of high-activity RGS proteins with Gαi1 and Gαo, we examined which substitutions of R4 residues in the R12 subfamily might hinder interactions with Gα subunits. Our analysis predicted that the contribution of all S&C residues is conserved across these R4 subfamily members (Fig. 3A) and across the three R12 subfamily members (Fig. 3B). Although most of the RGS modulatory residues were predicted to contribute similarly across the R12 subfamily and were thus classified as conserved modulatory residues (Fig. 3B), four modulatory positions were markedly different. Three of these residues, which we termed putative “disruptor” residues, are located in modulatory positions, are specific to the R12 subfamily, and were predicted to perturb the favorable interactions found in the R4 subfamily complexes with Gα subunits. In RGS16, these three modulatory positions (Glu135, Glu164, and Lys165) interact with the Gα helical domain (Fig. 2B). The residues corresponding to RGS16 Glu164 and Lys165 are conserved in RGS4 and RGS1 (Fig. 3A). However, our structural analysis indicated that there is no contribution in RGS4 from Cys132 (Fig. 2C), which is the position that corresponds to RGS16 Glu135. Rather, Arg166, which is unique to RGS4, is predicted to contribute to interactions with the helical domain. We note that an arginine in this position is unique to RGS4 among all RGS family members. The last RGS modulatory position that, according to our calculations, interacts with the Gα helical domain (Lys173 in RGS16 and Lys170 in RGS4) is identical across the R4 and R12 subfamilies (Fig. 3) and is therefore not likely to contribute to their specificity toward Gα subunits.

Fig. 3 R12 RGS residues in critical modulatory positions are predicted to disrupt intermolecular interactions with the Gα helical domain.

(A) Residue-level sequence map summarizing our structure-based energy calculations of the R4 representative complexes with Gα subunits. Colored boxes mark RGS residues that contribute substantially to interactions with Gα subunits, according to the type of energy contribution: nonpolar (np), electrostatic (elec) contributions from the residue side chain (sc elec) or the main chain (mc elec). S&C residues for all four R4 structures are marked with red asterisks above the alignment, and modulatory positions that contribute in any of the four R4 structures are marked with purple triangles. An open star symbol marks modulatory RGS residues that contribute to interactions with the Gα helical domain across all of the analyzed high-activity R4 subfamily members; a black circle marks a contribution found only in some R4 high-activity representatives. The numbering above the alignment is according to the human RGS16 sequence. (B) Sequence-based prediction of R12 subfamily residue-level contributions, classified into three groups: Conserved S&C residues and conserved modulatory residues are residues that are identical across the relevant R4 and R12 subfamilies members; putative disruptor residues are R12-specific residues in modulatory positions that are different than their counterparts in the high-activity R4 subfamily and are predicted to impair the interaction with the Gα helical domain.

Analyzing the structural differences between these R12 residues and the corresponding residues in high-activity R4 subfamily members suggested a mechanistic basis for the lower GAP activity of the R12 subfamily. We superimposed the structures of RGS16-Gαi1 and RGS16-Gαo and visualized the modulatory RGS16 residues Glu135, Glu164, and Lys165 and their interacting partner residues in the Gα subunit (Fig. 4A). In the high-activity RGS16 protein, Lys165 forms an electrostatic and hydrogen bond network with multiple residues on both sides of the interface—an intramolecular salt bridge with RGS16 Glu164, an intermolecular salt bridge with Gαi1 Glu116 (located in the αB-αC loop), and a hydrogen bond with Gαi1 Ser75 (Fig. 4A, top). We observed a similar interaction network in the complexes of Gαi1 with RGS1 and RGS4 (fig. S2). One interaction that we observed in the structural analysis that was not predicted from the energy calculations was the electrostatic contributions of Lys165 to the interactions with Gαi1. Because we calculated the energy difference between the monomers and the complex structures, which correspond to the net electrostatic contributions to affinity in the complex, we interpreted the former result to mean that the Lys165-mediated electrostatic network plays an allosteric role in RGS GAP activity rather than contributes to an increased RGS-Gα affinity. Moreover, the conformation of the αB-αC loop differs between Gαi1 and Gαo. Nevertheless, the corresponding residue to Gαi1 Glu116 in Gαo, Asp116, forms a similar electrostatic interaction with RGS16 Lys165 (Fig. 4A, top).

