Abstract

INTRODUCTION

In vertebrates, gonadotropins and TSH are essential for the differentiation and growth of gonads and thyroid gland, respectively (1–4). Among these glycoprotein hormones with common α- and specific β- subunits, LH, FSH, and TSH are secreted by the anterior pituitary whereas the human chorionic gonadotropin (hCG) is secreted by placenta cells. These heterodimers bind specific plasma membrane receptors on target cells and signal mainly through the cAMP-dependent pathway. The receptors for these glycoprotein hormones belong to the large G protein- coupled, seven-transmembrane protein superfamily but are unique in having a large N-terminal extracellular (ecto-) domain important for interaction with the large glycoprotein hormone ligands (5,6). Hallmarks of this subgroup of G protein-coupled receptors (GPCRs) are the leucine-rich repeats in the ectodomain that have been postulated to form a horseshoe-shaped interaction motif for ligand binding (7–9). Recently, putative receptors homologous to the mammalian glycoprotein hormone receptors were found in sea anemone (10,11), nematode (12), pond snailLymnaea stagnalis (13), andDrosophila (14), suggesting that this subgroup of GPCR evolved early during evolution and that these invertebrate receptors represent ancient homologs of mammalian glycoprotein hormone receptors.

Based on the conserved sequences of mammalian glycoprotein hormone receptors and invertebrate homologs, we and others have recently isolated two novel mammalian leucine-rich repeat-containing, G protein-coupled receptors (LGRs) based on a homologous sequence search of the expressed sequence tags (ESTs) database (15–17). Because phylogenetic analysis showed that sea anemone LGR shares a closer relatedness to mammalian glycoprotein hormone receptors than to an LGR isolated from pond snailLymnaea stagnalis (13,15), one can predict that there are additional LGRs in mammalian genomes. Indeed, a recent search of the EST and genomic databases and subsequent characterization revealed that there are at least two additional mammalian LGRs. These two genes were named as LGR6 and LGR7 based on the chronological order of discovery. Analysis of primary sequences and domain arrangement in these LGRs showed that LGR6 is closely related to LGR4 and LGR5; whereas LGR7 and snail LGR are likely derived from a common ancestor. Together with the three known glycoprotein hormone receptors, these studies define the existence of three subgroups of LGRs in mammals. Based on the conserved mechanisms identified in constitutively activated LH and TSH receptors (18–20), studies of putative gain-of-function point mutants of LGR7 showed that this orphan receptor could mediate signaling through the protein kinase A-dependent pathway. Thus, site-directed mutagenesis of key residues in the functional domains of seven-transmembrane receptors provided a novel approach to reveal the signal transduction pathway of selective orphan GPCRs and to facilitate future identification of their cognate ligands.

RESULTS

cDNA Cloning of Two Novel Leucine-Rich Repeat-Containing GPCR, LGR6 and LGR7

Using the known invertebrate and mammalian LGRs as queries to search for homologous EST sequences, two candidate sets of sequences with homology to known LGRs were selected. cDNAs corresponding to these two novel genes, designated as LGR6 and LGR7, were obtained by reverse transcription-PCR using gene-specific primers and mRNA from human ovary and testis as templates. Using 5′- and 3′-rapid amplification of cDNA ends (RACE), a full-length LGR6 cDNA encoding an open reading frame (ORF) of 846 amino acids with a calculated molecular mass of 91.6 kDa was obtained (Fig. 1). Analysis of multiple cDNA clones showed that the deduced methionine translation start site is preceded with an in-frame stop codon at position −177. Using a similar approach, two splicing variants of LGR7 were obtained which differ in the N terminus of the coding region. The long form of LGR7, LGR7(1), is 3759 nucleotides long and encodes an ORF of 757 amino acids with a calculated molecular mass of 87 kDa; whereas the short form of LGR7, LGR7(2), encodes an ORF of 723 amino acids with a corresponding mass of 82.9 kDa (Fig. 2). The LGR7(2) variant is 34 amino acids shorter than the LGR7(1), and these two variants could be derived from alternative splicing events at the N terminus of the coding region.

Figure 1.

Comparison of Deduced Amino Acid Sequences for LGR6, LGR4, and LGR5 Sequence alignment for different regions of LGR6 with that of two closely related LGRs, LGR4 and LGR5. A, The ectodomain includes the multiple leucine-rich repeats (LRR) and a C-flanking cysteine-rich hinge region. Individual leucine-rich repeats are indicated byarrowheads. Putative N-glycosylation sites in the ectodomain are inbold letters andunderlined. Conserved PYAYQCC and GXFKPCE motifs preceding transmembrane I are inbold italics. B, The seven-transmembrane region and the C-terminal tail. The transmembrane (TM) domain, intracellular loop (IL), and extracellular loop (EL) are indicated byarrowheads. Potential protein kinase C phosphorylation sites in the intracellular C-terminal region are inbold letters andunderlined. The two conserved cysteines found in extracellular loops 1 and 2, and believed to form a disulfide bond, are marked bysolid circles. Residue numbers are shown on theright andasterisks indicate the position of the stop codon.Shaded residues are identical in the three receptor proteins aligned, and gaps indicated bydashes are included for optimal protein alignment.

Open in new tabDownload slide

Figure 2.

