Molecular Cloning and Characterization of a Novel Regulator of G-protein Signaling from Mouse Hematopoietic Stem Cells*

A novel regulator of G-protein signaling (RGS) has been isolated from a highly purified population of mouse long-term hematopoietic stem cells, and designated RGS18. It has 234 amino acids consisting of a central RGS box and short divergent NH 2 and COOH termini. The calculated molecular weight of RGS18 is 27,610 and the isoelectric point is 8.63. Mouse RGS18 is expressed from a single gene and shows tissue specific distribution. It is most highly expressed in bone marrow followed by fetal liver, spleen, and then lung. In bone marrow, RGS18 level is highest in long-term and short-term hematopoietic stem cells, and is decreased as they differentiate into more committed multiple progenitors. The human RGS18 ortholog has a tissue-specific expression pattern similar to that of mouse RGS18. Purified RGS18 interacts with the a subunit of both G i and G q subfamilies. The results of in vitro GTPase single-turn-over assays using G a i indicated that RGS18 accelerates the intrinsic GTPase activity of G a i m M by mixing 0.25 ml of pre-warmed with 0.25 ml of varying concentration of II at 37 °C. 30 min, 0.5 ml of 20% trichloroacetic was added and were centrifuged at 3 g for min. was five times neutralized and to 0.5-ml was five times M myo were M ammonium formate, M were was O chloroform and m ScintiVerse to determine lipid associated radioactivity. of RGS4 diminished its biological potency by 10,000-fold (54). It has been demonstrated that different RGS can differentially inhibit Ca 2 1 mobilization induced by carbachol, bombesin, and cholecytokinin, whose receptors are coupled to G q . The pattern of inhibition did not change regardless of G a q gene deletion (55), and deletion of the NH 2 -terminal region of RGS4 abolished receptor selectivity by carbachol and cholecytokinin (54). These results indicate that the specificity of RGS functions depend on their interaction with the G-protein-coupled receptor complex rather than a specific G a q .

and ␤␥ subunits then transmit signals through various signal transduction pathways. The activated ␣ subunit has slow intrinsic GTPase activity. When the ␣ subunit is in the GDPbound form, it re-associates with the ␤␥ subunits, leading to an inactive form. The duration of the G-protein signal depends on the rate of GTP hydrolysis and the rate of subunit re-association. For small GTP-binding proteins such as ras, there are GAP 1 proteins (GTPase activating protein), which increase the GTP hydrolysis rate. Recently, functional homologs of the ras-GAP have been identified for the heterotrimeric G-protein.
These are called RGS (regulator of G-protein signaling) proteins. The first RGS identified, Sst2 (supersensitivity to pheromone) in yeast, is a negative regulator of pheromone signaling (2). Later, the SST2 gene product was shown to function as a GAP for Gpa1, a molecule involved in pheromone desensitization (3). So far ϳ20 RGS have been identified (4 -12), and more could be anticipated. All RGS proteins have a highly conserved domain consisting of 120 amino acid residues, the RGS box, with varying lengths of NH 2 and COOH termini. RGS4 can be expressed in bacteria, and it has been co-crystallized with G␣ i1 as the GDP-AlF4 Ϫ -bound form (13). It was shown that RGS binds to G␣ i through the switch region, and that site-directed mutagenesis of the contact residues lead to loss of interaction (14 -16).
In vitro most purified native or recombinant RGS proteins can bind G␣ q and/or G␣ i via the RGS box (4,6,(17)(18)(19). Overexpression by transient transfection of a RGS into mammalian cells can attenuate signaling from G i and/or G q -linked receptors (20 -22). The very recently discovered RGS protein, p115RhoGEF, can act as a GTPase activator for G␣ 12 and G␣ 13 (23). No RGS that can act on G␣ s in mammals has been found so far. However, in yeast, Rgs2 was shown to function as a negative regulator of glucose-induced cAMP signaling through direct GTPase activation of the G␣ s protein Gpa2 (24).
In this paper, we describe cloning of a novel RGS from a long-term hematopoietic stem cell cDNA library. The new RGS, designated as RGS18, is highly expressed in long-term as well as short-term hematopoietic stem cells, and less in more committed hematopoietic populations. RGS18 can bind both G␣ i and G␣ q in vitro, enhance GTPase activity of G␣ i , and attenuate signals from G q -coupled receptors.
Materials-Dulbecco's modified Eagle's medium, glutamine, penicillin, streptomycin, and Trizol were obtained from Life Technologies, Inc., and RPMI 1640 was from BioWhittaker. Fetal bovine serum was from HyClone. Leupeptin  Myristylated G␣ i1 was expressed in Escherichia coli (JM109) and purified to homogeneity by anion exchange and hydrophobic interaction chromatography as described (35). Specific activity was 16 nmol/mg of protein as determined by [ 35 S]GTP␥S binding.
