The Thyrotropin Receptor Hinge Region as a Surrogate Ligand: Identification of Loci Contributing to the Coupling of Thyrotropin Binding and Receptor Activation

Abstract

The TSH receptor (TSHR) hinge region, the least well understood component, bridges the leucine-rich repeat and transmembrane domains. We report data on clusters of hinge charged residues the mutation of which to Ala is compatible with cell surface expression and normal, or near normal, TSH binding affinity yet with a relative reduction in receptor activation. Mutation to Ala of E409 at the junction with the transmembrane domain was the most potent in uncoupling TSH binding and signal transduction (∼22-fold less sensitive than the wild-type TSHR) and was unique among the residues studied in reducing both the amplitude and the sensitivity of the ligand-induced signal. Unexpectedly, a dual E409A/D410A mutation partially corrected the major suppressive effect of TSHR-E409A. The combined Ala substitution of a cluster of positively charged hinge residues (K287, K290, K291, R293; termed “K3R1”) synergistically reduced sensitivity to TSH stimulation approximately 21-fold without altering the TSH binding affinity. Simultaneous Ala substitutions of a cluster of acidic hinge residues D392, E394, and D395 (termed “DE392-5A”) partially uncoupled TSH binding from signal transduction (4.4-fold reduction in sensitivity), less than for E409A and K3R1A. Remarkably, the combination of the K3R1A and DE392-5A mutations was not additive but ameliorated the major uncoupling effect of K3R1A. This lack of additivity suggests that these two clusters contribute to a common signaling pathway. In summary, we identify several TSHR hinge residues involved in signal transmission. Our data support the concept that the hinge regions of the TSHR (and other glycoprotein hormone receptors) act as surrogate ligands for receptor activation.

The TSH receptor (TSHR) plays a vital role in metabolic homeostasis as well as in disease pathophysiology. For the former function, the TSHR is an integral component in the hypothalamo-pituitary-thyroid axis that regulates the exposure level of all tissues to thyroid hormones. In terms of disease pathogenesis, stimulatory autoantibodies to the TSHR are the proximal cause of hyperthyroidism in Graves' disease, one of the most common autoimmune diseases with a prevalence of approximately 1%. Understanding the structure and function of the TSHR is, therefore, an important investigative goal. Indeed, there is strong evidence that posttranslational intramolecular cleavage of the TSHR into disulfide-linked A- and B-subunits (1, 2) with shedding of the former (3, 4), a process unique among the glycoprotein hormone receptors, contributes to the pathogenesis of Graves' disease (5–7).

Of the three TSHR structural components, 1) a leucine-rich repeat domain (LRD) linked by a 2) hinge region to the 3) seven helical transmembrane domain (TMD), only the structure of the hinge region remains unknown. The TSHR LRD, possibly missing an 11th leucine-rich repeat (8), has been crystallized in complex with thyroid-stimulating (9) and TSH-blocking (10) autoantibodies, and the TSHR TMD can be modeled on an increasing number of crystal structures for other G protein-coupled receptors (11–13). Despite these major achievements, the steric relationship between the three TSHR structural components as an integral unit remains unknown; consequently, information necessary for understanding the mechanism by which ligand binding leads to receptor activation is lacking. The importance of the structurally undefined TSHR hinge region is that it is not simply an inert scaffold between the LRD and TMD. Mutagenesis studies have revealed that the TSHR hinge region contains a portion of the TSH binding site (14–20). In addition, a few amino acids in the hinge region have been identified, the mutation of which does not alter TSH binding affinity yet reduces TSHR signal transduction (17, 21, 22). Hinge residue E409 and D410 mutations are also reported to diminish signal transduction to a single maximal TSH dose despite normal or near-normal TSHR expression levels as determined by TSH binding or flow cytometry (18, 23, 24). Consequently, the hinge region contains components in the signaling pathway as well as residues involved in ligand binding.

Without information on the TSHR hinge structure, or a homologous protein whose known structure would permit modeling of the TSHR, mutagenesis studies are difficult to design and to interpret. Indeed, even the boundary between the TSHR LRD and TSHR hinge region is uncertain. The recombinant TSHR LRD protein generated for crystallization, residues 22–260 (1–21 being the signal peptide), implies a hinge region of approximately 150 residues (261–410) before entry into the plasma membrane. However, some TSHR hinge residues can be excluded from consideration. Thus, spontaneous intramolecular cleavage of the TSHR at the cell surface into disulfide-linked A- and B-subunits results in a functionally normal minimal hinge despite deletion of 65 residues (307–371) (Fig. 1). Further, homology with the FSH receptor LRD (25), as well as molecular modeling of an additional leucine-rich repeat at the C terminus of the TSHR LRD (26), suggests that the upstream boundary of the TSHR hinge region is at residue 277 rather than 261. In the present report we present findings on the signal-coupling function of three clusters of charged residues in the TSHR hinge region. The data obtained, together with previous observations, provide new insight into the role of the TSHR hinge region in ligand-mediated receptor activation.

