Functional Characterization of the G Protein Regulator RGS13*

The signaling cascades evoked by G protein-coupled receptors are a predominant mechanism of cellular communication. The regulators of G protein signaling (RGS) comprise a family of proteins that attenuate G protein-mediated signal transduction. Here we report the characterization of RGS13, the smallest member of the RGS family, which has been cloned from human lung. RGS13 has been found most abundantly in human tonsil, followed by thymus, lung, lymph node, and spleen. RGS13 is a GTPase-activating protein for Gαi and Gαo but not Gαs. RGS13 binds Gαq in the presence of aluminum magnesium fluoride, suggesting that it bears GTPase-activating protein activity toward Gαq. RGS13 blocks MAPK activity induced by Gαi- or Gαq-coupled receptors. RGS13 also attenuates GTPase-deficient Gαq (GαqQL) mediated cAMP response element activation but not transcription evoked by constitutively active Gα12 or Gα13. Surprisingly, RGS13 inhibits cAMP generation elicited by stimulation of the β2-adrenergic receptor. These data suggest that RGS13 may regulate Gαi-, Gαq-, and Gαs-coupled signaling cascades.

G protein signaling is an effective mechanism of cellular communication during both physiological and pathological conditions (1)(2)(3). Regulators of G protein-signaling (RGS) 1 proteins are a relatively new family of proteins that attenuate G protein-mediated pathways by acting as GTPase-activating proteins (GAPs) for G␣ subunits (4). RGS binding stabilizes a G␣ conformation that favors hydrolysis of GTP to GDP, which hastens the termination of active G␣ (5). The second mechanism of G protein inhibition by RGS proteins is effector antagonism in which the RGS protein binds GTP-bound G␣ q and prevents G␣ q /effector interaction (6).
RGS proteins share a common domain, the RGS box. This motif of 120 amino acids is highly conserved in all RGS family members and conveys the ability of RGSs to bind G proteins (7). Although many RGS proteins exhibit GAP activity toward members of the G␣ i (8,9) and G␣ q (10) families, only one RGS protein has been shown to possess GAP activity specifically toward G␣ s (11). In addition to the classical RGS family, an additional group of proteins exists termed the RhoGEF RGSs (rgRGS) (12,13). These proteins, which contain a highly di-verged RGS homology domain, act as GAPs specifically for G␣ 12/13 and also stimulate GTP binding by Rho family members.
Although all RGS proteins contain the conserved RGS box, they display significant variability in sequences outside of this region. For example, a cysteine string found in RGS-GAIP and RGS20 is a site of palmitoylation, which may assist in membrane anchorage (14). The G␣ s GAP RGS-PX1 also contains a Phox domain that may be involved in intracellular trafficking (11).
Unique motifs in certain RGSs may modify the G protein GTPase cycle or G protein effector stimulation in additional ways. RGS14, for example, contains a "GoLoco" motif that inhibits guanine nucleotide dissociation on G␣ i (15). Most RGS proteins do not exhibit GAP activity toward G␣ s ; however, a short form of RGS3 (RGS3T) blocks calcitonin gene-related peptide-induced cAMP generation (16). More recently, it was shown that RGS2 directly inhibits some isoforms of adenylyl cyclase, independently of G␣ s (17), suggesting that RGS proteins can affect signaling through interactions with proteins downstream of G␣.
Here we report the characterization of RGS13, the smallest RGS found in mammalian tissues. Human RGS13 was cloned from lung cDNA and is expressed most prominently in immune tissues such as tonsil, thymus, lymph node, and spleen. As with other RGS proteins, recombinant RGS13 exhibits GAP activity toward G␣ i family members and not G␣ s . RGS13 binds G␣ q in the presence of AMF, implying that it acts as a GAP for G␣ q as well. Transfected RGS13 blunts MAPK activation evoked by either G␣ q -or G␣ i -coupled receptors. In addition, it blocks G␣ q QL-induced CRE-dependent transcription, consistent with a potential role for RGS13 as an effector antagonist of G␣ q . RGS13 inhibits receptor-stimulated cAMP generation, suggesting that it regulates G␣ s signaling despite its lack of G␣ s GAP activity. Thus, RGS13 may regulate G protein-mediated processes in the lung and immune system.