Comparison of Deduced Amino Acid Sequences between Two Human LGR7 Variants and Snail LGR Sequence alignment for different regions of two splicing variants (1,2 ) of LGR7 and snail LGR (LsLGR) was performed. A, Ectodomain with LDL receptor cysteine-rich motifs and leucine-rich repeats (LRR). Residues that are identical in LGR7(1 ) and LGR7(2 ) are shown asdouble lines, and residues that are missing in LGR7(2 ) are indicated bybold asterisks. The consensus LDL receptor cysteine-rich motif in LGR7 and the corresponding region of snail LGR are indicated bybold italics. C-terminal cysteine-rich hinge regions in the two receptors areoverlined. Putative N-linked glycosylation sites are inbold (N). B, The seven-transmembrane region and the C-terminal tail. Different structural motifs including transmembrane (TM) region, intracellular loop (IL), and extracellular loop (EL), are indicated byarrowheads;asterisks indicate the stop codon. Potential phosphorylation sites for protein kinase C are inbold letters andunderlined.Shaded residues are identical in the LGR7 and snail LGR sequences. Residue numbers are shown on theright and gaps are included for optimal protein alignment.

Open in new tabDownload slide

Comparison with known sequences in the GenBank showed that LGR6 and LGR7 are novel LGRs with closest relatedness to the newly isolated LGR4 and LGR5, and snail LGR, respectively. A motif search indicated that LGR6 and LGR7 contain an N terminus ectodomain composed of variable leucine-rich repeats and a seven-transmembrane region followed by unique C-terminal intracellular tail. The similarity between these LGRs and other known LGRs extends from the ectodomain to the seven-transmembrane region, and LGR6 shares approximately 45–47% identity with LGR4 and LGR5, as compared with 33% identity with the glycoprotein hormone receptors. However, LGR6 is distinct from LGR4 and LGR5 in having only 13 leucine-rich repeats instead of the 17 repeats found in LGR4 and LGR5 (Fig. 1A). Among the 13 leucine-rich repeats of LGR6, the first seven repeats align perfectly with leucine-rich repeats 1–3 and leucine-rich repeats 8–11 of LGR4 and LGR5, but the remaining leucine-rich repeats showed disparity in these receptors. In addition, distinctive sequence motifs, including a PYAYQCC motif and a GXFKPCE motif in the hinge region of the ectodomain, and the highly conserved cysteine residues for disulfide bond formation in extracellular loops 1 and 2, were completely conserved in LGR4, LGR5, and LGR6 (Fig. 1, A and B). A motif search using the ScanProsite program showed that both LGR6 and LGR7 contain N-glycosylation sites at the ectodomain and that LGR7 carries a protein kinase C phosphorylation site at the intracellular C-terminal sequences [STR, residues 750–752 of LGR7(1)]. Unlike LGR4 and LGR5 (15), both LGR6 and LGR7 lack protein kinase A phosphorylation sites in their C termini.

Among all known LGRs, LGR7 shares the highest identity with snail LGR (33%) (13) and less with the three mammalian glycoprotein hormone receptors (24%). As shown inFig. 2, LGR7 and snail (Ls) LGR shared similar primary sequences and common domain arrangement as shown by the presence of the N-terminal low density lipoprotein (LDL) receptor cysteine-rich motif followed by leucine-rich repeats and the seven transmembrane region. However, the predicted tertiary structure of the LGR7 ectodomain differed from that of snail LGR; the ectodomain of snail LGR is bulkier and contains approximately 760 amino acids instead of the 410 amino acids found in LGR7(1) (Fig. 2A). In addition to the 10 leucine-rich repeats at the C terminus of the ectodomain, snail LGR contains 12 LDL-receptor cysteine-rich motifs at the N terminus. In contrast, both LGR7 variants have only one such motif[ residue 40–62 of LGR7(1), CLPQLLHCNGVDDCGNQADEDNC] preceding the conserved leucine-rich repeat domain. In addition, these two receptors are distinct in their hinge region of the ectodomain. In LGR7 and snail LGR, the hinge regions are approximately 30 amino acids long as compared with 72–123 amino acids found in other LGRs. The unique PYAYQCC and GXFKPCE motifs found in this region of other LGRs are absent in these two receptors. Thus, the overall structural features of LGR7 are similar to that of snail LGR.

Mammalian LGRs Have Diverged into Three Distinct Subtypes

Phylogenetic analysis using the neighbor-joining and parsimony methods showed that the 11 known LGRs from vertebrates and invertebrates can be divided into three distinct subgroups (Fig. 3A). The first subgroup contains the mammalian gonadotropin and TSH receptors, and LGRs from sea anemone,Caenorhabditis elegans, andDrosophila. The second branch contains only vertebrate receptors including LGR4, LGR5, and LGR6, whereas snail LGR and LGR7 belong to the third subgroup. To gain insight into the evolution of LGRs, phylogenetic analysis was performed together with diverse GPCRs with a polypeptide or neurotransmitter ligand using full-length receptor sequences. As shown inFig. 3B, LGRs share a branch with diverse GPCRs known to have a peptide ligand, including a subgroup of angiotensin receptor-like GPCRs[ angiotensin receptor, platelet-activating factor (PAF) receptor, and formyl-methionyl-leucyl-phenylalanine (FMLP) receptor] and another subgroup of bombesin receptor-like GPCRs (bombesin, gastrin, thrombin, and neuropeptide Y receptors). In contrast, the relatedness to other family 1 GPCRs (21), such as receptors for various kinins, amine derivatives, and somatostatin (SST), is more remote.

Figure 3.

