Molecular chaperoning function of Ric-8 is to fold nascent heterotrimeric G protein α subunits

Abstract

We have shown that resistance to inhibitors of cholinesterase 8 (Ric-8) proteins regulate an early step of heterotrimeric G protein α (Gα) subunit biosynthesis. Here, mammalian and plant cell-free translation systems were used to study Ric-8A action during Gα subunit translation and protein folding. Gα translation rates and overall produced protein amounts were equivalent in mock and Ric-8A–immunodepleted rabbit reticulocyte lysate (RRL). GDP-AlF₄⁻–bound Gαi, Gαq, Gα13, and Gαs produced in mock-depleted RRL had characteristic resistance to limited trypsinolysis, showing that these G proteins were folded properly. Gαi, Gαq, and Gα13, but not Gαs produced from Ric-8A–depleted RRL were not protected from trypsinization and therefore not folded correctly. Addition of recombinant Ric-8A to the Ric-8A–depleted RRL enhanced GDP-AlF₄⁻–bound Gα subunit trypsin protection. Dramatic results were obtained in wheat germ extract (WGE) that has no endogenous Ric-8 component. WGE-translated Gαq was gel filtered and found to be an aggregate. Ric-8A supplementation of WGE allowed production of Gαq that gel filtered as a ∼100 kDa Ric-8A:Gαq heterodimer. Addition of GTPγS to Ric-8A–supplemented WGE Gαq translation resulted in dissociation of the Ric-8A:Gαq heterodimer and production of functional Gαq-GTPγS monomer. Excess Gβγ supplementation of WGE did not support functional Gαq production. The molecular chaperoning function of Ric-8 is to participate in the folding of nascent G protein α subunits.

The proper biosynthesis and attainment of plasma membrane residence is crucial for heterotrimeric G proteins to serve as transducers of the signals received from G protein–coupled receptors. G-protein biosynthesis begins with the translation of the Gα, Gβ, and Gγ subunits by free ribosomes. Subunit folding ensues in the cytosol. The G-protein heterotrimer is assembled on the outer leaflet of an endomembrane (endoplasmic reticulum and/or Golgi) and trafficked to the plasma membrane (1, 2). Lipidation of Gα and Gγ precedes trafficking (2–4). It is quite clear that nascent Gβ subunits are folded by the concerted action of chaperonin-containing t-complex polypeptide 1 (CCT/TriC/TCP-1) and phosducin-like protein-1 (PhLP-1) (5–8). Gγ subunit folding is assisted by dopamine receptor–interacting protein 78 (DRIP78) (9). Evidence in support of chaperone-mediated Gα subunit folding exists, although the details are less clear. CCT was suggested to fold Gαt in a manner similar to its action toward actin and tubulin (10). Heat shock cognate chaperones are important for functional expression of several different Gα subunits, including Gα12 and Gαo (11–13). We recently demonstrated that resistance to inhibitors of cholinesterase 8 (Ric-8) proteins mediate an early event during Gα subunit biosynthesis (14). Ric-8 was discovered in Caenorhabditis elegans and implicated to genetically interact with various Gα subunits (15–18). Mammalian Ric-8 proteins were then defined as Gα subunit guanine nucleotide exchange factors (GEFs) (19, 20). Ric-8A and Ric-8B collectively stimulate nucleotide exchange of all Gα subunit classes by stabilizing the Gα nucleotide-free transition state. Ric-8A acts upon Gαi/q/13 and Ric-8B is a GEF for Gαs/αolf.

Several lines of evidence have shown Ric-8 positive influence of the cellular abundances of G proteins. Genetic ablation or RNAi-knockdown of Ric-8 in model organisms and in mammalian cultured cells reduced Gα steady-state abundances and levels at the plasma membrane (14, 21–25). Overexpression of Ric-8 proteins in HEK293, NIH 3T3, or Sf9 cells enhanced cellular Gα levels (26–28). In Ric-8A– or Ric-8B–deleted ES cells, newly produced Gα subunits were defective in the initial association with an internal membrane. These G proteins were turned over ∼10-fold faster than in WT cells (14).

