Characterization of a novel mammalian RGS protein that binds to Galpha proteins and inhibits pheromone signaling in yeast.

Genetic studies of molecules that negatively regulate G-coupled receptor functions have led to the identification of a large gene family with an evolutionarily conserved domain, termed the RGS domain. It is now understood that RGS proteins serve as GTPase-activating proteins for subfamilies of the heterotrimeric G-proteins. We have isolated from mouse pituitary a full-length cDNA clone encoding a novel member of the RGS protein family, termed RGS16, as well as the full-length cDNA of mRGS5 and mRGS2. Tissue distribution analysis shows that the novel RGS16 is predominantly expressed in liver and pituitary, and that RGS5 is preferentially expressed in heart and skeletal muscle. In contrast, RGS2 is widely expressed. Genetic analysis using the pheromone response halo assay and FUS1 gene induction assay show that overexpression of the RGS16 gene dramatically inhibits yeast response to α-factor, whereas neither RGS2 nor RGS5 has any discernible effect on pheromone sensitivity, pointing to a possible functional diversity among RGS proteins. In vitro binding assays reveal that RGS5 and RGS16 bind to Gαi and Gαo subunits of heterotrimeric G-proteins, but not to Gαs. Based on mutational analysis of the conserved residues in the RGS domain, we suggest that the G-protein binding and GTPase-activating protein activity may involve distinct functional structures of the RGS proteins, indicating that RGS proteins may exert a dual function in the attenuation of signaling via G-coupled receptors.

comotion in Caenorhabditis elegans, have also led to the isolation of a new gene, referred to as EGL-10 (6). The protein product of EGL-10 has an opposing effect of a G␣ protein encoded by GOA-1 (7), implying that EGL-10 exerts an inhibitory role on G-protein signaling. Subsequent cDNA library screening under low stringency, degenerate PCR reactions, as well as data base search of expressed sequence tags, have identified over 15 different cDNA species of mammalian RGS members (6,8,9). In addition, a two-hybrid screen has isolated another mammalian RGS member, known as GAIP (G ␣ interacting protein) that binds to G␣ i3 known to play a part in protein trafficking in the Golgi apparatus (10). They all share sequence similarity to a yeast gene, SST2, the founding member of the RGS family, isolated by genetic analysis from a strain that exhibits supersensitivity to the yeast mating pheromone (11,12). It has been demonstrated that introduction of mammalian RGS family members (RGS 1 to 4) can substitute the yeast SST2 gene product in blunting pheromone signaling (8,9), emphasizing a functional significance of the sequence similarity among RGS family members.
Strong genetic evidence has suggested that RGS proteins exert their function through G-proteins (6,13,14). The biochemical nature of their inhibitory effect remained unclear until the very recent demonstrations that at least some of the RGS members (RGS1, -4, -10, and GAIP) can physically interact with G␣ i and G␣ o proteins and accelerate the hydrolysis of GTP by their intrinsic GTPase activity (10, 14 -18). Whether this mode of action is prototypic of other RGS functions remains to be seen. It is also noteworthy that the mammalian RGS4 protein inhibits G-protein function in yeast (8).
Tissue distribution analysis of RGS family members shows that they seem to be differentially expressed. RGS4, for example, is mainly expressed in the brain regions (8). It is likely that each RGS protein only regulates the signal transduction pathway of a subset of receptors or a subtype of G-proteins, either by temporal and spatial restriction of expression or by having diverse intrinsic biochemical properties. We have previously cloned the seven-transmembrane, G s -coupled receptor for the growth hormone releasing factor, which is expressed exclusively in the anterior pituitary and is required for the cell proliferation of somatotrophs (19,20). We reasoned that a tissue-specific RGS protein may exist in the anterior pituitary that modulates certain aspects of endocrine activity and cell growth of the gland. We present in this report the identification and cloning of a novel RGS member, referred to as RGS16, from mouse pituitary. We have also isolated a full-length cDNA for mouse RGS5, a mouse homologue of human RGS2, and have analyzed the tissue distribution of these RGS mRNAs. Functional studies based on their ability to inhibit pheromone signaling in yeast and their affinity for different G ␣ subunits point to functional diversity among different RGS proteins. Furthermore, mutational analysis of conserved residues in the RGS domain underscores the functional significance of the sequence conservation, and suggests also that binding and GAP activity may involve separable structures.

