Structural Determinants for G Protein Activation and Selectivity in the Second Intracellular Loop of the Thyrotropin Receptor
http://www.100md.com
内分泌学杂志 2005年第1期
Third Medical Department, University of Leipzig (S.N., M.C., R.P.), D-04103 Leipzig, Germany; and Institute for Molecular Pharmacology (G.K.), D-13125 Berlin, Germany
Address all correspondence and requests for reprints to: Dr. Ralf Paschke, Third Medical Department, University of Leipzig, Ph. Rosenthal Strasse 27, 04103 Leipzig, Germany. E-mail: ralf.paschke@medizin.uni-leipzig.
Abstract
The TSH receptor (TSHR) activates mainly two signal transduction pathways, cAMP production and phosphoinositide turnover, mediated by Gs and Gq coupling, respectively. Several activating deletion and point mutations within intracellular loop 3 (ICL3) and the adjacent portion of transmembrane domain 6 (TM6) support a direct G protein activation by this receptor domain. The ICL3, however, is predicted by modeling to interact with other receptor domains, primarily ICL2, to form a pocket for G protein binding and to allow optimum interaction. Systematic mutagenesis was used to identify important sites within ICL2 and potential interactions between ICL2 and ICL3 of the TSHR required for G protein coupling. Deletions of four or five residues and their corresponding multiple alanine substitutions were introduced into ICL2. Residues I523-D530, comprising mainly the N-terminal half of ICL2, appeared to be critical for Gs- and Gq-mediated signaling. A single alanine substitution screening within ICL2 revealed hydrophobic residue M527 in particular and, to lesser extents, F525, R528, L529, and D530 as residues that selectively abolished or strongly impaired Gq activation. Molecular modeling suggests that F525 interacts with ICL3. To test this hypothesis, ICL2/ICL3 double mutants introducing strong complementary properties were constructed and tested for functional rescue of Gq-mediated signaling. Our results indicate that ICL2 interacts with ICL3 in close vicinity to F525 and T607, suggesting a conformational cooperation between ICL2 and ICL3 during Gq activation by TSHR.
Introduction
STIMULATION OF HETEROTRIMERIC G proteins by cell surface receptors activates signaling pathways that mediate specific responses to hormones and neurotransmitters. However, it is unknown how seven transmembrane (TM)-spanning receptors (7TMRs) activate G proteins and lead to this intricate and finely tuned process of downstream signaling. Selectivity of G protein recognition is determined by intracellular domains of 7TMRs (1). Numerous studies have established the pivotal role of intracellular loop 2 (ICL2) and ICL3 in G protein coupling (2). Most findings in the field of receptor/G protein coupling indicate that ICL3, particularly its N- and C-terminal regions, is important for G protein recognition (3). However, ICL3 is predicted to interact with other receptor domains, primarily with ICL2, to allow for optimum coupling efficiency and selectivity (1). The proposed movement of TM6 relative to TM3 indicated by spectroscopic analysis (4, 5) is consistent with this suggested scenario.
ICL2 seems to be important for the selectivity of receptor/G protein interactions and the efficiency of G protein activation (3). Investigation of ICL2 in the m5 muscarinic receptor by a random mutagenesis approach has shown that ICL2 contains a group of residues necessary for holding the receptor in an inactive conformation and another group necessary for G protein coupling. Therefore, it was hypothesized that ICL2 could act as a switch that enables G protein coupling (6). Moreover, in ICL2 of the m3 muscarinic receptor, four specific amino acids (aa) were identified as important for enabling the receptor to couple selectively to the Gq/11 protein (7). Recent findings for metabotropic receptors, which belong to class 3 of 7TMRs, underline the importance of ICL2 by demonstrating that the central portion of ICL2 is responsible for the selective recognition of the C-terminal end of the G-subunit (8).
The TSH receptor (TSHR) together with the LH and FSH receptors belong to the glycoprotein hormone receptor family. TSHR activation by several activating point and deletion mutants within ICL3 and the lower portion of TM6 supports a direct G protein activation by this region (9, 10). Interestingly, naturally occurring, activating TSHR mutations, which are one of the molecular causes of autosomal dominant, nonautoimmmune hyperthyroidism and toxic thyroid nodules (9, 11) (OMIM 603372), were predominantly localized to TM domains, extracellular loops, or ICL3 and rarely in the first or second ICL (12). For this reason many studies have focused on residues in ICL3, but only two studies have focused on ICL2 of the TSHR (13, 14). The first study concluded that the carboxyl-terminal domain of ICL2 (aa 528–537) is important for activation of Gs-mediated cAMP production (13). TSHR mutants created by substituting sequences from the 1- or ?2-adrenergic receptor demonstrated that the middle portion of ICL2 (residues 525–527) of the TSHR seems to be important for agonist-induced Gs interaction, whereas residues 528–532 were determined to be more critical for agonist-induced Gq activation (14). Mutagenesis studies of LH and FSH receptors have also provided evidence for an important functional role of several aa residues within ICL2 (15, 16). However, only single selected residues in ICL2 were substituted in these studies.
Less is known about the specific function of single residues of ICL2 and potential interaction with ICL3 in dual G protein activation by glycoprotein hormone receptors. Therefore, the first aim of our study was to identify and characterize specific intracellular aa of the TSHR involved in G protein activation and selectivity. In contrast to both former studies (13, 14), in which several adjacent aa within ICL2 of the TSHR were replaced simultaneously, systematic single site mutagenesis of the entire ICL2 was carried out. The second aim was to identify potential intramolecular interactions between ICL2 and ICL3 that might facilitate G protein activation. Deletions of four or five consecutive residues and multiple alanine mutations of the same residues (Fig. 1) localized the ICL2 region important for TSHR activation to residues 523–530. Our data reveal that M527 and, to a lesser extent, F525, R528, L529, and D530 are residues with selective influence on Gq activation. Double mutants between ICL2 and ICL3 suggest interaction between these loops in the vicinity of F525 and T607 during TSHR activation. Moreover, our data exclude an interaction of M527 with ICL3, supporting the idea of direct selective interaction of M527 with the Gq-subunit, which has already been suggested for other 7TMRs (17, 18).
FIG. 1. TSHR mutations within the second intracellular loop. A putative arrangement of the TSHR is shown, highlighting ICL2 between aa residues 522 and 534. Residue numbers are determined by counting from the methionine start site of the TSHR. Deleted aa are indicated by dashes. The aa substitutions are highlighted in bold letters. Residues in ICL3, which were involved in double mutants between ICL2 and ICL3, are also highlighted.
Materials and Methods
Site-directed mutagenesis
The TSHR mutations were created by standard PCR mutagenesis techniques (19) using the human TSHR plasmid, TSHR-pSVL (20), as a template. PCR products with the corresponding mutants were incompletely digested with ScaI (there is an additional ScaI site within pSVL) and subsequently with Eco91I (MBI Fermentas, Vilnius, Lithuania). The mutated TSHR constructs were generated by replacing the ScaI/Eco91I fragment in the wild-type (wt) TSHR cloned in pSVL with the corresponding mutated fragment. For the construction of all double mutants, the single mutants were also created by standard PCR mutagenesis technique and subsequently subcloned by replacement of an Eco81I/Eco91I fragment. Sequences of mutated TSHR were confirmed by dideoxy sequencing with Big Dye Terminator Cycle Sequencing chemistry (Applied Biosystems, Foster City, CA). Sequencing reactions were analyzed on a Genetic Analyzer ABI 310 (Applied Biosystems).
Cell culture and transfection
COS-7 cells were grown in DMEM supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen Life Technologies, Inc., Paisley, UK) at 37 C in a humidified 7% CO2 incubator. Cells were transiently transfected in 12-well plates (1 x 105 cells/well) with 1 μg DNA/well using the FuGene 6 reagent (Roche, Basel, Switzerland).
Radioligand binding assay
Competitive binding studies were performed as previously described (10). Data were analyzed assuming a one-site binding model using the fitting module of SigmaPlot 2.0 for Windows (21).