EXPERIMENTAL PROCEDURES
Cell Lines and Plasmids-Stable transfectants of the m1 muscarinic receptor into Chinese hamster ovary (CHO) cells and m2 muscarinic receptor into A9L cells were the generous gift of Jurgen Wess (NIDDK, National Institutes of Health). Cells were grown in Dulbecco's modified Eagle's medium, 10% fetal calf serum, penicillin/streptomycin, glutamine (Invitrogen) (A9L and 293T) ϩ NEAA (CHO). Plasmids directing expression of G␣ s QL, G␣ 12 QL, G␣ 13 QL, or G␣ q QL were obtained from Silvio Gutkind (NIDCR, National Institutes of Health). CRE-Luciferase was purchased from Stratagene, and SRE L-Luciferase was the generous gift of Koza Kaibuchi (Nara Institute of Science and Technology, Ikoma, Nara, Japan).
Real-time TaqMan Polymerase Chain Reaction (PCR) for Quantification of cDNA-Real-time TaqMan PCR was performed using primers 5Ј-TCAATTCTGGATGGCATGTGA-3Ј and 5Ј-TCTTGTCGAACTGTCA-ATGTTAATCTCT-3Ј with a TaqMan probe of 5Ј-ACCTATAAGAAATT-GCCTCACGGTGGAGC-3Ј. The amount of RGS13 cDNA present was normalized to glyceraldehyde-3-phosphate dehydrogenase values and expressed as n-fold greater than the lowest detectable level using the comparative cycle threshold method. The identity of PCR products was verified by automated sequencing.
Generation of Polyclonal Antisera and Immunoblotting-An N-terminal peptide (RDESKRPPSNLTLEEV) and C-terminal peptide (LK-SEMYQKLLKTMQSNNSF) were synthesized and conjugated to keyhole limpet hemocyanin before injection into rabbits for the generation of polyclonal antibodies. Antibodies were affinity-purified using recombinant peptide (Quality Controlled Biochemicals/BIOSOURCE International). To evaluate the specificity of RGS13 peptide antibodies, immunoblot analysis was performed. Recombinant RGS13, RGS4, or RGS16 (250 ng) were separated by SDS-PAGE and transferred to Immobilon-P membranes (Millipore). Blots were blocked in 10% milk in Tris-buffered saline-Tween prior to incubation with 1:1000 dilution of purified anti-RGS13 peptide antibody. A secondary, horseradish peroxidase-conjugated goat anti-rabbit antibody (Santa Cruz Biotechnology) was used to amplify positive signal. Immunodetectable protein was revealed upon the addition of a horseradish peroxidase substrate (Pierce) according to the manufacturer's instructions.
Localization of RGS13-293T cells were grown in 10-cm dishes to 90% confluency. Cells were transfected with 10 g of GFP-RGS13 or GFP using SuperFect (Qiagen) according to the manufacturer's instructions. For analysis of nuclei, cells were lysed in 20 volumes of lysis buffer (10 mM NaCl, 10 mM Tris-HCl, pH 7.4, 3 mM MgCl 2 , 0.5% Nonidet P-40, and complete protease inhibitor mixture (Roche Molecular Biochemicals)) on ice for 15 min. Cells were centrifuged at 500 ϫ g for 5 min. The pellet was washed twice, resuspended in lysis buffer, and quantified by Bradford assay. For preparation of membranes and cytosol, cells were homogenized in 25 mM HEPES pH 7.5, 250 mM sucrose, 2 mM EDTA, and complete protease inhibitor mixture (Roche Molecular Biochemicals). Homogenates were centrifuged at 2,000 ϫ g for 5 min. The supernatant was removed and centrifuged at 50,000 ϫ g for 30 min. The supernatant was precipitated with 9 volumes of cold acetone, resuspended in lysis buffer, and quantified by Bradford assay. The 50,000 ϫ g pellet was washed in lysis buffer, resuspended in the same buffer, and quantified by Bradford assay. Protein from each cellular fraction (100 g) was boiled in Laemmli buffer, separated by SDS-PAGE, and immunoblotted as above with RGS13-specific antibodies.
GAP Assays-GAP assays were performed as previously described (9). In short, recombinant G␣ i , G␣ o , or G␣ s (20 pmol) were labeled with [␥-32 P]GTP in the absence of magnesium. MgS0 4 (10 mM) and unlabeled GTP (100 M) with or without RGS (400 nM) were added to initiate GTP hydrolysis at 4°C. Free 32 P was separated from non-hydrolyzed 32 P by centrifugation with activated charcoal. GTP hydrolysis was measured as quantity of free 32 P in supernatants at various time points.
