Molecular Cloning and Expression of a 5=Hydroxytryptamine, Serotonin Receptor Subtype*

We have utilized the polymerase chain reaction tech- nique to selectively amplify a G protein-coupled receptor cDNA from rat kidney proximal convoluted tubule mR.NA, which exhibits high homology with previously cloned serotonin receptors. Sequencing of a full-length clone isolated from a rat hippocampal cDNA library re- vealed an open reading frame of 1,212 base pairs encoding a 404-residue protein with seven hydrophobic re- gions predicted to represent transmembrane-spanning domains. Within the transmembrane regions, this receptor was found to be 4“50?6 identical with var- ious members of the 5-HT1, 5-HTa, and 5-HTs subfamilies with lower (3740%) homology to the 5-HTz-like re- ceptors. Northern blots revealed a -3.6-kilobase transcript localized in various brain regions with the following rank order of abundance: hypothalamus > hippocampus = mesencephalon > cerebral cortex = olfactory bulb > olfactory tubercle. Expression of this clone in COS-7

only 5-HT completely inhibited radioligand binding. The inhibition of radioligand binding by other serotonergic agents revealed a pharmacological profile that does not correlate with any previously described serotonin receptor subtype. In addition, this receptor exhibits high affinity for a number of tricyclic antipsychotic and antidepressant drugs including clozapine, loxapine, and amitriptyline. In HEK-293 cells stably transfected with this receptor, serotonin elicits a potent stimulation of adenylylcyclase activity. The distinct structural and pharmacological properties of this receptor suggests that it represents a completely novel serotonin receptor subtype, which we propose to designate 5-HT7. Based on its pharmacology and its localization to limbic and cortical regions of the brain, it is likely that this receptor may play a role in several neuropsychiatric disorders that involve serotonergic systems.
* The costs of publication of this article were defrayed in part by the "advertisement" in accordance with 18 U.S.C. Section 1734 solely to payment of page charges. This article must therefore be hereby marked indicate this fact.

to the GenBankmIEMBL Data Bank with accession numberfs) L15228.
The nucleotide sequencefs) reported in this paper has been submitted )I Supported by the Department of Veterans Af€airs. ** % whom correspondence should be addressed Molecular Neuropharmacology Section, Experimental Therapeutics Branch, National Health, Bldg. 10, Rm. 5C-108, Bethesda, MD 20892.
Institute of Neurological Disorders and Stroke, National Institutes of Serotonin is a ubiquitous neurotransmitter, which is found in both the central and peripheral nervous systems of many species and is involved in a wide variety of behavioral and physiological functions. The transduction of serotonergic signals across cellular membranes is mediated by a diversity of receptor subtypes, which, in mammals, appear to fall into six pharmacologically distinct classes designated 5-hydroxytryptaminel (5-HT1)' through 5-HT6. With the exception of the 5-HT3 receptor, which is a ligand-gated ion channel (l), all of the known serotonin receptor subtypes belong to the large super gene family of G protein-linked receptors. The primary structures for a number of these receptors have been elucidated by molecular cloning, including the 5 -H T 1~-~ subtypes (2-151, the 5-HTZA4 types (16)(17)(18)(19)(20)(21), the 5-HT5A & receptors (22,23), and the 5-HTs receptor (24). In addition, three different Droso-(261, have been cloned and sequenced as well as a serotonin receptor from Lymnaea stagnalis (27).
We now report the cloning from rat brain of a cDNA encoding a novel serotonin receptor, which contains seven predicted transmembrane domains but is distinct from any class of previously described 5-HT receptor in both primary structure and pharmacology. This receptor appears to be localized predominantly in the central nervous system, exhibiting high expression in various limbic and cortical brain regions. When expressed in mammalian cells, this receptor shows high affinity for serotonin as well as a number of tricyclic antipsychotic and antidepressant drugs including clozapine, loxapine, and amitriptyline. Based on its afflnity for these psychotropic drugs and its localization to limbic and cortical regions of the brain, it is likely that this receptor may play a role in several neuropsychiatric disorders which involve serotonergic systems. We propose to designate this novel serotonin receptor the 5-HT7 subtype.

