Structure of the porcine thyrotropin receptor: a 200 kilodalton glycoprotein heterocomplex

Cleavage of the human thyrotropin receptor by ADAM10 is regulated by thyrotropin

The thyrotropin receptor (TSHR) has a unique 50 residue (317-366) ectodomain insertion that sets it apart from other glycoprotein hormone receptors (GPHRs). Other ancient members of the leucine-rich repeat G protein-coupled receptor (GPCR) (LGR) family do exhibit ectodomain insertions of variable lengths and sequences. The TSHR-specific insert is digested, apparently spontaneously, to release the ectodomain (A-subunit) leaving the balance of the ectodomain attached to the serpentine (B-subunit). Despite concerted efforts for the last 12 years by many laboratories, the enzyme involved in TSHR cleavage has not been identified and a physiologic role for this process remains unclear. Several lines of evidence had suggested that the TSHR protease is likely a member of the a disintegrin and metalloprotease (ADAM) family of metalloproteases. We show here that the expression of ADAM10 was specific to the thyroid by specially designed DNA microarrays. We also show that TSH increases TSHR cleavage in a dose-dependent manner. To prove that ADAM10 is indeed the TSHR cleavage enzyme, we investigated the effect of TSH-induced cleavage by a peptide based on a motif (TSHR residues 334-349), shared with known ADAM10 substrates. TSH increased dose dependently TSHR ectodomain cleavage in the presence of wild-type peptide but not a scrambled control peptide. Interestingly, TSH increased the abundance of non-cleaved single chain receptor, as well higher molecular forms of the A-subunit, despite their enhancement of the appearance of the fully digested A-subunit. This TSH-related increase in TSHR digested forms was further increased by wild-type peptide. We have identified for the first time ADAM10 as the TSHR cleavage enzyme and shown that TSH regulates its activation.

Evolutionary divergence of thyrotropin receptor structure

The availability of 18 thyrotropin receptor (TSHR) sequences, including two recent entries for primates and seven from fish, have allowed us to investigate diversification of residues or domains during evolution. We used a likelihood ratio test for evolutionary rate shifts [Proc. Natl. Acad. Sci. 98 (2001) 14512] using LH/CGR sequences as an out-group. At each residue in the alignment, a statistical test was performed for a rate shift at the divergence between mammals and fish. Eighty-two rate shift sites were found, significantly more than was expected (p < 0.0001). The occurrence of rate shifts was highest in the intracellular tail, lowest in the transmembrane serpentine and intermediate in the ectodomain. In 52 mammalian sites, the rates were significantly faster than for the corresponding sites in fish. We have identified rate shift in sites important to TSHR function or in intimate proximity to such regions. The former category includes residues 53 and 55 (of LLR1 beta strand) and 253 and 255 (of LLR9 beta strand), crucial to TSH thyrotropic activity, residue 113, the site of N-linked glycosylation limited to humans, residue 310, an important switch in the hinge region for receptor binding and constitutive activity and residue 382 which centres a motif important for TSH-mediated receptor activation. The rate shifts positions close to functional region include a site proximal to a TSHR-specific motif on LLR3 beta strand, sites important in TM helix structure and homodimerization as well as, in the case of the third intracellular loop, to TSHR/G protein coupling. Rate shift analyses have identified residues whose manipulation in the human TSHR may lead to better understanding of receptor functions and help in the creation of designer analogues.

Evolution of the thyrotropin receptor: a G protein coupled receptor with an intrinsic capacity to dimerize