Fig. 4 Putative disruptor residues of RGS10, RGS14, and RGS18 are predicted to perturb favorable interactions with the helical domain of Gαi/o.

(A) RGS16 residues that contribute to favorable electrostatic interactions with the Gα helical domain, shown as sticks. Electrostatic interactions or hydrogen bonds are marked with dashed lines. The complexes of human RGS16–Gαi1 and mouse RGS16–Gαo are superimposed and shown as ribbon diagrams: RGS16 (cyan) and Gα GTPase domain (green; bottom) and helical domain (orange; top and bottom). (B) Corresponding RGS10 and RGS14 putative disruptor residues, shown in stick form as in (A). RGS10 (yellow) and RGS14 (tan) were superimposed onto the complexes of human RGS16–Gαi1 and mouse RGS16–Gαo, with Gα colored as described in (A). (C) Electrostatic interaction between Arg166, a unique modulatory residue in RGS4 (blue), and the Gαi1 helical domain. (D) RGS10 (yellow) and RGS14 (tan) residues corresponding to RGS4-Arg166, shown with the Gαi1 helical domain residue that interacts with RGS4 Arg166. (E) RGS18 (pink) putative disruptor residues (sticks), shown with the helical domain of Gαi/o with Gα colored as in (A). RGS18 was superimposed onto RGS16-Gαi1 and RGS16-Gαo, as described in (B).

Modeling the corresponding residues in RGS10 and RGS14 suggested that they perturb interactions with the Gα helical domain. We superimposed the structures of RGS10 and RGS14 on the structures of the RGS16-Gαi1 and RGS16-Gαo complexes (Fig. 4B, top). This revealed that the KY (RGS10) and KF (RGS14) motifs cannot form the electrostatic and hydrogen-bond network found in the Gα complexes with high-activity RGS domains (Fig. 4B, top; see also Fig. 4A, top, and fig. S2). The third RGS16 modulatory residue, Glu135, forms a salt bridge with Gα Arg90 (Fig. 4A, bottom). The corresponding residues in RGS10 (Gln103) and RGS14 (Gln86) cannot form such an intermolecular salt bridge (Fig. 4B, bottom). In Gαo, the basic residue (Arg86) that interacts with the RGS16 Glu135 originates one helix turn earlier than does Arg90 in Gαi1. However, the guanidino groups of both basic Gα residues occupy essentially the same position and interact similarly with RGS16 Glu135. The corresponding R12 glutamine residue is predicted to disrupt this interaction with both Gαi1 and Gαo (Fig. 4B, bottom). The unique Arg166 residue in RGS4 forms a salt bridge with Gαi1 Glu116 (Fig. 4C). The corresponding RGS10 Ser136 and RGS14 Ala119 residues cannot form such an intermolecular salt bridge (Fig. 4D). Together, these data suggest that the three putative disruptor residues in either RGS10 or RGS14 are predicted to have a similar negative effect on the interactions of these RGS domains with both Gαo and Gαi1.

Looking across the R4 subfamily, we identified RGS18 as a unique member with putative disruptor residues in the same positions as the R12 putative disruptor residues. In the position that corresponds to RGS16 Lys165 (Fig. 4A), RGS18 has a glutamine (Gln186). Similar to the R12 KY and KF motifs, Gln186 is predicted to perturb the electrostatic network with the Gα helical domain residues (Fig. 4E, left; see also Fig. 4B, top). In the position that corresponds to RGS16 Glu135 (Fig. 4A, bottom), RGS18 has a histidine (His156) that, similar to the R12 residues in this position, cannot form an intermolecular salt bridge with the Gα helical domain residues (Fig. 4E, right). These results suggest that RGS18 contains the same putative disruptor elements as those of the R12 subfamily.