This collective evidence suggests that the early biosynthetic event mediated by Ric-8 proteins is either to aid in Gα subunit folding or to direct folded, cytosolic Gα subunits to an endomembrane. Here, we have used cell-free translation systems to ascertain the role of Ric-8 in Gα subunit translation and folding. Ric-8A and Ric-8B were identified as endogenous components of rabbit reticulocyte lysate (RRL). Gα subunits were produced efficiently in Ric-8A–immunodepleted RRL. Use of an established Gα subunit trypsin protection assay showed that Gα subunits were folded improperly in Ric-8A–depleted RRL. Reconstitution of Ric-8A–depleted RRL with Ric-8A protein before or after Gα translation restored proper Gα folding. Striking results were obtained using wheat germ extract (WGE), which has no Ric-8 component. WGE-produced Gαq was aggregated. Ric-8A, but not Gβγ, supplementation of WGE enabled production of functionally folded Gαq. We conclude that Ric-8 directly participates in the posttranslational folding of nascent Gα subunits to promote G-protein biosynthesis.

Results

Ric-8 Proteins Are Endogenous RRL Components That Do Not Affect Gα Subunit Translation.

Ric-8 proteins mediate an early event of heterotrimeric Gα subunit biosynthesis in cells (14). We propose that Ric-8 could act upon nascent Gα subunits in three nonexclusive capacities. Ric-8 could (i) promote Gα translation, (ii) act with cochaperones to fold nascent Gα subunits, or (iii) mediate endomembrane association of newly folded, soluble Gα subunits.

To discriminate among these possibilities, we used the RRL and WGE cell-free translation systems to examine Ric-8–dependent Gα subunit translation and folding. Ric-8A and Ric-8BFL (full-length) were first identified as endogenous RRL components by immunoblotting and immunoprecipitation (Fig. 1A). The Ric-8BΔ9 (deleted exon 9) isoform was not detected in RRL. RRL was depleted of Ric-8A or Ric-8B by immunoprecipitation using rabbit polyclonal antisera 1184 and 2414 that were raised against the full-length versions of each protein (14, 29). Mock-depleted RRL was prepared using normal rabbit serum. The resultant Ric-8–depleted RRLs were immunoblotted and the relative levels of Ric-8 proteins that remained in the respective lysates were reduced (Fig. 1A).

Fig. 1.

Ric-8 proteins are components of RRL that are not required for in vitro Gα translation. (A) Polyclonal Ric-8A or Ric-8B antisera and normal rabbit serum (Mock) were preadsorbed to Protein A agarose and used to immunodeplete Ric-8A and Ric-8B from RRL. The relative levels of Ric-8A, Ric-8BFL, and the Ric-8BΔ9 isoforms that were immunoprecipitated (I.P.), and retained in RRL were immunoblotted alongside purified recombinant protein standards (2.5 ng each). (B) Capped Gαq and Gαolf mRNAs were generated in vitro and added to the mock- and Ric-8–depleted lysates with a [³⁵S]-methionine and [³⁵S]cysteine mixture to initiate in vitro translation reactions. Aliquots of the reactions were quenched at the indicated times and processed for SDS/PAGE fluorography. Pixel densitometry measurements were made with Image J (National Institutes of Health) to quantify the levels of synthesized G proteins. The data represent the mean ± SEM of three experiments. The indicated boxed protein bands were excised from the dried SDS gels and subjected to scintillation counting.

The kinetics of Gαq and Gαolf translation were examined in the Ric-8A– and Ric-8B–depleted lysates, respectively. These Ric-8/Gα combinations were chosen because of the known specificity of Ric-8–regulated Gα biosynthesis in cells and Gα subunit guanine nucleotide exchange stimulatory activities (14, 19, 20). The rate of translation and final abundances of Gαq and Gαolf were found to be nearly identical among the Ric-8A–, Ric-8B–, and mock-depleted RRL preparations (Fig. 1B). We conclude that Ric-8 proteins are not required for efficient Gα subunit translation.

Ric-8A Is Required for Proper Folding of RRL-Produced Gα Subunits.