Fig. 1.

Schematic representation of amino acids in a minimal TSHR hinge region, so named because spontaneous intramolecular cleavage removes a C-peptide region of 50–65 residues with poorly defined borders (48). Experimental deletion of residues 307–371 is the largest deletion compatible with normal TSHR function (22). The TSHR LRD is at the hinge N terminus, and the hinge enters the membrane at residue 411. Positively and negatively charged residues are shown in red and blue letters, respectively. Residues whose mutation increases ligand-independent constitutive activity (CAM) are circled in red. The thickness of the circles approximates the functional strength of the individual mutations at these sites. Hinge residues (selected) reported to contribute to the TSH binding site (14–20) are circled in green. Hinge residues E409 and D410 whose mutations to K suppress TSH stimulation disproportionately to the relatively normal level of maximal TSH binding and TSHR cell surface expression (E409K greater effect than D410K) (24) are circled in blue. The disulfide bonds (filled orange circles), deduced from experimental evidence (reviewed in Ref. 49) are splayed out (joined by the straight dashed lines) in order depict the hinge region amino acid residues in two dimensions. The 283–408 and 284–398 bonds shown are uncertain and may be reversed. The superficial perspective shown depicts the disulfide bonds behind the intervening residues, curving the latter outwards toward the viewer. This curvature, rather than the reverse, is likely because it draws the CAM residues (red circles), as well as E409 and D410, close to one another (large circles connected by curved dashed line). The major functional effects of the residues within these joined circles suggest close apposition to the three extracellular loops of the TMD, and also brings the TSH binding residues to the surface of the hinge region, necessary for accessibility to ligand. The thick ovals enclose the three groups of charged amino acids (large font) selected for study in the present report. We re-examined E409A and D410A (28) because of limited data on these mutations, and also investigated the two other clusters of charged residues that have not previously been studied by simultaneous alanine substitutions of all residues within the cluster.

Materials and Methods

TSHR cDNA mutations

Introduction of the wild-type human TSHR (hTSHR) cDNA (27)(with the H601 polymorphism converted to Y601) into the vector pcDNA5/FRT was described previously (17). Amino acid numbering includes the signal peptide. The TSHR cDNA mutations described below in the text were generated using the QuikChange Site-Directed Mutagenesis kit (Stratagene, San Diego, CA). Multiple amino acid substitutions were performed sequentially. All mutations were confirmed by nucleotide sequencing.

TSHR expression

TSHR cDNA were transfected into Flp-In-CHO cells (Invitrogen, Carlsbad, CA) using Fugene HD (Roche, Indianapolis IN). Cell lines stably expressing the TSHR were obtained by selection with hygromycin B (Invitrogen; ∼300 μg/ml). Cells were cultured in Ham's F12 medium supplemented with 10% fetal calf serum, penicillin (100 U/ml), gentamycin (50 μg/ml), and fungizone (2.5 μg/ml).

Cultured cell cAMP assays

Chinese hamster ovary (CHO) cells stably expressing the wild-type or mutated TSHR were transferred into 96-well plates (0.8–1.0 × 10⁵ cells per well). For bioassay, the culture medium described above was replaced with F12 medium supplemented with 1 mm isobutyl methylxanthine, 10 mm HEPES, 0.3% BSA and, where indicated in the text, bovine (b) TSH (Sigma-Aldrich, St. Louis MO). Untransfected CHO cells were included as controls. After 60 min at 37 C, the medium was aspirated and intracellular cAMP was extracted with 0.2 ml 95% ethanol. The extracts were evaporated to dryness, resuspended in 0.1 ml of PBS, pH 7.5, and samples (12 μl) were assayed using the LANCE cAMP kit according to the protocol of the manufacturer (PerkinElmer, Shelton CT). The effective dose of TSH required for half-maximal stimulation of intracellular cAMP levels (EC₅₀₎ was calculated using GraphPad Prism software (La Jolla, CA).