Inserting R12 putative disruptor residues into high-activity R4 subfamily members impairs their GAP activity

To examine whether the R12 putative disruptor residues that we identified are responsible for the lower GAP activities of RGS10 and RGS14, we inserted them into high-activity R4 subfamily members. We replaced all three or pairs of the corresponding RGS16 residues (Glu135, Glu164, and Lys165) with their counterparts in RGS10 (Gln103, Lys131, and Tyr132) or RGS14 (Gln86, Lys114, and Phe115) and measured the GAP activities of these mutants using single-turnover GTPase assays. On the basis of our structural comparison (Fig. 4, A and B), we hypothesized that substituting the RGS16 Glu164-Lys165 motif would have a more substantial effect on GAP activity than would substituting RGS16 Glu135 with the corresponding R12 glutamine motif. Substituting the RGS16 Glu164-Lys165 motif with the RGS10 putative disruptor residues [to generate the RGS16 E164K-K165Y (EK>KY) mutant] reduced its GAP activity (Fig. 5A). The additional E135Q substitution (to generate the EEK>QKY mutant) did not further reduce GAP activity. When we substituted the RGS16 E164-K165 motif with the corresponding putative disruptor residues from RGS14, the RGS16 E164K-K165F (EK>KF) mutant exhibited a larger impairment in GAP activity than did the EK>KY mutant (Fig. 5A). The larger effect of the RGS14 disruptor residues, compared to that of the RGS10 residues, recapitulated the lower GAP activity of RGS14 compared to that of RGS10 (Fig. 1). In contrast with the EEK>QKY mutant, adding the E135Q mutation to the E164K-K165F mutant (EEK>QKF) further reduced GAP activity. As expected from the similarity of the disruptor motifs in the structures of RGS10, RGS14, and RGS18 (Fig. 4, B and E), substituting RGS16 Glu135 and Lys165 with their RGS18 putative disruptor counterparts (to generate the RGS16 mutant E135H-K165Q, E-K>H-Q) also impaired GAP activity (Fig. 5A). In contrast, mutation of adjacent residues that were not predicted to contribute as much to the interaction between RGS16 and Gα, such as RGS16 residues (Tyr168 and Pro169 to Ala), did not impair GAP activity. Tyr168 is conserved across most RGS domains and is present in all high-activity R4 subfamily members and in all R12 subfamily members. Pro169 is conserved across all high-activity R4 subfamily members, with the exception of RGS4, in which the corresponding Arg166 is a unique contributing modulatory residue (Fig. 4C). In RGS14, the corresponding residue is an alanine (Fig. 4D). As expected from our energy calculations indicating that these two conserved residues do not contribute to interactions with Gα subunits (Fig. 3), the Y168A and the P169A mutants exhibited similar GAP activities to that of wild-type (WT) RGS16 (Fig. 5A).

Fig. 5 Replacement of RGS16 and RGS4 modulatory residues that interact with the Gα helical domain with the corresponding putative disruptor residues impairs GAP activity toward Gαo.

(A) kGAP constants for WT RGS16 and the following RGS16 mutants: RGS16-to-RGS10 mutants E164K-K165Y (EK>KY) and E135Q-E164K-K165Y (EEK>QKY); RGS16-to-RGS14 mutants E164K-K165F (EK>KF) and E135Q-E164K-K165F (EEK>QKF); the RGS16-to-RGS18 mutant E135H-K165Q (E-K>H-Q); and the RGS16 mutants in noncontributing residues adjacent to Glu164 and Lys165, Y168A and P169A. The kGAP values are means ± SEM of at least three independent biological replicates. (B) Dose-response analysis of WT RGS16, the EEK>QKY mutant, and the EEK>QKF mutant GAP activity toward Gαo. EC50 values (RGS16 = 7 ± 1 nM; EEK>QKY = 26 ± 2 nM; EEK>QKF = 74 ± 3 nM) are means ± SEM of experiments performed in triplicate and are representative of at least three independent biological replicates each. (C) kGAP constants for WT RGS4 and the following RGS4 mutants: RGS4-to-RGS14 mutants E161K-K162F and E161K-K162F-R166A and the RGS4 S&C residue mutant N128S. kGAP values are means ± SEM of at least three independent biological replicates.

In a reciprocal experiment, we replaced the RGS10 or the RGS14 putative disruptor residues with their RGS16 counterparts, which reduced GAP activity (fig. S3). The R12 subfamily has an adjacent heterogeneous structural region in the extended αV-to-αVI loop (37). We predict that this region will likely affect the local conformation of the R12 subfamily members. Therefore, we hypothesized that this heterogeneity does not enable substituted residues to reach the exact conformation seen in RGS16 or RGS4 (Fig. 4A and fig. S2). Thus, substituting only the two or three modulatory residues would be insufficient to convert the RGS10 and RGS14 proteins into high-activity GAPs. However, an RGS14 chimera in which we substituted all RGS14 residues that can affect regions αV to αVI and αVII, which contain the three RGS14 disruptor residues Gln86, Lys114, and Phe115, with their RGS16 counterparts, resulted in a protein with increased GAP activity compared to that of WT RGS14 (fig. S3).