Identification of Novel RGS Members in the Mouse Pituitary-Mouse
pituitary cDNA was generated by reverse transcription using avian myeloblastosis virus reverse transcriptase (Amersham). The reverse transcription was carried out in a 25-l reaction mixture using 5 g of total cellular RNA and 0.1 g of oligo(dT) primer (Boehringer Mannheim), according to the instructions of Life Technologies, Inc. Degenerate primers, synthesized based on conserved amino acid residues in the RGS domain, were used in polymerase chain reactions (PCR) according to Koelle and Horvitz (6). The primers that generated the probe inserts for mRGS2, mRGS5, and mRGS16 were: "5R", G(G/A)IGA(G/A)AA(T-/C)(A/T/C)TI(A/C)GIT(T/C)TGG and "3T", G(G/A)TAIGA(G/A)T(T/C)IT-T(T/C)T(T/C)CAT. Amplification reactions were performed using 2 l of mouse pituitary cDNA generated as described above in a final volume of 50 l containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl 2 , 200 M each of dATP, dCTP, dGTP, dTTP, and 2.5 units of Taq DNA polymerase (Boehringer Mannheim). Reaction cycles were as follows: following an initial denaturation of reaction mixture at 94°C for 3 min, 35 cycles of 94°C for 45 s, 40°C for 1 min, and 72°C for 2 min. Subcloning of the PCR products into pBluescript (Stratagene) and subsequent sequencing analysis were carried out using standard techniques (21). The PCR fragments (240 base pairs) were then used as probes to screen a mouse pituitary gt11 cDNA library (a gift from Drs. B. Andersen and G. Rosenfeld, University of California San Diego) for full-length cDNA clones according to manufacturer's instructions (Stratagene).
Northern Blot Analysis and RNase Protection Assay-A Northern blot containing eight different mouse tissue poly(A) ϩ RNAs (2 g each lane, Clontech) was probed successively with 32 P-labeled RGS2, -5, and -16 cDNA. The DNA probes were generated from full-length cDNA inserts using a random priming reaction kit (Boehringer Mannheim) and [␣-32 P]dCTP (DuPont NEN) to specific activities of approximately 2 ϫ 10 9 cpm/g. Northern hybridization was carried out at 42°C overnight using a buffer containing 5 ϫ SSPE, 10 ϫ Denhardt's solution, 1% SDS, 0.5 mg/ml herring sperm DNA (Sigma), and 50% formamide. The blot was then thoroughly rinsed and washed with 1 ϫ SSC at 65°C for 20 min, followed by washing with 0.2 ϫ SSC at 65°C for 30 min. A ␤-actin probe (Clontech) was used for RNA loading normalization.
For further analysis of the RGS16 expression in different tissues, RNase protection analysis was performed. An RGS16 SacI-EcoRI 157base pair fragment (101 to 257 nucleotides downstream of the initiation ATG codon) was blunt-ended with Klenow polymerase and ligated into the SmaI site of pBluescript. The antisense strand RNA probe was transcribed using T7 RNA polymerase (Promega) and hybridized to total tissue RNAs, and the RNase protection assay was performed as described previously (19).
Yeast Pheromone Response Assays-A bioassay was used to measure the sensitivity of pheromone response in yeast cells that express different mammalian RGS proteins. Both bar1⌬ mutant cells, US356 (MATa, ade2-1, trp1-1, can1-100, leu2-3, his3-11, bar1⌬), and W303 cells were transformed with each RGS cDNA in the pMW29 vector under a galactose-inducible promoter (22) and selected on ura Ϫ dropout plates. Halo assays were carried out as described (8,23), except that transformants were grown on ura Ϫ dropout plates containing 2% galactose. Pheromone response was also monitored by measuring the pheromone-inducible levels of the FUS1 transcript in cells transformed with different RGS cDNA constructs. Cell culture, pheromone treatment, and Northern hybridization with the FUS1 probe were carried out as described (24). Briefly, a single colony of each yeast transformant was grown at 30°C in 40 ml of ura Ϫ dropout medium supplemented with 2% galactose to an OD of 0.8. Each culture was then divided into halves. The one half-culture was added with ␣-factor (Sigma) to a final concentration of 5 M and incubated for an additional 20 min, while the other half was untreated. Cells were then pelleted and subjected to RNA isolation as described (25). Equal amounts of total cellular RNA (15 g) isolated from each sample was separated on a formaldehyde-denaturing agarose gel, transferred to a nitrocellulose filter, and hybridized with 32 P-labeled FUS1 probe (gift from Dr. L Lim, Glaxo-IMCB), as described above.