We then examined our second postulate that Ric-8 proteins function to fold newly produced Gα subunits. Properly folded Gα subunits adopt a characteristic resistance to limited trypsinolysis in the GDP-AlF₄⁻– or GTP(γS)-bound conformations (30–32). An N-terminal ∼2 kDa fragment is cleaved from Gα-GTP, leaving a protected ∼37–43 kDa C terminus. The amount of protected C-terminal fragment correlates with the amount of properly folded Gα subunit capable of adopting an active conformation.

Gα mRNAs were added to mock- and Ric-8A–depleted RRLs to initiate the translation and limited trypsinization scheme as shown in Fig. 2A. Partial protection of AlF₄⁻-bound Gαq and Gα13, and nearly full protection of Gαi₂ and Gαs_L was observed in mock-depleted RRL at 10 and 30 min of trypsin treatment (Fig. 2B). Translated WT Gα subunits were not protected in the GDP alone condition. Only Gαq-Q209L was trypsin resistant in the GDP condition because it became bound to GTP present in RRL and remained GTP-bound because it is GTPase-defective. A dramatic reduction in the amount of trypsin-protected Gαq, Gαq-Q209L, Gαi₂, Gα13, but not Gαs_L, was observed when these subunits were translated in Ric-8A–depleted RRL. Interestingly, the Gαq-Q209L mutant was no longer trypsin resistant in the GDP (or AlF₄⁻) condition. This shows that Gα subunits produced in RRL in the absence of Ric-8A are not capable of binding to GTP and/or adopting the GTP-bound conformation. We conclude that endogenous RRL Ric-8A is required to produce functionally folded Gα subunits. Importantly, Ric-8A does not bind members of the Gαs class and would not be expected to act in a capacity to fold Gαs.

Fig. 2.

Gαi/q/13 subunits translated in Ric-8A–depleted RRL are not folded properly. (A) Schematic shows the procedure for Gα in vitro translation and trypsin protection assay. Dashed arrows indicate the timing of Ric-8A protein addition in the reconstitution experiments. (B) Gαq, Gαq Q209L, Gα13, Gαi₂, and Gαs_L were translated in mock- and Ric-8A–depleted RRL with a [³⁵S]-methionine and [³⁵S]cysteine mixture. Translation was stopped by treatment with cycloheximide (CHX) and RNase A, and the reactions were incubated with GDP or GDP and AlF₄⁻. Trypsin was added for 10 or 30 min and quenched. The reaction products were resolved by SDS/PAGE and visualized by fluorography. (C) Recombinant Ric-8A protein (10 nM or 1 µM) was added to the mock and Ric-8A–depleted RRL pre- or posttranslationally for 30 min as indicated. Gαq was translated and subjected to the trypsin protection assay. (D) Pixel densitometry measurements were made with Image J (National Institutes of Health) to quantify the levels of AlF₄⁻-bound Gαq that were protected from limited trypsinization. The data are presented as the amount of protected ∼37 kDa Gαq fragment normalized to the amount of intact Gαq without trypsin treatment (no Trp). The molar ratio of methionines and cysteines present in the fragment and intact protein was accounted for. The data represent the mean ± SEM from three experiments.

We then asked if reconstitution of Ric-8A–depleted RRL with recombinant Ric-8A could fulfill the requirement for endogenous Ric-8A to support efficient Gαq folding. The concentration of Ric-8A in RRL was estimated to be approximately ∼5–10 nM by quantitative immunoblotting (Fig. S1). Ric-8A–depleted RRL was supplemented with 10 nM and 1 µM doses of Ric-8A before or after Gα mRNA translation as depicted in Fig. 2A. Purified Ric-8A restored production of functionally folded Gαq in Ric-8A–depleted RRL (Fig. 2C). Ric-8A addition before or after translation led to quantitatively similar results (Fig. 2D). This suggests that the action of Ric-8A as a Gα folding chaperone is elicited after translation. Interestingly, the 10 nM Ric-8A dose restored Gαq trypsin protection to levels that were similar to those of mock-depleted RRL. The 1-µM Ric-8A dose produced levels of trypsin-protected Gαq that exceeded those of unsupplemented mock-depleted RRL. The capacity of normal RRL to produce functionally folded Gαq may be limited, in part, by the amount of Ric-8A. Far less Gα may be produced at any given time in cellular cytosol compared with the RRL system that was supplied with high amounts of Gα mRNAs. Cellular Ric-8A levels may be sufficient to keep pace with the level of endogenous Gα subunit production.