Dose-response analysis of the GAP potencies of the RGS16 EEK>QKY and EEK>QKF mutants indicated that the GAP potency of these mutants was reduced (Fig. 5B). The EC50 of the EEK>QKY mutant was 26 nM, and the EC50 of the EEK>QKY mutant was 73 nM, whereas the EC50 of WT RGS16 was 7 nM (Fig. 5B). Therefore, these RGS16 mutants had a comparably reduced GAP potency, which was similar to that of RGS10 and RGS14. Similar to the RGS16 mutants, replacing the putative R12 subfamily disruptor residues in the corresponding positions of RGS4 reduced GAP activity (Fig. 5C). Substituting the RGS4 Glu161-Lys162 motif, which corresponds to RGS16 Glu164-Lys165, with the corresponding RGS14 putative disruptor residues (to generate RGS4 E161K-K162F, the EK>KF mutant) substantially reduced GAP activity. Adding a substitution of the unique RGS4 Arg166 with its RGS14 counterpart to generate the triple mutant, RGS4 E161K-K162F-R166A (EKR>KFA), further reduced GAP activity. The kGAP of this RGS4 triple mutant (0.2 min−1) is similar to the kGAP that we measured for WT RGS14 (0.3 min−1; Fig. 1A). In comparison, mutating the RGS4 S&C residue Asn128, which is essential for RGS catalytic activity (18, 21, 48), completely abolished RGS4 GAP activity (Fig. 5C).

To test whether RGS16 residues Glu135, Glu164, Lys165, and Lys173, which are the modulatory positions of RGS16 that interact with the Gα helical domain (Fig. 2B), are necessary for its high GAP activity, we mutated all four residues to alanines. The resulting RGS16-Ala4 mutant had similar GAP activity to that of WT RGS16 (Fig. 6), suggesting that these residues are not essential for the high GAP activity of RGS16. This result indicates that alanines in these positions are permissible for high GAP activity; however, as we showed earlier, substituting the disruptor residues from low-activity RGS proteins, such as R12 family members, was not permissible for high GAP activity.

Fig. 6 Mutating all four RGS16 modulatory positions that interact with the Gα helical domain into alanines does not reduce the GAP activity of RGS16 toward Gαo.

Representative single-turnover GTPase assays of Gαo (400 nM) with RGS16 (20 nM; black circles and solid line) and the RGS16-Ala4 mutant E135A-E164A-K165A-K173A (20 nM; black triangles and dashed line). The reaction rate constant (k) for RGS16 was 1.5 ± 0.3 min−1, and that for the RGS16-Ala4 mutant was 1.6 ± 0.2 min−1. Data are means ± SEM of at least three independent biological replicates.

Replacing the putative RGS18 disruptor residues with their RGS16 counterparts increases GAP activity

To test whether substituting disruptor residues in the low-activity protein RGS18 with their high-activity RGS counterparts removes the negative effect on GAP activity, we replaced the RGS18 disruptor residues with their RGS16 counterparts (Fig. 7). The RGS18 Q186K mutant had increased GAP activity toward Gαo. The RGS18 H156E-Q186K double mutant had further increased GAP activity, which was comparable to that of WT RGS16. Therefore, these data suggest that replacing these RGS18 disruptor residues with their RGS16 counterparts is sufficient for a complete gain of function.

Fig. 7 Replacement of RGS18 disruptor residues with the corresponding RGS16 modulatory residues led to a gain of function.

The kGAP constants for WT RGS18, the RGS18-to-RGS16 mutants Q186K and H156E-Q186K, and WT RGS16 were calculated as in described in Fig. 1. Data are means ± SEM of at least three independent biological replicates.

DISCUSSION

The broad challenge of deciphering protein-protein interaction specificity is particularly relevant to the interactions of RGS proteins with Gα subunits. Because numerous RGS proteins are usually coexpressed in a given cell, identifying the molecular design principles that determine selective recognition of Gα subunits by RGS proteins is essential for understanding which RGS proteins mediate particular physiological functions and for manipulating these interactions with drugs. However, which RGS-Gα interactions occur and which residues encode the specificity of these interactions are mostly unclear, which is partially due to the low sequence identity among RGS domains (as low as 30%). Such low sequence identity increases the difficulty of pinpointing which residues contribute to similar interactions and which residues determine interaction selectivity. However, quantitative structure-based approaches, such as the approach we used here, can pinpoint which residues contribute substantially to interactions across a protein family and guide mutagenesis to redesign RGS interactions with Gα subunits.