Construction of a cDNA Library from Long-term HSCs of C57 Bl/ Ka-Thy1.1 Mice-Twenty-eight thousand twice-sorted long-term HSCs were resuspended in 100 l of Trizol reagent containing 20 g of glycogen. Total RNA was isolated as described by manufacturer's instructions except the sample was re-extracted with 100 l of Trizol. Total RNA was precipitated with isopropyl alcohol followed by ethanol and then resuspended in water. cDNAs were synthesized using Cap-Finder cDNA synthesis kit (CLONTECH) with modifications. There were 7.5 million clones in the original library. To test the quality of the library, plasmid DNA from 150 random clones were isolated and sequenced using ABI 3700 sequencer.
Northern Analysis-Total RNA was isolated from various mouse tissues and human cell lines using Trizol reagent according to the manufacture's instructions. Poly(A) ϩ RNA was then isolated from total RNA using oligo(dT) paramagnetic beads (Dynal). Two micrograms of poly(A) ϩ RNA per tissue were separated on a 1% agarose/formaldehyde gel, and transferred to Hybond-XL (Amersham Pharmacia Biotech) or Zeta-Probe (Bio-Rad). The membrane was blocked with salmon sperm DNA at 0.1 mg/ml in ExpressHyb buffer (CLONTECH) for 1 h and then hybridized with 32 P-labeled NotI-EcoRI fragment of RGS18 for 1 h at 68°C. The membrane was washed twice with 2 ϫ SSC, 0.1% SDS at room temperature and then with 0.1 ϫ SSC, 0.1% SDS at 50°C, and exposed to film for 2 to 4 days. Human Multiple Tissue Northern blot and Human Immune System Multiple Tissue Northern blot II were obtained from CLONTECH, and hybridized with 32 P-labeled 0.5-kb EcoRI fragment of human expressed sequence tag clone za69c05. The probes were stripped and the membranes were reprobed with mouse ␤-actin cDNA.
Southern Analysis-Genomic DNA was isolated from mouse spleen according to Maniatis et al. (38). Fifteen micrograms of genomic DNA was digested with various restriction endonucleases, separated on a 0.7% agarose gel, and then transferred onto Hybond-N ϩ . The membrane was blocked with 100 g/ml salmon sperm DNA and hybridized with the 32 P-labeled 0.5-kb NotI-EcoRI fragment of RGS18 cDNA at 55°C overnight at 2 ϫ 10 6 cpm/ml of hybridization buffer (10 ϫ Denhardt's, 6 ϫ SSC, 0.1% SDS). The membrane was washed twice with 2 ϫ SSC, 0.1% SDS at room temperature and then with 0.2 ϫ SSC, 0.1% SDS at 65°C, and exposed to XAR-5 film (Kodak) at Ϫ70°C for 20 h.
RGS18 Antibody Production and Western Blotting of Mouse Tissue Extracts-Rabbit polyclonal sera were raised against bacterially expressed RGS18 containing the first 202 amino acids, and purified using Protein A column. Various tissues of a BA mouse were homogenized in Buffer B containing 0.2% ␤-mercaptoethanol with 10 strokes in Potter-Elvehjem tissue grinder, and centrifuged for 10 min in a microcentrifuge at 4°C. Thirty micrograms of protein were separated on a 12% SDS-polyacrylamide gel, and Western blot was performed as described (39) using anti-RGS18 antibody at 5 g/ml.
RT-PCR Analysis of RGS18 in Hematopoietic Progenitor Cells-Long-term and short-term HSCs, common lymphoid progenitors, common myelocyte progenitors, granulocyte macrophage progenitors, and megakaryocyte erythroid progenitors were isolated as described previously (36,40,41). Five thousand cells of each of the above populations were double-sorted on a Vantage fluorescence-activated cell sorter. To ensure the correct populations were isolated to sufficient purity, day 12 spleen colony assays were performed on 100 long-term and short-term HSCs from the above sort. Similarly, common myelocyte progenitors, granulocyte macrophage progenitors, and megakaryocyte erythroid progenitors were functionally assayed in methocellulose cultures according to Ref. 41. Total RNA was prepared from these fluorescenceactivated cell sorter-purified cells using the Qiagen RNeasy miniprep kit. RNA was treated with DNase I to eliminate residual DNA contamination prior to reverse transcription reaction. cDNA was obtained using Superscript II (Life Technologies, Inc.) according the manufacture's recommendations. PCR was performed using 32 P-labeled primers and KlenTaq-1 (CLONTECH). Hypoxanthine-guanine phosphoribosyltransferase control PCR was performed to normalize the amount of cDNAs to be used in the PCR reaction. Generally, 50 cells worth of cDNA was used to long term and short term-HSCs, 30 cells worth of cDNA for common myelocyte progenitors and granulocyte macrophage progenitors, and 20 cells worth for megakaryocyte erythroid progenitors. After 28 cycles of PCR, one-fifth of the products were run on polyacrylamide gels. Gels were dried and exposed to x-ray films or a PhosphorImager for data acquisition and analyses. RT-PCR of RGS18 from 2 g of total RNA isolated from thymus was negative (data not shown).