Epitope Tagging of RGS Proteins-To compare expression levels of different wild-type RGS proteins and RGS16 mutants in yeast cells, PCR was employed to generate an XhoI site immediately upstream of the ATG start codon of all RGS cDNAs. A pBluescript derivative vector containing a synthetic DNA fragment coding for the flag tag (DYKD-DDDKH) was constructed and used to fuse in-frame to the translation initiation codon of each RGS cDNA. After confirmation by sequencing analysis, the flag-tagged RGS cDNAs were released by EcoRV and BamHI from pBluescript and ligated to SmaI/BamHI-treated yeast expression vector pMW29. Flag-tagged RGS constructs in the pMW29 vector were separately introduced to the US356 yeast cells as described above. Single colonies were inoculated into 2 ml of ura Ϫ dropout medium supplemented either with 2% galactose or 2% glucose, and were grown at 30°C to an identical density (A 600 ϭ 1.0). Cells (1.5 ml) were pelleted and lysed in 100 l of SDS protein sample buffer (2% SDS, 50 mM Tris-HCl, pH 6.8, 100 mM dithiothreitol, 0.1% bromphenol blue, and 10% glycerol). Protein was extracted by freezing in liquid nitrogen followed by thawing and boiling for 5 min. The cell lysates were then sonicated for 20 s and spun in a microcentrifuge for 5 min. The protein concentration of each extract was determined using the Bio-Rad DC Protein Assay (Bio-Rad). Proteins (150 g each) were separated on 10% SDS-PAGE gels, transferred to a Hybond TM -C Extra filter (Amersham), and flag-tagged RGS proteins were detected using the M2 anti-flag antibody precisely according to the manufacturer's instructions (Kodak) and the ECL system (Amersham). In parallel, the ability of flag-tagged wild-type RGS members and RGS16 mutants to attenuate pheromone signaling was re-assessed by halo assays as described above.
Site-directed Mutagenesis of the RGS16 cDNA-In vitro site-directed mutagenesis was performed using the Transformer TM Site-directed Mutagenesis Kit (Clontech). The oligonucleotides used to create RGS16 mutants, G74R, L82S, E85G, EN89/90GA, I116D, LM161/162SK, RF169/170SC, were TGAACAGTAAAAATcGGGTGGCTGCCTTCC, CTTCCATGCCTTCtcAAAGACGGAATTCAG, TTCCTAAAGACGGgA-TTCAGTGAGGAGAAC, GAATTCAGTGAGGgGgcCCTGGAGTTCTG-GTTG, GGGCTCACCACgaCTTTGACGAGTACATCC, AGACCCGCA-CATcGAaGGAGAAGGACTCCT, and GACTCCTATCCGaGCTgCCTC-AAGTCACCA, respectively. The mutant oligonucleotide for the XbaI site of the pBluescript vector was GGGATCCACTAGTtCTAGAGCGG-CCGCCAC. Double-stranded RGS16 cDNA carrying an XhoI site immediately upstream of the translation initiation site in pBluescript was heat-denatured and annealed to each of the specific mutant oligonucleotides and the common mutant primer for the XbaI site. The mutagenesis procedures were carried out according to the manufacturer's instructions (Clontech). The resulting mutants were confirmed by sequencing analysis.