FACS analysis
Forty-eight hours after transfection, nonpermeabilized cells were detached from the dishes using 1 mM EDTA and 1 mM EGTA in PBS and transferred in Falcon 2052 tubes. Before incubation with the primary antibody, cells were washed once with PBS containing 0.1% BSA and 0.1% NaN3. Afterward, cells were incubated with a mouse antihuman TSHR antibody (2C11, Serotec, 10 μg/ml) in the same buffer. Tubes were washed and incubated for 1 h on ice in the dark with fluorescein-conjugated F(ab')2 rabbit antimouse IgG (Serotec, Oxford, UK; dilution, 1:1000). Before FACS analysis (FACScan, BD Biosciences), cells were washed twice and fixed with 1% paraformaldehyde. Receptor expression was determined by the fluorescence intensity, whereas the percentage of signal-positive cells corresponds to the transfection efficiency.
cAMP accumulation assay
Measurement of cAMP accumulation was performed 48 h after transfection, as previously described (10).
Stimulation of inositol phosphate (IP) formation
Transfected COS-7 cells were incubated with 2 μCi/ml [myo-3H]inositol (18.6 Ci/mmol; Amersham Biosciences, Braunschweig, Germany) for 8 h. Thereafter, cells were preincubated with serum-free DMEM, without antibiotics, containing 10 mM LiCl for 30 min. Stimulation with TSH was performed with the same medium supplemented with 100 mU/ml TSH for 1 h. Basal and TSH-induced increases in intracellular IP levels were determined by anion exchange chromatography as previously described (22). IP values are expressed as the percentage of radioactivity incorporated from [3H]IP (IP1–3) over the sum of radioactivity incorporated in IPs and phosphatidylinositols.
Molecular modeling
The packing of the seven-helix backbone of the rhodopsin structure (23) was used as a template for the TM regions of the TSHR. The TSHR structure model was computed with special emphasis on the TM and intracellular portions, without the large amino-terminal ectodomain, but including the extra- and intracellular loops. The starting conformation of ICL1 and ICL2, the first portion of the C-terminal tail formed as helix 8, and the overall fold of the second extracellular loop were also adopted from the rhodopsin structure. For the remaining parts of the ICLs as well as for the extracellular loops, fragments of four to seven residues were selected and tested against the three-dimensional protein database. Only fragments occurring more than once with a similar backbone conformation in the database were used for assembling the loops. All model components were assembled with the biopolymer module of the SYBYL program package (TRIPOS, Inc., St Louis, MO). Simulated annealing runs were performed by heating up to 1,000 K 100 times, equilibrating for 2,000 psec, and cooling to 0 K during 10,000 psec. The C- atoms of the TM helices were restrained in place by a punctual harmonic potential. For ICL2 and ICL3, two main conformational clusters were sampled. For one of the two clusters, molecular dynamic simulations were performed at 300 K for 1 nsec, where only the helix stability was maintained by restraints for hydrogen bonds of the TM backbones. For all energy and dynamics calculations, the AMBER 5.0 force field (24) was used. The geometrical quality of the two resulting models, A and B, was controlled using the program PROCHECK (25).
Results
Deletion mutants
The effects of all mutations within ICL2 on TSH binding, cell surface expression, basal and TSH-stimulated cAMP, and IP production are summarized in Table 1.
TABLE 1. Functional characterization of the TSHR mutations in the second ICL
The close proximity of ICL2 and ICL3 and the highly conserved DRY motif at the junction of TM3 and ICL2 suggest that ICL2 plays a fundamental role in G protein coupling, which is particularly underlined by studies of the muscarinic receptor (6). In an effort to investigate the specific role of ICL2 in the process of TSHR activation and G protein coupling, deletion mutants within ICL2 were designed. Three deletions were introduced into ICL2: 1 from aa positions 522–525, 2 from aa positions 526–530, and 3 from aa positions 531–534 (Fig. 1).
Although characterized by a slightly increased cell surface expression compared with the wt TSHR, all three deletion mutants were inactive for TSH-stimulated IP production and strongly impaired for cAMP production (Table 1). All three deletion constructs had decreased basal cAMP activity compared with wt and TSH-stimulated activity that was substantially lower than wt. These findings suggest that G protein coupling was impaired in each deletion construct.
Multiple alanine substitutions (ASU)
To clarify whether the impaired G protein coupling is caused by deletion-related conformational changes or by specific aa residues within ICL2, ASU were introduced in place of the deletions. The ASUs span the following regions: ASU1 from aa 522–525, ASU2 from aa 526–530, and ASU3 from aa 531–534, corresponding to 1–3 (Fig. 1). The ASUs showed gradually decreased impairment of TSH-stimulated receptor activity for the cAMP pathway as they progressed toward the region of ICL2 nearest TM4 (Fig. 2 and Table 1). In contrast, the Gq-mediated IP pathway was not activated by TSH in any of the ASUs (Fig. 2 and Table 1). Because the ASU3 mutant showed slightly decreased cell surface expression compared with the wt TSHR, the TSH-stimulated cAMP accumulation for this mutant may be underestimated. These findings indicate that the N-terminal and middle portions of ICL2 seem to be more important for Gs activation than the region nearest TM4 and, moreover, reveal a critical importance of ICL2 for Gq activation.
FIG. 2. Basal and TSH-induced cAMP and IP accumulation in ASU mutants within ICL2. A, cAMP accumulation assays were performed with transiently transfected COS-7 cells. Forty-eight hours after transfection, COS-7 cells were incubated in the absence () or presence () of 100 mU/ml bovine TSH. cAMP levels were determined as described in Materials and Methods. Data are presented as the mean ± SEM of three independent experiments. B, Forty-eight hours after transfection, cells were labeled with 2 μCi/ml [myo-3H]inositol for 8 h and subsequently incubated in the absence () or presence () of 100 mU/ml bovine TSH. IP accumulation was determined as described in Materials and Methods. Data are expressed as the percent radioactivity incorporated in IPs over the sum of radioactivity in IPs and phosphoinositide (PI) and are presented as the mean ± SEM of three independent experiments.
Single ASU
Based on these findings, we hypothesized that specific aa residues might be involved in either intramolecular interactions between the ICLs or intermolecular interactions between TSHR and G proteins. To investigate this hypothesis, we substituted all single residues within the aa region from 523–534 with alanine (Fig. 1 and Table 1). These substitutions within the region from position R531 to R534 revealed no influence on Gs protein coupling (Table 1), confirming the finding of increased cAMP accumulation for ASU3 compared with ASU1 and ASU2. A lower IP accumulation for R531A after stimulation with TSH and a 2.5-fold higher EC50 for both R531A and K532A (Table 1) may partially account for the inactivation of IP production in the ASU3 mutant.
Single ASUs within the region from I523 to D530 revealed stronger differences between Gs- and Gq-dependent signaling, which could be pinpointed to specific aa. TSHR mutant I523A exhibited a weak TSH-induced cAMP accumulation, a 5.6-fold higher EC50 than wt TSHR, and no IP stimulation (Table 1 and Fig. 3, A and B). COS-7 cells transfected with I523A showed cell surface expression of 70% relative to wt TSHR, which does not explain the strongly decreased cAMP and inactive IP production of this mutant. R534A, for example, showed a comparable cell surface expression without significant impairment of the cAMP and IP signals (Table 1). ASUs within the region of 1 (aa 522–525), in which only one wt aa residue was preserved (AIAA, AATA, and AAAF; Fig. 1) confirm the importance of I523, because the AIAA substitution in which I523 was intact showed the best Gs coupling of these three mutants (Table 1).