TSH binding

CHO cells expressing the TSHR were cultured in 24-well plates (∼4 × 10⁵ cells per well). Highly purified bovine TSH (5 μg; ∼30 U/mg protein) was radiolabeled with ¹²⁵I to a specific activity of approximately 50 μCi/μg protein using the Bolton-Hunter reagent (2200 Ci/mmol; PerkinElmer) according to the protocol of the manufacturer. Medium was aspirated and replaced with 250 μl binding buffer (Hanks buffer with 250 mm sucrose substituting for NaCl to maintain isotonicity and 0.25% BSA) containing about 30,000 cpm [¹²⁵I]TSH and the indicated concentrations of unlabeled bTSH. After incubation for 4 h at room temperature, cells were rapidly rinsed three times with binding buffer (4 C) and solubilized with 0.2 ml 1 n NaOH, after which radioactivity was then measured in a γ-counter. Calculating the TSH binding affinity from these data is difficult (at least in our experience) because the physiological high-affinity TSH binding site overlaps extensively with a nonspecific, low affinity-high capacity binding site with which TSH interacts (unlike the other glycoprotein hormones). Data reported in the present study were obtained using a nonlinear regression, one-site model GraphPad Prism (La Jolla, CA), excluding TSH data points above 1–3 mU/ml. The two-site nonlinear regression model does not sufficiently separate the high- and low-affinity binding sites. The data obtained were comparable to those determined by Scatchard analysis, also excluding the nonspecific low-affinity binding site.

Statistics

Significance of differences between receptors was assessed by Student's t test (SigmaPlot; Systat Software, San Jose, CA).

Results

TSHR hinge residues E409 and D410

These residues, the most C terminal in the TSHR ectodomain, are critically involved in TSH-mediated signal transduction. Mutation of E409 to Ala (28) and residue D410 to Lys (24) or Asn (23) substantially reduced cAMP responses to a single maximal TSH dose despite TSH binding comparable to the wild-type TSHR. We reexamined TSHR with the E409A and D410A mutations individually as well as together (not previously done) to obtain detailed information on their functional responses to a range of TSH concentrations. With TSHR-D410A, consistent with the previous report (28), the maximal TSH-induced cAMP response was 36.9 ± 6.2% (mean ± sem) of the wild-type TSHR (Fig. 2 and Table 1). However, the maximal TSH binding (B_max) to TSHR-D410A was 47.0 ± 7.7% (mean ± sem) of the wild-type TSHR, indicating a relatively normal maximal cAMP response given the lower level of receptor expression. In addition, the TSH concentration required to attain a half-maximal response (EC₅₀) for TSHR-D410A was not significantly higher than for the wild-type TSHR (Table 1).

Fig. 2.

TSH stimulation of cAMP with TSHR mutants E409A and D410A. In this and subsequent experiments, CHO cells stably expressing TSHR with the indicated mutations, as well as the wild-type TSHR, were incubated for 1 h in the indicated concentrations of TSH and intracellular cAMP measured in duplicate wells (Materials and Methods). Each point represents the mean ± sem of cAMP values from the following number of experiments; wild-type TSHR, n = 7; CHO-K1 (untransfected cells), n = 7; E409A, n = 6; D410A, n = 4; and E409A+D410A, n = 4. The wild-type TSHR data were from parallel dishes of cells in the same experiment. The concentrations of TSH required for half-maximal stimulation of cAMP (EC₅₀) and the maximal cAMP values attained in this and subsequent experiments are provided in Table 1 and were calculated by nonlinear regression analysis (GraphPad Prism). WT, Wild type.

Table 1.

Summary of data on TSH stimulation of cAMP generation by, and [¹²⁵I]TSH binding to, TSHR mutants in stably transfected CHO cells