The rapidly escalating number of genome sequences has emphasized the basic tenants of the schema of life. By the same token comparisons according to specialized function or niche within nature expose genomic strategies to optimize the use of resources and ensure biological success. Increasing complexity may result from diversification, shuffling, and re-arrangement of an otherwise limited functional genomic complement. To further test the concept of relative structural plasticity of the TSH receptor we sequenced the TSHR gene of two Old World monkey species Macaca mulatta and Cercopithecus aethiops, evolutionary removed from Homo sapiens by >20Myr. Both genes encoded a protein of 764 residues. This structure was 99% homologous between the two species of Old World monkeys while C. aethiops was 97% and M. mulatta was 96% homologous to H. sapiens. TSHR sequence comparisons were sought for an additional eight mammals as well as four (two Salmon, Tilapia, and Sea Bass) from teleosts. The amino-acid sequences of the 14 TSH receptors were similar. The most variable sequences were those of the intracellular tail and the distal cysteine-rich C-terminus flanking region of the ectodomain, whereas the trans-membrane domain was most preserved. Some sequences were decidedly H. sapiens specific, while others were primate specific or showed the changes expected of evolutionary descent. Others, however, exhibited "cross-species polymorphism," sometimes at quite remarkable evolutionary distances. As opposed to H. sapiens the sequence differences may have subtle influences on TSHR function or may affect long-range compensation for radical changes in adducts. The two Old World monkeys share with other lower mammals the absence of a glycosylation site at 113-115. Sea Bass and Tilapia have four glycosylation sites, whereas the two salmon receptors have only three. Changes in some critical residues raise questions about variation in function: thus S281 is conserved in all mammals and an important determinant of negative agonist function of TSHR is replaced by R in Sea Bass. Likewise the K183, found at an important transitional region at LRR 6 conserved in all mammals, is represented by M in fish and may contribute to TSHR lutenization in fish. There is no evidence that evolutionary changes in primate receptors are more rapid than that in other mammals and the separation times of different mammals based on silent nucleotide changes of TSHR are closely parallel to archaeological estimates. Results of correlated mutation analysis, referenced to the rhodopsin crystal structure, affirms dimerization of TSHR transmembrane helices. In addition, it suggests the involvement of critical lipid-facing residues in the helices in receptor dimerization and oligomerization. We highlight the value of evolutionary informatics and set the stage for dissecting out potential subtle differences in TSHR function associated with structural variations.

Use of anti-idiotypic antibodies as probes for the interaction of TSH subunits with its receptor

TSH is a heterodimeric glycoprotein hormone, whose dissociated subunits are without biological activity. This has precluded the assessment of the relative contribution of each subunit to hormone action. We have raised anti-idiotypes to monoclonal antibodies specific, respectively, for the alpha and beta hTSH subunits. The anti-beta anti-idiotype inhibited 125I-hTSH binding to the beta subunit-specific monoclonal quantitatively, whereas 125I-hTSH binding to the alpha subunit-specific monoclonal was not inhibited by anti-alpha anti-idiotypes, suggesting that only the former is an "internal image" anti-idiotype. Neither of the two anti-idiotypes nor equimolar mixtures thereof inhibited 125I-bTSH binding to thyroid membranes, even though radiolabelled anti-idiotypes showed saturable binding to thyroid plasma membrane which was inhibited 41-65% by bTSH. Each anti-idiotype alone caused 9% inhibition (compared to 50% by NRIgG) of thyroid plasma membrane adenylate cyclase. Equimolar mixtures (125 micrograms/ml IgG of each anti-idiotype) induced enzyme activity equivalent to 85% of that of 250 mU/ml of TSH. The TSH-like action of the two anti-idiotypes was also reflected in their capacity to increase (450% by 250 micrograms/ml IgG compared to normal rabbit IgG) the uptake of 131I into isolated thyrocytes and to promote the organization of such cells into follicular structures. At 250 micrograms/ml, anti-beta anti-idiotype promoted the organization of small follicles and only at a concentration of 500 micrograms/ml did it enhance 131I uptake.

An anti-idiotypic antibody against Graves' IgG

Rabbit antibodies were raised against Graves' IgG adsorbed onto a TSH-receptor affinity and eluted thereof by [3H]NaCl. The rabbit serum absorbed against normal human IgG (ARI) still bound to Graves', control and Hashimoto's IgG preparations but in the latter two, binding was inhibited by bTSH (10 mU/ml). In addition, ARI stimulated thyroid cell cyclic AMP accumulation in both human and rat thyroid cells. The ARI preparation may, therefore, contain an "internal image" anti-idiotype causing thyroid (Ab2) stimulation, a Graves' disease specific anti-idiotype whose binding with Ab2 inhibits its ability to bind TSH and anti-anti-idiotype (Ab3) to "internal image" Ab2. In further studies, Graves' specific cross-reactive idiotype was found in 10/11 IgGs from patients with active Graves' disease. This study emphasizes the workings of Jerne's immunologic network and the complexity of polyclonal "anti-idiotypic" antibodies.