EXPERIMENTAL PROCEDURES PCR and Sequencing-%tal
RNA was prepared from rat kidney proximal convoluted tubules (PCT) using RNAzol B (CinnaBiotecx Laboratories International), and poly(A)+ RNA was isolated using poly(A) Quick columns (Stratagene). First strand cDNA synthesized from the rat PCT mRNA was PCR amplified as described (24) with 1 VM of each of the following primers: 5'-GTCGACCC(GT)T(GT)(CG)GCC(AC)TCA-Individual bands were excised, electroeluted, concentrated by ultrafiltration, and ligated into SalIIHindIII-digested pBluescript I1 SK(+) (Stratagene). Competent DH5a cells (Life Technologies, Inc.) were transformed and minipreparations of plasmid DNA prepared for insert sequencing. Nucleotide sequence analysis was performed using the Sanger dideoxy nucleotide chain termination method with Sequenase (U.S. Biochemical Corp.) on denatured doubled-stranded plasmid templates. Primers were synthetic oligonucleotides that were either vectorspecific or derived from prior sequence information. cDNA sequence was confirmed through the sequencing of both strands.
Northern Blot Analysis-Poly(A)+ RNAwas prepared from rat tissues and Northern blotting performed as described previously (24). Northern blots were probed with the -500-bp PCR amplification product, which was "P-labeled by the random primer method, and exposed for 15 days a t -70 "C to x-ray film with an intensifying screen.
Library Screening-Recombinants (1 x 109 from a rat hippocampal cDNA library, constructed in the hgtll vector (Stratagene), were screened with the PCR fragment that was 32P-labeled via nick translation. Library screening and plaque purification was performed as described previously (24). One positive clone containing a 2.9-kb insert was identified, from which a 1.5-kb fragment was excised using EcoRI and subcloned into pBluescript SK II(+). Nucleotide sequence was determined as described above.
Expression and Radioligand Binding Assays-For transient expression of the PCT-65 receptor, COS-7 cells were transfected with the pSRa-PCT65 construct using the calcium phosphate precipitation method (for PHILSD binding assays) as previously described (24) or the DEAE-dextran method (for C3H15-HT binding assays) as described by Cullen (28). Stably transfected HEK-293 cells were generated and screened as described in Ref. 24. ["HILSD (67 Ci/mmol, Du Pont-New England Nuclear) binding experiments were performed as previously described for 12sI-labeled LSD (24), except that the assay volume was 1 ml containing 30 pg of membrane protein. 5-HT (100 PM) was utilized to determine nonspecific binding, which represented between 2-20% of the total binding over the range of concentrations employed in saturation experiments.
[3H]LSD (2 nM) was utilized for competition experiments. [3H15-HT (24-27 Cilmmol, Du Pont-New England Nuclear) binding experiments were performed as previously described (24) using a final protein concentration of 5-10 pg/ml. Nonspecific binding was determined in the presence of 10 p~ 5-HT or 10 PM methiothepin and represented between 2 and 10% of the total binding over the range of concentrations employed. Intracellular CAMP levels in stably transfected HEK-293 cells were determined as described in Ref.

RESULTS AND DISCUSSION
Following the original report of Libert et al. (29), we have used the polymerase chain reaction (PCR) in an attempt to clone novel members of the G protein-coupled receptor family. Degenerate oligonucleotide primers were prepared using sequences from the third and sixth transmembrane (TM) regions of previously cloned biogenic amine receptors and used to amplify cDNA synthesized from poly(A)+ RNA purified from rat kidney proximal convoluted tubules. This resulted in the amplification of a number of cDNA fragments which were determined by DNA sequence analysis to represent portions of previously cloned receptors, some novel receptors as well as unknown DNA sequences (data not shown). One putatively novel receptor PCR fragment, termed PCT-65, exhibited high homology with previously cloned members of the serotonin receptor family and was selected for further study.
