Benzylimidazolines as h5-HT1B/1D serotonin receptor ligands: a structure-affinity investigation.

Benzylimidazolines may represent a class of 5-HT1D ligands that has yet to be exploited. On the basis of a previous report that the 2-(substituted-benzyl)imidazoline alpha-adrenergic agonist oxymetazoline (8) binds with high affinity at calf brain 5-HT1D receptors, we explored the structure-affinity relationships of a series of related derivatives. Each of the aromatic substituents was removed and then reinstated in a systematic manner to determine the influence of the individual substituents on binding. It was found that all of the aromatic substituents of 8 act in concert to impart high affinity. However, although the 3-hydroxy group could be removed without significantly reducing affinity for h5-HT1D (i.e., human 5-HT1Dalpha) receptors, this modification reduced h5-HT1B (i.e., human 5-HT1Dbeta) receptor affinity by nearly 50-fold. The 2, 6-dimethyl groups also contribute to binding but seem to play a greater role for h5-HT1B binding than h5-HT1D binding. With the appropriate structural modifications, several compounds were identified that display 20- to >100-fold selectivity for h5-HT1D versus h5-HT1B receptors. Preliminary functional data suggest that these compounds behave as agonists. Given that 5-HT1D agonists are currently being explored for their antimigraine action and that activation of h5-HT1B receptors might be associated with cardiovascular side effects, h5-HT1D-selective agents may offer a new lead for the development of therapeutically efficacious agents.

A unitary non-NMDA receptor short subunit from Xenopus: DNA cloning and expression.

A high-affinity homomeric, non-NMDA glutamate receptor was previously purified from the amphibian Xenopus laevis. We have obtained nine peptide sequences from its subunit, applied in cDNA cloning. The cDNA encodes a subunit (XenU1) containing all nine sequences. The 51,600-dalton mature subunit has four hydrophobic domains homologous to the four in the C-terminal half of mammalian non-NMDA receptor subunits. Transient expression in COS cells showed 1:1 binding (at Bmax) of [3H] kainate (KD = 9.1 nM) and of [3H] AMPA (alpha-amino-3-hydroxy-5-methyl-isoxazole-4-propionic acid; KD = 62 nM). The competitive binding series domoate > kainate > AMPA > NBQX > glutamate was established (where NBQX is 2,3-dihydroxy-6-nitro-7-sulphamoyl-benzo (f) quinoxaline). Each agonist shows the same KI value against [3H] kainate and [3H] AMPA binding, suggesting a common agonist site, but two conformations thereof are distinguishable by their different affinities for the antagonist NBQX and by the allosteric effect of thiocyanate anion (greatly potentiating AMPA binding, inert with kainate). XenU1 is exceptional among non-NMDA receptor subunits because it lacks most of the large N-terminal domain found in those of mammals and it has high affinity for both kainate and AMPA. It differs from the similarly-short "kainate-binding proteins" (KBPs), in binding AMPA and in forming glutamate receptor channels when the native protein is reconstituted. Moreover, whereas a full-length kainate receptor of mammals, GluR6, is shown here (from a partial cDNA sequence) to exist also in Xenopus, with approximately 97% sequence identity to rat GluR6, XenU1 is much less homologous to any rat kainate or AMPA receptor and also to the KBPs, even from another amphibian, Rana. Another difference is that a potential concensus sequence ("EF hand") for Ca2+ binding is present in the N-terminal domain of XenU1, but not in the chicken (glial) KBP. XenU1 is deduced to be in a new family of non-NMDA receptors.

cDNA cloning and functional properties of human glutamate receptor EAA3 (GluR5) in homomeric and heteromeric configuration.

We have isolated a new member of the human glutamate receptor family from a fetal brain cDNA library. This cDNA clone, designated EAA3a, shares a 90% nucleotide identity with the previously reported rat GluR5-2b cDNA splice variant and differed from human GluR5-1d in the amino and carboxy terminal regions. Cell lines stably expressing EAA3a protein formed homomeric ligand-gated ion channels responsive, in order of decreasing affinity to domoate, kainate, L-glutamate and (RS)-alpha-amino-3-hydroxy-5- methylisoxazole-propionate (AMPA). Kainate-evoked currents showed partial desensitization that was reduced on incubation with concanavalin A (conA) but not cyclothiazide and were attenuated by the non-N-methyl-D-aspartate (NMDA) receptor antagonist CNQX (6-cyano-7-nitro-quinoxalinedione). Coexpression of EAA3a and human EAA1 cDNAs in HEK 293 cells formed a heteromeric channel with unique properties. Kainate and AMPA activated the heteromeric channel with significantly higher affinities than observed for EAA3a alone. Ligand binding studies with the recombinant EAA3a receptor expressed in mammalian cells indicated a high affinity kainate binding site (Kd = 120 +/- 15.0 nM). The relative potency of compounds in displacing [3H]-kainate binding to EAA3a receptor was: domoate > kainate > L-glutamate = quisqualate > 6,7-dinitroquinoxaline-2,3-dione (DNQX) = CNQX > AMPA > dihydrokainate > NMDA.