RGS-G-protein Binding Assays-Wild-type and mutant RGS cDNAs were fused in-frame to XhoI site of a derivative of the bacterial expression vector pGex2TK (Pharmacia); the fusion proteins were expressed in the Escherichia coli strain BL21(DE3)-pLysS. Transformant cells were grown to an OD of 1.0 and induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside (Life Technologies, Inc.) at 37°C for 2 h to express fusion RGS proteins. Cells were then pelleted and lysed in a lysis buffer containing 0.4 M NaCl, 50 mM Tris-HCl (pH 7.6), 1 mM EDTA, 1% Triton X-100, and 10% glycerol, supplemented before use with 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, 4 mM benzamidine, 10 g/ml leupeptin, and 10 g/ml aprotinin. The lysates were then sonicated for 3 ϫ 30 s on ice and centrifuged at 4°C at 30,000 rpm for 20 min in an SW41 rotor. The supernatants (from 400 ml of cell culture) were incubated at 4°C for 30 min with 1 ml of 1:1 suspension of glutathione-agarose beads (Sigma). The beads were washed six times each with 10 volumes of lysis buffer. The purified glutathione S-transferase fusion proteins were stored in the beads in the same lysis buffer except 1% Triton was omitted. Protein was eluted with the lysis buffer without Triton containing 10 mM glutathione (Sigma) and quantified by the Bradford method (Bio-Rad).
Full-length cDNAs encoding G-proteins, G␣ s , G␣ i2 (26), and G␣ o (27) were obtained by PCR amplification using Pfu DNA polymerase (Stratagene) and total cDNA derived from mouse pituitary; and G␣ i3 (28) was amplified from rat pituitary cDNA. The oligonucleotide sequences for G␣ i2 were: CCATGGGCTGCACCGTGAGCGCC and CCCTCAGAAGA-GGCCACAGTCCTT; for G␣ i3 , CATATGGGCTGCACGTTGAGCGC-CGA and AGTCGACTCAGTAAAGCCCACATTCCT; for G␣ o : CATAT-GGGATGTACGCTGAGCGCA and TAGGTTGCTATACAGGACAAGA-GG. Amplified cDNA fragments were first cloned into the pBluescript vector (Stratagene) and multiple recombinant clones were isolated for each cDNA species for analyzing PCR fidelity by restriction digestions and DNA sequencing. For G␣ s , due to the high G/C content of its cDNA sequence at the 5Ј-end, it took two steps to obtain the full-length cDNA. The 3Ј-end oligonucleotide for PCR amplification of G␣ s cDNA was: TGAATTCGGGTGTTCCCTTCTTAGAGC. The 5Ј-oligonucleotide was designed from 45 nucleotides downstream of the ATG codon (GGCCC-AGCGCGAGGCCAACAAAAA) to avoid the G/C-rich region. The am-plified fragment was treated with Klenow DNA polymerase and polynucleotide kinase and ligated to Klenow-treated (blunt-end) ApaI site of pBluescript. In this way, ligation of the vector (providing a G nucleotide) to the insert (GGCCC) restored the ApaI site (GGGCCC) at the 5Ј-end. The recombinant vector was then cut with ApaI and ligated to the following synthetic double-stranded DNA fragment encoding the remaining N-terminal region. The two synthetic oligonucleotides were: cATGGCGGCGCGGGGCGCGGCCGGGCTGCGGGGCGGCGGGGAG-AAGGCC and TTCTCCCCGCCGCCCCGCAGCCCGGCCGCGCCCCG-CGCCGCCATGGGCC.
The G-protein cDNAs were then transferred to the pMet vector under the control of T7 promoter (29); and the G-proteins were generated and 35 S-labeled by in vitro transcription and translation reactions using the TNT Coupled Reticulocyte Lysate Systems (Promega) and [ 35 S]methionine. Assays for RGS interactions with different G-proteins were carried out essentially in the same way as described (16). Briefly, 10 l of the in vitro translated G-proteins (approximately 20 ng) were incubated for 30 min at room temperature by adding 90 l of buffer A (50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM MgSO 4 , 20 mM imidazole, 10 mM ␤-mercaptoethanol, 10% glycerol) supplemented with either GDP (10 M) or GDP and AlF 4 Ϫ (10 and 30 M, respectively). RGS fusion proteins (approximately 1 g) and glutathione S-transferase (3 g), as a control, bound to agarose beads, were equilibrated with buffer A, and then separately incubated with the G-protein mixture for 30 min at 4°C. The agarose beads were washed 4 times with buffer A containing either GDP or GDP with AlF 4 Ϫ . The bound G-protein and RGS fusion protein were eluted in 2 ϫ SDS protein sample buffer and separated on SDS-PAGE gels. After electrophoresis, gels were stained with Coomassie Brilliant Blue, dried, and exposed to x-ray film.