Binding of RGS to G␣-Three and a half million 293T cells were transfected with 10 g of various RGS constructs by the calciumphosphate method (42). Twenty-four hours after transfection, medium was changed, and cells were further grown for another 24 h. Cells were rinsed once with cold phosphate-buffered saline, and resuspended in Buffer A containing 50 mM HEPES, pH 8.0, 150 mM NaCl, 1% Triton X-100, 0.5 mM phenylmethylsulfonyl fluoride, 2 mM benzamidine, 5 g of pepstatin A, TLCK, and TPCK, and 10 g of leupeptin and soybean trypsin inhibitor per ml. After 15 min on ice, cell lysates were centrifuged in a microcentrifuge for 15 min at 4°C. The RGS proteins were immunoprecipitated from the supernatant with 20 l of anti-FLAG M2 affinity gel for 1 h at 4°C. The immunoprecipitates were washed three times with Buffer A. Jurkat cell extracts (50 million cells per point) were prepared as described (28) using Buffer B (Buffer A plus 1 mM MgCl 2 and 0.3 M NaCl). The extracts were treated with 30 M GDP alone or 30 M GDP, 30 M AlCl 3 , and 0.1 M NaF for 30 min at 30°C, and then incubated with immunoprecipitated RGS proteins for 1 h at 4°C. Bound proteins were washed once with Buffer B without Triton X-100, eluted by boiling in 30 l of SDS sample buffer, and separated on a 12% SDS-polyacrylamide gel. Proteins were transferred electrophoretically to nitrocellulose membrane (Schleicher & Schuell). Western blot was performed using polyclonal antibodies against the ␣ subunits of G i1ϩ2 (AS7 from PerkinElmer Life Sciences), G q/11 , G 12 , G 13 , and G s or FLAG (Santa Cruz).
Single Turnover GTP Hydrolysis Assay-[␥-32 P]GTP (1 M) was al-lowed to bind to 50 nM myristylated G␣ i1 for 15 min at 30°C in Buffer E consisting of 50 mM HEPES, pH 8, 5 mM EDTA, 100 mM NaCl, 0.1% Lubrol, and 1 mM dithiothreitol. After lowering the temperature to 4°C, single turnover GTP hydrolysis was initiated by mixing equal volumes of G␣ i1 preloaded with [␥-32 P]GTP and Buffer E plus 30 mM MgSO 4 , 400 M unlabeled GTP, and FLAG-tagged RGS proteins bound to M2agarose beads. The hydrolysis reaction was terminated by adding 1 ml of 15% (w/v) charcoal solution containing 50 mM NaH 2 PO 4 , pH 2.3, and placing samples on ice at the indicated time points. The charcoal was removed by centrifugation for 20 min at 4,000 ϫ g, and [␥-32 P]P i release was assessed by liquid scintillation counting of a 250-l aliquot of the supernatant in 4 ml of ScintiVerse.
Receptor Binding Assay-Angiotensin binding by whole cells was determined as described previously (43). Briefly, cells were harvested, washed, and resuspended in Buffer C (Opti-MEM, 0.1% bovine serum albumin, and 0.1 mg/ml bacitracin). The binding reaction was initiated by adding [ 125 I]angiotensin II at a final concentration of 100 nM into each cell suspension, and incubating at 37°C for 1 h. Unbound radioligands were filtered through a GF/B filter and the filters were washed three times with Buffer C. Cell-bound radioligands on filter was quantitated by ␥-counting. Nonspecific binding (less than 5% of the total) was determined by adding 1 M unlabeled Saralasin. Protein assay was performed on each sample according to Bradford (44). Total specific binding of angiotensin II was normalized to protein content.
Measurement of Inositol Phosphate Release-Inositol phosphate measurements were carried out as described (43). Seven hours after transfection, cells were incubated with myo-[ 3 H]inositol (10 Ci/ml) in Dulbecco's modified Eagle's medium for 24 h at 37°C. Cells were harvested in phosphate-buffered saline containing 0.02% EDTA, washed twice with ice-cold Buffer D (142 mM NaCl, 30 mM HEPES, pH 7.4, 5.6 mM KCl, 3.6 mM NaHCO 3 , 2.2 mM CaCl 2 , 1 mM MgCl 2 , and 1 mg/ml D-glucose), and then resuspended in ice-cold Buffer D containing 60 mM LiCl. The reaction was initiated by mixing 0.25 ml of pre-warmed cell suspension with 0.25 ml of varying concentration of angiotensin II at 37°C. After 30 min, 0.5 ml of 20% trichloroacetic acid was added and the samples were centrifuged at 4100 ϫ g for 20 min. The supernatant was extracted five times with ethyl ether, neutralized with sodium bicarbonate, and adsorbed to 0.5-ml Dowex AG1-X8 formate resin (50:50 slurry). Resin was washed five times with 2.5 ml of unlabeled 5 mM myo-inositol and inositol phosphates were eluted with 1 ml of 1.2 M ammonium formate, 0.1 M formic acid mixture. The eluates were counted by liquid scintillation counting in 10 ml of ScintiVerse. Released [ 3 H]inositol phosphates were normalized to the amount of [ 3 H]inositol incorporated into cellular lipids. The pellet after centrifugation was resuspended in 0.5 ml of H 2 O and 1.5 ml of chloroform/methanol, and vortexed vigorously. An additional 0.5 ml of H 2 O and 1.5 ml of chloroform were added, and a 200 l-aliquot of the organic phase was counted by liquid scintillation spectrophotometer in 10 ml of Scinti-Verse to determine lipid associated radioactivity.