Human glutamate receptor hGluR3 flip and flop isoforms: cloning and sequencing of the cDNAs and primary structure of the proteins.

Several cDNA clones encoding the human glutamate receptor subunit GluR3 flip and flop isoforms, were isolated from human hippocampus and fetal brain libraries. DNA sequence analysis revealed overlapping clones permitting the reconstruction of full-length GluR3-flip and GluR3-flop cDNAs. The GluR3 cDNAs demonstrated an 94.1-94.7% nucleotide (nt) identity with the corresponding rat cDNAs. The nt sequence of the GluR3 cDNAs would encode 894 amino acid proteins that have a 99.4% identity with the rat GluR3 isoforms. The human GluR3 cDNAs predict an additional 6 amino acid in the N-terminal signal peptide as compared to the rat GluR3.

Cloning and sequence analysis of additional splice variants encoding human N-methyl-D-aspartate receptor (hNR1) subunits.

Two cDNA clones representing previously unidentified human N-methyl-D-aspartate receptor (hNR1) subunit polypeptides were isolated and sequenced. Clone hNR1-4 was isolated from a human hippocampus cDNA library and was presumably generated by alternative RNA splicing in the 3' amino acid (aa) coding regions. The hNR1-4 cDNA demonstrated an 85.7% nucleotide (nt) identity to the corresponding rat NR1 (rNR1) cDNA. The nt sequence of hNR1-4 would encode a protein that has a 99.8% identity with the corresponding rNR1 subunit. Clone hNR1N was isolated by polymerase chain reaction (PCR)-mediated amplification of a 0.6-kb DNA fragment from human cerebellum cDNA. The nt sequence of this DNA fragment was identical to previously isolated hNR1 cDNA clones, except for the presence of a 63-bp DNA insertion that would encode an additional 21 aa. This DNA insertion occurs in the 5' aa coding regions of hNR1 and presumably represents an exon that is subject to alternative splicing. The nt and aa sequences of this exon are identical between human and rat.

Functional expression and pharmacological characterization of the human EAA4 (GluR6) glutamate receptor: a kainate selective channel subunit.

A cDNA encoding an ionotropic glutamate receptor subunit protein humEAA4 (GluR6), has been cloned from a human fetal brain library. This cDNA when expressed in COS or HEK-293 cells is associated with high-affinity kainate receptor binding and ion channel formation. We have successfully established cell lines stably expressing humEAA4 in HEK-293 cells This is the first report of the establishment of stable cell lines expressing a glutamate receptor channel. The relative potency of compounds for displacing [3H]-kainate binding to humEAA4 receptors expressed in COS or HEK-293 cells is domoate > kainate > quisqualate > 6-cyano-7-nitroquinoxaline-2,3-dione > L-glutamate = 6,7- dinitroquinoxaline-2,3-dione > dihydrokainate. Applications of kainate, glutamate, and domoate but not AMPA evoked rapidly desensitizing currents in cells expressing homo-oligomeric humEAA4 in a concentration dependent manner. The order of potency was: domoate > kainate > L-glutamate. Although AMPA did not itself activate humEAA4 receptors it did reduce, to a limited extent, kainate-evoked responses. AMPA may therefore be a weak partial agonist for this receptor. To date this effect has not been demonstrated with rat GluR6. It is possible that subtle species differences may exist in the nature of agonist receptor interaction. Kainate evoked currents were attenuated by the quinoxalinediones CNQX and DNQX but not by DAP5. The receptor desensitization was attenuated on application of concanavalin A. Ion-permeability studies indicated that the receptor-linked ion channel is permeable to both Na+ and Ca2+ ions.

Molecular characterization of the human EAA5 (GluR7) receptor: a high-affinity kainate receptor with novel potential RNA editing sites.

Several cDNA clones encoding EAA5 receptor polypeptides were isolated from a human fetal brain library. The EAA5 cDNAs demonstrated an 88.7-90.1% nucleotide identity with rat GluR7 cDNAs. The nucleotide sequence of EAA5 would encode a 919-amino acid protein, that has a 97.7-98.9% identity with the rat GluR7 receptor. Two variation of the EAA5 cDNA were identified which result in amino acid substitutions in the predicted extracellular amino-terminal region; Ser310-->Ala and Arg352-->Gln. These variations can be attributed to RNA editing involving T-->G and G-->A substitutions. Both the location (with respect to glutamate receptors), and the nucleotides involved, in this putative RNA editing are novel and may therefore involve novel mechanisms. Ligand binding studies with membranes of transfected COS-1 cells expressing EAA5 polypeptides demonstrate a rank order of ligand affinity similar to that observed with the rat GluR7 receptor, and a dissociation constant for kainate (2.72 +/- 0.12 nM (n = 3)) that is approximately 20- to 30-fold higher than that observed for the rat GluR7 receptor. All of the ligands tested had a higher affinity for the human EAA5 receptor as compared to the rat GluR7 receptor. This report provides another example of pharmacological differences for similar receptors across species.