Three Subgroups of LGRs from Diverse Species and the Evolutionary Relatedness between LGRs and Family 1 GPCRs with a Peptide or Neurotransmitter Ligand A, Phylogenetic relatedness of diverse LGRs from mammals and invertebrates. Full-length amino acid sequences of 11 LGRs from mammals (LH, FSH, and TSH receptors plus LGR4 to LGR7), sea anemone, nematode, pond snail, andDrosophila were analyzed using the Blocks program. Sequence blocks predicted from this alignment were then used to construct the neighbor-joining tree for establishing possible subfamily relationships. Comparable results were obtained using the parsimony method in the Phylogeny Inference Package. In the three major branches of LGR evolution, the three classical glycoprotein hormone receptors as well as LGRs from sea anemone, nematode, andDrosophila belong to the first subgroup whereas LGR6 is most closely related to the second subgroup containing LGR4 and LGR5. In contrast, LGR7 belongs to a third subgroup together with snail LGR. B, Diagrammatic representation of phylogenetic relatedness between mammalian LGRs and diverse family 1 GPCRs (21 ). Different subgroups of family 1 GPCRs with known ligands were first grouped based on sequence similarity, and full-length consensus sequences for each subgroup of family 1 GPCRs were generated using the Blocks program. Subsequently, the individual full-length amino acid sequences of different LGRs, and the consensus sequences of different subtypes of family 1 GPCRs with a peptide or neurotransmitter ligand, were analyzed together. The GPCRs included in the phylogenetic tree are adenosine receptors, cholecystokinin receptors, tachykinin receptors, opsins, neurotensin receptors, TSH-releasing hormone receptor, angiotensin receptors, PAF receptor, FMLP receptor, bombesin receptors, gastrin receptors, thrombin receptors, neuropeptide Y receptors, 5-HT type 1 serotonin receptors, dopamine receptors, β-adrenergic receptors, GH secretogue receptor (GHS receptor), galanin receptors, somatostatin (SST) receptors, and endothelin receptors. R, Receptor.

Open in new tabDownload slide

Tissue Expression Pattern of LGR6 and LGR7

To investigate the tissue expression pattern of LGR6 and LGR7, Northern blot analyses of different tissues were performed using multiple tissue blots containing rat poly (A)+ RNA and specific radiolabeled probes for LGR6 or LGR7. Specific LGR6 transcripts of 4.0 kb were detected in multiple tissues including testis, ovary, oviduct, uterus, thymus, small intestine, colon, spleen, kidney, adrenal, brain, and heart (Fig. 4A). In addition, a minor LGR6 transcript of >5.5 kb was also detected in the oviduct and heart, whereas a smaller transcript was found in the spleen. Likewise, LGR7 transcripts were also found in multiple tissues. As shown inFig. 4B, a major LGR7 transcript of 5.5 kb was detected in all tissues tested with the exception of spleen. Furthermore, multiple smaller transcripts (2 or 3 kb) were detected in oviduct, uterus, colon, and brain.

Figure 4.

Tissue Expression Pattern of LGR6 and LGR7 For Northern blot analysis, 2 μg of poly (A)+-selected RNA from different tissues of immature rats were probed with a³²P-labeled LGR6 or LGR7 cRNA probe. After washing under highly stringent conditions (0.1 × SSC, 1% SDS at 75 C), the blots were exposed to x-ray films with intensifying screens at −80 C. Subsequent hybridization with a β-actin cDNA probe was performed to estimate nucleic acid loading. A, LGR6 Northern blot. A major message of 4.0 kb was found in testis, ovary, oviduct, uterus, small intestine, colon, spleen, kidney, adrenal, brain, and heart. A transcript of higher mol wt (>5.5 kb) was also detected in the oviduct and heart. B, LGR7 Northern blot. A major message of 5.5 kb was found in multiple tissues while multiple transcripts of lower sizes were detected in oviduct, uterus, colon, and brain.Lower panels show the expression of β-actin in the same blots. The sizes of mol wt markers are shown on theleft.

Open in new tabDownload slide

Isolation of Human LGR6 and LGR7 Genes and their Chromosomal Localization

Using an LGR6 cDNA fragment corresponding to the transmembrane region as a probe, genomic clones for LGR6 were isolated from a bacterial artificial chromosome-based human genomic DNA library. For LGR7, a genomic DNA clone (AQ053279) from the genomic survey sequence database was found to contain the C terminus of the LGR7 gene and was obtained from Genome Systems (St. Louis, MO). To identify the chromosomal localization of LGR6 and LGR7 genes in the human genome, genomic fragments (>50 Kb) derived from these BAC clones were used as probes in fluorescencein situ hybridization (FISH) analysis. As shown inFig. 5, A and B, LGR6 and LGR7 genes were localized to banded DNA in chromosome 1q32 and 4q32, respectively.

Figure 5.

Chromosomal Localization of LGR6 and LGR7 in Humans Using DNA fragments of bacterial artificial chromosome containing human genomic fragments of LGR6 and LGR7 as probes, LGR6 (upper panel) and LGR7 (lower panel) genes were localized to chromosome 1q32 and 4q32 regions, respectively, by the fluorescencein situ hybridization method. Denatured chromosomes from synchronous cultures of human lymphocytes were hybridized with biotinylated probes for localization. Assignment of the mapping data was achieved by superimposing fluorescence signals with the 4,6-diamdino-2-phenylindole-banded chromosome. Specific signals on the banded chromosome are indicated byarrows.