Transcriptional Activation Assay-HEK293T cells (3.5 ϫ 10 6 ) were transfected with 2 g of pCMV-M1, 2 g of pCRE/␤-gal, and 8 g of control or FLAG-tagged RGS proteins. After 24 h, cells were serum starved for additional 24 h. Cells were stimulated with 1 mM carbachol for 6 h. Cell extracts were prepared and luciferase activity was measured using a luciferase assay kit (Promega) according to the manufacture's instruction.
Plasmids-FLAG-tagged RGS18 was generated by PCR using two primers (5Ј-CGGGTCATGAGATATGTCACTGGTTTTCTTCTC-3Ј and T3 primer) and RGS18 cloned in pBlueScript. The PCR reaction consists of 1 cycle of 2 min at 94°C, 30 cycles of 30 s at 94°C/30 s at 60°C/1 min at 72°C and 1 cycle of 10 min at 72°C. The PCR product was cleaned using Qiaspin mini-prep kit (Qiagen). After PCR, the DNA was digested with XbaI and ApaI, and ligated to pcFLAG. Human RGS2 was FLAG-tagged at the C terminus by PCR using primers (5Ј-TTCAG-GATCCAAGAGAGATACCACCATGCAAAGTGCTATGTTCTTG-3Ј and 5Ј-CTTCTCGAGTGTAGCATGAGGCTCTGTGGTG-3Ј). The PCR product was digested with BamHI and XhoI and ligated to pcFLAG. FLAGtagged rat RGS4 was provided by Dr. Robert McKenzie (Parke Davis, Ann Arbor, MI). The angiotensin receptor 1a cDNA in pCDM8 has been described previously (43). pCRE/␤-gal was provided by Dr. Roger Cone (Oregon Health Sciences University, Portland, OR), and pCMV-M1 was by Dr. J. Silvio Gutkind (National Health Institute, Bethesda, MD).

Cloning of a Novel Regulator of G-Protein Signaling from
Mouse Hematopoietic Stem Cell-The hematopoietic cells are constantly replenished by a self-replicating common precursor called the HSC. A large body of data on the biology of these cells has been accumulated. However, due to their rarity (less than 0.01% of the bone marrow cells; Ref 36) and the inability to grow these cells in vitro, there is little information regarding the molecular mechanisms that regulate stem cell functions. To better understand long-term self-renewing hematopoietic stem cells on the molecular level, a cDNA library was constructed from small numbers of highly purified long-term self-renewing hematopoietic stem cells. Approximately 150 clones were randomly chosen for sequencing to evaluate the quality of this library. The results of DNA sequencing indicated that the library contained ϳ50% previously unknown genes that are not present in the expressed sequence tag or GenBank TM data base (data not shown). One of the unknown clones showed limited homology to RGS (regulator of G-protein signaling), and this clone was further analyzed. The novel RGS will be referred to as RGS18. Complete sequencing and translation of the cDNA clone indicated that the clone contained the entire coding sequence (Fig. 1). The first ATG codon in the sequence is at nucleotide position 187, and conforms to the consensus sequence of Kozak (45). The base composition of the entire 1399 base pairs is 65.9% A ϩ T. In the 3Ј-untranslated region, a single polyadenylation signal sequence, AATAAA, is present at nucleotide position 1122, and three ATTTA or ATTTTA sequence motifs (46) are indicated (Fig. 1). In addition, a TTTT-GAT sequence motif followed by an AT-rich sequence is present in the 3Ј-untranslated region. This motif is present in immediate early genes and suggested to play a role in transcriptional activation (47,48). Translation of cDNA showed that RGS18 has 234 amino acids containing a central RGS box. A data base search using NCBI BLAST generated many nonredundant clones. The homology lies mostly within the RGS box (data not shown). Among the clones, RGS2 and RGS5 are most closely related to RGS18 ( Fig. 2A). RGS2 has 51% identity and 67% homology, and RGS5, 49% identity and 66% homology.