Molecular cloning, expression, and pharmacological characterization of humEAA1, a human kainate receptor subunit.

Kainate is a potent neuroexcitatory agent; its neurotoxicity is thought to be mediated by an ionotropic receptor with a nanomolar affinity for kainate. In this report, we describe the cloning of a cDNA encoding a human glutamate ionotropic receptor subunit protein from a human hippocampal library. This cDNA, termed humEAA1, is most closely related to rat and human cDNAs for kainate receptor proteins and, when expressed in COS or Chinese hamster ovary cells, is associated with high-affinity kainate receptor binding. We have successfully established cell lines stably expressing humEAA1. This is the first report of establishment of stable cell lines expressing a glutamate receptor subunit. The relative potency of compounds for displacing [3H]kainate binding of humEAA1 receptors expressed in these stable cell lines was kainate > quisqualate > domoate > L-glutamate >> (RS)-alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid > dihydrokainate > 6,7-dinitroquinoxaline-2,3-dione > 6-cyano-7-nitroquinoxaline-2,3-dione. Homooligomeric expression of humEAA1 does not appear to elicit ligand-gated ion channel activity. Nevertheless, the molecular structure and pharmacological characterization of high-affinity kainate binding of the humEAA1 expressed in the stable cell line (ppEAA1-16) suggest that the humEAA1 is a subunit protein of a human kainate receptor complex.

Cloning and sequence analysis of cDNAs encoding human hippocampus N-methyl-D-aspartate receptor subunits: evidence for alternative RNA splicing.

Several cDNA clones encoding human N-methyl-D-aspartate receptor (hNR1) subunit polypeptides were isolated from a human hippocampus library. Degenerate oligodeoxyribonucleotide (oligo) primers based on the published rat NR1 (rNR1) amino acid (aa) sequence [K. Moriyoshi et al. Nature 354 (1991) 31-37] amplified a 0.7-kb fragment from a human hippocampus cDNA library, via the polymerase chain reaction (PCR). This fragment was used as a probe for subsequent hybridization screening. DNA sequence analysis of 28 plaque-purified clones indicated three distinct classes, designated hNR1-1, hNR1-2 and hNR1-3, presumably generated by alternative RNA splicing. One of these clones, hNR1-1(5A), was isolated as a full-length cDNA. The hNR1-2 and hNR1-3 cDNAs represented 66.8 and 98.9%, respectively, of the total aa coding information predicted for the polypeptides. The hNR1 cDNAs demonstrated an 84-90.8% nucleotide (nt) identity with the corresponding rodent cDNAs. The nt sequences of hNR1-1, hNR1-2 and hNR1-3 would encode 885-, 901- and 938-aa proteins, respectively, that have 99.1-99.8% identity with the corresponding rodent NR1 (roNR1) subunits. The changes between the predicted aa sequences of hNR1 and the corresponding roNR1 subunits are confined to the extracellular N-terminal regions. We have also identified two possible allelic variations of the hNR1-3 cDNA that result in aa substitutions in the extracellular N- and C-terminal regions. One of these naturally occurring aa variations is situated within a potential glutamate-binding site.

Bovine elastin cDNA clones: evidence for the occurrence of a new elastin-related protein in fetal calf ligamentum nuchae.

Poly (A+) mRNA was isolated from fetal calf ligamentum nuchae and used for the construction of cDNA libraries. A fraction highly enriched in elastin mRNA was used to prepare the cDNA probes for screening the libraries. A 2 kb clone, pRE1, gave the most positive signal in colony hybridization. It hybridized to a mRNA of the same size as reported for elastin mRNAs from chick and sheep. Hybrid-arrested translation showed that translation of mRNAs for proteins other than elastin doublet was not inhibited by pRE1. Southern blot analysis showed that pRE1 has sequence homology with pVE6 and pVE10, which were tentatively identified as elastin-related cDNA clones representing two distinct mRNAs. DNA sequence data from the 5' end of pRE1 show that the translated amino acid sequence is not typical of known elastin sequences but contains some elastin-like sequences. All of this evidence strongly suggests the occurrence in fetal calf nuchal ligament of a mRNA which codes for a previously unknown elastin-related protein.