A G protein-coupled receptor for UDP-glucose.

Uridine 5'-diphosphoglucose (UDP-glucose) has a well established biochemical role as a glycosyl donor in the enzymatic biosynthesis of carbohydrates. It is less well known that UDP-glucose may possess pharmacological activity, suggesting that a receptor for this molecule may exist. Here, we show that UDP-glucose, and some closely related molecules, potently activate the orphan G protein-coupled receptor KIAA0001 heterologously expressed in yeast or mammalian cells. Nucleotides known to activate P2Y receptors were inactive, indicating the distinctly novel pharmacology of this receptor. The receptor is expressed in a wide variety of human tissues, including many regions of the brain. These data suggest that some sugar-nucleotides may serve important physiological roles as extracellular signaling molecules in addition to their familiar role in intermediary metabolism.

G protein-coupled receptors (GPCRs) 1 are a large family of receptors uniquely characterized by their ability to detect a diverse variety of extracellular signals such as single photons, odorants, inorganic ions, nucleotides, biogenic amines, chemokines, lipids, proteases, amino acids, and peptides. In recent years, a large number of DNA sequences have been identified that encode novel proteins with many of the sequence motifs characteristic of GPCRs but for which no natural ligand has been identified. These putative GPCRs have been termed "orphan" receptors. Recently, naturally occurring ligands have been identified for a number of orphans, using the recombinant orphan receptor as the specific sensor component of a bioassay. Tissue extracts have often provided the source of these ligands, but some orphans have been identified by mass screening of large libraries of known or putative GPCR ligands (1)(2)(3). Here, we describe how we have employed this latter strategy to identify a naturally occurring ligand for the orphan receptor KIAA0001 (4). KIAA0001 (GenBank TM accession number D13626) was originally cloned from an immature human myeloid cell line (4) and contains a number of features typical of members of the GPCR superfamily, including a DRY motif at the boundary of transmembrane domain 3 and the second intracellular loop, consensus sites for asparagine-linked glycosylation on extracellular sequences, and consensus sites for protein kinases A and C phosphorylation on the third intracellular loop. Recently, a putative rat ortholog, termed VTR 15-20, with 81% amino acid identity to KIAA0001 and conserved substitutions was cloned from ventral tegmental tissue (5). This orphan is expressed widely throughout the mammalian nervous system and in rat primary microglial and astrocyte cultures. Furthermore, expression of VTR 15-20 in the brain and spleen is regulated by immunologic challenge. These data have implicated this orphan in neuroimmune function (5). We demonstrate that KIAA0001 and VTR 15-20 are likely to be orthologs and that KIAA0001, like VTR 15-20, is expressed widely throughout the brain and in many other regions of the body.

EXPERIMENTAL PROCEDURES
Bioinformatics-The phylogenetic tree (length ϭ 2167, rescaled consistency index ϭ 0.27, retention index ϭ 0.41) was generated from polypeptide sequences obtained from GenBank TM and Swissprot, using PAUP* (6) with 10,000 replicates of an heuristic search. Identities and similarities were calculated from hand-adjusted alignments.
Screening Yeast Strains-Lawns of yeast were prepared by pouring 1 A 600 unit of cells resuspended in 10 ml of SD-Leu-Trp, pH 6.8, 0.7% agar, onto a solid base of SD-Leu-Trp, pH 6.8, growth media. 1-l aliquots of 1 mM solutions of putative agonists were transferred from 96-well plates to yeast lawns. After an overnight incubation at 30°C, the assay was developed by the addition of 10 ml of lysis/developer solution, pH 7.0 (0.5% agar, 0.58 M Na 2 HPO 4 , 0.42 M NaH 2 PO 4 , 0.1% * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. b These authors contributed equally to this work. c To whom correspondence should be addressed. p Current address: Myriad Genetics, Inc., 320 Wakara Way, Salt Lake City, UT 84108. 1 The abbreviations used are: GPCR, G protein-coupled receptor; HPLC, high pressure liquid chromatography; Pipes, 1,4-piperazinediethanesulfonic acid; RT-PCR, reverse transcriptase-polymerase chain reaction. sodium dodecyl sulfate, and 0.05% 5-bromo-4-chloro-3-indolyl ␤-D-galactopyranoside). The presence of an agonist among the test compounds was evidenced by the development of a blue color at the point of deposition of that compound. A library of over 700 known and putative natural GPCR agonists was assembled, mostly from commercial sources, and screened in this assay. The library included neuropeptides, bioactive lipids (leukotrienes, prostaglandins, platelet-activating factor, etc.), steroids (aldosterone, testosterone, etc.), amines (catecholamines, etc.), cannabinoids (anandamide, delta-9-tetrahydrocannabinol), all commonly occurring L-amino acids, and nucleotides (ATP, ADP, UDP, UTP, etc.) and chemically related substances. UDP-glucose, UDP-glucuronic acid, and CDP-glucose were obtained from Sigma and were a minimum of 98% pure by HPLC. UDP-galactose and UDP-N-acetylglucosamine, from the same source, were approximately 98% pure by HPLC.