G␣ q Binding-200 -500 ng of recombinant G␣ q (R183C) (purified from Sf9 cells infected with a G␣ q baculovirus as described in Ref. 19) was incubated in buffer A (50 mM Tris, pH 8, 100 mM NaCl, 1 mM MgSO 4 , 10 mM ␤-mercaptoethanol, 20 mM imidazole, 10% glycerol, and 0.025% C 12 E 10 ) plus GDP (10 M) and 30 l of a 50% slurry of Ni 2ϩagarose (Invitrogen). The mixtures were incubated in the presence or absence of His-tagged RGS (5 g) and in the presence or absence of aluminum fluoride (30 M AlCl 3 and 10 mM NaF). Beads were pelleted and washed three times with buffer A with or without AMF. Proteins bound to Ni 2ϩ -agarose were eluted by the addition of 20 l of Laemmli buffer and boiling for 5 min. Proteins were separated by SDS-PAGE and analyzed by immunoblot analysis with anti-G␣ q antibodies (Santa Cruz Biotechnology).
Assay of Adenylyl Cyclase Activity-3 ϫ 10 5 293T cells were transiently transfected with 0.2 g of ␤ 2 -adrenergic receptor plus 1.5 g of GFP or GFP-RGS13 using SuperFect (Qiagen). 24 h post-transfection, media was replaced with 1 mM isobutylmethylxanthine in the presence or absence of 5 M isoproterenol for 10 min at 37°C. Rinsing cells with cold phosphate-buffered saline stopped the reaction. Cells were scraped into 4 mM EDTA, boiled to aggregate protein, and centrifuged at 14,000 ϫ g for 10 min. Supernatants were assayed for cAMP using the Biotrak radioimmunoassay (Amersham Biosciences).
Statistical Analysis-Statistical analysis was performed using GraphPad InStat software. Paired Student's t test or repeated measures analysis of variance with a Tukey-Kramer post-hoc test were used to determine two-tailed p value. p values Ͻ 0.05 were considered significant.

RGS13 Is Enriched in Lung and Immune
Tissues-To evaluate the expression of RGS13, we performed TaqMan analysis using cDNA from various tissues. We found RGS13 most abundantly in the tonsil followed by thymus, lymph node, lung, and spleen (Fig. 1A). Liver and pancreas exhibited intermediate levels of RGS13, whereas heart, skeletal muscle, kidney, and placenta expressed low levels. No RGS13 was detected in peripheral blood mononuclear cell, fetal liver, and brain. Thus, it appears that RGS13 is enriched in lymphoid tissues. To determine which cells within tonsil and other lymphoid compartments expressed RGS13, cDNAs from peripheral blood cells isolated with specific surface markers were assessed for RGS13 transcripts. Resting CD19 ϩ (B cells) and CD14 ϩ (monocytes) cells expressed the highest levels of RGS13, and similar to other RGS transcripts (1, 20, 21) RGS13 increased upon stimulation in CD19ϩ cells. Activated CD8 ϩ (T cells) cells expressed very low levels of RGS13 (Fig. 1B).
Expression and Subcellular Localization of RGS13 Protein-To facilitate detection of RGS13 protein, we raised rabbit polyclonal antibodies against two distinct RGS13 peptides, one at the N terminus and a second at the C terminus. We utilized each antibody to reveal recombinant RGS proteins by immunoblotting. Both RGS13 peptide antibodies detected as little as 250 ng of recombinant RGS13 protein but failed to detect either recombinant RGS4 or RGS16 ( Fig. 2A). We then utilized these antibodies to probe an immunoblot of detergent lysates derived from various human tissues including lung. However, we failed to detect a specific band of the predicted molecular mass of RGS13 with either antibody (data not shown).
Several recent publications have addressed the issue of cellular localization of RGS proteins. To determine where exogenous RGS13 was expressed in mammalian cells, we constructed a vector directing expression of an RGS13-GFP fusion protein and transfected the plasmid into 293T cells. We lysed cells in hypotonic lysis buffer and fractionated membranes, cytosol, and nuclei by differential centrifugation. Equal amounts of protein from each fraction were immunoblotted with anti-RGS13 antibodies. RGS13 was localized predominantly in the membrane and nuclear fractions with a very small amount in the cytosolic fraction (Fig. 2B).
RGS13 Is a G␣ i GAP and Binds G␣ q -To examine the function of RGS13 and its interaction with various G proteins we performed GAP assays. Recombinant His-tagged RGS13 was purified from bacteria by Ni 2ϩ affinity chromatography and tested for GAP activity. RGS13 increased the rate of GTP hydrolysis by both G␣ i and G␣ o but not G␣ s (Fig. 3, A-C). In addition, recombinant RGS13 bound purified G␣ q only in the presence of AMF (Fig. 3D), suggesting that RGS13 possesses GAP activity toward G␣ q as well.