Assessment of the shape and molecular size of TSH-TSH receptor complexes

Photoaffinity labelling and analysis under denaturing conditions (SDS-PAGE) have shown that the porcine TSH receptor contains an A subunit (Mr = 47,000) which forms the binding site for TSH and a B subunit (Mr = 25,000) linked to the A subunit by a disulphide bridge. In order to assess the size and shape of the receptor under non-denaturing conditions we have solubilized photoaffinity-labelled porcine TSH receptors using the small micelle-sized detergent sodium deoxycholate and analysed the preparations by sucrose density gradient centrifugation and gel filtration. Under these conditions, the cross-linked TSH-TSH receptor complex showed an S20,w of 6.4 S and a frictional ratio f/f0 of 1.8. These values were consistent with those which might be expected from an elongated protein complex with a molecular weight of about 100,000 (the value obtained by SDS-PAGE). Analysis of another thyroid membrane protein, human thyroid microsomal antigen (Mr = 110,000 by SDS-PAGE) under the same conditions gave an S20,w of 6.0 S and f/f0 = 1.3, suggesting that this protein has a compact structure. The TSH receptor A subunit cross-linked to TSH (Mr = 70,000 by SDS-PAGE) gave an S20,w of 4.6 S and f/f0 = 1.8 and these values could be compared with those obtained for the A subunit alone (S20,w = 3.6 S; f/f0 = 1.4; Mr by SDS-PAGE = 47,000) and TSH alone (S20,w = 2.6 S; f/f0 = 1.6; Mr = 28,000.(ABSTRACT TRUNCATED AT 250 WORDS)

The thyroid "microsomal" antigen is an epitope on the thyrotropin receptor

Antimicrosomal antibodies are present in the sera of most patients with autoimmune thyroiditis, and Graves' disease. It has, in general, been difficult to separate antimicrosomal activity from that directed against the thyrotropin (TSH) receptor in Graves' IgG preparations. The "microsomal" antigen has been localized to the endoplasmic reticulum and microfollicular aspect of thyrocytes; its structure is however unknown. In an attempt to identify the thyroid microsomal antigen, we studied the interaction of Hashimoto's IgG with high microsomal antibody titre and negative for thyroglobulin with purified thyroid plasma and light microsomal membranes. We allowed Hashimoto's, Graves', and control IgGs to bind to protein blots of thyroid plasma membranes resolved on SDS-PAGE under non-reducing conditions. All seven Hashimoto's IgG at a concentration of 2 mg/ml interacted with an M approximately 197,000 polypeptide corresponding to the TSH holoreceptor. By contrast to Graves' IgG (which were positive at 1 mg/ml), however, this binding was not blocked by pretreatment of the protein blots with TSH. Normal IgGs showed no binding at concentrations of up to 2 mg/ml. Both Hashimoto's and Graves' IgG interacted with TSH-affinity column-purified receptor preparations. Two of the Hashimoto's IgGs induced adenylate cyclase activation in thyroid plasma membranes, three inhibited TSH-stimulated enzyme activation, and two were without effect. Two classes of autoantibodies, other than TSH receptor directed, were encountered; one class raised to antigens common to all seven patients and another class unique to individual patients, eg, Mr 210,000 and Mr 20,000 polypeptides. We propose that the TSH receptor has multiple epitopes (functional domains), and the one to which antimicrosomal antibody bind is likely to be spatially separated from that with which Graves' IgG and TSH interact. Differences in affinity or number of sites allows for the demonstration of Graves' IgG against a background of antimicrosomal antibody.