It could be argued that recombinant Ric-8A protein-dependent Gα trypsin protection in reconstituted RRLs was a consequence of the Gα:Ric-8A complex being more resistant to trypsinization than Gα alone. We reject this idea for theoretical and experimental reasons. No Ric-8A–dependent WT Gα subunit trypsin protection was observed when GDP alone was used. Ric-8A readily binds Gα-GDP, initiates rapid GDP release, and forms a nucleotide-free Ric-8A:Gα complex that is stable unless Mg⁺² and GTPγS (or AlF₄⁻) are provided to disrupt the complex. Ric-8 proteins have greatly diminished affinity for Gα-GTP(γS) or Gα-GDP-AlF₄⁻ (19, 20). To test this directly, purified Gαq-GTPγS (100 nM) was prepared as described and incubated with or without 1 µM Ric-8A (33). These proteins were treated with increasing doses of trypsin. No appreciable enhancement of Gαq trypsin protection was observed whether Ric-8A was included with Gαq-GTPγS or not (Fig. S2).

Ric-8A Reconstitutes WGE Production of Functional Gα Subunits.

We sought to corroborate results obtained in RRL using the WGE translation system. Importantly, no Ric-8 ortholog is documented to exist in any plant (34). No obvious Ric-8–like gene is present in Arabidopsis (35). Gαi₂, Gαq, and Flag-tagged Gβ₁ mRNAs were translated in WGE for 0 to 90 min. The radiolabeled G proteins were visualized by fluorography. The G proteins were produced with similar abundances as in RRL, although the rates of production were significantly slower (compare Fig. S3 and Fig. 1B). Gαi₂ and Gαq were then translated for 60 min in WGE that was supplemented with or without recombinant Ric-8A (1 µM). The G proteins were subjected to trypsin protection analysis (Fig. 3A). After 30 min of limited trypsinization, ≤5% of AlF₄⁻-bound Gαq and Gαi₂ were protected in unsupplemented WGE. Ric-8A supplementation afforded ∼30% and ∼50% protection to Gαq and Gαi₂, respectively (Fig. 3B). The dramatic improvement of WGE-folded Gα levels upon Ric-8A supplementation indicates that Ric-8A works efficiently with WGE chaperone components to fold Gα subunits.

Fig. 3.

Ric-8A supplementation of WGE restores translated Gα subunit folding. (A) Gαq and Gαi₂ were translated in WGE alone or WGE supplemented with 1 µM Ric-8A and [³⁵S]-methionine and [³⁵S]cysteine. Translation was quenched with cycloheximide and RNase, and the reactions were incubated with GDP or GDP and AlF₄⁻. Trypsin was added for 10 or 30 min and quenched. The reaction products were resolved by SDS/PAGE and subjected to fluorography. (B) Pixel densitometry measurements were made with Image J to quantify the levels of WGE-produced, AlF₄⁻-bound Gαq and Gαi₂ that were protected from limited trypsinization. The amount of protected Gα fragment was normalized to the amount of intact Gα without trypsin treatment (No Trp). The molar ratios of methionines and cysteines present in the fragments and intact proteins were accounted for. The data represent the mean ± SEM from three experiments.

The Ric-8A–dependent folding status of WGE-produced Gαq was evaluated further using a size-exclusion chromatography-based assay. Gαq was translated in WGE supplemented with or without Ric-8A. Translation was stopped and the reactions were incubated for 40 min at 30 °C and then chromatographed over a Superdex 200 gel filtration column. The majority of Gαq produced in WGE without Ric-8A addition eluted as an apparent high-molecular-weight aggregate in the column void volume (Fig. 4). Strikingly, Ric-8A addition resulted in nearly complete disappearance of the Gαq aggregate and the appearance of a different Gαq species with a ∼100-kDa estimated molecular weight. The elution position of this species was coincident to the elution position of purified Ric-8A:Gαq complex (∼102.1 kDa, actual molecular weight) (Fig. S4).

