Cloning, sequencing and invitro functional expression of recombinant ...

2 downloads 0 Views 884KB Size Report
Nov 14, 1996 - porcine ribonuclease inhibitor (Kobe & Deisenhofer. 1993), This analogy has led to several models of the three-dimensional (3D) structure of ...
Cloning, sequencing and in vitro functional expression of recombinant donkey follicle-stimulating hormone receptor: a new insight into the binding specificity of gonadotrophin receptors F

Richard,

N

Martinat, J-J Remy,

R Salesse and Y Combarnous

'Unité Récepteurs et Communication Cellulaire, Bâtiment des Biotechnologies, Institut National de la Recherche Agronomique, 78352 Jouy-en-Josas, France 2Equipe 'Gonadotropines', Unité INRA-CNRS no. 1291 de Physiologie de la Reproduction des Mammifères Domestiques, 37380 Nouzilly, France (Requests for offprints should be addressed to F Richard at Unité Récepteurs et Communication Cellulaire) ABSTRACT

Among all mammalian FSH receptors (FSH-R; including donkey (dk) FSH-R), only horse (hs) FSH-R does not bind hsLH/chorionic gonadotrophin (CG). In order to delineate the structural origin of hsFSH-R specificity precisely, we have cloned dkFSH-R cDNA from donkey testis mRNA by RT-PCR. Transiently expressed dkFSH-R endowed COS-7 cells with both hsLH/CG- and FSH-binding activity, as well as FSH-induced cAMP production. INTRODUCTION

The deduced dkFSH-R amino acid sequence shares 96% identity with the hsFSH-R: notably, in the hormone-binding domain, the specificity of hsFSH-R may be ascribed to only four divergent amino acids: Thr 173, Asp 202, Asn 268 and Pro 322. Interestingly, hsAsn 268 could bear an additional N-glycosylation. According to receptor negative specificity, these amino acids could be implicated in preventing LH/CG binding to FSH-R.

Journal of Molecular Endocrinology (1997) 18, 193-202

gonadotrophin receptors discriminate be¬ LH/CG and FSH with a very high specificity, even in interspecies hormone-receptor associations (0-001-0-1% cross-reaction (Pierce & Parsons 1981)), except for horse (hs) LH/CG, so that hsLH/CG is used as a surrogate of FSH in inducing folliculogenesis and superovulation in non-horse livestock females. Among all mammalian FSH receptors (FSH-R), only hsFSH-R discrimi¬ nates hsLH/CG from FSH (Stewart & Allen 1979, Combarnous & Henge 1981). This outstanding characteristic could be expected by the ectodomain of the hsFSH-R, since the gonadotrophin receptor The

tween

The

glycoprotein hormone family includes one thyrotrophin (TSH) and two gonadotrophins (lutropin/chorionic gonadotrophin (LH/CG) and follitropin (FSH)). LH, FSH and TSH are of pitu¬ itary origin, while CG is a placental LH secreted by pregnant primates and equidae. These hormones are composed of two non-covalently associated subunits, a and ß. The former is common to all glycoprotein hormones within one species whereas the ß subunit brings about receptor-binding specificity. The glycoprotein hormone receptors belong to the large family of seven transmembrane segment receptors coupled to G-proteins and possess a very large extracellular domain (ectodomain) consisting of leucine-rich repeat sequences (McFarland et al. 1989) similar to those found in the structure of porcine ribonuclease inhibitor (Kobe & Deisenhofer 1993), This analogy has led to several models of the three-dimensional (3D) structure of the glyco¬ protein hormone-receptor ectodomain (Jiang et al. 1995, Kajava et al. 1995, Moyle et al. 1995).

ectodomain alone is sufficient to bind its cognate hormone with high affinity and specificity (TsaiMorris et al. 1990, Xie et al. 1990, Braun et al. 1991, Ji & Ji 1991, Davis et al. 1995). However, the discrimination between LH/CG and FSH is thought to rely on a 'negative specificity' mechan¬ ism, in which the receptor-binding site would be a mosaic of positive binding region(s), presumably largely common to both receptors, and of nega¬ tive, discriminating region(s) which would prevent

binding of the non-cognate hormone (Combarnous & Henge 1981, Moyle et al. 1994). The search for such sequences has been illustrated by the construc¬

tion of chimeras of the rat LH- and FSH-R, some of which chimeras exhibit both LH and FSH binding (Braun et al. 1991, Moyle et al. 1994). In comparing the hsFSH-R amino acid sequence with those known so far from other species, Robert et al. (1994) found horse-specific amino acids in only ten positions of the ectodomain. In order to reduce the number of potentially discriminating residues further, we turned to an equine species close to the horse, i.e. the donkey (dk), whose FSH-R behaves as all non-horse FSH receptors. Since the donkey (Equus asinus) is a species phylogenetically very close to the horse (Equus equus), we anticipated that there would be only a few differences between the amino acid sequences of the FSH-R ectodomain of the two species and that these differences would lead to locat¬ ing the amino acids supporting the strict specificity ofthehsFSH-R. MATERIALS AND METHODS