	TSH stimulation of cAMP					TSH binding
	Basal level (pmoles per well)	Maximum (pmoles per well)	Maximum (% of wild type)	EC50 (mU/ml)	EC50 (fold-wild type)	Kd (mU/ml)	Bmax (fmoles TSH per well)	Bmax (% of wild type)
Wild type	0.43 ± 0.04	20.6 ± 0.9	100	0.31 ± 0.05	1.0	2.1 ± 0.58	245.1 ± 55.2	100
E409A	0.31 ± 0.07	1.1 ± 0.4	5.3 ± 1.8	6.75 ± 0.41	21.8 ± 2.8a	1.4 ± 0.13	113.9 ± 12.9	46.5 ± 11.3
D410A	0.38 ± 0.10	7.4 ± 1.3	36.0 ± 6.2	0.71 ± 0.22	2.3 ± 0.38	1.2 ± 0.11	115.2 ± 8.8	47.0 ± 7.7
ED409-410A	0.43 ± 0.04	3.3 ± 0.9	16.2 ± 4.5	1.72 ± 0.05	5.5 ± 0.09a	1.0 ± 0.14	115.3 ± 12.5	47.0 ± 10.8
Wild type	0.53 ± 0.10	37.1 ± 3.4	100	0.30 ± 0.02	1.0	1.7 ± 0.3	216.1 ± 49.1	100
K287A	0.53 ± 0.12	36.5 ± 2.3	98.4 ± 6.2	0.50 ± 0.05	1.7 ± 0.03b	2.1 ± 0.4	160.8 ± 47.1	74.4 ± 29.3
K290A	0.49 ± 0.04	31.7 ± 3.1	85.4 ± 8.4	0.74 ± 0.06	2.5 ± 0.06c	1.8 ± 0.3	94.0 ± 20.2	43.5 ± 21.5
K291A	0.63 ± 0.06	33.6 ± 4.2	90.6 ± 11.3	0.70 ± 0.11	2.3 ± 0.09b	1.5 ± 0.2	121.4 ± 36.3	56.2 ± 29.9
R293A	0.56 ± 0.04	30.8 ± 2.6	83.0 ± 7.0	0.47 ± 0.03	1.6 ± 0.02c	2.6 ± 0.5	229.7 ± 74.5	106.3 ± 32.4
K3R1	0.43 ± 0.09	24.1 ± 3.3	65.0 ± 8.9	6.34 ± 0.93	21.1 ± 6.8a	2.6 ± 0.1b	99.1 ± 6.9	45.9 ± 6.9
Wild type	0.43 ± 0.03	25.2 ± 5.1	100	0.26 ± 0.02	1.0	2.1 ± 0.13	226.8 ± 41.7	100
D392A	0.48 ± 0.21	32.5 ± 4.5	129.0 ± 17.9	0.24 ± 0.02	0.9 ± 0.01	1.8 ± 0.24	183.5 ± 23.4	80.9 ± 23.4
E394A	0.52 ± 0.05	26.0 ± 4.0	103.2 ± 15.9	0.34 ± 0.05	1.3 ± 0.01	2.7 ± 0.30	257.4 ± 57.2	113.4 ± 22.2
D395A	0.39 ± 0.07	25.8 ± 5.5	102.4 ± 21.8	0.39 ± 0.03	1.5 ± 0.01b	1.7 ± 0.27	136.2 ± 37.1	60.1 ± 27.2
DE392-5A	0.42 ± 0.08	25.1 ± 2.9	99.6 ± 11.5	1.14 ± 0.11	4.4 ± 0.13c	3.1 ± 1.01	164.3 ± 34.5	72.5 ± 21.0
Wild type	0.52 ± 0.05	28.0 ± 1.8	100	0.33 ± 0.03	1.0	1.5 ± 0.16	209.7 ± 29.7	100
K3R1 + 392	0.47 ± 0.03	22.3 ± 2.7	79.6 ± 9.8	2.90 ± 0.80	8.7 ± 2.3b	1.4 ± 0.01	84.7 ± 2.7	40.4 ± 1.5
K3R1 + 394	0.52 ± 0.03	21.7 ± 2.0	77.5 ± 7.3	3.2 ± 0.87	9.8 ± 2.8b	2.5 ± 0.29	123.03 ± 8.3	58.7 ± 4.5
K3R1 + 395	0.53 ± 0.06	17.1 ± 1.6	61.1 ± 5.7	5.6 ± 2.05	16.9 ± 11.4	2.7 ± 0.38	128.9 ± 15.9	61.5 ± 8.8
K3R1 + 392–5	0.61 ± 0.05	21.1 ± 3.6	75.4 ± 12.9	1.7 ± 0.39	5.0 ± 0.7b	1.9 ± 0.23	117.6 ± 4.8	56.1 ± 2.7

Data represent the mean ± sem of three to seven separate experiments (with each datum point determined in duplicate). For comparison with the wild-type TSHR, values are normalized to the latter expressed as 1.0 or 100%. TSH stimulation of cAMP generation and [¹²⁵I]TSH binding were performed on frozen aliquots stored from the same cell line. For the same individual cell line, maximum levels of cAMP attained that are less than half maximum TSH binding (B_max) values are underlined and shown in bold. Intercell line comparison of maximum cAMP and TSH binding values is less reliable than sensitivity to TSH stimulation (EC₅₀ values) or TSH binding affinities (K_d values) because the latter values are less dependent on cell number or level of TSHR expression in stably transfected cells lines, for the following reasons: 1) Estimates of cell number plated may vary by 10–20%, and not all cells plated will remain viable and adherent; 2) Although all cell lines contain a single transgene at the same genomic site and are largely clonal, there is variability in the level of TSHR expression (dependent on receptor synthesis and trafficking efficiency); and 3) Cells are plated the day before experiments are performed during which time the rate of cell division may be influenced by the transgene. Significantly different from wild type (t test):

^a, <0.001;

^b, <0.05;

^c, <0.01.