By searching the expressed sequence tag data base, we have found a human fetal lung expressed sequence tag clone (Gen-Bank TM accession number N98410), showing 85% identity spanning from nucleotide position 203 to 501. IMAGE clone 297800, from which the sequence was derived, was obtained and completely sequenced. In the coding sequence, the human clone has 86% identity at the nucleotide level and 82% identity and 90% homology at the protein level (Fig. 2B), strongly suggesting that the human clone is a RGS18 ortholog. The human ortholog had a longer 3Ј-untranslated region than its mouse counterpart (Fig. 1), and there are two polyadenylation signal sequences and three ATTTA or ATTTTA sequences in the 3Јuntranslated region. RGS18 proteins from both species contained putative phosphorylation sites for casein kinase II, protein kinase C, and protein kinase A (Fig. 2B) Expression of RGS18 mRNA in Tissues and Cells-Poly(A) ϩ RNAs isolated from different mouse tissues were analyzed by Northern using the RGS18 cDNA. The NotI-EcoRI fragment containing the 5Ј-untranslated region and the partial coding region detected a 2.4-kb transcript (Fig. 3A). The highest level of RGS18 expression was observed in bone marrow followed by spleen, fetal liver, and then lung. RGS18 was undetectable in brain, thymus, liver, kidney, and skeletal muscle. A very faint signal was seen in heart. The expression pattern of human RGS18 was also analyzed. In tissues, human RGS18 is highest in peripheral leukocytes followed by bone marrow, spleen, and fetal liver (Fig. 3B). Thymus, as well as lymph nodes, did not express human RGS18, similar to the mouse RGS18 expression pattern. No signal was detected in other tissues tested. In cultured cell lines, RGS18 was expressed only in the monocytic line U937, but not in Molt3 (acute lymphoblastic T-cell leuke-mic line), K562 (chronic myelogenous leukemic line), and Ramos (B-lymphocytes).
A rabbit antibody against recombinant RGS18 containing the first 202 amino acids was generated and used to test RGS18 expression in mouse tissue extracts (Fig. 3C). Anti-RGS18 recognized a specific protein with an apparent molecular mass of 26 kDa on a SDS-polyacrylamide gel (Fig. 3C, Ϫ) but not when the antibody was preincubated with the recombinant RGS18 polypeptide (Fig. 3C, ϩ). As predicted from the Northern blot, RGS18 was most highly expressed in the bone marrow.
To confirm expression of RGS18 in long-term self-renewing hematopoietic stem cells, cells at various stages of hematopoiesis were purified from bone marrow by fluorescence-activated cell sorter, and RT-PCR was performed (Fig. 4). Compared with hypoxanthine-guanine phosphoribosyltransferase control, RGS18 signal was highest in long-term and short-term HSCs, and the level was lower in common lymphoid progenitors, common myeloid progenitors, granulocyte macrophage progenitors, and megakaryocyte erythroid progenitors. This indicates that RGS18 is expressed more in the primitive cells, and is downregulated as cells differentiate to more committed linages.
Southern Analysis of RGS18 -Mouse genomic DNA was digested with BamHI, EcoRI, or HindIII, and transferred to the membrane, and hybridized with the 0.5 kb of 5Ј end of RGS18 cDNA (Fig. 5). BamHI, EcoRI, and HindIII generated single bands of 9, 2.8, and 6 kb, respectively, suggesting that there is a single copy for RGS18.
Binding of RGS18 to G␣ i and G␣ q from Jurkat T Leukemic Cell Extracts-From sequence comparison, RGS18 showed the most homology to RGS2 and RGS5. RGS2 has been shown to selectively bind and inhibit G␣ q function (20). RGS5 can bind both G␣ i and G␣ q (9). To determine which G-protein signaling pathway RGS18 might act on, binding of RGS18 to endogenous G␣ protein was analyzed. HEK293T cells were transfected with the plasmids carrying FLAG-tagged RGS2, RGS4, and RGS18 cDNA. RGS proteins were immunoprecipitated with anti-FLAG M2 antibody coupled to agarose beads, and incubated with Jurkat cell extracts to facilitate binding to endogenous G␣ proteins (Fig. 6). In has been shown that RGS binds G␣ with high affinity when G␣ is complexed with GDP-AlF 4 Ϫ , which mimics the transition state during GTP hydrolysis. As shown in Fig. 6, the RGS proteins bound G␣ only in the transition state (Fig. 6, ϩ AlF 4 Ϫ ). No binding was observed in the GDPbound state (Fig. 6, Ϫ AlF 4 Ϫ ). The amount of different RGS proteins used in the reaction was similar (Fig. 6, FLAG). No bound G␣ protein was seen with the immunoprecipitates prepared from cells transfected with control plasmid (Fig. 6, pc-FLAG). As previously shown (20), RGS2 did interact with G␣ q but not with G␣ i , and RGS4 was able to bind both G␣ i and G␣ q . RGS18 was also able to interact with G␣ i and G␣ q . However, RGS18 did not bind G␣ 12 , G␣ 13 , or G␣ s (data not shown).