A quantitative assay was performed essentially as described (7) except that the assay was developed by the addition of 20 l of substrate/lysis solution (0.92 mM fluorescein di-␤-D-galactopyranoside, 2.3% Triton X-100, and 0.127 M Pipes, pH 7.2). Plates were incubated for 1 h at 37°C and the reactions stopped by the addition of 20 l of 1 M Na 2 CO 3 . Plates were read in a Spectrafluor fluorimeter (Tecan U.S., Research Triangle Park, NC) at 485 (excitation) and 535 (emission) at optimal gain.
Expression in Mammalian Cells-Receptor expression in HEK-293 cells was carried out as follows. Receptors and G␣ 16 were subcloned into the mammalian expression vector pCDN (9) and transiently transfected into HEK-293 cells using lipofectAMINE Plus (Life Technologies, Inc.), according to the manufacturer's instructions with minor modifications. Thus, a mixture of 15 g of receptor DNA and 15 g of G␣ 16 DNA was used to transfect a single 175-cm 2 culture flask of cells at 80% confluency. For transient transfection of receptor DNA alone, 30 g of plasmid DNA was used. Stable, clonal cell lines were produced by serial dilution into selective media containing 400 g/ml G418, and message-positive clones from Northern blot analysis were assayed further. A single clone responded most sensitively to UDP-glucose in a GTP-␥-S assay and was used in all other experiments. Stable clonal HEK-293 cells co-expressing G␣ 16 and KIAA0001 were generated by transfecting cells with a derivative of pCDN-G␣ 16 in which the neomycin coding sequence had been replaced with a histidinol selectable marker (10). A single histidinol-resistant clone (mRNA-positive for G␣ 16 by Northern analysis) was transfected with KIAA0001 as described above, and a single G418resistant clonal cell line most sensitive to UDP-glucose in an intracellular Ca 2ϩ assay was used in all further work.
Functional Assays with Mammalian Cells-Intracellular Ca 2ϩ assays were carried out as follows. HEK-293 cells were seeded (50,000 cells/well) into poly-D-lysine-coated 96-well black-wall, clear-bottom microtiter plates (Becton Dickinson) 24 h prior to assay. Cells were loaded for 1 h with 1 M Fluo-4-AM fluorescent indicator dye (Molecular Probes) in assay buffer (Hanks' balanced salt solution, 10 mM HEPES, 200 M Ca 2ϩ , 0.1% bovine serum albumin, 2.5 mM probenecid), washed three times with assay buffer, and then returned to the incubator for 10 min before assay on a fluorimetric imaging plate reader (FLIPR, Molecular Devices). The maximum change in fluorescence over baseline was used to determine agonist response. The same library of putative ligands used for screening with yeast cells was used for Ca 2ϩ assays, except that peptides, lipids, steroids, and cannabinoids were screened at final concentrations of Ͼ100 nM and other ligands at Ͼ1 M. GTP-␥-S binding was carried out essentially as described previously (11). Concentration response curve data were fitted to a four-parameter logistic equation using GraFit (Erithacus Software Ltd.).
Radioligand Binding-In saturation binding studies, increasing concentrations of uridine diphospho-D-6[ 3 H]glucose (2.5-150 nM, Amersham Pharmacia Biotech, specific activity ϭ 22.0 Ci/mmol) were added to membranes and incubated in a total volume of 200 l (approximately 100 g of membrane proteins/ml) for 60 min at 25°C. Specific binding was defined as that displaced by excess (1 M) unlabeled UDP-glucose and represented 85-90% of the total binding at a ligand concentration of 25 nM. In competition binding studies, the membranes were incubated with increasing concentrations (1 pM to 10 M) of unlabeled competing ligand and 20 -30 nM of uridine diphospho-D-6[ 3 H]glucose for 60 min at 25°C. Bound and free radioligand were separated using Skatron filters (FilterMat 11734), presoaked in 0.1% polyethylenimine. Bound radioactivity was quantitated in a beta scintillation counter. Protein was determined using BCA (Pierce). Analyses of all binding data (i.e. the determination of K d , B max , and K i values) were performed by computer-assisted nonlinear least square fitting using GRAPHPAD PRIZM (Graphpad Software, Inc., San Diego).