TSHR-E409A responded to TSH stimulation in a different manner to TSHR-D410A (Fig. 2). The maximal cAMP response (only ∼5% of the wild type) was far lower than the TSH binding B_max (∼47% of wild-type), the latter similar to that with TSHR-D410A (detailed values in this and subsequent experiments provided in Table 1). Remarkably, combining the two mutations (ED409-410A) partially reversed the suppressive effect of the E409A mutant on the maximal cAMP response (16% of wild-type), again with a TSH binding B_max comparable to the individual E409A and D410A mutants (Fig. 2 and Table 1). For all three mutants studied (E409A, D410A, and ED409-410A), TSH binding affinities were not reduced relative to the wild-type TSHR (Table 1). It is important to note that, despite the low levels attained, the cAMP responses to TSH of E409A and ED409-410A reached or approached a plateau (Fig. 2), permitting nonlinear regression analysis of their TSH EC₅₀ values; 21.8-fold and 5.5-fold higher than for the wild-type TSH, respectively (Table 1).

TSHR hinge residues K287, K290, K291, and R293

Another cluster of positively charged amino acids in the TSHR hinge region includes K287, K290, K291, and R293 (Fig. 1). Although these residues have previously been mutated individually to alanine, with increased constitutive activity reported for K291A and increased cAMP responses to a single maximum TSH dose reported for K287A, K290A, and R293A (28), we wished to reexamine the functional effects by performing full TSH dose-responses with these individual mutations, as well as with a receptor combining these four mutations (for brevity termed “K3R1A”). Using CHO cells stably expressing the receptors with individual mutations, we confirmed that for the four individual mutations the maximal cAMP responses to TSH were comparable to the wild-type TSHR (Fig. 3 and Table 1). Nevertheless, these receptors were significantly less sensitive to TSH stimulation, with EC₅₀ values approximately 2-fold greater than for the wild-type TSHR (Table 1). TSH binding affinities [dissociation constant (K_d)] for the individual mutations were comparable to those of the wild-type TSHR (Table 1).

Fig. 3.

TSH stimulation of cAMP with TSHR mutants K287, K290, K291, and R293 and with all four mutations combined (K3R1A). Data were obtained as described in the legend to Fig. 2. For the wild-type TSHR and for the TSHR with all four mutations combined (K3R1A), each point represents the mean ± sem of cAMP values from six separate experiments. The data for the individual mutations were obtained in three separate experiments. The concentrations of TSH required for half-maximal stimulation of cAMP (EC₅₀) and the maximal cAMP values (including for K3R1A which did not attain a plateau) were calculated by nonlinear regression analysis and are indicated in Table 1.

For the TSHR with all four mutations (K3R1A), the cAMP response to the highest concentration of TSH (∼65% of the wild-type TSHR) was not diminished relative to the TSH binding B_max (∼46% of wild-type)(Table 1). On the other hand, the EC₅₀ for TSH stimulation was much greater (∼21-fold) than for the wild-type TSHR (Table 1). The foregoing functional responses for TSHR-K3R1A could be calculated by linear regression analysis despite cAMP values not attaining a plateau (Fig. 3). Although the TSH binding affinity for K3R1A was slightly lower (K_d 2.6 ± 0.1 mean ± sem) than for the wild-type TSHR in parallel wells of cells (K_d 1.7 ± 0.3; P < 0.05), this reduction was disproportionately less than the reduction in the functional cAMP response (reflected by an increase in the EC₅₀).

Although not the goal of the present study (coupling of TSH binding and function), it was of interest that the basal cAMP levels in cells expressing the K291A mutant (0.63 ± 0.06 pmol/well) were the highest among the group of mutants (Table 1), although not significantly different than the wild-type TSHR. Correction for the lower level of expression relative to the wild type would be consistent with the previous report of TSHR-K291A being a weak constitutively activating mutation (28).