Fig. 4.

WGE-translated Gαq aggregation is rescued by Ric-8A supplementation. Capped Gαq mRNA was translated for 30 min in WGE alone or in WGE supplemented with 1 µM Ric-8A or 1 µM Gβγ as indicated. Cycloheximide and RNase A were added to stop translation and the reactions were incubated for an additional 40 min at 30 °C. In the GTPγS experiments, MgCl₂ (10 mM) and GTPγS (100 µM) were added for the final 15 min of the 40-min incubation. The translation reactions were chromatographed over a Superdex 200 HR 10/300 column. The column eluate was fractionated and portions of the fractions were analyzed by SDS/PAGE fluorography. Mean pixel densitometry measurements of the radiolabeled Gαq bands in all fractions were summed (the Gβγ experiment was not plotted). The data are presented as the percentage of total Gαq that was present in each individual fraction and represent the mean ± SEM of three experiments.

If Gαq was indeed bound to the exogenous Ric-8A and functionally folded, it should be dissociable from Ric-8A with GTPγS. Mg⁺² and GTPγS were added to WGE Gαq translation reactions during the last 15 min of the 40-min posttranslation incubation period. GTPγS alone was unable to prevent Gαq aggregation. In the reaction containing Ric-8A, GTPγS inclusion resulted in a ∼50% reduction in the amount of Ric-8A:Gαq complex with a concomitant ∼50% increase in the amount of a monomeric Gαq species that eluted coincident to the 44-kDa standard. We concluded that this Gαq species was a functionally folded monomer that had been released from Ric-8A. Ric-8B exerted similar action toward WGE-produced Gαs (Fig. S5).

WGE was supplemented with purified Gβγ (1 µM) to test the possibility that Gβγ might facilitate production of functional Gαq. Gβγ did not prevent the nearly complete aggregation of Gαq. This result supports the idea that Ric-8A has an active role in production of functional Gαq in WGE.

Nascent Gα and Gβ Use Unique Subsets of Molecular Chaperones.

The identities of other cellular chaperones that Ric-8A might work with to promote Gα folding were sought. Both the CCT and HSC70/90 chaperone systems have been suggested to participate in Gα folding (10–13). CCT and PhLP-1 fold nascent Gβ subunits before Gβγ heterodimer assembly (2, 6–8). Gαq and Gβ₁ were translated in untreated RRL for 10 min. Translation was stopped and the reactions were chased for 40 min in the absence or presence of EDTA to trap G-protein subunits bound to Mg⁺²-dependent chaperones. The reactions were then gel filtered.

At least three distinct Gαq and Gβ₁ species were identified (peaks 1–3, Fig. 5 A and B). Peak 1 represented a small portion of Gαq present in a high-molecular-weight, EDTA-dependent complex. The majority of Gβ₁ was found in these same fractions, and EDTA enhanced the amount. The elution patterns of native CCT purified from bovine brain (Fig. S6) and of RRL CCT that was detected with a CCTɛ antibody coincided with the Gαq and Gβ₁ species of peak 1 (Fig. 5C). These results are consistent with the known association of nascent Gβ with CCT (5, 7, 36). Our results suggest that nascent RRL-produced Gαq may weakly interact with CCT.

Fig. 5.

In vitro–translated Gαq and Gβ₁ associate with distinct sets of endogenous RRL chaperones. (A) Gαq and (B) Flag-Gβ₁ were translated in RRL for 10 min and chased for 40 min in the absence (No Addition) or presence of EDTA. The reactions were gel filtered over a Superdex 200 HR 10/300 column and the eluates were fractionated. The presence of G-protein subunits in the fractions was visualized by SDS/PAGE fluorography. Mean pixel densitometry measurements of the radiolabeled G proteins in all fractions were summed. The data are presented as the percentage of total G protein present in each individual fraction and represent the mean ± SEM of three experiments. The elution positions of three distinct populations of Gαq and Gβ₁ are indicated as peaks 1, 2, and 3. (C) Native CCT multisubunit complex purified from bovine brain was gel filtered over the Superdex 200 column and visualized by Coomassie-stained SDS/PAGE. The gel filtration elution profiles of native RRL cellular chaperones, CCTɛ, HSP/HSC70, and Ric-8A were identified by immunoblotting. The positions of gel filtration calibration standards are denoted along the column elution profiles.