Looking beyond the RGS family, specific protein-protein interactions between families of signaling proteins are crucial for the wiring of signaling networks. Useful terms to define structural elements that determine such specific interactions were coined by the protein design field: “Positive design elements” stabilize favorable interactions that strengthen particular protein pairings, whereas “Negative design elements” introduce unfavorable interactions that limit selected interactions between some family members (66, 67). In particular, negative design elements are critical specificity determinants among well-studied examples in protein-protein interactions, such as heterodimeric coiled-coil pairs (68, 69), colicin-immunity protein interactions (70, 71), β-lactamase and its protein inhibitors (72), and Bcl-2 receptors binding to BH3-only proteins (73).

Previous studies identified RGS residues that are crucial positive design elements for RGS GAP function (18, 21, 47, 48, 51, 74). We classified most of these previously studied RGS residues as S&C residues (Fig. 2A). These RGS residues interact with Gα residues adjacent to its catalytic site and presumably have a primary role in accelerating the GTPase activity of a Gα subunit by stabilizing a conformation that is optimal for GTP hydrolysis (1, 12, 21, 48, 49, 74). However, to achieve high GAP activity, RGS domains also require sufficient modulatory residues that function as positive design elements. We had also suggested that modulatory residues encode specific interactions with particular Gα subunits (18).

Here, we found that R12 subfamily “disruptor residues” function as negative design elements by means of interacting with the Gα helical domain. Our results demonstrate that the R12 subfamily members RGS10 and RGS14 have reduced GAP activity toward Gαo compared to the high-activity RGS proteins, such as RGS4 and RGS16. Our structure-based analysis identified three R12 positions as critical for this phenotype. Mutating the corresponding RGS16 and RGS4 modulatory residues to their R12 counterparts reduced the GAP activities of RGS16 and RGS4 to those of RGS10 and RGS14. Because the S&C and modulatory residues of RGS12 are essentially identical to those of RGS14, and in particular, because their putative disruptor residues are identical, we suggest that these RGS domains will have similar GAP activity and specificity and that our results may be applicable to the entire R12 subfamily. Moreover, because mutating these residues in RGS16 to alanines did not reduce GAP activity, we conclude that these positions are not positive design elements in RGS16 but rather play a specific role as negative design elements in the R12 subfamily. Finally, we showed that replacing these negative design elements in RGS18 with their counterparts from RGS16 led to a complete gain of function, further supporting our conclusions that these negative design elements attenuate GAP activity through interactions with the Gα helical domain.

More generally, our energy calculations and structural analysis provide insights into how RGS domains encode specificity toward Gα subunits. Although it was known that RGS interactions with the Gα GTPase domain are central to RGS GAP activity (1, 21, 48, 49) because the GTPase domain contains essentially all of the GTP-binding site, the functional roles of RGS interactions with the Gα helical domain were not clear. Our results show how contacts with the Gα helical domain can govern RGS-Gα interactions; a small number of key disruptor residues that interact with the Gα helical domain can fine-tune RGS GAP activity and determine specificity. Nevertheless, the interaction of these RGS modulatory residues with the Gα helical domain is not essential for high GAP activity, showing that these positions function as negative design elements only. Finally, the approach we used here can be leveraged to uncover additional positive and negative design elements across the larger RGS family and provide tools to rewire RGS interaction in cells and in vivo.