FIG. 3. Effects of key residues in ICL2 of TSHR on Gs- and Gq-dependent signaling. TSH-induced cAMP formation (A and C) or IP formation (B and D) was determined in single ASU mutants within the region of ASU1 (A and B) or ASU2 (C and D). For cAMP measurements, COS-7 cells were incubated 48 h after transfection with various concentrations of TSH. Data represent one of three experiments, each performed in duplicate. For IP measurements, COS-7 cells were labeled 48 h after transfection with 2 μCi/ml [myo-3H]inositol for 8 h and subsequently incubated with various concentrations of TSH. Data are expressed as the percent radioactivity incorporated in IPs over the sum of radioactivity in IPs and phosphoinositide (PI) and represent one of three experiments, each performed in duplicate.
One of the more interesting single ASUs was M527A. This mutant showed a decreased basal cAMP activity compared with the basal wt TSHR activity despite a slightly increased cell surface expression (Table 1 and Fig. 3C). After maximal stimulation with TSH, this mutant showed a cAMP accumulation 45% that of wt TSHR, with an EC50 that was 2.5-fold higher than the wt value. However, no TSH-stimulated IP production was measured (Table 1 and Fig. 3, C and D). Thus, M525A is inactive for the Gq-coupled pathway and loses the dual G protein signaling characteristic of the TSHR.
A selective impairment of TSH-induced phosphoinositide hydrolysis was also observed for the mutants F525A, R528A, L529A, and D530A. cAMP production was near wt TSHR levels in these mutants, with only slightly lower maximum TSH-induced cAMP production and modest increases in EC50 values (Table 1 and Fig. 3C). In contrast, IP production of these mutants was between 19–42% of the wt TSHR, and the EC50 values increased 2.4- to 3.3-fold (Fig. 3D). Mutant T524A, which does not show reduced maximum cAMP or IP production, did exhibit a 2.7-fold increased EC50 for Gq activation (Table 1 and Fig. 3, A and B). Because cell surface expression for T524A was only 75% compared with that of wt TSHR, the observed reduced cell surface expression of R528A and D530A (Table 1) most likely does not account for the reduced TSH-stimulated IP production of those mutants. Taken together, these data demonstrate that the most important residues for Gs/Gq coupling and, in particular, for selective loss or impairment of Gq coupling are localized in the N-terminal and middle portions of ICL2 at position I523 and within the region from F525 to D530.
Molecular modeling
To study the ICLs, TSHR lacking the large N-terminal ectodomain was modeled based on the rhodopsin structure template. However, the ICL2 conformation of rhodopsin could only be used as a starting point, because the ICL2 of the TSHR is one residue longer and lacks sequence similarity in its hydrophobic residues compared with rhodopsin. Concerning ICL2 and ICL3, simulated annealing runs provided two different clusters of conformations. The conformational stability of each was confirmed using molecular dynamic simulations of one representative. The main difference in the resulting two models is the orientations of F525 and M527 (Fig. 5). In model A, F525 and M527 are constituents of a large hydrophobic core stabilizing the fold of the TM3/ICL2 junction, whereas in model B, F525 is interacting with the ICL3 in the region of T607. Residue M527 points to the cytosol toward ICL3. In our simulated annealing studies the M527 side chain was observed to be rather flexible, moving between the hydrophobic ICL2 core, an intracellular orientation, and an orientation toward ICL3 (Y613). Common to both models is the conformation of the TM3/ICL2 junction, which is dominated by a hydrophobic core formed by W520 and Y521, which are still members of TM3, and the residues I529 and I533, which only escort this cluster. The residue I523 is the last hydrophobic residue of TM3 pointing to a hydrophobic patch between TM5 and TM6. The C-terminal portion of ICL2, comprised of residues 528–534, forms an extended conformation toward TM4.
FIG. 5. Two molecular models of the TSHR. A (upper panel) and B (lower panel) resulted from simulated annealing runs. For a better overview, only ICL2 and ICL3, their attached TM portions, and discussed side-chains are visualized. Fragments of deletion and multiple ASU within ICL2 are highlighted by colors: blue: 1, ASU1 (522–525); red: 2, ASU2 (526–530); and yellow: 3, ASU3 (531–534). Fragments 1 and 2 mainly cooperate with hydrophobic residues of TM3 (Y521 and W520), a hydrophobic cluster of the TM3/ICL2 junction fold. This cluster is flanked by residues L529 and I533. The other hydrophobic residues (green) of ICL2 are involved in different interactions. Residue I523, the last residue of the extended TM3 helix, is oriented toward hydrophobic residues (not shown) of TM5 and TM6, stabilizing the helical fold among the three helices. The main differences in the two models are the orientations of F525 and M527. In model A, F525 and M527 participate at the large hydrophobic core stabilizing the fold of TM3/ICL2 junction, whereas in model B, F525 interacts with ICL3 in the region of T607; M527 is pointing to the cytosol. Thereby other aa will be localized in the right position for Gq-protein coupling, such as R528 and D530, which are orientated toward the intracellular region. The double mutant approach revealed model B as the more likely tertiary fold for ICL2 and ICL3.
Testing of the ICL2-ICL3 interaction
To distinguish between models A and B we approached the question of whether ICL2 stabilization or the ICL2-ICL3 interaction accounts for a more reliable tertiary structure. To investigate whether ICL2 interacts with ICL3, a lysine at positions F525 and M527 in ICL2 and a glutamate with a strong complementary property were introduced at positions T607 and Y613 in ICL3. The single mutant M527K showed a much stronger inactivating phenotype than M527A. In addition to the disrupted Gq protein coupling previously observed with M527A, M527K strongly abolished cAMP production despite a normal cell surface expression compared with the wt TSHR (Table 2). This inactive phenotype of M527K for both signaling cascades was maintained in the double mutants with T607E and Y613E (Table 2), suggesting that M527 (ICL2) does not interact with these residues in ICL3. In contrast to position M527, the phenotypes of F525A and F525K were comparable for Gs and Gq signaling (Tables 1 and 2). The double mutants, F525K/Y613E and F525K/T607E, showed functional properties similar to those of the wt TSHR for cAMP production (Table 2 and Fig. 4, A and B). However, strong differences between both double mutants were observed regarding Gq-mediated signaling. The double mutant F525K/T607E showed a rescue of the Gq-coupled pathway by a 24% higher maximal IP accumulation compared with F525K alone (Table 2 and Fig. 4C). The EC50 shift for this double mutant was comparable to that with F525K alone. In contrast, F525K/Y613E exhibited a 14% lower IP production and a 2.2-fold higher EC50 compared with F525K alone (Fig. 4D). These data suggest that ICL2 interacts with ICL3 in the vicinity of F527 and T607 and provides support for the tertiary fold proposed in model B above (Fig. 5B).
TABLE 2. Functional characterization of TSHR double mutants between second and third ICL
FIG. 4. TSH-induced cAMP or IP formation in double mutants between position 525 in ICL2 and position 607 or 613 in ICL3. For cAMP measurements (A and B), COS-7 cells were incubated 48 h after transfection with various concentrations of TSH. For IP measurements (C and D), COS-7 cells were labeled 48 h after transfection with 2 μCi/ml [myo-3H]inositol for 8 h and subsequently incubated with various concentrations of TSH. Data represent one of three experiments, each performed in duplicate.
Discussion
In this study we provide evidence that the N-terminal and central portions of ICL2 from I523 to D530 participate in dual Gs/Gq protein activation by the TSHR. Moreover, we show that mutant M527A loses Gq-dependent phospholipase C activation, but maintains Gs-mediated cAMP formation, suggesting a functional role for this methionine in G protein activation selectivity. This finding is supported by alanine mutants of residues F525 and R528 to D530 in the vicinity of M527, which also show selective inhibition of Gq activation. The partial restoration of the Gq-mediated response of F525K in a double mutant with T607E in ICL3 supports an intramolecular interaction between ICL2 and ICL3 in the close vicinity of F525 and T607 and thus a conformational cooperation between ICL2 and ICL3.