TSHR hinge residues D392, E394, and D395

Another notable cluster of charged residues in the TSHR hinge region comprises negatively charged D392, E394, and D395 (Fig. 1). Of these residues, E394 (previously mislabeled D394) and D395 have been mutated to alanine with relatively normal cell surface expression and maximal TSH binding (18). To evaluate all three residues in detail, we performed full TSH dose-response curves on stably expressed TSHR mutants with D392A, E394A, and D395A, as well as a receptor with all three mutant combined (for brevity termed “DE392-5A”). With the individual mutations, TSH elicited a robust cAMP response with maximal cAMP levels attained and sensitivity to TSH stimulation (EC₅₀ values) comparable to the wild-type TSHR studied in parallel dishes of cells (Fig. 4 and Table 1). Consistent with these normal functional responses, their TSH binding affinity and B_max values were not significantly different from the wild-type TSHR (Table 1). However, the combination of all three mutations decreased the sensitivity of the cAMP response to TSH stimulation with an EC₅₀ 4.4 ± 0.1 (mean ± sem; P < 0.01) greater than the wild-type TSHR (Table 1). With TSH binding kinetics not being significantly different than the wild-type TSHR (Table 1), these data indicate modest (compared with E409A and K3R1A) uncoupling of TSH binding and signal transduction.

Fig. 4.

TSH stimulation of cAMP with TSHR mutants D392, E394, and D395 and with all three mutations combined (DE392-5A). Data were obtained as described in the legend to Fig. 2. Each point represents the mean ± sem of cAMP values from the following number of experiments; wild-type TSHR, n = 4; D392A, n = 6; E394A, n = 3; D395, n = 4; and DE392-5, n = 4. The concentrations of TSH required for half-maximal stimulation of cAMP (EC₅₀) and the maximal cAMP were calculated by nonlinear regression analysis and are indicated in Table 1.

Combination of multiple TSHR mutants K3R1A and DE392-5A

As described above, uncoupling of TSH binding and function (induction of cAMP generation) was moderately severe with TSHR combined mutant K3R1A and modest with combined mutant DE392–395A. Therefore, we hypothesized that the addition to K3R1A of the individual mutations (K3R1A+D392A; K3R1A+E394A; K3R1A+D395A), or with all seven mutations combined (K3R1A+DE392-5A) would be additive, if not synergistic, and result in a very poorly responsive receptor. Contrary to our hypothesis, the individual mutations added to K3R1A did not reduce, and even slightly improved, the sensitivity (reduced EC₅₀) of the K3R1A cAMP response to TSH (Table 1). Dramatically, however, complementing K3R1A with the combination of all three mutations (DE392-5A) markedly improved the K3R1A sensitivity to TSH, with an EC₅₀ comparable to DE392-5A on its own (Fig. 5 and Table 1). For all TSHR mutants studied, the TSH binding affinities were comparable to the wild-type TSHR.

Fig. 5.

TSH stimulation of cAMP with TSHR with combined K3R1A and DE392-5A mutations. Data were obtained as described in the legend to Fig. 2. Each point represents the mean ± sem of duplicate cAMP values from three experiments. The concentrations of TSH required for half-maximal stimulation of cAMP (EC₅₀) and the maximal cAMP were calculated by nonlinear regression analysis and are given in Table 1. WT, Wild type.

Discussion

The hinge region is the least well-understood structural component of the TSHR yet plays an important role in TSH binding and subsequent transmission of a signal to the TMD. A number of hinge residue mutations activate the TSHR in the absence of ligand (increased constitutive activity), in particular mutations involving S281 (29, 30) and a cluster of residues, D403, E404, and N406, close to the hinge-TMD (31)(Fig. 1, circled in red). Other TSHR hinge residue mutations are associated with reduced ligand-induced signal transduction either because of diminished ligand binding, e.g. Y385 (14, 16), E297, D382, and D386 (18)(Fig. 1, circled in green), or because of improper folding and trafficking to the cell surface. The present study focuses on TSHR hinge residues the mutation of which is compatible with cell surface expression and normal, or near-normal, TSH binding affinity yet with a relative reduction in receptor activation. Partial uncoupling of ligand binding from signal transduction (as observed with mutation of TSHR E251 and adjacent residues at the C terminus of the LRD (32, 33) is reported to occur with mutation of TSHR E409 and D410 at the hinge-TMD junction (23, 24, 28) (Fig. 1, circled in blue). These two residues are conserved among the glycoprotein hormone receptors (34). Spontaneous mutations of D410N in the human TSHR (23) and E354K in the human LH receptor (homologous to TSHR E409) (35) lead to in vivo loss of function.