MATERIALS AND METHODS

Protein structures and sequences

We used the following 3D structures in our analysis and visualization of RGS-Gα complexes [with Protein Data Bank (PDB) codes for each structure]: rat RGS4–Gαi1 (PDB ID: 1AGR) (46), human RGS16 (hRGS16)–Gαi1 (PDB ID: 2IK8) (37), mouse RGS16 (mRGS16)–Gαo (PDB ID: 3C7K) (51), human RGS1–Gαi1 (PDB ID: 2GTP) (37), human RGS10–Gαi3 (PDB ID: 2IHB) (37), and the monomeric nuclear magnetic resonance (NMR) structures of human RGS14 (PDB ID: 2JNU) and RGS18 (PDB ID: 2JM5) (37). For visualization of the latter structures, we used the first model of the 20 NMR models in the PDB. 3D structural visualization and superimposition were performed with the molecular graphics program PyMOL (http://pymol.org). We used the following RGS domain sequences from the UniProt database (www.uniprot.org/): O43665 (RGS10), O14924 (RGS12), and O43566 (RGS14). Sequences were aligned using the T-Coffee server (http://tcoffee.vital-it.ch/apps/tcoffee/index.html), followed by manual adjustments in Jalview (www.jalview.org) according to our 3D structural comparisons.

Energy calculations to identify RGS residues that contribute substantially to the RGS-Gα interactions

We followed a methodology described previously (18) to analyze the per-residue contributions of RGS residues to their Gα partner for the following crystal structures of RGS-Gα complexes (with PDB codes): human RGS16 (hRGS16)–Gαi1 (PDB ID: 2IK8), mouse RGS16 (mRGS16)–Gαo (PDB ID: 3C7K), RGS4-Gαi1 (PDB ID: 1AGR), and RGS1-Gαi1 (PDB ID: 2GTP). We used the FDPB method to calculate the ΔΔGelec of each residue that is within 15 Å of the dimer interface. Residues that contributed substantially to the interaction were defined as those contributing ΔΔGelec ≥ 1 kcal/mol to the interactions (twice the numerical error of the electrostatic calculations) (75). ΔΔGnp were calculated as a surface-area proportional term by multiplying the per-residue surface area buried upon complex formation using surfv (76) by a surface tension constant of 0.05 kcal/mol per Å2 (75). Residues that contributed substantially to binding were defined as those contributing ΔΔGnp ≥ 0.5 kcal/mol to the interactions (namely, bury more than 10 Å2 of each protein surface upon complex formation).

Protein expression, purification, and mutagenesis

The RGS4, RGS10, RGS14, RGS16, and RGS18 domains were expressed in the pLIC-SGC1 vector as N-terminally His6-tagged fusion proteins (Addgene). The N-terminally His6-tagged rat Gαo clone was a gift from V. Arshavsky (Duke University). RGS16 mutants were generated with the QuikChange site-directed mutagenesis kit (Invitrogen) with primers designed using the Primer Design Program (www.genomics.agilent.com). An RGS14-to-RGS16 chimera was also generated with the QuikChange site-directed mutagenesis kit by inserting the mutations T29E, E30A, K34T, A38E, R103D, Q105A, A119P, and V122L and by replacing regions Gln76 to Pro80, Arg85 to Asp101, and Leu107 to Asn111 with their RGS16 counterparts. Proteins were expressed in Escherichia coli BL21 (DE3) cells and grown in 0.5 or 1 liter of LB broth at 37°C for RGS or Gα proteins, respectively, until OD600nm (optical density at 600 nm) ≥ 1.4 was reached. The temperature was then reduced to 15°C, and protein expression was induced by addition of 0.5 mM or 100 μM isopropyl-d-thiogalactopyranoside for RGS or Gα proteins, respectively. After 16 to 18 hours, cells were harvested by centrifugation at 6000g for 30 min at 4°C, which was followed by freezing the pellets at −80°C. Bacterial pellets were suspended in lysis buffer [50 mM tris (pH 8.0), 50 mM NaCl, 5 mM MgCl2, 5 mM β-mercaptoethanol, protease inhibitor cocktail (Roche), and 0.5 mM phenylmethylsulfonyl fluoride for G proteins only], and the cells were lysed with a Sonics Vibra-Cell sonicator, which was followed by centrifugation at 24,000g for 30 min at 4°C. The supernatants were equilibrated to 500 mM NaCl and 20 mM imidazole and loaded onto HisTrap FF 1-ml columns (GE Healthcare Life Sciences). The columns were washed with >20 volumes of wash buffer [20 mM tris (pH 8.0), 500 mM NaCl, and 20 mM imidazole] at 4°C, and the tagged proteins were eluted with elution buffer [20 mM tris (pH 8.0), 500 mM NaCl, and 100 mM imidazole (pH 8.0)]. The elute was loaded onto a HiLoad 16/600 Superdex 75 PG gel filtration column (GE Healthcare Life Sciences) at 4°C with ≥1.5 volumes of GF elution buffer [50 mM tris (pH 8.0), 50 mM NaCl, 5 mM β-mercaptoethanol, and 1 mM MgCl2 added for Gα subunits only]. The eluate was dialyzed against a dialysis buffer [GF elution buffer with 40% (v/v) glycerol]. All purified proteins were estimated to be >95% pure, as assessed by SDS–polyacrylamide gel electrophoresis and Coomassie staining. Protein concentrations were determined by measuring their absorption at 280 nm using the predicted extinction coefficients (ProtParam, Swiss Institute for Bioinformatics) of the sequence of each expressed protein.