Identification of key residues in ICL2 for the dual G protein activation by TSHR
All three deletions within ICL2 are expressed comparable to the wt TSHR, but Gs/Gq protein coupling was abolished or strongly diminished. It is conceivable that these deletions result in TSHR conformations that allow the receptors to reach the cell surface, but do not allow productive contact with G proteins. To investigate the cause of the disturbance in intermolecular signal transduction in more detail and to gain information at the level of single aa, we introduced alanine residues within the regions of the deletion mutants. The stepwise increase in reconstitution of cAMP activity from ASU1 to ASU3 toward TM4 was a decisive indication that specific residues in the N-terminal half of ICL2 were probably involved in Gs coupling. Moreover, the complete inhibition of Gq-mediated IP production by ASU1–3 indicated a functional impact of ICL2 on this pathway. The substitution of single aa with alanine confirmed these observations, showing that the region of F525 to D530 has a slight effect on Gs signaling, but plays a more important role in Gq signaling.
The strongest change regarding selective loss of G protein activation was observed with M527A. Indeed, the efficiency of Gs coupling was affected only slightly, whereas this mutant lost the ability to stimulate Gq-dependent phospholipase C activation despite a normal cell surface expression. Interestingly, Y601H in TM5 of TSHR near the border with ICL3 is also a determinant for Gq coupling, because this mutant does not signal via Gq, but cAMP formation remains unaltered (26). Similarly, Y601N (27) confirms the functional role of Y601 in the activity and dual signaling capability of the TSHR. Taken together, these and our data are consistent with the observation that residues involved in G protein coupling and selectivity are primarily localized at the TM/cytoplasmic borders between TM3/ICL2, TM5/ICL3, and ICL3/TM6 (1).
The extension of this selective effect on Gq activation up- and downstream of M527 (F525A and R528A to D530A) fits well with the idea that receptor domains, rather than single aa, participate in forming a G protein binding pocket. Our findings are in part consistent with the data reported by Kosugi et al. (14) showing that the most important determinant for agonist-increased cAMP production is the region around residues 525–527 of ICL2 of the TSHR. However, they observed a total loss of TSH-induced cAMP response and a lesser loss in phosphoinositide signaling for a mutant in which residues 525–527 of the TSHR were substituted with a comparable ?2-adrenergic receptor sequence. Furthermore, single ASUs at positions 528–534 do not or only slightly affect Gs activation, which seems contradictory to the findings by Chazenbalk et al. (13). One explanation for these differences might be that simultaneous substitution of three or more residues has a stronger effect than single substitutions, which we also observed with our deletions and ASU mutants.
Different tasks of hydrophobic residues F525 and M527
F525: experimental indication for an interaction between ICL2 with ICL3.
The main differences in the proposed models are the orientations of F525 and M527. In the ICL2 stabilization model (A), F525 participates in a large hydrophobic core, stabilizing the fold of the TM3/ICL2 junction, whereas in the ICL2-ICL3 interaction model (B), F525 in ICL2 points toward ICL3 to residue T607 and in the vicinity of Y613. Drastic alteration of hydrophobic properties at position F525 by alanine and lysine has only a weak impact on TSH-induced cAMP formation, but, in contrast, the stimulated IP accumulation is clearly decreased. Interestingly, only the double mutant F525K/T607E achieved a partial functional rescue of stimulated IP production. In contrast, with the double mutant F525K/Y613E, TSH-induced IP formation was further decreased. Despite increased cell surface expression, Gs- and Gq-dependent signaling was strongly inhibited by M527K. Double mutants between M527 at ICL2 and Y613 and T607 at ICL3 did not lead to a functional rescue for cAMP or IP production. Taken together, a cooperative effect between ICL2 and ICL3 in the vicinity of F525 and T607 is indicated, supporting the tertiary fold proposed in model B.
M527 and its vicinity from F525 to D530 play pivotal roles in Gq-dependent signaling
Our findings suggest that the M527 side-chain is essential for Gq-mediated signaling and confirm previous findings with other 7TMRs that a methionine or another hydrophobic residue at the position of M527 could play a key role in selective signaling toward Gq (26, 28). M145 in the V2 vasopressin receptor, which is located at a position identical to M527 in the TSHR, was identified as a residue with pronounced effects on receptor/G protein coupling selectivity (18). The presence of a relatively large hydrophobic aa side-chain such as leucine or tryptophan at position 145 favored Gq signaling and suggests a recognition of the C terminus of Gq by this residue (18). This fits with our observation that alterations of the hydrophobic properties at M527 in the TSHR have the strongest selective effect on Gq protein activation. The corresponding position to M527 in the GnRH receptor also has a hydrophobic residue (L147), and its mutation led to impaired Gq signaling (17).
We propose that M527 stabilizes the hydrophobic core of ICL2, thus positioning other aa in suitable positions for Gq-protein coupling. Support for this possibility can be derived from the inhibition of Gq activation by F525A, R528A, L529A, and D530A in the vicinity of M527. Furthermore, its selectivity for Gq, its failure to interact experimentally with ICL3, and its likely cytosolic orientation in modeling provide additional support for a direct interaction of the hydrophobic residue M527 with Gq upon activation. Subsequent studies will provide examinations of these hypotheses by homology modeling of TSHR/G? complexes and related experiments. This will allow additional understanding of the dual G protein coupling of TSHR and particularly of the selectivity of activation.
Acknowledgments
We thank Mrs. Eileen B?senberg for her excellent technical assistance, BRAHMS Diagnostica (Berlin, Germany) for providing [125I]bovine TSH, Dr. Gilbert Vassart for supplying the plasmid TSHR-pSVL, and Dr. Bruce Raaka for critical reading of the manuscript.
References
Wess J 1998 Molecular basis of receptor/G-protein-coupling selectivity. Pharmacol Ther 80:231–264
Gether U 2000 Uncovering molecular mechanisms involved in activation of G protein-coupled receptors. Endocr Rev 21:90–113
Wess J 1997 G-protein-coupled receptors: molecular mechanisms involved in receptor activation and selectivity of G-protein recognition. FASEB J 11:346–354
Farrens DL, Altenbach C, Yang K, Hubbell WL, Khorana HG 1996 Requirement of rigid-body motion of transmembrane helices for light activation of rhodopsin. Science 274:768–770
Gether U, Lin S, Ghanouni P, Ballesteros JA, Weinstein H, Kobilka BK 1997 Agonists induce conformational changes in transmembrane domains III and VI of the ?2 adrenoreceptor. EMBO J 16:6737–6747
Burstein ES, Spalding TA, Brann MR 1998 The second intracellular loop of the m5 muscarinic receptor is the switch which enables G-protein coupling. J Biol Chem 273:24322–24327
Blin N, Yun J, Wess J 1995 Mapping of single amino acid residues required for selective activation of Gq/11 by the m3 muscarinic acetylcholine receptor. J Biol Chem 270:17741–17748
Havlickova M, Blahos J, Brabet I, Liu J, Hruskova B, Prezeau L, Pin J-P 2003 The second intracellular loop of metabotropic glutamate receptors recognizes C termini of G-protein -subunits. J Biol Chem 278:35063–35070
Paschke R, Ludgate M 1997 The thyrotropin receptor in thyroid diseases. N Engl J Med 337:1675–1681
Wonerow P, Schoneberg T, Schultz G, Gudermann T, Paschke R 1998 Deletions in the third intracellular loop of the thyrotropin receptor. A new mechanism for constitutive activation. J Biol Chem 273:7900–7905
Fuhrer D, Lachmund P, Nebel IT, Paschke R 2003 The thyrotropin receptor mutation database: update 2003. Thyroid 13:1123–1126
Wonerow P, Neumann S, Gudermann T, Paschke R 2001 Thyrotropin receptor mutations as a tool to understand thyrotropin receptor action. J Mol Med 79:707–721
Chazenbalk GD, Nagayama Y, Russo D, Wadsworth HL, Rapoport B 1990 Functional analysis of the cytoplasmic domains of the human thyrotropin receptor by site-directed mutagenesis. J Biol Chem 265:20970–20975