The gel filtration elution profile of RRL HSC70 was examined and a monomeric species and higher ordered complex that eluted between the 158 and 670 kDa standards were identified. Portions of translated Gαq and Gβ₁ were present in the same fractions as the apparent HSC70 complex, but there was no EDTA enhancement. It could not be concluded whether translated Gαq or Gβ were engaged by HSC70 chaperones in RRL.

The Gαq population of intermediate molecular weight (100 kDa, peak 2) was enhanced modestly with EDTA and eluted coincidently with the major portion of RRL Ric-8A (Fig. 5C) and the purified Ric-8A:Gαq complex (Fig. S4). When Gαq was translated in Ric-8A–depleted RRL, Gαq Peak 2 was absent (Fig. S7). Interestingly, the RRL Ric-8A peak in the Gβ₁ translation experiment was of lower molecular weight and consistent with the gel filtration pattern of monomeric Ric-8A (Fig. S4). The EDTA enhancement in RRL Ric-8A:Gαq complex levels was consistent with the requirement of Mg-GTP to efficiently dissociate Ric-8 from Gα (Fig. 4) (19, 20). An intermediate-sized Gβ₁ species was modestly enriched by EDTA (Fig. 5B, peak 2). This Gβ₁ species appeared to be larger than the Gαq:Ric-8A complex and was likely nascent Gβ₁ bound to an unidentified RRL protein.

Levels of Gαq monomer (44 kDa, pool 3) and a lower molecular weight Gβ₁ species (> 44 kDa, peak 3) were enhanced in the absence of EDTA, corroborating the premise that EDTA treatment entrapped client Gαq and Gβ proteins in higher ordered complexes with chaperones.

Discussion

Information about the identity and workings of the molecular chaperones that fold newly synthesized Gα subunits has been limited. Using the rabbit reticulocyte lysate and WGE cell-free translation systems, we provide mechanistic data showing that Ric-8 proteins are chaperones required for Gα subunit folding. Our results explain observations made in mammalian cell and model organisms in which Ric-8 homologs are required to support proper abundances of heterotrimeric G proteins. In embryonic stem cells derived from Ric-8A–null mice, Gαi and Gαq subunits were defective in endomembrane association shortly after synthesis and were degraded rapidly (14). The portion of overexpressed Gαs that was multiubiquitinated in NIH 3T3 cells was reduced upon coexpression of Ric-8B (28). These results suggested two possible mechanisms for the Ric-8–chaperoning function. Ric-8 could fold newly produced Gα subunits or could facilitate the initial membrane association of folded Gα subunits as a trafficking chaperone. In either scenario, insufficient Ric-8 (function) would result in enhanced Gα degradation because misfolded and/or cytosolic Gα subunits are expected to be readily degraded. The results of our present study have clarified these possibilities. Ric-8 proteins are required to fold Gα subunits during biosynthesis. G-protein degradation and defective endomembrane targeting are most likely secondary manifestations of the folding defect.

Ric-8 probably does not act alone to fold Gα subunits. We envision a mechanism of Ric-8–mediated Gα folding that has some parallels to PhLP1-mediated folding of Gβ subunits. Nascent Gβ is folded within the CCT pore in a tripartite complex with PhLP-1. PhLP-1 then mediates Gβ:Gγ heterodimer formation with the Gγ chaperone, DRIP78 (5, 6, 8, 9). Fig. 5 shows that CCT binds most of the Gβ produced in RRL, and is suggestive that Gαq may have a more transient interaction with the CCT. The Gαq and CCT association that we observed was quantitatively modest in comparison with a study that showed strong Mg⁺²- and ATP-dependent CCT binding to Gαt in RRL (10). We can only speculate that these differences are due to the use of Gαq versus Gαt or to batch variations of Promega RRL over 2 decades. We also note that the Farr et al. study (10) did not uncover the large population of nascent Gα:Ric-8A heterodimer (Fig. 5A, peak 2) that we observed in RRL.