RESULTS AND DISCUSSION
Sequence Comparisons-Phylogenetic analysis ( Fig. 1) showed that KIAA0001 is most closely related to the orphan receptors H963 (13), GPR34 (14), and EBI2 (15) and that thrombin and platelet-activating factor receptors are the closest relatives with identified functions. Functionally identified receptors with nucleotide ligands (P2Y 1,2,4,6,11 ) also appear to share a recent common ancestor with these receptors but are clearly more distantly related (Fig. 1). Phylogenetic analysis also placed KIAA0001 and VTR 15-20 as closest relatives, and a rigorous search of data bases revealed no other potential orthologs. Furthermore, Southern blot analysis using either the KIAA0001 or the VTR 15-20 coding regions under low stringency hybridization conditions revealed only one hybridization band in genomic samples from human, rat, and mouse that was identical with both probes (data not shown). These data are consistent with the notion that KIAA0001 and VTR 15-20 are orthologs and that no close paralogs of this receptor exist.
Activation of KIAA0001 Expressed in Yeast Cells-As part of a large program to identify the natural ligands for orphan GPCRs, we expressed a number of known and orphan GPCRs, including KIAA0001, in yeast strains engineered to respond to agonist activation with increased expression of a pheromone signaling pathway-inducible FUS1-lacZ reporter gene (7,8); expression of this gene is easily monitored by a simple colorimetric assay of enzyme activity. As one cannot predict the specific G protein that will transduce signals from a given orphan receptor, we expressed receptors in several yeast FIG. 1. A phylogenetic tree showing the relationships among a number of human GPCRs, illustrating that KIAA0001 is most closely related to the orphan receptors H963, GPR34, and EBI2 (34.2, 32.1, and 28.4% uncorrected sequence identity and 47.1, 42.9, and 39.6% uncorrected sequence similarity, respectively) but is also a close relative of proteinase-activated receptors (PAR-1, PAR-2, and PAR-3) and the platelet-activating factor receptor (PAF-R). KIAA0001 is a more distant relative of a number of orphan receptors, which have been termed P2Y-like orphans (P2y 5, P2y 9 , and P2y 10 ) on the basis of similarity to the functionally characterized P2Y receptors. Functionally identified receptors with nucleotide ligands (P2Y 1,2,4,6,11 ) also share a common ancestor with KIAA0001 but are clearly more distantly related. GalR1, mt 1 and NK 1 receptor sequences were used as outgroups. Note that the International Union of Pharmacology Nomenclature Committee has recommended that the notation P2Y be used to denote cloned GPCRs functionally characterized as extracellular nucleotide receptors and that lowercase (P2y) be used to denote cloned orphan GPCRs with similarity to P2Y receptors but that have not yet been unambiguously functionally characterized as such (28). strains, each carrying a different G protein ␣-subunit. We then screened multiple receptors, each expressed in several such strains, in parallel, against a large library of over 700 known and putative natural GPCR agonists. UDP-glucose, which was included in this library because it has been reported to possess pharmacological activity (16 -20), was the only substance that specifically activated KIAA0001. Two yeast strains responded to UDP-glucose. The first expressed KIAA0001 in conjunction with the endogenous yeast G␣ protein, Gpalp (21). The second expressed KIAA0001 and a hybrid of Gpalp and the promiscuous mammalian G␣ protein, G␣ 16 (22). A control yeast strain transformed with empty vector did not respond to UDP-glucose nor did yeast strains transfected with the same G proteins and a number of other receptors.
In a follow-up screen of substances related to UDP-glucose, it was discovered that UDP-galactose, UDP-glucuronic acid, and UDP-N-acetylglucosamine, exhibited agonist activity (Table I). However, a large number of related substances were inactive as agonists (listed in Table I), including various other sugar-nucleotides and the naturally occurring nucleotides known to activate GPCRs of the P2Y class: ATP, ADP, UTP, and UDP.