Kosugi S, Kohn LD, Akamizu T, Mori T 1994 The middle portion in the second cytoplasmic loop of the thyrotropin receptor plays a crucial role in adenylate cyclase activation. Mol Endocrinol 8:498–509
Fernandez LM, Puett D 1997 Evidence for an important functional role of intracellular loop II of the lutropin receptor. Mol Cell Endocrinol 128:161–169
Timossi C, Maldonado D, Vizcaino A, Lindau-Shepard B, Conn P, Ulloa-Aguirre A 2002 Structural determinants in the second intracellular loop of the human follicle-stimulating hormone receptor are involved in Gs protein activation. Mol Cell Endocrinol 189:157–168
Arora KK, Sakai A, Catt KJ 1995 Effects of second intracellular loop mutations on signal transduction and internalization of the gonadotropin-releasing hormone receptor. J Biol Chem 270:22820–22826
Erlenbach I, Kostenis E, Schmidt C, Serradeil-Le Gal C, Raufaste D, Dumont ME, Pausch MH, Wess J 2001 Single amino acid substitutions and deletions that alter the G protein coupling properties of the V2 vasopressin receptor identified in yeast by receptor random mutagenesis. J Biol Chem 276:29382–29392
Higuchi R 1989 Using PCR to engineer DNA. In: Ehrlich HA, ed. PCR technology. New York: Stockton Press; 61–70
Libert F, Lefort A, Gérard C, Parmentier M, Perret J, Ludgate M, Dumont JE, Vassart G 1989 Cloning, sequencing and expression of the human thyrotropin (TSH) receptor: evidence for binding of autoantibodies. Biochem Biophys Res Commun 165:1250–1255
Swillens S 1995 Interpretation of binding curves obtained with high receptor concentrations: practical aid for computer analysis. Mol Pharmacol 47:1197–1203
Berridge MJ 1983 Rapid accumulation of inositol trisphosphate reveals that agonists hydrolyse polyphosphoinositides instead of phosphatidylinositol. Biochem J 212:849–858
Palczewski K, Kumasaka T, Hori T, Behnke CA, Motoshima H, Fox BA, Le Trong I, Teller DC, Okada T, Stenkamp RE, Yamamoto M, Miyano M 2000 Crystal structure of rhodopsin: a G protein-coupled receptor. Science 289:739–745
Weiner S, Kollmann P, Nguyen D, Case T 1986 An all atom forcefield for simulation of proteins und nucleic acids. J Comp Chem 7:230–252
Laskowski RA, MacArthur MW, Moss DS, Thornton JM 1993 PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Cryst 26:283–291
Biebermann H, Schoneberg T, Schulz A, Krause G, Gruters A, Schultz G, Gudermann T 1998 A conserved tyrosine residue (Y601) in transmembrane domain 5 of the human thyrotropin receptor serves as a molecular switch to determine G-protein coupling. FASEB J 12:1461–1471
Arseven OK, Wilkes WP, Jameson JL, Kopp P 2000 Substitutions of tyrosine 601 in the human thyrotropin receptor result in increase or loss of basal activation of the cyclic adenosine monophosphate pathway and disrupt coupling to Gq/11. Thyroid 10:3–10
Yang M, Wang W, Zhong M, Philippi A, Lichtarge O, Sanborn BM 2002 Lysine 270 in the third intracellular domain of the oxytocin receptor is an important determinant for Gq coupling specificity. Mol Endocrinol 16:814–823
Cetani F, Tonacchera M, Vassart G 1996 Differential effects of NaCl concentration on the constitutive activity of the thyrotropin and the luteinizing hormone/chorionic gonadotropin receptors. FEBS Lett 378:27–31(Susanne Neumann, Gerd Kra)
Address all correspondence and requests for reprints to: Dr. Ralf Paschke, Third Medical Department, University of Leipzig, Ph. Rosenthal Strasse 27, 04103 Leipzig, Germany. E-mail: ralf.paschke@medizin.uni-leipzig.
Abstract
The TSH receptor (TSHR) activates mainly two signal transduction pathways, cAMP production and phosphoinositide turnover, mediated by Gs and Gq coupling, respectively. Several activating deletion and point mutations within intracellular loop 3 (ICL3) and the adjacent portion of transmembrane domain 6 (TM6) support a direct G protein activation by this receptor domain. The ICL3, however, is predicted by modeling to interact with other receptor domains, primarily ICL2, to form a pocket for G protein binding and to allow optimum interaction. Systematic mutagenesis was used to identify important sites within ICL2 and potential interactions between ICL2 and ICL3 of the TSHR required for G protein coupling. Deletions of four or five residues and their corresponding multiple alanine substitutions were introduced into ICL2. Residues I523-D530, comprising mainly the N-terminal half of ICL2, appeared to be critical for Gs- and Gq-mediated signaling. A single alanine substitution screening within ICL2 revealed hydrophobic residue M527 in particular and, to lesser extents, F525, R528, L529, and D530 as residues that selectively abolished or strongly impaired Gq activation. Molecular modeling suggests that F525 interacts with ICL3. To test this hypothesis, ICL2/ICL3 double mutants introducing strong complementary properties were constructed and tested for functional rescue of Gq-mediated signaling. Our results indicate that ICL2 interacts with ICL3 in close vicinity to F525 and T607, suggesting a conformational cooperation between ICL2 and ICL3 during Gq activation by TSHR.
Introduction
STIMULATION OF HETEROTRIMERIC G proteins by cell surface receptors activates signaling pathways that mediate specific responses to hormones and neurotransmitters. However, it is unknown how seven transmembrane (TM)-spanning receptors (7TMRs) activate G proteins and lead to this intricate and finely tuned process of downstream signaling. Selectivity of G protein recognition is determined by intracellular domains of 7TMRs (1). Numerous studies have established the pivotal role of intracellular loop 2 (ICL2) and ICL3 in G protein coupling (2). Most findings in the field of receptor/G protein coupling indicate that ICL3, particularly its N- and C-terminal regions, is important for G protein recognition (3). However, ICL3 is predicted to interact with other receptor domains, primarily with ICL2, to allow for optimum coupling efficiency and selectivity (1). The proposed movement of TM6 relative to TM3 indicated by spectroscopic analysis (4, 5) is consistent with this suggested scenario.
ICL2 seems to be important for the selectivity of receptor/G protein interactions and the efficiency of G protein activation (3). Investigation of ICL2 in the m5 muscarinic receptor by a random mutagenesis approach has shown that ICL2 contains a group of residues necessary for holding the receptor in an inactive conformation and another group necessary for G protein coupling. Therefore, it was hypothesized that ICL2 could act as a switch that enables G protein coupling (6). Moreover, in ICL2 of the m3 muscarinic receptor, four specific amino acids (aa) were identified as important for enabling the receptor to couple selectively to the Gq/11 protein (7). Recent findings for metabotropic receptors, which belong to class 3 of 7TMRs, underline the importance of ICL2 by demonstrating that the central portion of ICL2 is responsible for the selective recognition of the C-terminal end of the G-subunit (8).
The TSH receptor (TSHR) together with the LH and FSH receptors belong to the glycoprotein hormone receptor family. TSHR activation by several activating point and deletion mutants within ICL3 and the lower portion of TM6 supports a direct G protein activation by this region (9, 10). Interestingly, naturally occurring, activating TSHR mutations, which are one of the molecular causes of autosomal dominant, nonautoimmmune hyperthyroidism and toxic thyroid nodules (9, 11) (OMIM 603372), were predominantly localized to TM domains, extracellular loops, or ICL3 and rarely in the first or second ICL (12). For this reason many studies have focused on residues in ICL3, but only two studies have focused on ICL2 of the TSHR (13, 14). The first study concluded that the carboxyl-terminal domain of ICL2 (aa 528–537) is important for activation of Gs-mediated cAMP production (13). TSHR mutants created by substituting sequences from the 1- or ?2-adrenergic receptor demonstrated that the middle portion of ICL2 (residues 525–527) of the TSHR seems to be important for agonist-induced Gs interaction, whereas residues 528–532 were determined to be more critical for agonist-induced Gq activation (14). Mutagenesis studies of LH and FSH receptors have also provided evidence for an important functional role of several aa residues within ICL2 (15, 16). However, only single selected residues in ICL2 were substituted in these studies.