RGS13 Inhibits Both G␣ i and G␣ q -mediated MAPK Activation-We evaluated the ability of RGS13 to block G proteinmediated signaling in intact cells by measuring G proteincoupled receptor-stimulated MAPK activation in the presence or absence of transfected RGS13. Carbachol stimulation of stably transfected G␣ i -coupled m2 muscarinic receptors in A9L cells resulted in an 8-fold increase in MAPK activity as measured by phosphorylated Erk. Co-transfection of RGS13 blunted this response by almost 50% (Fig. 4A). Similarly, carbachol evoked an increase in phospho-Erk in 293T cells transiently transfected with the m1 muscarinic receptor, which is coupled to G␣ q . This increase was also blocked by expression of RGS13, indicating that RGS13 inhibits G␣ q -mediated MAPK activation as well (Fig. 4B).
RGS13 May Act as an Effector Antagonist for G␣ q -Because some RGS proteins block the interaction between G␣ q and its effector, PLC␤, we examined the ability of RGS13 to regulate transcription stimulated by GTPase-deficient G proteins. To assess signaling via G␣ q , 293T cells were transiently trans- fected with a CRE-Luciferase reporter gene in the presence or absence of G␣ q QL. Co-transfection of RGS13 blocked G␣ q QLmediated CRE activation (Fig. 5A). However, RGS13 did not significantly inhibit transcription of a serum-response element  luciferase reporter gene (SRE-Luciferase) induced by either G␣ 12 QL or G␣ 13 QL (Fig. 5, B and C).

RGS13 Inhibits Isoproterenol-induced cAMP Generation-
Because it has been recently reported that certain RGS proteins inhibit some isoforms of adenylyl cyclase directly, we examined the ability of RGS13 to block cAMP production. 293T cells were transfected with the ␤ 2 -adrenergic receptor in the presence of GFP or RGS13. Stimulation of these cells with isoproterenol evoked a 4-fold increase in cAMP. Co-transfection of RGS13 attenuated the increase in cAMP by approximately 50% (Fig. 6). Because RGS13 does not act as a G␣ s GAP (Fig.  3C), RGS13 may block this signaling pathway at the level of adenylyl cyclase.

DISCUSSION
This study represents the initial characterization of RGS13 expression and function. RGS13 is most abundant in human tonsil followed by thymus, lymph node, lung, and spleen with low levels or a lack of expression in various other tissues. Ectopically expressed GFP-RGS13 localizes in both membrane and nuclear fractions of 293T cells with a very small portion of RGS13 in the cytosol. Similar to other classical RGS proteins, RGS13 possesses GAP activity toward G␣ i family members and inhibits signaling evoked by G␣ i -coupled receptors. RGS13 binds to G␣ q in the presence of AMF, suggesting that RGS13 is a GAP for G␣ q . In addition to its GAP activity, RGS13 likely blocks G␣ q signaling by effector antagonism as well because GTPase-deficient G␣ q QL-mediated CRE stimulation is attenuated by RGS13. RGS13 is not a GAP for G␣ s , although it inhibits cAMP generation induced by stimulation of a G␣ scoupled receptor.
The expression pattern of RGS family members is highly variable and may contribute to physiological specificity of RGS proteins that share similar biochemical properties. RGS16 is expressed predominantly in the retina, pituitary, and liver (22)(23)(24). RGS18 is expressed in hematopoietic tissues and lung (25,26), whereas RGS2 and RGS3 are ubiquitously expressed (20,27). RGS13 expression is highest in tonsil (about 6.8 ϫ 10 6 times greater than CD4ϩ cells). There is 10-fold less RGS13 in thymus, lymph node, lung, and spleen, with much lower levels of RGS13 detected in other tissues such as brain, which is a reservoir for many of the RGS family members (28,29). In contrast to earlier reports describing RGS13 expression in brain (29, 30), we failed to detect any RGS13 in brain cDNA. This discrepancy could be explained by localization of RGS proteins within the brain; that is, certain areas of the brain enriched with RGS13 may be underrepresented in total brain cDNA.
Using the RGS13 peptide antibodies, we evaluated endogenous RGS13 expression in human lung, spleen, liver, pancreas, kidney, heart, and brain (data not shown). We concluded that although the peptide antibodies were able to detect recombinant RGS13 protein and not other closely related RGS proteins of similar size, the sensitivity of the peptide antibodies was insufficient to identify endogenous protein in tissue lysates by immunoblotting. To address this issue, we have initiated generation of knockout mice in which a LacZ reporter gene is placed under control of the RGS13 promoter, which should allow us to better define the anatomical and developmental expression of RGS13.