We investigated the concentration dependence of UDP-glucose on KIAA0001-mediated activity using the Gpalp yeast strains with a modification to the original screen that permitted quantitative pharmacology. In this assay, UDP-glucose activated KIAA0001 with a potency of 67.9 Ϯ 6.8 nM (Ϯ S.E., n ϭ 3) (Fig. 2), but no significant response was detected when ATP, ADP, UTP, and UDP were tested at concentrations up to 1 M, demonstrating that this receptor exhibits a distinctly different pharmacology from that of the P2Y nucleotide receptor family. In addition, UDP-glucose was inactive at concentrations up to 1 M when tested against control cells carrying an empty vector (Fig. 2).
Activation of KIAA0001 in Mammalian Cells-In parallel studies, we transiently co-transfected mammalian HEK-293 cells with both KIAA0001 and G␣ 16 to promote coupling to intracellular Ca 2ϩ . We screened these cells against the same ligand bank in an intracellular Ca 2ϩ assay, and UDP-glucose was the only substance observed to clearly elicit KIAA0001mediated Ca 2ϩ responses. Furthermore, we observed that only cells transiently transfected with both KIAA0001 and G␣ 16 were responsive to UDP-glucose (Fig. 3A); cells transfected with either receptor or G protein alone were unresponsive to UDP-glucose in this assay (Fig. 3A). Co-transfected cells responded to UDP-glucose and UDP-galactose with EC 50 values (Ϯ S.E., n ϭ 3) of 104 Ϯ 22 and 421 Ϯ 43 nM, respectively (Fig.  3B), but the related sugar-nucleotide, CDP-glucose, was inactive when tested at concentrations up to 10 M (Fig. 3B). We carried out similar studies with the related human orphan receptors (Fig. 1) H963 (13), GPR34 (14), and P2y 5 (23) but did not observe orphan-mediated responses to either UDP-glucose or UDP-galactose. The specificity of the interaction of UDP-  2. KIAA0001 expressed in yeast cells is activated by UDPglucose in a concentration-dependent manner. A yeast strain containing the native G␣-encoding gene, GPA1, and a pheromone pathway-responsive FUS1-lacZ gene was transformed with an expression vector containing the sequence for KIAA0001 or with empty vector. The two resulting strains were assayed for ␤-galactosidase activity after exposure to a range of concentrations of UDP-glucose. The strain transformed with KIAA0001 (open circles) responded to UDP-glucose in a concentration-dependent manner, whereas the strain transfected with empty vector (filled circles) did not. Data shown are from a single representative experiment. Each point was determined in triplicate and is given as a mean Ϯ S.E. F.U., fluorescence units. glucose and UDP-galactose with KIAA0001 was further confirmed by the observation that HEK-293 cells transiently cotransfected with G␣ 16 , and a variety of GPCRs with known ligands (e.g. melanin-concentrating hormone (2) and urotensin II receptors (3)) responded with large transient elevations of intracellular Ca 2ϩ when challenged with the cognate ligands for these receptors but not when challenged with UDP-glucose or UDP-galactose at a concentration of 10 M (data not shown).
A stable HEK-293 cell line expressing both recombinant G␣ 16 and KIAA0001 responded to UDP-glucose, UDP-galactose, UDP-glucuronic acid, and UDP-N-acetylglucosamine with large, concentration-dependent, transient, intracellular Ca 2ϩ responses ( Fig. 4) with EC 50 values (Ϯ S.E., n ϭ 3) of 80 Ϯ 31, 124 Ϯ 17, 370 Ϯ 33, and 710 Ϯ 27 nM, respectively. Control HEK-293 cell lines expressing G␣ 16 alone were unresponsive to these ligands when tested at concentrations up to 10 M, and none of the additional compounds listed in Table I exhibited detectable KIAA0001-mediated activity in this assay (data not shown). The presence of endogenous P2Y receptors coupled to Ca 2ϩ mobilization in HEK-293 cells (24) prevented us from assessing in this assay whether nucleotides such as ATP and UTP interact with KIAA0001.