Open in new tabDownload slide

Signal Transduction by Mutant LGR7 as well as by Chimeric Receptors with the Ectodomain from the LH Receptor and the Transmembrane Region from LGR6 or LGR7

To investigate the signal transduction pathway of LGR6 and LGR7, wild-type and mutated receptors were constructed to analyze receptor binding and signal transduction. Earlier studies using mutant LH and TSH receptors found in patients with male-limited precocious puberty and nonimmune hyperthyroidism, respectively, indicated that these diseases are associated with constitutive receptor activation as a result of point mutations of key residues in the transmembrane VI region of these receptors (18–20,22–25). To investigate the possibility that the orphan LGRs could be rendered constitutively active in the absence of their unknown ligands, detailed comparisons of this region were conducted between gain-of-function LH and TSH receptors and the two new receptors. It was found that a tripeptide motif (FTD, residue 576–578 in the LH receptor) is completely conserved in the three glycoprotein hormone receptors and LGR7, but not in LGR6. Previous studies have shown that mutations of D578 in the LH receptor and the corresponding D633 in the TSH receptor cause constitutive increases of basal cAMP production in cells expressing these mutant receptors (20,22,26). Accordingly, we constructed mutants for the two LGR7 variants, LGR7(1) and LGR7(2), by generating a point mutation at the corresponding residue in transmembrane VI. These mutants were named as LGR7(1) D637Y and LGR7(2) D603Y.

As shown inFig. 6, transfection of 293T cells with increasing concentrations of the expression plasmid encoding the mutant LGR7 receptors [LGR7(1) D637Y or LGR7(2) D603Y], similar to those transfected with the plasmid encoding the D578Y gain-of-function mutant LH receptor, resulted in dose-dependent increases of basal cAMP production by transfected cells. In contrast, cAMP levels in cells transfected with different amounts of wild-type LH receptor or LGR7 did not show an increase in basal cAMP production. To allow quantitative comparison of basal cAMP production by different receptors, cAMP levels in transfected cells were normalized based on the level of cell surface expression of an N-terminally tagged FLAG epitope in different LGRs (Fig. 6, shown as percentage changesvs. wild-type LH receptor). Although the D637Y mutation caused significant increases of basal cAMP levels in both LGR7(1) and LGR7(2) constructs, cells expressing mutant LGR7(1) consistently showed greater levels of cAMP increase as compared with those expressing the mutant LGR7(2) construct. To evaluate the specificity of the activating LGR7 mutation for the Gs protein, the effect of this mutation on a different signal transduction pathway, phosphatidyl inositol (PI) turnover, was measured. As shown inTable 1, the Gs-activating mutants, LGR7(1) D637Y and LGR7(2) D603Y, did not stimulate inositol phosphate (IP) turnover by transfected 293T cells whereas hCG treatment significantly increased the IP content in cells expressing either wild-type or constitutively active mutant LH receptors.

Figure 6.

Gain-of-Function Mutants of LGR7(1 ) and LGR7(2 ) Leads to Constitutive Increases of Basal cAMP Production by Transfected 293T Cells Based on the gain-of-function point mutation (LHR D578Y) found in the LH receptor gene of patients with familial male-limited precocious puberty, LGR7(1 ) and LGR7(2 ) with homologous point mutation [LGR7(1 ) D637Y and LGR7(2 ) D603Y] were generated. The mutated residues in the TM VI of these two forms of LGR7 were identical, but the numbering of this mutation differs in these two LGR7 variants [D637 in LGR7(1 )vs. D603 in LGR7(2 )] due to dissimilar length in their N-termini. After transfection of expression constructs encoding wild-type or mutant receptors into 293T cells, basal cAMP levels were monitored using an RIA. Transfection of 293T cells with increasing concentrations (0–500 ng/well) of expression vectors encoding LGR7(1 ) D637Y, LGR7(2 ) D603Y, or LHR D578Y led to increases in basal cAMP levels in transfected cells. In contrast, cAMP levels in cells transfected with wild-type (WT) receptors were negligible. Production of cAMP by different receptors was normalized based on cell surface expression of FLAG epitope tagged at the N terminus of all receptor cDNAs determined by immunodetection.Numbers in parentheses denote percentage of receptor protein expression by individual constructs as compared with expression in cells transfected with the wild-type LH receptor construct (LHR WT), which was arbitrarily set as 100% (n = 3; mean ±se).

Open in new tabDownload slide

Because chimeric receptors among the known glycoprotein hormone receptors have been successfully used to study signal transduction by these proteins (27), we further generated chimeric LH/LGR6 and LH/LGR7 receptors by fusing the ectodomain of the human LH receptor with the transmembrane region and C-terminal tail of either LGR6[ LH(EC)/LGR6(TM)] or LGR7(1) [LH(EC)/LGR7(TM)] to further investigate the putative signal transduction pathway of these orphan receptors. Although the chimeric LH(EC)/LGR6(TM) receptor is expressed on the cell surface as reflected by FLAG epitope expression, and is binding to¹²⁵I-hCG as detected by the LH receptor binding assay (data not shown), no stimulation of cAMP was detected after incubation with hCG. In contrast, the same hCG treatment increased cAMP production by cells expressing the wild-type LH receptor. Because activation of the LH receptor by hCG elicits increased PI turnover (28,29), hydrolysis of PI in cells transfected with the chimeric receptor was also performed. Measurement of PI turnover showed a negligible increase of IP accumulation in cells expressing the chimeric LH(EC)/LGR6(TM) receptor after stimulation with hCG, whereas wild-type LH receptor-expressing cells showed an approximately 1.4-fold increase in PI hydrolysis after hCG treatment. In contrast to the chimeric receptor LH(EC)/LGR6(TM), the LH receptor ectodomain and LGR7 transmembrane region appeared to be incompatible because the chimeric LH(EC)/LGR7(TM) receptor was not found on the cell surface based on the assay of FLAG epitope expression.