In an initial characterization of PCR fragment PCT-65, we examined the tissue distribution of its corresponding mRNA. Fig. 1 shows Northern blots of poly(A)+ RNA prepared from a variety of rat brain regions as well as other peripheral tissues and probed with the PCT-65 fragment. A single transcript of -3.6 kb was observed in various brain regions with highest expression appearing to occur in the hypothalamus. PCT-65 mRNA was also observed in high abundance in the hippocampus and mesencephalon with lower amounts found in the cortex, olfactory bulb, and olfactory tubercle, whereas no transcript was observable in the cerebellum, pituitary, or retina. In peripheral tissues, this transcript was only seen in the spleen, whereas it was absent, o r present in undetectable levels, in the The gel locations of the RNA(kb) size marker are indicated on the right. Northern blots were prepared and analyzed as described under "Experimental Procedures." retina, pituitary, testis, stomach, prostate, ovary, skeletal muscle, lung, liver, kidney, and gut. The inability to detect PCT-65 mRNA in kidney via Northern blotting, despite this being the tissue of origin, suggests that PCT-65 must be a relatively minor transcript in the kidney and/or is restricted to a small population of cells.
Based on the Northern blot analysis, we decided to screen a rat hippocampal cDNA library using the PCT-65 PCR fragment in order to obtain a full-length clone. One positive clone was isolated, which contained a cDNA insert of -2.9 kb. A 1.5-kb EcoRI fragment was excised from this insert and subcloned into pBluescript for characterization. The complete nucleotide sequence of the EcoRI fragment was determined and is shown in Fig. 2. This cDNA exhibits a long open reading frame containing sequences that show homology to TM regions of previously cloned biogenic amine receptors. Careful inspection of the sequence also revealed the presence of putative intronic sequence containing an in-frame stop codon. This putative intron is located downstream of the third predicted TM domain of the receptor and suggests that the cDNA was derived from an incompletely or incorrectly spliced mRNA. This possibility was further investigated by using PCR to amplify this region from cDNA prepared from rat hippocampal mRNA. Amplification across this region using primers P1 and P2 (Fig. 2) resulted in two equally abundant cDNA products of approximately 460 and 390 bp in size (data not shown). The larger fragment is the size expected based on the sequence of the isolated cDNA clone, whereas the smaller fragment is the size predicted from the postulated splice sites shown in Fig. 2. Sequencing of the smaller PCR product confirmed the predicted splice sites, which indicates that there are two forms of PCT-65 mRNA in the hippocampus, only one of which is correctly spliced. Splicing at nucleotide 416, as shown in Fig. 2, results in an uninterrupted Arg codon and a 1,212-bp open reading frame encoding a protein of 404 amino acids. It should be noted that without identifying an in-frame stop codon prior to the initiator methionine, the assignment of the NH2 terminus of this protein is tentative until additional 5' sequence is acquired. The calculated molecular mass of the 404-residue PCT-65 receptor protein is 45 kDa.
Hydropathy analysis (data not shown) of the deduced amino acid sequence of the PCT-65 protein suggests the existence of seven hydrophobic domains, which are predicted to represent putative TM regions. When compared to previously cloned G protein-coupled receptors, the TM regions of PCT-65 exhibit high homology to various serotonin (5-HT) receptors, suggesting that it may be a member of this receptor family. Within the TM regions, the following homologies to PCT-65 were calcu- The predicted third cytoplasmic loop of the PCT-65 receptor is relatively short, consisting of about 60 residues, whereas the approximately 65-residue-long carboxyl-terminal tail is intermediate in length in comparison to other cloned 5-HT receptors. The PCT-65 receptor protein also contains one potential N-linked glycosylation site at Asn-25 in the extracellular amino terminus, in addition to several potential sites for phosphorylation by CAMP-dependent protein kinase or protein kinase C in the cytoplasmic loops and carboxyl terminus.
In order to definitively establish the identity of the receptor encoded by PCT-65, we proceeded with its expression in mammalian cells. As the intron-containing clone proved unsuitable for expression, we decided to construct an appropriately spliced cDNA. We first utilized the correctly spliced cDNA fragment, generated using PCR primers P1 and P2 as described above, and digested it with BcZI, the sites of which are indicated by solid triangles in Fig. 2. The resulting BclI fragment was subsequently ligated into the full-length BcZI-digested cDNA clone. We were thus able to excise the intron-containing BcZI fragment from the cDNA isolated from the hippocampal library and replace it with the corresponding fragment from a PCR-amplitied correctly spliced cDNA. This construct was subcloned into the mammalian expression vector pCD-SRa (30) yielding pSRa-PCT65.