Transfected RGS13 localizes in the cell membrane and the nucleus. Several RGS proteins have been localized to the cell membrane (31) and have undergone post-translational modifications such as palmitoylation, which may assist in membrane anchorage (32,33). RGS13 lacks several of the N-terminal residues that adopt an ␣-helical conformation in closely related RGSs such as RGS4, 5, and 16 and facilitate direct interaction with anionic membrane phospholipids (34,35). It will be interesting to determine whether palmitoylation of RGS13 is required for its membrane localization or whether another mechanism of membrane targeting is responsible. It has been reported that RGS2 and RGS10 are localized in the nucleus (36). We show here that transfected RGS13 is also located in the nucleus, although the functional significance of nuclear RGS expression has yet to be determined.
RGS13 functions similarly to many RGS proteins in that it bears GAP activity toward members of the G␣ i family and likely has GAP activity toward G␣ q as well. RGS13 may be an effector antagonist for G␣ q as demonstrated by its ability to block G␣ q QL-mediated CRE activation. Although one could postulate that the capacity of RGS13 to block cAMP generation may have influenced the CRE-Luciferase reporter assay, the lack of significant decrease in either basal cAMP levels or basal CRE-Luciferase activity suggests that this mechanism would not account for the inhibition of G␣ q QL-stimulated CRE-Luciferase by RGS13.
RGS13 blunted isoproterenol-evoked increases in cAMP but exhibited no GAP activity toward G␣ s . There are several possible explanations for this result. An interaction between RGS9 and a phosphodiesterase has been reported (37). RGS13 may enhance cAMP metabolism by stimulating a phosphodiesterase, resulting in a net decrease in cAMP concentration. In that case one would expect to see decreased basal levels of cAMP, which we failed to detect. A truncated form of RGS3 (RGS3T) was shown to decrease receptor-mediated cAMP generation without decreasing basal cAMP levels (16). The authors suggested that it was unlikely that RGS3T blocked adenylyl cyclase directly. Given the similarity in our results, RGS13 and RGS3T may be acting via the same uncharacterized mechanism. Another possibility is that like RGS2 (17) RGS13 may block adenylyl cyclase directly. It has been demonstrated that RGS1, RGS2, and RGS3 (but not RGS4 or RGS5) block GTP␥S/ odorant-mediated cAMP generation in olfactory epithelium membranes. In addition, RGS2 was shown to block forskolinevoked increases in cAMP (17), indicating that some RGS proteins regulate certain isoforms of adenylyl cyclase directly. Interestingly, RGS2 failed to attenuate cAMP production induced by stimulation of endogenous adenylyl cyclase in HEK-293 cells (17). In contrast we found that RGS13 inhibited cAMP generation in these cells, suggesting that RGS2 and RGS13 may regulate different isoforms of adenylyl cyclase. Given that the only identified domain in RGS13 is the RGS box, it will be of interest to examine whether the RGS box is the region that confers the ability of RGS proteins to inhibit cAMP production. If so, why then do RGS4 and RGS5 lack the ability to regulate this process?
Functional analysis of RGS13 reported here was derived from cells transfected with GFP-RGS13. We utilized GFP-RGS13 only after several failed attempts to express RGS13 in untagged or His-tagged forms. Although we used the GFP vector or other GFP-tagged RGS proteins as controls, the difficulty in expressing untagged RGS13 is of interest. Future experiments are required to determine whether the difficulty in detecting untagged RGS13 is a result of such anomalies as unstable transcription or translation. This explanation seems unlikely because we could easily detect in vitro-translated untagged RGS13 in rabbit reticulocyte lysates (data not shown). Alternatively, RGS13 may require a binding partner or modification for stable expression in mammalian cells. An example of this type of protein instability is G␤ 5 , which requires expression of RGS9 in the brain (18). These studies might also explain the difficulty in detecting endogenous levels of RGS13 by immunoblotting.
Through the characterization of RGS13 we have added another piece to the puzzle of the RGS family. Unlike many RGS proteins RGS13 has no identified domain other than the RGS box, but because of its high expression in the immune system and lung and its ability to block G␣ i , G␣ q , and cAMP generation, the biological niche of RGS13 might be to regulate specific G protein-dependent signal transduction pathways in these regions. The physiological relevance of RGS13 and other RGS proteins is slowly being evaluated through the use of targeted gene disruption and expression of transgenes in mice. Through these lines of experimentation we hope to better understand the function of RGS13 and its relevance to health and disease.