DISCUSSION

Based on database analysis, we have isolated two novel GPCR genes belonging to the mammalian leucine-rich repeat-containing GPCRs. In addition to the subgroup of classical glycoprotein hormone receptors, LGR6 is evolutionarily related to two other mammalian LGRs (LGR4 and LGR5) whereas LGR7 shares a similar ancestry with snail LGR, thus indicating the presence of three subgroups of LGRs in the mammalian genome. The LGR7 was found to signal through the cAMP-dependent pathway, as reflected by increases in basal cAMP production in cells transfected with mutant LGR7 constructs carrying a point mutation in transmembrane VI. These orphan receptors are expressed in diverse tissues, and the identification of the signaling mechanism for LGR7 could facilitate the identification of its cognate ligand(s).

Recent expansion of the nucleic acid sequence database for diverse organisms have provided opportunities to identify novel mammalian gene paralogs through pairwise sequence comparison (30). Although the physiological roles of most new genes are not known, alignment of their primary or secondary structures have allowed their preliminary grouping. Studies on the entire nematode genome indicated that the GPCR protein superfamily represents one of the most abundant signaling molecules that allows the cell to communicate with its environment (31,32), and the GPCR family proteins could account for more than 1% of human genes. Among the known mammalian GPCRs, the glycoprotein hormone receptors represent a unique group of family 1 GPCRs with a large ectodomain for interaction with their ligands. Based on modeling with a prototypic leucine-rich repeat- containing polypeptide, ribonuclease inhibitor (8,9), it was envisioned that the multiple leucine-rich repeats in the ectodomain of mammalian glycoprotein hormone receptors are arranged in a horseshoe-shaped binding motif and the inwardly β-sheet structures interact with the specific ligand (7–9,33). Unlike the glycoprotein hormone receptors with nine leucine-rich repeats, newly isolated mammalian LGRs contain varying numbers of repeats, indicating that the binding domain in these receptors could have the same configuration, but with distinct curvature and size. The glycoprotein hormones have been widely used in the treatment of diverse diseases; the current finding of new LGRs and the possibility of finding additional hormones as ligands for these orphan LGRs could reveal novel endocrine regulatory mechanisms.

The physiological and pathophysiological actions of glycoprotein hormone receptors are mediated mainly through interaction with the Gs protein. Recent characterization of constitutively activated LH and TSH receptors, based on patients with specific etiology (18–20,25), has allowed analyses of structural requirements for signaling by these receptors. Taking advantage of these observations, mutant LGR7 constructs were made to explore the putative signaling pathways of LGR7. Similar to LH and TSH receptors, mutation of a single structure-determining amino acid in the transmembrane VI region of LGR7 caused constitutive signaling, as reflected by increases in basal cAMP levels in transfected cells. It has been proposed that mutations causing constitutive activation alter the receptor conformation from an inactive to an active state, mimicking the ligand stimulation of these receptors (34–36). Thus, the increase of basal cAMP production by mutant LGR7 receptors could be the result of conformational changes as found for constitutively active LH and TSH receptors. The present findings suggest that wild-type LGR7 may signal through the cAMP-protein kinase A pathway, a mechanism similar to the related glycoprotein hormone receptors. Although the Gs activating mutants LGR7(1) D637Y and LGR7(2) D603Y do not stimulate IP turnover by transfected 293T cells, it is important to note that the coupling of LGR7 to other G proteins cannot be excluded. Other examples of signaling by orphan GPCRs include a wild-type orphan ACCA (adenylate cyclase constitutive activator) receptor found to constitutively activate adenylate cyclase in transfected cells (37).

Recently, endogenous ligands for several orphan receptors have been isolated by monitoring the stimulation of downstream signaling transduction (38–44). With no knowledge of the G protein-signaling mechanism, several of these studies were made possible by coexpressing the orphan receptor with chimeric G proteins or G proteins that are promiscuous in receptor coupling. The present study demonstrates that mutagenesis of key residues in the transmembrane domain of GPCR is a novel approach to characterize signaling pathways for orphan GPCRs. While the present study demonstrated that mutagenesis of conserved residues is a useful approach for the characterization of signaling by the orphan LGR7, previous studies indicated that an FSH receptor mutant equivalent to D578Y in the LH receptor does not lead to receptor activation (27), suggesting this approach is only applicable to selective orphan GPCRs.

Because chimeric receptors with the ectodomain and transmembrane region from different glycoprotein hormone receptors have been shown to be functional (5,27), we have attempted to characterize the potential coupling of LGR6 to Gs or Gq pathways by fusing the transmembrane region of LGR6 with the ectodomain of the LH receptor in a chimeric construct. Although the chimeric receptor could be expressed on the cell surface, no activation of the chimeric receptor by the LH receptor ligand was observed. Previous studies on similar chimeric receptors containing the ectodomain domain of the LH receptor and the transmembrane domain of LGR4 or LGR5 also suggest that such chimeric receptors do not react to ligand stimulation (15). Thus, either the ectodomain of the LH receptor and the transmembrane region of these orphan receptors are incompatible, or they could signal through other unknown mechanisms. Indeed, a similar construct of the LH receptor ectodomain plus the LGR7 transmembrane domain failed to allow expression of the chimeric receptor on the cell surface of transfected cells.

The newly identified LGR7 shows close sequence homology with the only known snail LGR (13). Interestingly, the putative ligand-binding domains of these two receptors appear to have diverged. The most obvious difference is the number of LDL receptor cysteine-rich motifs in the N terminus of these two receptors. In the ectodomain of snail receptor, there are 12 LDL receptor cysteine-rich motifs, each encoding three conserved cysteine residues (13,45) important for disulfide bond formation and ligand binding (46). Unlike snail LGR, LGR7 contains only one typical LDL receptor cysteine-rich motif in its N terminus. Assuming the motifs in these two related receptors comprise part of their ligand-binding domain, the respective ligands for these two receptors could have diverged during evolution. Future studies on the LGR7 gene could reveal additional LGR7 variants that are closer to snail LGR, and uncover the evolutionary relationship of these two receptors.