RESULTS
Identification of Novel RGS Members-To identify new RGS members that are enriched in the anterior pituitary, degenerate PCR primers were designed corresponding to conserved amino acid residues among previously identified RGS members and used to perform PCR amplification using mouse pituitary cDNA as the template as described previously (6). We identified cDNA fragments corresponding to RGS2, -4, -5, -6, -7, -9, and -11 (members numbered according to Ref. 10, data not shown), which had been previously identified by other groups or as an expressed sequence tag in the data base (4,6). In addition, we obtained a novel one, named clone 16 (Fig. 1). We used the novel clone 16 cDNA fragment, together with those of RGS2 and RGS5 to screen the mouse gt11 pituitary cDNA library. In previous limited tissue distribution analysis (5), RGS2 was found to be absent in all tissues, and we thus wondered if it might be enriched in the pituitary. As for RGS5, only a partial sequence was previously identified as both an EST partial sequence (human) and a degenerate PCR product (rat) (6).
Sequence Analysis and Comparison to Known RGS Members-Nucleotide sequences and conceptually translated amino acid sequences derived from the isolated cDNA inserts were aligned with all other known RGS members. It was found that the cDNAs of RGS2, -5, and -16 all contained an initiation Met codon in a position corresponding to that of many previously characterized RGS proteins. Furthermore, RGS5 has an inframe upstream stop codon. The cDNAs for RGS2, -5, and -16, encode open reading frames of 211, 181, and 201 amino acids, respectively. These sizes are in good agreement with those of other RGS proteins such as RGS1, hRGS2, and GAIP. Together with the presence of stop codons at the end of each open reading frame, these new cDNAs are full-length clones. The three mouse RGS full-length protein sequences, together with the human full-length RGS2 and the partial rat RGS5 sequences, are aligned as shown in Fig. 1A. The most conserved region, the RGS domain, is underlined (Fig. 1A). Sequence alignment for the RGS domains from various known full-length RGS proteins are aligned in Fig. 1B. The mouse RGS2 shares 95% overall identity with its human homologue (Fig. 1A), with only one conservative change (Thr 156 to Ser 156 , from human to mouse) in the RGS domain. When the mouse RGS5 was compared with the rat counterpart (6), which contained only a large portion of the RGS domain (Fig. 1B), it was found that this region has one amino acid variation (Arg 149 to His 55 , from mouse to rat). A BLAST search indicated that RGS16 is a novel RGS member of the RGS protein family, sharing the highest degree of similarity to RGS3 (63%) between their RGS domains, and is highly divergent outside this region from all known RGS members.
Tissue Distributions of RGS2, RGS5, and RGS16 -As an initial step toward understanding the biological roles of our newly characterized RGS proteins, we analyzed the tissue distribution of the three RGS members by Northern hybridization and RNase protection assay. A Northern blot containing poly(A) ϩ RNA from eight different tissues was successively hybridized to the RGS2, RGS5, and RGS16 probes. The RGS2 probe hybridized to two transcripts, 1.5 and 1.8 kb, in length, present in all tissues except liver (Fig. 2). RGS5 and RGS16 probes hybridized to a single RNA species of 4.4 and 2.4 kb, respectively (Fig. 2). The RGS5 mRNA was expressed abundantly in heart and skeletal muscle, and at low levels in brain, liver, and kidney. In contrast, RGS16 was present predominantly in liver, and at low levels in heart and brain (Fig. 2). Since RGS16 expression was very tissue-restricted, we further performed an RNase protection assay using a collection of RNA samples from 14 mouse tissues. The RNase protection assay showed that RGS16 was only present in brain, pituitary, and liver (Fig. 3). This is in total agreement with the Northern hybridization results indicating that RGS16 is present in a limited number of tissues. The same RNase protection experiment was also carried out with the RGS5 probe, showing that it is expressed in all the tissues examined except liver, spleen, intestine, fat, and adrenal (data not shown).