Less is known about the specific function of single residues of ICL2 and potential interaction with ICL3 in dual G protein activation by glycoprotein hormone receptors. Therefore, the first aim of our study was to identify and characterize specific intracellular aa of the TSHR involved in G protein activation and selectivity. In contrast to both former studies (13, 14), in which several adjacent aa within ICL2 of the TSHR were replaced simultaneously, systematic single site mutagenesis of the entire ICL2 was carried out. The second aim was to identify potential intramolecular interactions between ICL2 and ICL3 that might facilitate G protein activation. Deletions of four or five consecutive residues and multiple alanine mutations of the same residues (Fig. 1) localized the ICL2 region important for TSHR activation to residues 523–530. Our data reveal that M527 and, to a lesser extent, F525, R528, L529, and D530 are residues with selective influence on Gq activation. Double mutants between ICL2 and ICL3 suggest interaction between these loops in the vicinity of F525 and T607 during TSHR activation. Moreover, our data exclude an interaction of M527 with ICL3, supporting the idea of direct selective interaction of M527 with the Gq-subunit, which has already been suggested for other 7TMRs (17, 18).
FIG. 1. TSHR mutations within the second intracellular loop. A putative arrangement of the TSHR is shown, highlighting ICL2 between aa residues 522 and 534. Residue numbers are determined by counting from the methionine start site of the TSHR. Deleted aa are indicated by dashes. The aa substitutions are highlighted in bold letters. Residues in ICL3, which were involved in double mutants between ICL2 and ICL3, are also highlighted.
Materials and Methods
Site-directed mutagenesis
The TSHR mutations were created by standard PCR mutagenesis techniques (19) using the human TSHR plasmid, TSHR-pSVL (20), as a template. PCR products with the corresponding mutants were incompletely digested with ScaI (there is an additional ScaI site within pSVL) and subsequently with Eco91I (MBI Fermentas, Vilnius, Lithuania). The mutated TSHR constructs were generated by replacing the ScaI/Eco91I fragment in the wild-type (wt) TSHR cloned in pSVL with the corresponding mutated fragment. For the construction of all double mutants, the single mutants were also created by standard PCR mutagenesis technique and subsequently subcloned by replacement of an Eco81I/Eco91I fragment. Sequences of mutated TSHR were confirmed by dideoxy sequencing with Big Dye Terminator Cycle Sequencing chemistry (Applied Biosystems, Foster City, CA). Sequencing reactions were analyzed on a Genetic Analyzer ABI 310 (Applied Biosystems).
Cell culture and transfection
COS-7 cells were grown in DMEM supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen Life Technologies, Inc., Paisley, UK) at 37 C in a humidified 7% CO2 incubator. Cells were transiently transfected in 12-well plates (1 x 105 cells/well) with 1 μg DNA/well using the FuGene 6 reagent (Roche, Basel, Switzerland).
Radioligand binding assay
Competitive binding studies were performed as previously described (10). Data were analyzed assuming a one-site binding model using the fitting module of SigmaPlot 2.0 for Windows (21).
FACS analysis
Forty-eight hours after transfection, nonpermeabilized cells were detached from the dishes using 1 mM EDTA and 1 mM EGTA in PBS and transferred in Falcon 2052 tubes. Before incubation with the primary antibody, cells were washed once with PBS containing 0.1% BSA and 0.1% NaN3. Afterward, cells were incubated with a mouse antihuman TSHR antibody (2C11, Serotec, 10 μg/ml) in the same buffer. Tubes were washed and incubated for 1 h on ice in the dark with fluorescein-conjugated F(ab')2 rabbit antimouse IgG (Serotec, Oxford, UK; dilution, 1:1000). Before FACS analysis (FACScan, BD Biosciences), cells were washed twice and fixed with 1% paraformaldehyde. Receptor expression was determined by the fluorescence intensity, whereas the percentage of signal-positive cells corresponds to the transfection efficiency.
cAMP accumulation assay
Measurement of cAMP accumulation was performed 48 h after transfection, as previously described (10).
Stimulation of inositol phosphate (IP) formation
Transfected COS-7 cells were incubated with 2 μCi/ml [myo-3H]inositol (18.6 Ci/mmol; Amersham Biosciences, Braunschweig, Germany) for 8 h. Thereafter, cells were preincubated with serum-free DMEM, without antibiotics, containing 10 mM LiCl for 30 min. Stimulation with TSH was performed with the same medium supplemented with 100 mU/ml TSH for 1 h. Basal and TSH-induced increases in intracellular IP levels were determined by anion exchange chromatography as previously described (22). IP values are expressed as the percentage of radioactivity incorporated from [3H]IP (IP1–3) over the sum of radioactivity incorporated in IPs and phosphatidylinositols.
Molecular modeling
The packing of the seven-helix backbone of the rhodopsin structure (23) was used as a template for the TM regions of the TSHR. The TSHR structure model was computed with special emphasis on the TM and intracellular portions, without the large amino-terminal ectodomain, but including the extra- and intracellular loops. The starting conformation of ICL1 and ICL2, the first portion of the C-terminal tail formed as helix 8, and the overall fold of the second extracellular loop were also adopted from the rhodopsin structure. For the remaining parts of the ICLs as well as for the extracellular loops, fragments of four to seven residues were selected and tested against the three-dimensional protein database. Only fragments occurring more than once with a similar backbone conformation in the database were used for assembling the loops. All model components were assembled with the biopolymer module of the SYBYL program package (TRIPOS, Inc., St Louis, MO). Simulated annealing runs were performed by heating up to 1,000 K 100 times, equilibrating for 2,000 psec, and cooling to 0 K during 10,000 psec. The C- atoms of the TM helices were restrained in place by a punctual harmonic potential. For ICL2 and ICL3, two main conformational clusters were sampled. For one of the two clusters, molecular dynamic simulations were performed at 300 K for 1 nsec, where only the helix stability was maintained by restraints for hydrogen bonds of the TM backbones. For all energy and dynamics calculations, the AMBER 5.0 force field (24) was used. The geometrical quality of the two resulting models, A and B, was controlled using the program PROCHECK (25).
Results
Deletion mutants
The effects of all mutations within ICL2 on TSH binding, cell surface expression, basal and TSH-stimulated cAMP, and IP production are summarized in Table 1.
TABLE 1. Functional characterization of the TSHR mutations in the second ICL
The close proximity of ICL2 and ICL3 and the highly conserved DRY motif at the junction of TM3 and ICL2 suggest that ICL2 plays a fundamental role in G protein coupling, which is particularly underlined by studies of the muscarinic receptor (6). In an effort to investigate the specific role of ICL2 in the process of TSHR activation and G protein coupling, deletion mutants within ICL2 were designed. Three deletions were introduced into ICL2: 1 from aa positions 522–525, 2 from aa positions 526–530, and 3 from aa positions 531–534 (Fig. 1).
Although characterized by a slightly increased cell surface expression compared with the wt TSHR, all three deletion mutants were inactive for TSH-stimulated IP production and strongly impaired for cAMP production (Table 1). All three deletion constructs had decreased basal cAMP activity compared with wt and TSH-stimulated activity that was substantially lower than wt. These findings suggest that G protein coupling was impaired in each deletion construct.