To investigate the natural G protein coupling specificity of KIAA0001 in mammalian cells, we generated a clonal HEK-293 cell line expressing KIAA0001 without additional recombinant G proteins. Membranes prepared from these cells responded to UDP-glucose in a concentration-dependent way with increased binding of radiolabeled GTP-␥-S, a nonhydrolyzable analog of GTP, with an EC 50 value of 234 Ϯ 17 nM (Ϯ S.E., n ϭ 3), but nontransfected control cells were unresponsive to UDP-glucose at concentrations up to 10 M (Fig. 5A). Overnight incubation of KIAA0001-expressing cells with pertussis toxin (25 ng/ml) abolished the response to UDP-glucose, suggesting that KIAA0001 is coupled to G proteins of the G i/o class (Fig. 5B). In addition, freshly prepared ATP, ADP, UTP, UDP, and CDPglucose exhibited no detectable KIAA0001-mediated increase in GTP-␥-S binding activity when tested at concentrations up to 10 M (Fig. 5B). Inhibition of basal levels of GTP-␥-S binding at higher concentrations of nucleotide triphosphates was identical in membranes prepared from pertussis toxin treated-and nontreated KIAA0001-transfected cells (Fig. 5B), and an identical effect was also observed in nontransfected cells (data not shown). These data demonstrate that KIAA0001 exhibits a pharmacology distinct from that of the P2Y family, confirming our conclusions drawn from studies with yeast cells.
Radioligand Binding Studies-We attempted to determine whether uridine diphospho-D-6[ 3 H]glucose bound KIAA0001 with high affinity. Membranes prepared from clonal HEK-293 cells expressing KIAA0001 bound uridine diphospho-D-6[ 3 H] glucose with an affinity of 8.1 nM and B max values of 7.8 pmol/mg protein. The specific binding was dependent on the protein concentration and was lost upon boiling membranes prior to assay. In competition assays, UDP-glucose and UDPgalactose competed for specific binding with K i values of 41.2 and 348 nM, respectively, whereas CDP-glucose, UTP, ATP, and ADP had little effect (K i Ͼ 10 M). Essentially identical specific binding was also observed in nontransfected HEK-293, 1321N1, and Chinese hamster ovary cells, suggesting the presence of endogenous binding sites on these cells. These data suggest that this radiolabel may not be a suitable radioligand for KIAA0001. Indeed, one might expect that the receptor expression levels required to observe a specific binding signal in transfected cells above nontransfected cells would need to be at least of the same order of magnitude as the B max of the background specific binding (i.e.: 7.8 pmol/mg protein), a level of cloned receptor expression unlikely to be achieved in mammalian cells. Likewise, other workers investigating the use of radioactive forms of the structurally similar molecule ATP to radiolabel P2Y-like receptors report similar technical problems (25).
Expression Pattern of KIAA0001-To investigate the physiological role of KIAA0001, we carried out studies to localize the expression of this receptor in the body. We performed quantitative RT-PCR analysis using KIAA0001-specific primers to determine the relative levels of KIAA0001 mRNA in a variety of human tissues from multiple individuals. We observed a widespread human tissue distribution (Fig. 6), with only some differences to that reported by Northern blot analysis for VTR 15-20 in the rat (5). Thus, the highest levels of expression were observed in placenta, adipose tissue, stomach, and intestine, with moderate levels noted in a wide variety of brain regions, spleen, lung, and heart. Unlike with VTR 15-20 in the rat, we detected only relatively low amounts in the kidney.
For many years the roles of UDP-glucose and UDP-galactose have been considered to be primarily those of metabolic intermediates that act as activated carriers of sugar moieties for the enzymatic biosynthesis of carbohydrates. However, the data presented here suggest that they also serve a role as extracellular signaling molecules. A limited literature suggests that UDP-glucose may have pharmacological activity. Thus, UDPglucose produces concentration-dependent depolarizations of the rat cervical ganglion (16) and contractions of the rat (17) and guinea pig (18) phrenic diaphragm and is effective in the treatment of certain liver ailments (19), an effect that may result from interaction with cell surface receptors on hepatocytes (20). However, it is premature to conclude that these effects are mediated by KIAA0001, especially because KIAA0001 expression in liver appears to be relatively low (Fig.  6) and because definitive evidence of the role of KIAA0001 in any physiological system must await the development of specific antagonists and evaluation of null alleles (e.g. in the knockout mouse).
Although the precise extracellular signaling roles of UDPglucose and its related molecules is unknown at present, the similarity of these molecules to the natural ligands for the P2Y receptors encourages speculation. For example, the well established extracellular transmitter ATP is both co-released with neurotransmitters, and nonlytically released from mechanically agitated cells (26). Similarly, UTP is released from cells by this second mechanism (27). Such a release mechanism for UDP-glucose and its related molecules, perhaps to register local tissue stress to neighboring cells, would fit well alongside the widespread distribution of this receptor and the reported regulation of its message levels upon immune challenge and neuronal insult (5).