Analysis of LGR6 sequences showed that LGR6 belongs to a subgroup of LGRs which includes LGR4 and LGR5, suggesting that these receptors may share similar ligand binding and signal transduction characteristics. However, the ectodomain of LGR6 is unique and contains only 13 leucine-rich repeats instead of the 17 repeats found in LGR4 and LGR5. It is unclear whether the cloned LGR6 cDNA is the only transcript encoded by the LGR6 gene. Observed differences in the number of leucine-rich repeats in the ectodomain of these receptors may be the result of alternative splicing. Recent studies on GPCRs have shown that within a given subfamily functional diversity is most often conferred by the existence of multiple receptor subtypes, each encoded by a distinct gene. Additional diversity results from alternative splicing of a given gene to form receptor variants (47). Indeed, our unpublished results showed that the LGR4 gene encodes multiple splicing variants, including one with only 14 leucine-rich repeats. Likewise, glycoprotein hormone receptor genes also encode multiple splicing variants with distinct functional characteristics (48), and two splicing variants of LGR7 were isolated here. Alternative splicing of these receptors, especially in the ectodomain, could result in alternative binding characteristics or specificity.

Sequence alignment showed that the three subgroups of LGRs could be distinguished merely by the amino acid sequences in their hinge regions between leucine-rich repeats and the transmembrane domain. In the glycoprotein hormone receptors, this region is flanked by the conserved YPSHCC and DXFNPCED motifs whereas LGR4, LGR5, and LGR6 contain YAYQCC and GXFKPCEX sequences in the corresponding regions, respectively. In contrast, these two motifs are absent in LGR7. Interestingly, recent studies on the TSH receptor have shown that point mutation of the serine residue in the conserved YPSHCC motif resulted in constitutive activation of the TSH receptor, leading to severe congenital hyperthyroidism in patients (49–51). Because LGR4, LGR5, and LGR6 share similar structural determinants in the hinge region, investigations on this region of the orphan receptors could provide insights toward the activation mechanisms of different LGRs.

Analysis using 11 known LGRs from diverse organisms indicated that LGRs from sea anemone, nematode, andDrosophila grouped under a single branch with mammalian glycoprotein hormone receptors while the newly isolated LGR4, LGR5, LGR6, and LGR7 diverged early during evolution. Assuming the evolutionary pressures on these receptors are constant, the sequence divergence in mammalian LGRs suggests that the ancestral gene giving rise to modern LGRs could have evolved before the emergence of cnidarians for cell-cell communication. Also of interest, analysis of the completely sequencedC. elegans genome showed that this nematode contains only one GPCR with LGR characteristics (12), suggesting a possible gene loss during the evolution of modern nematodes and that different LGRs in present day organisms evolved to serve adaptive functions in different phylogenies.

Findings of multiple LGRs allow a better comparison of the relatedness of the LGR subfamily with other GPCRs in the superfamily. Previous studies have shown that known GPCRs can be divided into six major families with distinct evolutionary origins, and the majority of GPCRs with a peptide or neurotransmitter ligand belong to family 1 (21). Phylogenetic analysis of LGRs with other GPCRs in family 1 indicated that LGRs belong to a distinct branch and share the closest relatedness with a subgroup of GPCRs including receptors for bombesin, gastrin, thrombin, neuropeptide Y, angiotensin, PAF, and FMLP. This classification would allow a better understanding of the ligand signaling mechanisms for these receptors through comparison of their structure-function relationship.

In conclusion, we have cloned two novel mammalian LGRs (LGR6 and LGR7) and identified the putative signal transduction pathway for LGR7. The constitutive activation of cAMP production by the mutant LGR7 suggests that this orphan receptor could signal through the cAMP-dependent pathway. This study represents the first demonstration of the elucidation of the signaling mechanism for an orphan GPCR based on single-point mutations to allow constitutive activation of the protein. The present study also defines the existence of three subclasses of leucine-rich repeat-containing GPCRs in the human genome and possibly other metazoans. Identification and functional characterization of these novel LGRs allow elucidation of the evolutionary relationship of this subfamily of GPCRs with leucine-rich repeats. It also facilitates future studies on the physiology of this expanding subgroup of GPCRs and the identification of their cognate ligands.

MATERIALS AND METHODS

Hormones and Reagents

Purified hCG (CR-129) was supplied by the National Hormone and Pituitary Program (NIDDK, NIH, Bethesda, MD). FLAG M1 antibody and FLAG peptide were purchased from Sigma (St. Louis, MO).¹²⁵I-Na and myo-[³H] inositol were purchased from Amersham Pharmacia Biotech (Arlington Heights, IL). Dowex AG1-X8 was from Bio-Rad Laboratories, Inc. (Hercules, CA).

Computational Analysis

cDNA sequences related to invertebrate LGRs and mammalian glycoprotein hormone receptors were identified from the EST and genomic survey sequences (GSS) database at the National Center for Biotechnology Information and an Incyte EST database using the BLAST and Gapped BLAST server with the BLOSUM62 comparison matrix (52). Initially, four independent human DNA entries (three ESTs and one GSS AQ053279) were found to encode sequences homologous, but not identical, to the known mammalian LGRs. Based on these sequences, RACE experiments were used to clone the full-length cDNA sequences. After RACE, the four original DNA entries were found to encode different portions of two novel LGRs. The alignment of primary sequences for genes in the LGR family was carried out by the Blocks WWW server (http://blocks.fhcrc.org/blocks/blockmkr). The Block Maker program also calculated the branching order and phylogenetic relatedness of aligned sequences by the Cobbler and Gibbs algorithms. To compare the phylogenetic relationship of LGRs with diverse peptide and neurotransmitter GPCRs, neighbor-joining (http://www.biophys.kyoto-u.ac.jp/maketree2.html) and parsimony methods (Phylogeny Inference Package) were used. In all studies, the full-length amino acid sequences of different GPCRs were used for phylogenetic analyses to provide the maximum possible information within families and subfamilies.