Transient transfection of COS-7 cells with pSRa-PCT65 resulted in the appearance of high affinity, homogeneous, and saturable binding sites for the serotonergic ligand l3H1LSD, which exhibited a KD of 4.9 * 0.78 ( n = 5) x m and B , , values ranging from 5-15 pmoVmg protein (data not shown). No specific binding of 13HlLSD was observed in untransfected COS-7 cells or in cells transfected with the pCD-SRa vector alone. Preliminary characterization of the PCT-65 pharmacology indicated that among several endogenous biogenic amines, including dopamine, melatonin, epinephrine, norepinephrine, or histamine, 5-HT was the most potent in competing for f3H1LSD binding sites, exhibiting a Ki of 1-2 nM ( Table I)

TABLE I1 Stimulation of CAMP accumulation by PCT-65 in transfected HEK-293 cells HEK-293 cells stably expressing the PCT-65 receptor were grown to confluence in 35-mm culture wells (approximately 4 x lo5 cells/well) and assayed for CAMP accumulation. The cells were preincubated with 2 ml of HEPES-buffered Earle's balanced salt solution containing 300
PM isobutylmethylxanthine for 20 min followed by drug addition and further incubation for 10 min. The incubations were terminated through the addition of 1 ml of 5% trichloroacetic acid and the CAMP produced was assayed as described under "Experimental Procedures." The data represent the mean f S.E. values for a single experiment conducted using six replicates per treatment condition. This experiment was performed twice with similar results. shown). Inclusion of 0.3 m~ GTP did not appreciably alter the binding of [3H]5-HT. These initial radioligand binding assays suggest that PCT-65 represents a 5-HT receptor subtype. The apparent absence of receptor-G protein coupling in the transiently transfected COS cells, as evidenced by the lack of guanine nucleotide regulation of agonist binding, may be due to either an excess of receptor expression relative to the endogenous G proteins or, alternatively, the cells may not express the appropriate G protein(s) that is normally linked to the PCT-65 receptor.
Further characterization of the PCT-65 pharmacology involved competition assays using a variety of drugs that exhibit specificity for various serotonergic receptor subtypes and other binding sites. The average K, values for compounds competing with better than -1 PM affinity are shown in Table I. Examination of the rank order of potency for a variety of serotonergic agents reveals that the pharmacology of clone PCT-65 does not correspond to any previously described serotonin receptor subtype. A number of drugs that exhibit high affinity for the 5-HT1 (1-(3-chlorophenyl)piperazine and 1-(2-methoxyphenyl)piperazine), 5-HTz (ketanserin and mianserin) 5-HT3 (quipazine and MDL 72222), and 5-HT4 (zacopride and DAU 6285) receptor subfamilies exhibit relatively low affinity for PCT-65. Ergot alkaloids, especially ergoline derivatives (ie. LSD, lisuride, or metergoline), display relatively high affinity for PCT-65 as does the non-selective serotonergic antagonist methiothepin. The 5-HT5 receptors similarly exhibit high affinity for ergot alkaloids but differ from PCT-65 in having low affinity for 5-HT and methiothepin (22,23). Interestingly, the atypical and typical anti-psychotics clozapine and loxapine, respectively, also exhibit high affinity for PCT-65, as does the tricyclic anti-depressant amitriptyline. In general, the drugs that exhibit the greatest affinity for PCT-65 (ie. Ki < 200 n~) are tricyclic, ergoline, or tryptamine derivatives. This is reminiscent of the recently described 5-HTs receptor (241, which exhibits a similar but distinctive pharmacology.