The CCT folds client proteins with substantial hydrophobic β-sheet content, like Gβ (37). Gα subunits have β-strand elements, but are mostly α-helical. This suggests that a series of molecular chaperone systems fold Gα subunits. Ric-8 may serve a role like PhLP-1 for Gβ and allow the Gα client polypeptide to sequentially engage the folding machineries of the HSC70/90 systems and the CCT complex. It will be important to dissect these molecular details and identify the cellular chaperones that Ric-8 works with. A comparison of Gαq produced in RRL versus WGE strongly suggests that other chaperones besides Ric-8A must be involved in Gα subunit folding. WGE has no Ric-8 ortholog and translated Gαq was completely aggregated (Fig. 4). In Ric-8A–depleted RRL, much of the produced Gαq was monomeric, although the monomer was not folded properly. This suggests that in Ric-8A absence, other RRL chaperones facilitate some degree of Gα folding and that these chaperones are absent from WGE or they have reduced activities toward mammalian Gα proteins.

Our work may explain long-standing enigmas regarding production of recombinant Gα subunits in bacteria and Sf9 cells. Milligrams of recombinant Gαi/o and Gαs can be purified from Escherichia coli, but functional Gαq/12/13 cannot (38). Like WGE, E. coli has no endogenous Ric-8. Reduced portions of Gαi₂ were folded in Ric-8A–depleted RRL and in WGE, but no functional Gαq or Gα13 could be made. Therefore, Gαi has a limited capacity to fold in systems that lack a Ric-8A chaperone, whereas Gαq and Gα13 do not. Sf9 cells likely possess an endogenous Ric-8 protein. Some overproduced soluble Gαq was functionally folded in Sf9, and coexpression of Gβγ with Gαq afforded a modest increase in functional Gαq levels with a more prominent relocalization of Gαq to the membrane (39). In Sf9 cells coexpressing mammalian Ric-8A and Gαq, a pronounced ∼50-fold induction of functional, soluble Gαq was realized (26). We conclude that Ric-8A acts on nascent Gα subunits in a protein-folding capacity before Gα binding to Gβγ. Gβγ binding may permit folded Gα subunits to become membrane associated.

It will be important moving forward to mechanistically dissect the significance of Ric-8 guanine nucleotide exchange stimulatory activity in cells. No published study has made a legitimate observation of in vivo Ric-8 GEF activity. Claims that Ric-8 acts as an amplifier of G-protein signaling that emanates from G protein–coupled receptors by reactivating Gα-GDP before it recouples to Gβγ lack experimental demonstration. Most gene manipulation experiments that raised, lowered, or ablated Ric-8 ortholog expression realized effects on G-protein signaling because the abundances of functional Gα subunits were altered. However, some data, particularly the localization of Ric-8A to mitotic structures, are not intuitively consistent with an exclusive role of Ric-8 as a Gα chaperone. Ric-8 may be a multifunctional protein. Further experimentation will address this hypothesis.

We propose that Ric-8 GEF activity and its function as a biosynthetic folding chaperone of Gα subunits are intertwined. GEF activity may be a consequence of the preferential affinity that Ric-8 has for molten-globule, nucleotide-free Gα state(s) over either nucleotide-bound conformation. Purified Ric-8A clearly induced nucleotide-free Gα conformation(s) with reduced definable tertiary structure, unlike the Gα-GDP or Gα-GTPγS conformations (40). Ric-8 may facilitate the transition of Gα from a prefolded globular state to its native state by promoting the first Gα guanine nucleotide–binding event. The Rab GTPase GEF Mss4/Dss4 elicits action by disordering the Rab guanine nucleotide–binding pocket to promote GDP release (41). Mss4 is now commonly thought to be a chaperone of exocytic Rab nucleotide-free states.

Materials and Materials

Materials.