Inhibition of Pheromone Signaling by Mouse RGS16 -Previous work has shown that mammalian RGS members can attenuate the pheromone response pathway in yeast (8,9). To ascertain that we had cloned functional RGS members, those RGS cDNAs were inserted into the yeast expression vector pMW29 under a galactose-inducible promoter, and were transformed into yeast strains US356 and W303. The former strain harbors a deletion mutation in the bar1 gene, which encodes a secreted proteinase capable of cleaving ␣-factor (30). Results obtained from the halo assays showed that in the bar1⌬ mutant overexpression of RGS16 almost completely blocked the pheromone response (Fig. 4A). On the other hand, RGS2 and RGS5 had very little effect on the pheromone response and were phenotypically indistinguishable from that of the mutant cells to which only the blank pMW29 vector was introduced (Fig.   FIG. 2. Northern blot analysis of three mouse RGS members in various tissues. A Northern blot with poly(A) ϩ RNA isolated from various tissues, indicated on the top, was successively hybridized with the cDNAs of mRGS2, mRGS5, and mRGS16 and, as a control of RNA loading, with a ␤-actin probe provided by Clontech. The hybridization probe used for each panel is indicated on the left; and estimated molecular weights of mRNA transcripts detected are indicated by arrows. The mRGS2 probe hybridized to two RNA species of 1.5 and 1.8 kb, while the mRGS5 and mRGS16 probes each detected a single transcript, 4.4 and 2.4 kb, respectively. As expected, the actin probe hybridized to two transcripts representing nonmuscle (upper band) and muscle (lower band) actins. 4A). Similar inhibitory effects on ␣-factor response were obtained with yeast strain W303 (not shown). The efficacy of the three RGS proteins in functionally resembling the endogeneous Sst2 protein was also assessed by measuring the mRNA levels of the pheromone-inducible marker, Fus1, in the presence or absence of the ␣-factor. As shown in Fig. 4B, the RGS16 protein almost entirely abolished FUS1 transcript induction, whereas RGS2 and RGS5 had no effect on FUS1 induction. To compare the expression levels of different RGS proteins in the yeast transformants, the flag-tag was introduced to the N terminus of each RGS. Expression levels of each RGS protein in yeast cells grown in galactose were determined by Western blotting analysis using the monoclonal antibody specifically against the flag epitope. Comparable levels of RGS2, RGS5, and RGS16 proteins were detected (Fig. 4A, right lower panel). The antibody specificity was tested by also using protein extracts from different transformants grown in glucose (data not shown).
Differential Binding of RGS Proteins to G␣ Subunits-It has been demonstrated that many RGS proteins bind to the ␣-subunits of heterotrimeric G-proteins (10, 14 -18). We carried out biochemical binding assays to determine the binding efficiencies of RGS2, RGS5, and RGS16 to different G␣ subunits. Assays were carried out using bacterially expressed glutathione S-transferase fusion RGS proteins and in vitro translated G␣ subunits in the absence or presence of GTP␥S, GDP, or GDP plus AlF 4 Ϫ . In the presence of GTP␥S alone, none of the RGS proteins bound to any G-protein tested (data not shown). RGS5 and RGS16 strongly interacted with G␣ i3 , G␣ i2 , and G␣ o that had been preincubated with both GDP and AlF 4 Ϫ (Fig. 5). Interestingly, as previously reported with RGS1 (16), RGS5 and RGS16 also interacted, albeit to lower degrees, with the GDPbound forms of G␣ i2 and G␣ o . RGS2 did not bind to any of the G-proteins (Fig. 5). RGS2/G0S8 was first isolated as a cDNA specifically expressed in monocytes (4). Although we have shown that it is also expressed in other tissues at low levels, it is conceivable that RGS2 protein may only recognize G-proteins of a hematopoietic origin. Our results showed that all of the RGS proteins did not bind to the stimulatory subunit G␣ s (data not shown), confirming the previous findings (15)(16)(17).