TSHR residues E409 and D410

Previously, E409K and D410K mutations introduced ex vivo were observed to reduce cAMP responses using a single maximal TSH dose to 22% and 25% of the wild-type TSHR response despite maximal TSH binding (B_max) values more closely aligned with the wild-type TSHR (90% and 63%, respectively) (36). However, substitution of a residue with a longer side chain, as well as of opposite charge (K), is more likely than substitution with Ala (single carbon, neutral side chain) to alter protein folding with possible allosteric effects. Indeed, with the D410A mutation, the maximal cAMP response (71% of the wild-type TSHR) was proportional to the TSH binding B_max (73% of wild type) (28). Because these previous studies were limited to examining TSH binding B_max values and maximal cAMP responses, we more fully explored the properties of the E409A and D410A mutations, singly and together, on TSH binding affinity (K_d) and the sensitivity of the cAMP response to TSH (EC₅₀). In addition, we generated the dual mutation receptor because mild Ala substitutions may reveal functional effects not detected with single mutations (for example, with the TSHR LRD (33).

To summarize the novel information obtained in the present study:

1. With the TSHR-E409A mutation, although we confirm blunting of the maximum cAMP response to TSH relative to the TSH binding B_max (5% and 47% of wild-type, respectively), this blunting is more severe than previously observed (38% and 83% of wild-type, respectively) (28). Moreover, not reported previously, there is also a major reduction in sensitivity to TSH stimulation of cAMP (EC₅₀ 22-fold greater than that of the wild-type TSHR)(Table 1).

2. Most unexpected, rather than having an additive or synergistic effect, the dual E409A/D410A mutation partially corrected the major suppressive effect of TSHR-E409A on both the sensitivity to TSH stimulation and the maximal cAMP level attained (Table 1). The potential implications of these findings are discussed below.

TSH receptor residues K287, K290, K291, and R293

Given the importance of charged residues E409 and D410 in coupling TSH binding and receptor activation, we focused on the most prominent cluster of charged residues in the TSHR hinge region (Fig. 1), a site at which proteolytic cleavage increases TSHR activity by reducing the inherent inverse agonist activity of ectodomain (37). Our data differ in some respects from previous, more limited data on individual Ala substitutions (28) and provide novel information, as follows:

1. In contrast to the conclusion that mutations to Ala of residues K287, K290, and R293 “have no effect on signal transduction” when tested with a single TSH dose (28), with full dose-response curves we find that the sensitivity to TSH of these mutations, as well as K291A, is reduced approximately 2-fold relative to the wild-type TSHR without significant reduction in the TSH binding affinity (Table 1).

2. The combined substitution of all four residues (termed “K3R1A”), not previously examined, has a major synergistic effect in increasing the TSH EC₅₀ approximately 20-fold, also without alteration in the TSH K_d.

Although the goal of our study was to examine coupling of TSH binding and signal transduction, we observed that the ligand-independent constitutive activity of K291A was the highest of the receptors in this group, although not significantly higher than the wild type. Relative to the lower level of expression of K291A, the data are consistent with the report that K291A has increased constitutive activity (28).

TSHR residues D392, E394, and D395

In this final cluster of charged TSHR hinge region residues investigated (in this case negatively charged), our detailed analysis determining bTSH functional sensitivity (EC₅₀) and binding affinity (K_d) are consistent with previous data limited to maximal parameters (20) that individual E394A and D395A mutations have little effect on receptor function with this ligand. However, the new or contrary data in our present study are:

1. The maximal TSH-mediated cAMP responses for TSHR with E394A and D395A mutations are similar to those of the wild-type, rather than being substantially reduced, as reported previously (20). Moreover, the EC₅₀ values for receptors for these mutations (not reported previously) are comparable to those for the wild-type TSHR.

2. Our data, including determination of TSH binding affinities (K_d), do not confirm that, based on Ala substitutions for D392 (20), E394, and D395 (18), these residues contribute to the TSH binding site.

3. Most important, in contrast to individual mutations, the simultaneous Ala substitutions for all three residues partially uncoupled TSH binding from signal transduction (TSH EC₅₀ 4.4-fold greater than that of wild type), although to a far lesser extent than E409A and K3R1A (both approximately 21-fold greater than wild type) (Table 1).

TSHR combining the K3R1 quadruple and the DE392-5 triple mutations

Such receptors have not been investigated previously. That the DE392-5A mutations are not additive, but rather ameliorate the major uncoupling effect of K3R1A, provides insight into the ligand-induced activation pathway transmitted through the TSHR hinge region to the TMD. Attempting to restore wild-type function by reversing the positively and negatively charged residues is a highly complex undertaking given the large and unequal number of residues at each locus. Moreover, we observed that substituting K3R1 with residues of the opposite charge led to loss of cell surface expression (data not shown).

Inconsistencies between some of the data obtained and conclusions made in the present and previous studies demonstrate the obvious need for independent experimental confirmation. Possible methodological explanations for these differences include:

1. Stable vs. transient transfections. Unlike with the latter, the expression system that we employed stably introduces a single transgene into a defined genomic site.