GAP Activity of RGS18 -The ability of RGS18 to stimulate GTPase activity of G␣ i1 was compared with other RGS proteins. To obtain large amounts of RGS18 protein, His-tagged RGS18 was expressed in bacteria. However, recombinant protein was insoluble. Therefore, HEK293T cells were transfected with FLAG-tagged RGS plasmids, and RGS proteins were immunoprecipitated as described before. To normalize the amount of RGS proteins in the assays, the immunoprecipitates were resolved on a SDS-polyacrylamide gel and stained with Coomassie Blue. RGS proteins were scanned with a densitometer, and the RGS was normalized with FLAG-agarose beads. Immunoprecipitates prepared from HEK293T cells transfected with pcFLAG showed no stimulation of GTPase activity (Fig. 7,  Vector). As shown before, RGS2 showed no GTPase activity toward G␣ i1 . RGS4 dramatically enhanced endogenous GTPase activity of G␣ i1 . RGS18 also stimulated GTPase activity but not as much as RGS4. RGS4 reduced the calculated t1 ⁄2 for P i release of G␣ i1 from 1.04 to 0.19 min and RGS18 reduced t1 ⁄2 to 0.56 min.
Inhibition of G q -mediated Signaling by RGS18 -Since RGS18 was able to bind the G␣ q subunit, biological assays were used to determine whether this interaction has functional significance. If RGS18 can modulate a signal from G q -coupled receptors, it will be indicative of a functional interaction with G␣ q . HEK293T cells were co-transfected with angiotensin 1a receptor plasmid and a FLAG-RGS or control plasmid. Angiotensin 1a receptor has shown to be coupled to the G q signaling  (Fig. 8A). RGS18 was also able to attenuate G q signaling mediated by angiotensin II. There was no difference in the amount of [ 125 I]angiotensin binding to the cells transfected with the RGS constructs or the empty vector (data not shown). Next, we tested whether RGS18 could attenuate G q -mediated transcriptional activity. HEK293T cells were transfected with RGS or control plasmid, and M1 muscarinic receptor and pCRE/␤-gal. It has been shown that activation of M1 muscarinic receptor, which couples G q protein, resulted in transcriptional activation through binding of cAMP responsive element-binding protein to cAMP responsive element (50). Carbachol treatment of cells transfected with control plasmid showed ϳ20-fold activation of transcription of the reporter gene (Fig. 8B). All RGS constructs inhibited transcriptional activation. RGS2 inhibited activation by 75%, RGS4 by 71%, and RGS18 by 77.5% of the control.
These results indicate that RGS18 can modulate signals from G q -coupled receptors. DISCUSSION In this paper, we report the cloning of a novel RGS from mouse long-term self-renewing hematopoietic stem cells. The sequence surrounding the third ATG located at nucleotide position 187 was in agreement with the Kozak's consensus se-FIG. 3. Tissue-specific expression of RGS18. A, Northern analysis of mouse RGS18. Two micrograms of poly(A) ϩ RNA isolated from various tissues were separated on a 1% agarose-formaldehyde gel, and then transferred to Hybond-XL membrane. The membrane was hybridized with mouse RGS18 cDNA as described under "Experimental Procedures." B, Northern analysis of human RGS18. Human Multiple Tissue Northern blot and Human Immune System Multiple Tissue Northern blot II were obtained from CLONTECH and hybridized with human RGS18 cDNA. Note a similar tissue expression pattern between the two species. C, Western analysis of mouse RGS18. Mouse tissue extracts were separated on a 12% SDS-polyacrylamide gel, transferred onto nitrocellulose membrane, and incubated with rabbit antibody raised against mouse RGS18. Ϫ and ϩ indicate antibody has been preincubated in the absense or presence of with 30-fold excess antigen protein, respectively. RGS18 was visualized with ECL reagent.
FIG. 4. RT-PCR of mouse RGS18 from early hematopoietic progenitors. Cells from early stages of hematopoiesis were isolated from mice bone marrow, and RT-PCR was performed as described under "Experimental Procedures." LT-HSC, long-term self-renewing hematopoietic stem cells; ST-HSC, short-term hematopoietic stem cells; CLP, common lymphoid progenitors; CML, common myeloid progenitors; GMP, granulocyte macrophage progenitors; MEP, megakaryocyte erythroid progenitors.