Multiple alanine substitutions (ASU)
To clarify whether the impaired G protein coupling is caused by deletion-related conformational changes or by specific aa residues within ICL2, ASU were introduced in place of the deletions. The ASUs span the following regions: ASU1 from aa 522–525, ASU2 from aa 526–530, and ASU3 from aa 531–534, corresponding to 1–3 (Fig. 1). The ASUs showed gradually decreased impairment of TSH-stimulated receptor activity for the cAMP pathway as they progressed toward the region of ICL2 nearest TM4 (Fig. 2 and Table 1). In contrast, the Gq-mediated IP pathway was not activated by TSH in any of the ASUs (Fig. 2 and Table 1). Because the ASU3 mutant showed slightly decreased cell surface expression compared with the wt TSHR, the TSH-stimulated cAMP accumulation for this mutant may be underestimated. These findings indicate that the N-terminal and middle portions of ICL2 seem to be more important for Gs activation than the region nearest TM4 and, moreover, reveal a critical importance of ICL2 for Gq activation.
FIG. 2. Basal and TSH-induced cAMP and IP accumulation in ASU mutants within ICL2. A, cAMP accumulation assays were performed with transiently transfected COS-7 cells. Forty-eight hours after transfection, COS-7 cells were incubated in the absence () or presence () of 100 mU/ml bovine TSH. cAMP levels were determined as described in Materials and Methods. Data are presented as the mean ± SEM of three independent experiments. B, Forty-eight hours after transfection, cells were labeled with 2 μCi/ml [myo-3H]inositol for 8 h and subsequently incubated in the absence () or presence () of 100 mU/ml bovine TSH. IP accumulation was determined as described in Materials and Methods. Data are expressed as the percent radioactivity incorporated in IPs over the sum of radioactivity in IPs and phosphoinositide (PI) and are presented as the mean ± SEM of three independent experiments.
Single ASU
Based on these findings, we hypothesized that specific aa residues might be involved in either intramolecular interactions between the ICLs or intermolecular interactions between TSHR and G proteins. To investigate this hypothesis, we substituted all single residues within the aa region from 523–534 with alanine (Fig. 1 and Table 1). These substitutions within the region from position R531 to R534 revealed no influence on Gs protein coupling (Table 1), confirming the finding of increased cAMP accumulation for ASU3 compared with ASU1 and ASU2. A lower IP accumulation for R531A after stimulation with TSH and a 2.5-fold higher EC50 for both R531A and K532A (Table 1) may partially account for the inactivation of IP production in the ASU3 mutant.
Single ASUs within the region from I523 to D530 revealed stronger differences between Gs- and Gq-dependent signaling, which could be pinpointed to specific aa. TSHR mutant I523A exhibited a weak TSH-induced cAMP accumulation, a 5.6-fold higher EC50 than wt TSHR, and no IP stimulation (Table 1 and Fig. 3, A and B). COS-7 cells transfected with I523A showed cell surface expression of 70% relative to wt TSHR, which does not explain the strongly decreased cAMP and inactive IP production of this mutant. R534A, for example, showed a comparable cell surface expression without significant impairment of the cAMP and IP signals (Table 1). ASUs within the region of 1 (aa 522–525), in which only one wt aa residue was preserved (AIAA, AATA, and AAAF; Fig. 1) confirm the importance of I523, because the AIAA substitution in which I523 was intact showed the best Gs coupling of these three mutants (Table 1).
FIG. 3. Effects of key residues in ICL2 of TSHR on Gs- and Gq-dependent signaling. TSH-induced cAMP formation (A and C) or IP formation (B and D) was determined in single ASU mutants within the region of ASU1 (A and B) or ASU2 (C and D). For cAMP measurements, COS-7 cells were incubated 48 h after transfection with various concentrations of TSH. Data represent one of three experiments, each performed in duplicate. For IP measurements, COS-7 cells were labeled 48 h after transfection with 2 μCi/ml [myo-3H]inositol for 8 h and subsequently incubated with various concentrations of TSH. Data are expressed as the percent radioactivity incorporated in IPs over the sum of radioactivity in IPs and phosphoinositide (PI) and represent one of three experiments, each performed in duplicate.
One of the more interesting single ASUs was M527A. This mutant showed a decreased basal cAMP activity compared with the basal wt TSHR activity despite a slightly increased cell surface expression (Table 1 and Fig. 3C). After maximal stimulation with TSH, this mutant showed a cAMP accumulation 45% that of wt TSHR, with an EC50 that was 2.5-fold higher than the wt value. However, no TSH-stimulated IP production was measured (Table 1 and Fig. 3, C and D). Thus, M525A is inactive for the Gq-coupled pathway and loses the dual G protein signaling characteristic of the TSHR.
A selective impairment of TSH-induced phosphoinositide hydrolysis was also observed for the mutants F525A, R528A, L529A, and D530A. cAMP production was near wt TSHR levels in these mutants, with only slightly lower maximum TSH-induced cAMP production and modest increases in EC50 values (Table 1 and Fig. 3C). In contrast, IP production of these mutants was between 19–42% of the wt TSHR, and the EC50 values increased 2.4- to 3.3-fold (Fig. 3D). Mutant T524A, which does not show reduced maximum cAMP or IP production, did exhibit a 2.7-fold increased EC50 for Gq activation (Table 1 and Fig. 3, A and B). Because cell surface expression for T524A was only 75% compared with that of wt TSHR, the observed reduced cell surface expression of R528A and D530A (Table 1) most likely does not account for the reduced TSH-stimulated IP production of those mutants. Taken together, these data demonstrate that the most important residues for Gs/Gq coupling and, in particular, for selective loss or impairment of Gq coupling are localized in the N-terminal and middle portions of ICL2 at position I523 and within the region from F525 to D530.
Molecular modeling
To study the ICLs, TSHR lacking the large N-terminal ectodomain was modeled based on the rhodopsin structure template. However, the ICL2 conformation of rhodopsin could only be used as a starting point, because the ICL2 of the TSHR is one residue longer and lacks sequence similarity in its hydrophobic residues compared with rhodopsin. Concerning ICL2 and ICL3, simulated annealing runs provided two different clusters of conformations. The conformational stability of each was confirmed using molecular dynamic simulations of one representative. The main difference in the resulting two models is the orientations of F525 and M527 (Fig. 5). In model A, F525 and M527 are constituents of a large hydrophobic core stabilizing the fold of the TM3/ICL2 junction, whereas in model B, F525 is interacting with the ICL3 in the region of T607. Residue M527 points to the cytosol toward ICL3. In our simulated annealing studies the M527 side chain was observed to be rather flexible, moving between the hydrophobic ICL2 core, an intracellular orientation, and an orientation toward ICL3 (Y613). Common to both models is the conformation of the TM3/ICL2 junction, which is dominated by a hydrophobic core formed by W520 and Y521, which are still members of TM3, and the residues I529 and I533, which only escort this cluster. The residue I523 is the last hydrophobic residue of TM3 pointing to a hydrophobic patch between TM5 and TM6. The C-terminal portion of ICL2, comprised of residues 528–534, forms an extended conformation toward TM4.
FIG. 5. Two molecular models of the TSHR. A (upper panel) and B (lower panel) resulted from simulated annealing runs. For a better overview, only ICL2 and ICL3, their attached TM portions, and discussed side-chains are visualized. Fragments of deletion and multiple ASU within ICL2 are highlighted by colors: blue: 1, ASU1 (522–525); red: 2, ASU2 (526–530); and yellow: 3, ASU3 (531–534). Fragments 1 and 2 mainly cooperate with hydrophobic residues of TM3 (Y521 and W520), a hydrophobic cluster of the TM3/ICL2 junction fold. This cluster is flanked by residues L529 and I533. The other hydrophobic residues (green) of ICL2 are involved in different interactions. Residue I523, the last residue of the extended TM3 helix, is oriented toward hydrophobic residues (not shown) of TM5 and TM6, stabilizing the helical fold among the three helices. The main differences in the two models are the orientations of F525 and M527. In model A, F525 and M527 participate at the large hydrophobic core stabilizing the fold of TM3/ICL2 junction, whereas in model B, F525 interacts with ICL3 in the region of T607; M527 is pointing to the cytosol. Thereby other aa will be localized in the right position for Gq-protein coupling, such as R528 and D530, which are orientated toward the intracellular region. The double mutant approach revealed model B as the more likely tertiary fold for ICL2 and ICL3.