2. The methodology used to determine the extent of cell surface expression, i.e. flow cytometry using monoclonal antibodies vs. TSH binding kinetics. Each of these approaches has limitations. In the present study, we used the latter because of the lesser sensitivity of flow cytometry (at least in our hands), but primarily because some of our hinge region mutations influenced binding by the most effective and commonly used monoclonal antibodies such as 2C11 (38).

3. Determining TSH binding kinetics and TSH EC₅₀ values as opposed to maximal TSH binding and maximal TSH stimulation values.

Another methodological approach requiring comment is the use of multiple, mild Ala mutations vs. individual mutations either to Ala, which may have little effect, or to residues with greatly different properties that may alter protein folding with possible allosteric effects. Our present findings demonstrate the value of this approach.

Initially overlooked as a functional component of the TSHR, there is now strong evidence that the hinge region linking the LRD to the TMD is a conduit for TSH-induced signal transduction. G protein-coupled receptor, of which the TSHR is a family member, are activated by shifts in the relative positions of the membrane-spanning helices that in turn impact upon intracellular G proteins (39). Small molecule ligands for the β-2-adrenergic (40), CXCR4 (41), dopamine D3 (42), and A2A adenosine (43) receptors enter the plasma membrane and directly contact the transmembrane helices. Synthetic small molecule ligands are also likely to interact with the TSHR in a similar manner (44–46). Larger ligands like the glycoprotein hormones do not interact directly with the TMD extracellular loops, and it has been suggested that the immediate agonist for the TSHR is the ectodomain itself (47). Our present data amplify this concept to suggest that it is the hinge regions of the TSHR (and other glycoprotein hormones) that act as surrogate ligands which undergo conformational changes thereby transmitting a signal to the TMD extracellular loops. Hinge region conformational changes can be induced by ligand binding, primarily to the LRD but also in the case of TSH to the hinge region itself. Unlike TSH, thyroid stimulating antibodies in Graves' disease do not directly contact the TSHR hinge region (9) but steric hindrance to their LRD binding site(s) may induce torsion in the hinge region (5, 22, 33).

Because the absence of a protein of known structure with sufficient homology to the TSHR hinge region precludes providing a reliable molecular model, we present our interpretation of our functional data schematically (Fig. 6). These data provide insight into some of the hinge residues involved in signal transmission. Although hydrophilic, the K3R1 and DE392-5 clusters could lie within the core of the hinge region as well as contributing to the outer surface. We recognize that other, particularly hydrophobic, residues could lie within the hinge core and contribute to signal transmission. The fact that the simultaneous mutation to Ala of both clusters is not additive (as might be expected) but, instead, partially restores signal transduction suggests that these two clusters contribute to a common signaling pathway (depicted schematically in Fig. 6). This remarkable phenomenon of signaling restoration may occur because the combination of Ala mutations permits receptor folding with more integrative residue interactions. It is logical that hinge residue E409 at the junction with the TMD is integral to the final pathway to signal transduction (28). Our data also reveal that E409 is unique among the residues studied in contributing to both the amplitude and the sensitivity of the ligand-induced signal.

Fig. 6.

Schematic concept of role of the hinge region in TSHR activation. In this model, the hinge region bridges the LRD and the extracellular loops (ECL) of the TMD that activates intracellular adenylate cyclase via Gs proteins. TSH binds primarily to the LRD but also makes contact with the hinge regions (see Fig. 1). LRD and hinge region residues involved in coupling ligand binding and signal transduction are indicated, with their influence proportional to the size of the type. The black ellipse indicates a proposed conduit by which conformational changes in the hinge region shift the alignment of the ECL and, thereby, alter the interrelationship of the transmembrane helices. Mutation of residue E251 (and, to a lesser degree, K250 and R255) at the base (C terminus) of the LRD contributes to coupling ligand binding and signal transduction (32, 33). Hinge residue E409 at the junction with the TMD is the final pathway to hinge signal transduction (28) and is unique among the residues studied in influencing both the amplitude and the sensitivity of the ligand-induced signal (present data). Two hinge region clusters of charged residues lie between E251 and E409, namely K287, K290, K291, R293 (K3R1) and D392, E394, D395 (DE392-5). The lack of additivity in the severity of ligand-signaling uncoupling when mutations of these clusters are combined suggests that these two clusters contribute to a common signaling pathway. Although hydrophilic, these residues could lie within the core of the hinge region or be on the outer surface. Other, particularly hydrophobic, residues could lie within the hinge core and be involved in signal transmission.