FIG. 5. Southern analysis of RGS18. Mouse genomic DNA (50 g) was digested with indicated restriction endonucleases and separated on a 0.7% agarose gel. DNA was transferred onto the Hybond-XL membrane and probed with the RGS18 cDNA fragment as described under "Experimental Procedures." quence for eukaryotic initiation codons (45). The 702-nucleotide open reading frame encodes a polypeptide of 234 residues. This new RGS protein was designated as RGS18. The entire RGS18 cDNA is A ϩ T-rich, and the 3Ј-untranslated region contains three ATTTA motifs. These features have been linked to mRNA stability (46) and translational control (51). The presence of these structures suggests that expression of RGS18 could be highly regulated. Both mouse and human RGS18 were expressed as a 2.4-kb transcript as determined by Northern hybridization, and showed a hematopoietic tissue-specific expression pattern, with the highest levels in peripheral leukocytes and bone marrow followed by fetal liver and spleen (Fig. 4). There was no RGS18 message detected in thymus and lymph nodes. In cultured cells, only monocytic U937 but not B, T, and myelocyte-derived cell lines expressed RGS18. RT-PCR of RGS18 from cells at the early stages of hematopoiesis indicated that RGS18 is highly expressed in both long-term and shortterm HSCs, and less so in cells with more committed lineages (Fig. 4). RGS18 protein can bind G␣ i and G␣ q (Fig. 6). In in vitro GTPase assays, RGS18 enhanced the intrinsic GTPase activity of G␣ i1 to a lower extent compared with RGS4 (Fig. 7). RGS2, which acts only on G q , was not able to stimulate the GTPase activity. Furthermore, RGS18 inhibited the inositol phosphates production mediated by angiotensin 1a receptor and transcriptional activation mediated by M1 muscarinic receptor in HEK293T cells (Fig. 8). Even though the RGS18 sequence is more homologous to RGS2 than RGS4, it clearly interacts with G␣ i as well as G␣ q . Effects of RGS18 on the G i pathway are contradictory. Therefore, it is necessary to study gain of function and/or loss of function mice to verify the role of RGS18 in G i pathway in vivo.
There are over 20 RGS genes cloned so far, but their in vivo regulation is not well understood. There are several ways in which cells can regulate RGS functions. First, many RGS proteins are expressed in tissue and cell-type specific manners. For example, some RGS proteins are abundant in lymphocytes FIG. 6. Interaction between RGS18 and G␣ proteins. RGS proteins were immunoprecipitated from transfected HEK293T cells using anti FLAG-M2-agarose beads and incubated with Jurkat cell extracts in the presence or absence of AlF 4 Ϫ formation. The bound proteins were separated on a 12% SDS-polyacrylamide gel and transferred for Western blot using antibodies against G␣ q , G␣ i1ϩ2 , G␣ 12 , G␣ 13 , G␣ s , and FLAG. Proteins were visualized with ECL reagents. Inositol phosphates were measured as described under "Experimental Procedures." Data are expressed as percentage of maximum inositol phosphate release by AT1R plus vector-transfected cells, and are means Ϯ S.E. from four independent experiments, each performed in duplicate. 125 I-Angiotensin II binding to each transfected cell was 2.2 Ϯ 1.7% for pcFLAG vector alone, 100% for AT1R and pcFLAG, 74 Ϯ 10% for AT1R and FLAG-RGS2, 94 Ϯ 8% for AT1R and FLAG-RGS4, and 98 Ϯ 16% for AT1R and FLAG-RGS18. Statistical analysis by ANOVA with Bonferroni corrected post-tests showed that p Ͼ 0.05 for differences between receptor alone and coexpression of RGS18 proteins. B, RGS inhibits carbachol-induced transcription of reporter gene. HEK293T cells were transfected with pCRE/ ␤-gal, pCMV-M1, and various FLAG-RGS plasmids. Twenty-four h after transfection, cells were starved for serum for 24 h, and then stimulated with 1 mM carbachol for 6 h. Cell extracts were prepared and ␤-galactosidase activity was measured as described under "Experimental Procedures." Data are expressed as means Ϯ S.E. from four separate experiments. and monocytes (8,25,26), brain (30 -33), and rods (34). Furthermore, the level of RGS can be modulated under certain conditions. For example, in antigen-activated B cells, RGS1 and RGS2 are up-regulated and RGS3 and RGS14 are downregulated (29). RGS1 and RGS2 are also up-regulated in phorbol ester-stimulated B cells and ConA-and cyclohexamidetreated human blood mononuclear cells (25). In vascular smooth muscle, RGS2 message was rapidly increased upon angiotensin stimulation (52). RGS16 expression is induced in human T cells by IL-2 and the induction was diminished by cAMP. RGS2 expression, however, was reciprocated (28). RGS18 also showed a tissue and cell-type specific expression pattern. It is expressed highly in long-term and short-term HSCs, and its level is decreased as these cells are more committed to differentiated pathways. In mature cells, RGS18 appears to be most highly expressed in peripheral blood leukocytes of myelomonocytic lineage.