Single-turnover GTPase assays

Single-turnover GTPase assays using Gαo and the various RGS proteins were performed, as detailed in previous studies (18, 48, 63, 64). Briefly, Gαo in reaction buffer [50 mM Hepes (pH 7.5), 0.05% (v/v) polyoxyethylene, 5 mM EDTA, bovine serum albumin (5 mg/ml), and 1 mM dithiothreitol] was incubated for 15 min at 20°C with 1 mM [γ-32P]GTP and cooled on ice for 5 min. GTP hydrolysis was initiated by increasing the magnesium concentration to 5 mM (with MgCl2), together with 100 μM cold GTP (final concentration) with or without RGS proteins at 4°C. Aliquots were taken at different time points and were quenched with 5% charcoal in 50 mM Na2H2PO4 (pH 3) on ice, which was followed by centrifugation at 12,000g for 5 min at room temperature. The supernatant (200 μl) was transferred to 3 ml of liquid scintillation liquid (PerkinElmer) and analyzed with a Tri-Carb 2810 TR scintillation counter (PerkinElmer). GTPase rates were determined from single-exponential fits to the time courses using SigmaPlot 10.0. We calculated kGAP rate constants by subtracting the basal GTPase rate (without RGS protein) from the GTPase rate that was measured in the presence of the RGS protein, as described previously (64).

RGS dose-response analysis

RGS dose-response analysis was performed as in previous studies (52, 65). Gαo was loaded with 1 mM [γ-32P]GTP for 15 min at 20°C in reaction buffer and then cooled on ice for 5 min. Each assay was initiated by adding 10 μl of RGS protein in different concentrations in assay buffer (5 mM MgCl2 and 100 μM cold GTP) to a tube containing 20 μl of Gαo subunit (500 nM) on ice. Each reaction was terminated after 45 s by adding 100 μl of 5% perchloric acid and quenching with 700 μl of 10% (w/v) charcoal slurry in 50 mM phosphate buffer (pH 7.5), which was followed by centrifugation at 12,000g for 5 min at room temperature. The supernatant (200 μl) was transferred to 3 ml of liquid scintillation and analyzed with a Tri-Carb 2810 TR scintillation counter (PerkinElmer). EC50 values were determined from three-parameter sigmoidal curves with SigmaPlot 10.0.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/11/534/eaan3677/DC1

Fig. S1. The RGS interface with Gα subunits is predominantly electrostatic and polar.

Fig. S2. A pair of RGS1 and RGS4 residues contributes favorably to electrostatic interactions with the Gα helical domain.

Fig. S3. A chimera of RGS14 and RGS16 exhibits increased GAP activity.

REFERENCES AND NOTES

Acknowledgments: We thank L. Barki-Harrington, E. Privman, and E. D. Levy for helpful comments and V. Arshavsky for the Gαo clone. Funding: This work was supported by grants from the Israel Science Foundation (grant numbers 1454/13, 1959/13, and 2155/15), the Israel Ministry of Science, Technology and Space, Israel, and the Italian Ministry of Foreign Affairs (3-10704). The authors acknowledge the contribution of COST Action CM-1207 (GLISTEN) to this work. Author contributions: A.A. designed the research, conducted most of the experiments and structural analysis, analyzed the results, and wrote the manuscript. I.S. conducted structural analysis and some of the experiments and analyzed the results. M.A.-S. conducted experiments, supervised the laboratory work, and analyzed the results. M.K. designed and supervised the research and wrote the manuscript. All authors were involved in the writing of the manuscript and approved the final version. Competing interests: The authors declare that they have no competing interests. Data and materials availability: Plasmids require a material transfer agreement from Addgene. All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article