Conserved Amino Acid Residues in the RGS Domain Are Critical for RGS Function-As the RGS domains among all the RGS proteins share significant identity in amino acid sequence, we generated seven RGS16 mutations within the RGS domain to alter residues that are absolutely conserved (Fig. 1B). These mutants will be informative in understanding whether those conserved residues are critical for either G-protein binding or the inhibition of G-protein signaling, or both. Halo assays showed that alterations of Gly 74 to Arg, Glu 85 to Gly, Glu-Asn 89/90 to Gly-Ala, Ile 116 to Asp, Leu-Met 161/162 to Ser-Lys, Arg-Phe 169/170 to Ser-Cys, respectively, abolished the inhibiting effect on pheromone signaling (Fig. 6A, and data not shown), although these mutants were expressed at a comparable level to that of the wild type and to that of L82S (Fig. 6B, and data not shown). The substitution of Leu 82 with Ser, however, had no effect on either the inhibition of pheromone signaling (Fig. 6), or its binding to the putative cognate G-proteins (Fig. 7). However, replacements of Glu-Asn 89/90 with Gly-Ala, or Arg-Phe 169/170 with Ser-Cys, eliminated binding to any Gprotein (Fig. 7). Unexpectedly, the mutant G74R and I116D FIG. 4. Yeast pheromone response assays. A, halo assays of pheromone response sensitivity of yeast US356 cells expressing different RGS members. pMW29 constructs each containing a full-length cDNA encoding RGS2, RGS5, or RGS16 were transformed into the US356 cells and assessed for their ability to attenuate the pheromone response pathway. Various amounts of ␣-factor (clockwise from the top of each dish, 300, 100, 30, 10, and 3 pmol) were applied to small discs of 3MM Whatmann paper. Relative pheromone sensitivity is measured by the size of the clear area elicited in response to a given dose of ␣-factor. The arrangement of the four dishes containing different yeast transformants is diagrammed on the top right. Identical results were obtained with flag-tagged RGS2, -5, and -16. On the bottom right is shown a Western analysis of RGS proteins in US356 cells grown in galactose. Three distinct bands (*) corresponding to relative molecular weights of RGS2, -5, and -16 were detected by anti-flag M2 antibody. B, FUS1 induction assay. Total cellular RNA was isolate from the untreated (Ϫ) and ␣-factor-treated (ϩ) yeast cells transformed with different RGS cDNAs, and subjected to Northern hybridization with the FUS1 probe. As a control for even RNA loading, the same blot was also probed with the act1 probe. proteins, which are no longer functional in the attenuation of pheromone signaling in yeast, had strong binding to the Gproteins in the presence of GDP and AlF 4 Ϫ , but did not bind to the G-proteins preincubated with GDP only (Figs. 6 and 7, and data not shown). Binding of mutant E85G and LM161/162SK proteins to G-proteins were not tested. These results indicate that the conserved residues are critical for RGS function, and yet the RGS domain may comprise distinct three-dimensional functional structures that are separately important for binding to cognate G-proteins and for GAP activity. DISCUSSION Employing a degenerate PCR approach, we have isolated a novel RGS member from the pituitary gland, as well as fulllength cDNA clones encoding the mouse RGS2 and rat RGS5. Tissue distribution analysis showed that these three genes are differentially expressed. RGS2 is most widely expressed, present in all the eight tissues examined except liver. RGS5 exhibits a very interesting expression pattern, mainly in heart and skeletal muscle, both of which are muscle tissues. However, extended analysis using the RNase protection assay shows that RGS5 is also present in pituitary, which correlates well with the fact that we obtained over 100 clones, out of one million screened, which hybridized to the RGS5 probe. In addition, it is also expressed in other tissues including all the 14 tissues shown in Fig. 3 except spleen, intestine, fat, and adrenal. In drastic contrast, RGS16 is only expressed abundantly in liver and pituitary, and at low levels in heart and brain, as assayed by both Northern blotting and RNase protection experiments.