Another way to regulate the RGS activity is by regulating specific interaction between the G␣. The RGS boxes interact with the switch regions of G␣, and these interactions are required for the GAP activity. Therefore, the specific interaction with G␣ would be determined by divergent sequences outside of the RGS box. The first evidence that the RGS box alone might not be enough to function normally in vivo comes from the Sst2 complementation assay in yeast (53). The full-length RGS16 protein could bind and function as a GAP for G␣ i and G␣ o in vitro, and attenuated pheromone signaling. The RGS16 core domain was also able to bind G␣ and enhance GTPase activity in vitro; however, the mutants lacking the NH 2 -terminal region were unable to attenuate pheromone signaling (53). Further evidence for the requirement of the non-RGS box is that deletion of the NH 2 -terminal domain of RGS4 diminished its biological potency by 10,000-fold (54). It has been demonstrated that different RGS can differentially inhibit Ca 2ϩ mobilization induced by carbachol, bombesin, and cholecytokinin, whose receptors are coupled to G q . The pattern of inhibition did not change regardless of G␣ q gene deletion (55), and deletion of the NH 2 -terminal region of RGS4 abolished receptor selectivity by carbachol and cholecytokinin (54). These results indicate that the specificity of RGS functions depend on their interaction with the G-protein-coupled receptor complex rather than a specific G␣ q . RGS proteins show differential subcellular distribution. Using confocal microscopy, Chatterjee and Fisher (56) have shown that RGS2 and RGS10 are present in the nucleus, that RGS4 and RGS16 are in the cytoplasm, and that RGSZ is localized to the trans-Golgi network. RGS-GAIP was also shown to be associated with Golgi membranes (57). The deletion of the NH 2terminal 15 residues of RGS4 and RGS16 resulted in nuclear accumulation of the RGS proteins. The deleted sequence contained a nuclear export signal, and mutations of the conserved leucine residues also resulted in nuclear accumulation. These data seem inconsistent with the report that RGS4 (58) and RGS16 (59) are associated with the membrane. Chen et al. (59) showed that amino acid residues 7 to 32 of RGS16 are required for RGS16 membrane association. Mutation of the second leucine led to the loss of RGS16 biological activity and membrane association. They proposed that the NH 2 -terminal domain contained an amphipathic structure that was responsible for membrane binding. Furthermore, Srinivasa et al. (58) showed that the first 33 amino acid residues are required for the membrane binding of RGS4. Since the NH 2 -terminal sequences of RGS4, RGS5, and RGS16 are conserved, they might share the same type of regulation. Indirect immunofluorescent staining of FLAG-tagged RGS18 showed that RGS18 is localized exclusively in the cytoplasm (data not shown). Further-more, there is a possible nuclear export signal (NES, L 5 XXFXXL, Ref. 60) in RGS18. Studies using the NH 2 -terminal deletion and point mutants of RGS18 would verify this sequence functions as a NES.
RGS proteins undergo post-translational modification. It has been shown that RGS-GAIP is phosphorylated at serine 24 by casein kinase II on clathrin-coated vesicles, and that the phosphorylated form is associated with the membrane (61). RGS18 has two putative casein kinase II sites, one protein kinase C site, and one protein kinase A site outside of the RGS box. It is thus possible that phosphorylation of RGS at these residues might regulate intracellular localization and/or functions of RGS18. Recently, it has been demonstrated that two cysteine residues at positions 2 and 12 of RGS16 are palmitoylated in vivo (62). Mutation of either of these residues decreased RGS16 activity in both G␣ i and G␣ q signaling pathways induced by isopreterenol/somatostatin and carbachol, respectively. Cysteine mutation did not significantly affect the cellular localization of RGS16 and in vitro GAP activity, suggesting that reversible palmitoylation of the protein might be important for biological activity of RGS16. This would also be true for RGS4 and RGS5 whose sequences are conserved at the NH 2 terminus. In RGS18 there is no amphipathic structure at the NH 2 terminus, and no possible palmitoylation site, suggesting a different mode of regulation for RGS18.
The fact that RGS18 is highly expressed in HSC and mature myelomonocyte compartment and that many other RGS proteins are lymphoid specific suggest that RGS proteins may have functions in regulation of hematolymphoid systems. For example, lymphocyte migration during inflammatory response is induced by a number of chemokines, whose receptors are coupled to G i . RGS1, RGS3, RGS4, and RGS14, which are expressed by the lymphoid system, can inhibit chemotaxis induced by various chemokines including proinflammatory factors, stromal cell-derived factor-1␣, and Epstein-Barr virusinduced molecule 1 ligand (26,29,63,64). RGS2, which acts on G q , showed no effect on chemotaxis. SDF-1␣ is the ligand for CXCR4 (65), which is expressed by various cells including HSCs (66). During the fetal development, hematopoiesis occurs in the fetal liver. As the fetus develops, hematopoiesis moves to the bone marrow. In adult, bone marrow is the primary site for hematopoiesis. It has been shown that in mice lacking SDF-1␣, bone marrow hematopoiesis was absent, even though fetal liver hematopoiesis was normal (67). This suggests that SDF-1␣, expressed by bone marrow stromal cells, is responsible for migration of HSCs from fetal liver to bone marrow. Even though many reports show that multiple RGS proteins could modulate SDF-1␣-induced chemotaxis, it is possible that there might be specificity of RGS to regulate different chemokine receptors. Since RGS18 can interact with G␣ i , it would be of interest to test the possible role of RGS18 in inflammatory response and HSC migration.