Isolation of dkFSH-R cDNA

Polyadenylated RNA was isolated from donkey and purified on oligo(dT) cellulose (Pharmacia, Orsay, France) or by using an mRNA isolation kit (PolyAttract III; Promega, Lyon, France). mRNA (100 ng) was used as template for the modified Moloney murine leukaemia virus reverse transcriptase (200 units SuperScript II; Gibco-BRL, Cergy-Pontoise, France) and PCR was performed with 0-25 unit Thermus aquaticus DNA polymerase (Appligene, Illkirch, France) in a Perkin Elmer temperature cycler. For the cloning of the FSH-R N-terminal domain (ectodomain), the synthesis of the first strand cDNA was primed with 4 ng random hexaprimers and PCR was performed with primers FSH-Rsp (5'-GATGGCCYTGCTCCTGGTCTC-3', Y=C or T) and Feel (5'-TTAAATGCATCTGGC testes

TTGGGGG-3'), according to the following pro¬ °C, 1 min; 60 °C, 1 min; 72 °C, 1 min; 35 cycles. For the cloning of the FSH-R C-terminal domain (endodomain), the synthesis of the first strand cDNA was primed with 100 ng oligo(dT) and PCR was performed with primers FSH-R5 (5'-GAA GTSGTTGATGTGACTTGC-3', S=G or C) and FSH-R7(5'-TTAGTTCTGAGCTAAATGGCT3'), according to the following programme: 94 °C, 1 min; 55 °C, 1 min; 72 °C, 1 min; 35 cycles. gramme: 94

Primer nucleotide sequences

were

based

on

horse

(Feel, FSH-R5 and FSH-R7) (Robert et al. 1994) and on rat/human (RFSHsp) (Minegishi et al. 1991, Heckert et al. 1992, Kelton et al. 1992) FSH-R sequences.

PCR fragments were gel purified, phosphorylated with T4 polynucleotide kinase (Biolabs, Montignyle-Bretonneaux, France), ligated into the Smal site of pUC18 or into the EcoRV site of BlueScript II KS- (Stratagene, Montigny-le-Bretonneaux, France) plasmid and amplified into E. coli DH5a competent cells (Gibco-BRL). Complete nucleotide sequences of all amplified fragments were determined three or four times on both strands by using an Applied Biosystems automated DNA sequencer. A full-length dkFSH-R cDNA was engineered by fusing the FSH-Rsp-Fecl and the FSH-R5FSH-R7 overlapping fragments at the unique Nsil site (at 1028 nt). The resulting cDNA was cloned into the Kpnl site of pCMV4, a eukaryotic expression plasmid (a kind gift from D W Russell, University of Texas Southwestern Medical Center, Dallas, TX, USA (Andersson et al. 1989)). Functional characterization of dkFSH-R in COS-7 cells

COS- 7 cell transfection COS-7 cells (African green monkey SV40transformed kidney cells) were cultivated in

1. Cloning of the dkFSH-R. (A) Schematic representation of the full-length dkFSH-R and amplified cDNA fragments. Two sets of primers (FSH-Rsp-Fecl and FSH-R5-FSH-R7, arrows) were used in RT-PCR experiments to amplify cDNA fragments encoding the 5' (1-1037 bp, the ectodomain) and the 3' (994—2064, the endodomain) moieties of dkFSH-R. The cDNA of the dkFSH-R was reconstituted and aligned on the structure of the rat FSH-R gene (F ,'ckert et al. 1992), assuming an identical genomic structure of ten exons (open boxes). The amplification of the 5' moiety yielded two variants. After sequencing, the shorter variant may be ascribed to the skipping of exon 2. Several clones corresponding to amplified cDNA fragments were sequenced on both strands (three clones for the figure

dkFSH-R ectodomain cDNA, two clones for the dkFSH-R ectodomain cDNA missing exon 2, and three clones for the dkFSH-R endodomain cDNA). (B) Complete nucleotide sequence of the dkFSH-R and its deduced amino acid sequence. Position +1 is assigned to the first nucleotide of the putative initiator codon ATG. The boundaries between exons (->