It is therefore clear that the mouse RGS members have overlapping yet distinct patterns of tissue distributions. It has been well established that each eukaryotic cell possesses different subsets of receptors, G-proteins, and effectors, responding to distinct signals in the form of chemicals, hormones, and cytokines (reviewed in Refs. [31][32][33]. The differential tissue distribution of various RGS members suggests that they each may only regulate a certain aspect of biological function. Based on pheromone response halo assays and measurement of FUS1 induction, we showed that only RGS16 has an inhibitory effect on the yeast pheromone response pathway, whereas mouse RGS2 and RGS5 do not, although previous results show that RGS2 has a minor desensitizing effect on the sst2⌬ AG57 mutant cells (8). Functional selectivity has been shown for RGS1-4 in their ability to interfere with interleukin-8 receptor signaling (8). The simplest explanation for the functional difference among various RGS proteins is that they are divergent outside the RGS domain, especially in their N terminus regions and may thus have different functional targets. However, when the N termini of the RGS2 and RGS16 were exchanged for each other, the resulting N2/C16 (RGS2 N terminus replacing the corresponding N terminus of RGS16 at amino acid residues 98 and 79, respectively) and N16/C2 chimeras did not inhibit pheromone signaling as does the wild-type RGS16 (data not shown). In fact, the N2/C16 chimera had some residual inhibiting activity, indicating that functional specificity may be conferred by the conserved RGS domain. Nevertheless, deletion of the N-terminal 13-amino acid residues completely abolished RGS16 function in the attenuation of pheromone signaling in yeast, suggesting at least that the N terminus is critical for a functional conformation (data not shown).
RGS5 and RGS16 bind with high affinity to all of the three G-proteins when they are incubated with GDP and AlF 4 Ϫ , but not when they are in the active (GTP␥S bound) form. These results are consistent with what has been observed with RGS1, RGS4, and GAIP (15)(16)(17)(18). As previously suggested, RGS proteins bind and stablize G␣ proteins in their transition state leading to GTP hydrolysis (15,16,18,34,35). It is therefore likely that RGS5 and RGS16 both act as a GAP for their cognate G-proteins. However, it is also clear that RGS proteins do not only bind to G-proteins in the GTPase transition state,  6. Mutational analysis of RGS16. A, shown here are the results of halo assay on RGS16 mutants. Yeast US356 cells were transformed with pMW29 constructs expressing the wild-type RGS16 (WT), L82S, EN89/90GA, I116D, or RF169/170SC mutant RGS16 proteins, and were subjected to pheromone response halo assays as described in the legend to Fig. 4. The arrangement of the dishes containing different transformants is diagrammed at the top; three dosages of ␣-factor were used: clockwise from top, 30, 100, and 300 pmol. Similar results were obtained in halo assays with transformants expressing different flagtagged RGS16 proteins (data not shown). B, Western analysis of wild type RGS16 and its mutants in US356 cells. Protein extracts from yeast cells expressing different flag-tagged RGS 16 were subjected to Western immunoblot analysis using anti-flag M2 antibody. A band of approximately 24 kDa was detected in all cells except those transformed with the pMW29 vector alone.
as one study has shown that RGS10 binds strongly to a GTPase-inactive form of G␣ i3 (17), in which Gln 204 necessary for transition state stablization (34,35) is changed to Leu. However, the binding assays did not indicate which form(s) of G␣ i3 (Q204L), GDP-bound or GTP-bound, associates with RGS10. Similarly to what has been observed with RGS1 (16), RGS5 and -16 also significantly interact with G␣ proteins that are GDPbound, which is particularly true for their binding to G␣ i2 or G␣ o . It is noteworthy that RGS proteins can bind to some G-proteins either in the transition state (GDP/AlF bound) or in the inactive state (GDP bound), but not to G-proteins in the active state (GTP␥S bound). While binding to the transition state is a prerequisite for GAP activity, binding of RGS to GDP-bound G-proteins may be of as yet unidentified physiological importance. One possible function is that RGS proteins, in addition to enhancement of GTP hydrolysis, may decrease dissociation of GDP from G-protein to prolong the inactive state. It is equally possible that RGS binding may render the GDP⅐G␣ complex inaccessible to G-protein-coupled receptors.
Mutational analysis showed that most of the conserved amino acid residues in the RGS domain are critical for RGS functions. It is perplexing, however, that while mutations in EN89/90GA and RF169/170SC abolish both G-protein binding and attenuation activity of pheromone signaling, G74R and I116D mutations retain their ability to bind G-proteins. Interestingly, binding of G74R and I116D to G␣ i2 and G␣ o is strictly GDP/AlF 4 Ϫ -dependent. They no longer bind to the GDP-bound form of the G-proteins. This finding suggests there may be several separable, yet not mutually exclusive, functional structures that are critical for either the physical interaction of RGS proteins with G-proteins or GAP activity on G-proteins.