Rabbit polyclonal antisera 2414 against Ric-8B and 1184 against Ric-8A were described (14, 29). Mouse monoclonal antibody 3E1 was raised against Ric-8A and used to detect Ric-8A by immunoblotting (SI Materials and Methods). The antibody against HSP/C70 was from Santa Cruz; the antibody against CCTα was from Stressgen; and the antibody against CCTɛ was from AbD Serotec. Human G-protein subunit cDNAs in pcDNA3.1+ (Gαq, Gαq Q209L, Gα13, Gαs_L, Gαi₂, Flag-Gβ₁) were obtained from the Unité Mixte de Recherche cDNA Resource Center. Expre³⁵S³⁵S protein-labeling mixture of [³⁵S]-methionine and -cysteine (11 mCi/mL) was from PerkinElmer. Protector RNase inhibitor was from Roche. Cycloheximide and RNase A were from Sigma. Nuclease-treated RRL and WGE were from Promega. Ric-8A and Ric-8BFL proteins were produced in Hi5 insect cells (19, 20). G-protein α subunits and prenylated Gβ₁γ₂ were purified as described (26, 42).

Immunodepletion of Endogenous RRL Ric-8 Proteins.

Nuclease-treated RRL was tumbled for 1 h at 4 °C with a 1/10th volume of clarified Ric-8A or Ric-8B whole polyclonal antisera that had been preadsorbed to Protein A agarose (Roche). Depleted RRLs were recovered as the supernatants after centrifugation at 2000 × g for 5 min.

In Vitro Transcription and Translation.

G-protein pcDNA3.1 plasmids were linearized with SmaI (Gαq, Gαolf, Gαi₂, Flag-Gβ₁) and SalI (Gαs_long, Gα13). Linearized plasmids were purified with a QIAquick gel extraction kit (Qiagen) and used as templates for in vitro transcription. Capped Gα mRNA transcripts were produced using the mMESSAGE/mMACHINE T7 Kit (Life Technologies). G-protein mRNAs (300 ng–1 µg) were translated in reactions containing 50 µL of nuclease-treated RRL or WGE, 40–60 µCi of EXPRE³⁵S³⁵S protein-labeling mixture and 1 µL of Protector RNase inhibitor for 10–30 min at 30 °C. Template was destroyed by addition of 10 µg RNase A and translation stopped by addition of 2 mM cycloheximide. Purified Ric-8 proteins (10 nM or 1 µM) were added to RRL or WGE before mRNA addition or immediately after the translation as indicated.

Trypsin Protection Assays.

In vitro translated G proteins from RRL or WGE were incubated with HEDG buffer (20 mM Hepes, pH 8.0, 1 mM EDTA, 1 mM DTT, 100 µM GDP, 0.1% (m/v) deionized polyoxyethylene 10 lauryl ether (C₁₂E₁₀) (Gαi₂, Gαs_long, Gα13), or 0.1% (m/v) Genapol (Gαq), or with HEDG buffer containing 50 mM MgCl₂, 30 µM AlCl₃, and 10 mM NaF at 4 °C for 15 min. Trypsin [0.002–0.0045% (m/v)] that had been pretreated with 25 ng/mL L-1-p-tosylamino-2-phenylethyl-chloromethyl ketone was added to the reactions and incubated for 10–30 min at 22 °C. Trypsin was quenched by addition of 40 µg/mL lima bean trypsin inhibitor and Laemmli sample buffer. The samples were resolved by SDS/PAGE and radiolabeled Gα proteins and fragments were visualized by sodium salicylate fluorography.

Size Exclusion Chromatography.

In vitro translated and radiolabeled G proteins in RRL or WGE were incubated in 40-min “chase” reactions at 30 °C containing 5 mM EDTA or an equivalent volume of water. Where indicated, 100 µM GTPγS and 10 mM MgCl₂ were added 25 min into the chase and incubated for 15 min. Reactions were centrifuged at 21,000 × g for 10 min at 4 °C before application to a Superdex 200 10/300 GL column (GE Healthcare). The column was resolved at 0.4 mL/min in gel filtration buffer (20 mM Hepes, pH 8.0, 100 mM NaCl, 2 mM MgCl₂, 1 mM EDTA, 0.5 mM DTT, 20 µM GDP). The column eluate was fractionated, and fractions were processed for immunoblotting and/or fluorography. The size exclusion chromatography was calibrated with Bio-Rad gel filtration standards.