Additionally, the analyses of primary and secondary structures of these novel LGRs were conducted using the BCM search launcher (http://gc.bcm.tmc.edu:8088/search-launcher/launcher.html), Biology Workbench (http://gc.bcm.tmc.edu:8088/search-launcher/launcher.html), Blocks WWW server (http://www.blocks.fhcrc.org/blocks/), the eMotif maker server (http://dna.stanford.edu/emotif/), or the ExPASy Molecular Biology Server (http://expasy.hcuge.ch/).

Identification and Isolation of the Full-Length cDNAs for LGR6 and LGR7

For the isolation of full-length cDNA fragments, specific primers with a design based on EST sequences were used to prepare cDNA pools enriched with the candidate cDNAs derived from human ovary and testis mRNA. Two micrograms of mRNA were reverse transcribed by using 25 U of avian myoblastosis virus reverse transcriptase with oligo(dT) primer, 0.5 mm deoxynucleoside triphosphate (dNTP), and 20 U of RNAse inhibitor. After second strand synthesis with T4 DNA polymerase, the cDNA pool was tailed at both ends with adaptor sequences to allow PCR amplification using specific primers. The tailed cDNA products were then employed as a template for 5′- and 3′-RACE with adaptor- and gene-specific primers. All PCR amplifications were performed under highly stringent conditions (annealing temperature >67 C) using Advantage DNA polymerase (CLONTECH Laboratories, Inc., Palo Alto, CA) or Pfu DNA polymerase (Stratagene, La Jolla, CA) to minimize mismatching and infidelity during PCR amplification. PCR products were fractionated using agarose electrophoresis, and specific bands showing hybridization with radiolabeled cDNA probes were subcloned into the pUC18 vector (Invitrogen, San Diego, CA) to further identify candidate clones. The PCR products were phenol/chloroform-extracted, precipitated with ethanol, phosphorylated with T4 polynucleotide kinase, and blunt-ended with the Klenow enzyme. The PCR products were then subcloned into theSmaI site in the pUC18 vector. At least two independent PCR clones were sequenced to verify the authenticity of the coding sequences of each cDNA. The resulting sequences were assembled into contigs using the Blast2 sequences server (http://www.ncbi.nlm.nih.gov/gorf/bl2.html) and ClustalW 1.7 at the BCM Search Launcher (http://dot.imgen.bcm.tmc.edu:9331/multi-align). After the initial round of RACE, it was determined that the multiple homologous sequences belong to two independent LGR genes, LGR6 and LGR7. The full-length sequences were obtained and confirmed after three rounds of RACE.

Northern Blot Analysis

For mRNA analysis, membranes containing poly(A)+ RNA from various rat tissues were hybridized with specific cRNA probes. The rat mRNAs were extracted from 27-day-old rats using Trizol solution (Life Technologies, Inc., Gaithersburg, MD) followed by the Oligotex mRNA purification columns (QIAGEN Inc. Chatsworth, CA) to select for poly(A)+ RNA. The cRNA probes were synthesized using a Riboprobe Combination System (Promega Corp., Madison, WI). For hybridization using cRNA probes, membranes were prehybridized for 1 h at 60 C in the ExpressHyb solution (CLONTECH Laboratories, Inc.). This was followed by hybridization under the same conditions for 2 h but with 1× 10⁶ cpm/ml of³²P-labeled LGR6 or LGR7 cRNA probes. After hybridization, the membranes were washed twice in 0.2 × sodium chloride/sodium citrate (SSC), 0.5% SDS at 60 C, followed by two washes under high stringency conditions (0.1 × SSC, 1% SDS at 75 C) before exposure to RX film (Eastman Kodak Co., Rochester, NY) with intensifying screens (Amersham Pharmacia Biotech, Buckinghamshire, UK). To monitor the loading of mRNA samples from different tissues, membranes were stripped and rehybridized with a³²P-labled β-actin cDNA probe. The cDNA probe was generated by random priming (Life Technologies, Inc.). For hybridization using cDNA probes, membranes were washed to a stringency of 0.1 × SSC, 1% SDS at 60 C.

Expression of LGR6 and LGR7 in Mammalian Cells

Wild-type and mutant LGR6 and LGR7 cDNAs were constructed by sequential PCR amplification and standard restriction digest and ligation procedures. To allow efficient targeting of receptors to the cell surface and immunodetectionin vitro, a lead cDNA sequence containing a PRL signal peptide for cell surface expression (MNIKGSPWKGSLLLLL-VSNLLLCQSVAP) and a FLAG (DYKDDDDK) epitope were added to the N terminus of the mature region of LGRs in all expression constructs (53). To construct chimeric LH/LGR6 and LH/LGR7 receptors, junctional amino acid sequences were designed to be CAPEPPDAFN/PCEYLFESWGIRL and CAPEPPDAFN/SCEDLMSNHVLRVS, respectively. For expression in eukaryotic cells, the receptor cDNAs were subcloned into the eukaryotic cell expression vector pcDNA3.1 Zeo (Invitrogen, San Diego, CA), and the plasmids were purified using the Maxi plasmid preparation kit (QIAGEN, Inc.). Each construct was sequenced on both strands using vector-derived primers and gene-specific primers before use in transfection experiments for the analyses of signal transduction and/or receptor binding.