To investigate possible effects on second messenger levels, we stably transfected HEK-293 cells with the pSRa-PCT65 construct. One cell line, expressing about 800 fmol of I3H15-HT binding activity/mg of protein, was selected for further analysis. Serotonin was found to elicit a potent dose-dependent stimulation of CAMP levels in these cells exhibiting an EC50 of 63 * 4 n~ (n = 2), whereas there was no response in nontransfected cells (data not shown). Maximum CAMP accumulation was observed using 10 PM serotonin, which produced a 10-15-fold trotonin Receptor 18203 stimulation over basal levels. The serotonin response was pharmacologically specific being blocked by methiothepin and clozapine (Table 11). These data indicate that the PCT-65 receptor is linked to activation of the adenylylcyclase signal transduction system. Taken together, our data indicate that we have identified and cloned a prototypical member of a seventh subfamily of 5-HT receptors. This receptor is structurally (<50% homology in the TM domains) and pharmacologically distinct from any previously cloned mammalian 5-HT receptor. On this basis, we propose to designate clone PCT-65 the 5-HT7 subtype. The 5-HT7 receptor does, however, exhibit significant structural and pharmacological similarities to two previously cloned non-mammalian receptors, the ~-H T D R o~ (25) and 5-HT~rn (271, suggesting that it may represent a mammalian homolog of one of these receptor subtypes. Another feature that distinguishes the 5-HT7 receptor from many previously cloned serotonin receptors is the presence of at least one intron in the coding region of its gene. Thus far, all of the 5-HT1 receptor subtypes appear to be encoded by intronless genes (131, whereas the genes for 5-HTz subfamily exhibit multiple introns (33,341. The 5-HT5~ B receptors each have a single intron interrupting their coding sequences (231, and the 5-HTs receptor also has at least one intron in its gene (24). Interestingly, the position of the intron in the 5-HT7 receptor gene (within the codon for Arg-139) does not correlate with the position of any of the introns within the genes encoding the 5-HTz and 5-HT5 receptor subfamilies nor with the only known intron in the 5-HTs receptor gene. There is, however, an intron at the identical position within the genes encoding the Dz (351, D3 (361, and D4 (37) dopamine receptors (the Dz subfamily). This might suggest that the 5-HT7 receptor is evolutionarily related to the Dz dopamine receptors. Interestingly, the 5-HT7 receptor is fairly homologous to Dz-like receptors exhibiting identities of 46% (Dz), 48% (D3), and 42% (D,) within the TM regions.
Another interesting feature concerning the intron identified in the 5-HT7 receptor cDNA is that it appears to exhibit a non-consensus 5' splice site. In our initial characterization of the intron-containing cDNA (Fig. 2), we attempted to express it in COS-7 cells. Using PCR analysis, however, we found that mRNA transcribed from this clone did not undergo further splicing, nor did it impart receptor binding activity to the cells (data not shown). This suggests that the 5' splice site in the cDNA clone is not capable of functional splicing, at least in the COS-7 cells. We are currently postulating that there is actually additional 5' sequence in this intron region of the gene and that this cDNA was derived through alternative splicing at an internal acceptor site producing a non-functional mRNA. Testing of this hypothesis will require the isolation and characterization of the 5-HT7 receptor gene. Confirmation of the PCR-generated, correctly spliced cDNA has recently been obtained from isolating and sequencing additional 5-HT7 receptor clones from both rat and human cDNA libraries (data not shown).
Although the pharmacology of the 5-HT7 receptor is distinct and does not correspond to any previously characterized subtype, it does exhibit high affinity for a number of ligands that interact with the 5-HT1 receptor family. For instance, its high affinity for 5-HT and 5-carboxamidotryptamine suggest that it may have been included in previous characterizations of "multiple'' 5-HT1 receptor binding sites using 13H]5-HT or 13H]5carboxamidotryptamine as radioligands (38). Similarly, the 5-HT7 receptor exhibits relatively high affinity for the classic 5-HT1~ receptor agonist, 8-hydroxy-Nfl-dipropyl-2-aminotetralin, suggesting that it might have contributed to previously described %-HTlA'' responses. Perhaps of greater interest is the demonstrated high affinity of the 5-HT7 receptor for a number

The Rat 5-HT7
Serotonin Receptor of therapeutically important psychotropic drugs including the atypical antipsychotic, clozapine. Clozapine is unique from other clinically used antipsychotic drugs due to its activity in patients resistant to other medications and by its lower propensity to cause extrapyramidal side effects. At -15 nM, the affinity of the 5-HT7 receptor for clozapine is similar to that of the D4 dopamine, 5-HTz subfamily, and 5-HTs serotonin receptors and further suggests that it is occupied at therapeutically employed concentrations. Given this consideration in addition to its demonstrated localization to limbic and cortical regions of the brain, it seems likely that the 5-HT7 receptor may play an important role in the therapy of psychotic disorders. The identification and cloning of this novel receptor subtype should enable testing of this hypothesis as well as further investigations of its role in normal physiology and behavior.