Mammalian 293T cells derived from human embryonic kidney fibroblast were maintained in DMEM/Ham’s F-12 (Life Technologies, Inc.) supplemented with 10% FBS, 100 μg/ml penicillin, 100μ g/ml streptomycin, and 2 mml-glutamine. The cells were transfected with expression plasmids using the calcium phosphate precipitation method (54). After 18–24 h of incubation with the calcium phosphate-DNA precipitates, media were replaced with DMEM/F12 containing 10% FBS. Forty-eight hours after transfection, cells were washed twice with Dulbecco’s PBS (D-PBS), harvested from culture dishes, and centrifuged at 400 ×g for 5 min. Cell pellets were then resuspended in DMEM/F12 supplemented with 1 mg/ml of BSA. Cells (2 × 10⁵/ml) were placed on 12-well tissue culture plates (Corning, Inc. Corning, NY) and preincubated at 37 C for 30 min in the presence of 0.25 mm 3-isobutyl-1-methyl xanthine (IBMX; Sigma) to prevent hydrolysis of cAMP before hormonal treatment for 16 h. To study basal cAMP production mediated by increasing numbers of receptors, each well was transfected separately with different amounts of the expression plasmid. To detect signaling by the wild-type and mutated LGRs, levels of cAMP production by transfected cells were measured by specific RIA using[¹²⁵I] cAMP (Amersham Pharmacia Biotech) (27). Cells transfected with the empty plasmid (mock) were routinely used as negative controls. At the end of incubation, cells and medium in each well were frozen and thawed once before heating at 95 C for 3 min to inactivate phosphodiesterase activity. Total cAMP in each well was measured in triplicate. For IP measurement, transfected cells were labeled for 24 h with myo-[³H]-inositol at 4 μCi/ml in inositol-free DMEM supplemented with 5% FBS. After washing three times with D-PBS, 2 × 10⁵cells were preincubated for 30 min in D-PBS containing 20 mm LiCl, and treated with or without hormones at 37 C for 1 h. Total IPs were extracted and separated as previously described (29). All experiments were repeated three times using cells from independent transfections. To monitor transfection efficiency, 0.5 μg of RSV-β-gal plasmid was routinely included in the transfection mixture, and theβ -galactosidase activity in the cell lysate was measured as previously described (55). Statistical analysis was performed using Student’st test.

Radioligand Binding Assays

For ligand binding analysis of the wild-type and mutant receptors, human CG (CR-129) was iodinated by the lactoperoxidase method (56) and characterized by a radioligand receptor assay using human LH receptors stably expressed in 293T cells. Specific activity and maximal binding of the labeled hCG were 100,000–150,000 cpm/ng and 40–50%, respectively. To estimate ligand binding to the cell surface, transfected cells were washed twice with D-PBS and collected in D-PBS before centrifugation at 400 ×g for 5 min. Pellets were resuspended in D-PBS containing 1 mg/ml BSA (binding assay buffer). Resuspended cells (2 × 10⁵/tube) were incubated with increasing doses (or a saturating dose) of labeled hCG at room temperature for 22 h in the presence or absence of unlabeled hCG (Pregnyl, 100 IU/tube; Organon, West Orange, NJ). At the end of incubation, cells were centrifuged and washed twice with the binding assay buffer. Radioactivity in the pellets was determined with a γ-spectrometer (53).

Determination of FLAG Epitope-Tagged Receptors on the Cell Surface

Transfected cells were washed twice with D-PBS, and resuspended cells (2 × 10⁶/tube) were incubated with FLAG M1 antibody (50 μg/ml) in Tris-buffered saline (pH 7.4) containing 5 mg/ml BSA and 2 mm CaCl₂ (assay buffer) for 4 h at room temperature in siliconized centrifuge tubes. Cells were then washed twice with 1 ml of assay buffer after centrifugation at 14,000 ×g for 15 sec. The¹²⁵I-labeled second antibody (antimouse IgG from sheep: ∼400,000 cpm/tube) was added to the resuspended cell pellet and incubated for 1 h at room temperature. Cells were again washed twice with 1 ml of assay buffer by repeated centrifugation before determination of radioactivity in cell pellets. Background binding was determined by adding excess amounts of the synthetic FLAG peptide at a concentration of 100 μg/ml.

Genomic Analysis and Chromosomal Localization of LGR6 and LGR7

To isolate genomic clones for LGR6, a human bacterial artificial chromosome (BAC) genomic DNA library was screened using the transmembrane region of LGR6 cDNA as a probe. The LGR7 genomic clone was identified by a sequence search of the GSS database and obtained from Genome Systems. These genomic fragments were then confirmed by Southern blot hybridization. For the identification of the chromosomal localization of LGR6 and LGR7 genes, genomic fragments (>50 kb) were used as probes for FISH of human metaphase chromosomes. Denatured chromosomes from synchronous cultures of human lymphocytes were hybridized with biotinylated probes for signal localization.

Acknowledgement

We are thankful to R. Slater (Department of Cytogenetics, Erasmus University, Rotterdam, The Netherlands) for the chromosomal localization of LGRs. We also thank Caren Spencer for editorial assistance. The GenBank accession numbers for LGR6 and LGR7 are AF190501 and AF190500, respectively.

This study was supported by NIH Grant HD-31566. S.Y.H. was supported by NIH Training Grant T32 DK-07217. The GenBank accession numbers for LGR6 and LGR7 are AF190501 and AF190500, respectively.

Email alerts

Citing articles via