Testing of the ICL2-ICL3 interaction
To distinguish between models A and B we approached the question of whether ICL2 stabilization or the ICL2-ICL3 interaction accounts for a more reliable tertiary structure. To investigate whether ICL2 interacts with ICL3, a lysine at positions F525 and M527 in ICL2 and a glutamate with a strong complementary property were introduced at positions T607 and Y613 in ICL3. The single mutant M527K showed a much stronger inactivating phenotype than M527A. In addition to the disrupted Gq protein coupling previously observed with M527A, M527K strongly abolished cAMP production despite a normal cell surface expression compared with the wt TSHR (Table 2). This inactive phenotype of M527K for both signaling cascades was maintained in the double mutants with T607E and Y613E (Table 2), suggesting that M527 (ICL2) does not interact with these residues in ICL3. In contrast to position M527, the phenotypes of F525A and F525K were comparable for Gs and Gq signaling (Tables 1 and 2). The double mutants, F525K/Y613E and F525K/T607E, showed functional properties similar to those of the wt TSHR for cAMP production (Table 2 and Fig. 4, A and B). However, strong differences between both double mutants were observed regarding Gq-mediated signaling. The double mutant F525K/T607E showed a rescue of the Gq-coupled pathway by a 24% higher maximal IP accumulation compared with F525K alone (Table 2 and Fig. 4C). The EC50 shift for this double mutant was comparable to that with F525K alone. In contrast, F525K/Y613E exhibited a 14% lower IP production and a 2.2-fold higher EC50 compared with F525K alone (Fig. 4D). These data suggest that ICL2 interacts with ICL3 in the vicinity of F527 and T607 and provides support for the tertiary fold proposed in model B above (Fig. 5B).
TABLE 2. Functional characterization of TSHR double mutants between second and third ICL
FIG. 4. TSH-induced cAMP or IP formation in double mutants between position 525 in ICL2 and position 607 or 613 in ICL3. For cAMP measurements (A and B), COS-7 cells were incubated 48 h after transfection with various concentrations of TSH. For IP measurements (C and D), COS-7 cells were labeled 48 h after transfection with 2 μCi/ml [myo-3H]inositol for 8 h and subsequently incubated with various concentrations of TSH. Data represent one of three experiments, each performed in duplicate.
Discussion
In this study we provide evidence that the N-terminal and central portions of ICL2 from I523 to D530 participate in dual Gs/Gq protein activation by the TSHR. Moreover, we show that mutant M527A loses Gq-dependent phospholipase C activation, but maintains Gs-mediated cAMP formation, suggesting a functional role for this methionine in G protein activation selectivity. This finding is supported by alanine mutants of residues F525 and R528 to D530 in the vicinity of M527, which also show selective inhibition of Gq activation. The partial restoration of the Gq-mediated response of F525K in a double mutant with T607E in ICL3 supports an intramolecular interaction between ICL2 and ICL3 in the close vicinity of F525 and T607 and thus a conformational cooperation between ICL2 and ICL3.
Identification of key residues in ICL2 for the dual G protein activation by TSHR
All three deletions within ICL2 are expressed comparable to the wt TSHR, but Gs/Gq protein coupling was abolished or strongly diminished. It is conceivable that these deletions result in TSHR conformations that allow the receptors to reach the cell surface, but do not allow productive contact with G proteins. To investigate the cause of the disturbance in intermolecular signal transduction in more detail and to gain information at the level of single aa, we introduced alanine residues within the regions of the deletion mutants. The stepwise increase in reconstitution of cAMP activity from ASU1 to ASU3 toward TM4 was a decisive indication that specific residues in the N-terminal half of ICL2 were probably involved in Gs coupling. Moreover, the complete inhibition of Gq-mediated IP production by ASU1–3 indicated a functional impact of ICL2 on this pathway. The substitution of single aa with alanine confirmed these observations, showing that the region of F525 to D530 has a slight effect on Gs signaling, but plays a more important role in Gq signaling.
The strongest change regarding selective loss of G protein activation was observed with M527A. Indeed, the efficiency of Gs coupling was affected only slightly, whereas this mutant lost the ability to stimulate Gq-dependent phospholipase C activation despite a normal cell surface expression. Interestingly, Y601H in TM5 of TSHR near the border with ICL3 is also a determinant for Gq coupling, because this mutant does not signal via Gq, but cAMP formation remains unaltered (26). Similarly, Y601N (27) confirms the functional role of Y601 in the activity and dual signaling capability of the TSHR. Taken together, these and our data are consistent with the observation that residues involved in G protein coupling and selectivity are primarily localized at the TM/cytoplasmic borders between TM3/ICL2, TM5/ICL3, and ICL3/TM6 (1).
The extension of this selective effect on Gq activation up- and downstream of M527 (F525A and R528A to D530A) fits well with the idea that receptor domains, rather than single aa, participate in forming a G protein binding pocket. Our findings are in part consistent with the data reported by Kosugi et al. (14) showing that the most important determinant for agonist-increased cAMP production is the region around residues 525–527 of ICL2 of the TSHR. However, they observed a total loss of TSH-induced cAMP response and a lesser loss in phosphoinositide signaling for a mutant in which residues 525–527 of the TSHR were substituted with a comparable ?2-adrenergic receptor sequence. Furthermore, single ASUs at positions 528–534 do not or only slightly affect Gs activation, which seems contradictory to the findings by Chazenbalk et al. (13). One explanation for these differences might be that simultaneous substitution of three or more residues has a stronger effect than single substitutions, which we also observed with our deletions and ASU mutants.
Different tasks of hydrophobic residues F525 and M527
F525: experimental indication for an interaction between ICL2 with ICL3.
The main differences in the proposed models are the orientations of F525 and M527. In the ICL2 stabilization model (A), F525 participates in a large hydrophobic core, stabilizing the fold of the TM3/ICL2 junction, whereas in the ICL2-ICL3 interaction model (B), F525 in ICL2 points toward ICL3 to residue T607 and in the vicinity of Y613. Drastic alteration of hydrophobic properties at position F525 by alanine and lysine has only a weak impact on TSH-induced cAMP formation, but, in contrast, the stimulated IP accumulation is clearly decreased. Interestingly, only the double mutant F525K/T607E achieved a partial functional rescue of stimulated IP production. In contrast, with the double mutant F525K/Y613E, TSH-induced IP formation was further decreased. Despite increased cell surface expression, Gs- and Gq-dependent signaling was strongly inhibited by M527K. Double mutants between M527 at ICL2 and Y613 and T607 at ICL3 did not lead to a functional rescue for cAMP or IP production. Taken together, a cooperative effect between ICL2 and ICL3 in the vicinity of F525 and T607 is indicated, supporting the tertiary fold proposed in model B.
M527 and its vicinity from F525 to D530 play pivotal roles in Gq-dependent signaling
Our findings suggest that the M527 side-chain is essential for Gq-mediated signaling and confirm previous findings with other 7TMRs that a methionine or another hydrophobic residue at the position of M527 could play a key role in selective signaling toward Gq (26, 28). M145 in the V2 vasopressin receptor, which is located at a position identical to M527 in the TSHR, was identified as a residue with pronounced effects on receptor/G protein coupling selectivity (18). The presence of a relatively large hydrophobic aa side-chain such as leucine or tryptophan at position 145 favored Gq signaling and suggests a recognition of the C terminus of Gq by this residue (18). This fits with our observation that alterations of the hydrophobic properties at M527 in the TSHR have the strongest selective effect on Gq protein activation. The corresponding position to M527 in the GnRH receptor also has a hydrophobic residue (L147), and its mutation led to impaired Gq signaling (17).
We propose that M527 stabilizes the hydrophobic core of ICL2, thus positioning other aa in suitable positions for Gq-protein coupling. Support for this possibility can be derived from the inhibition of Gq activation by F525A, R528A, L529A, and D530A in the vicinity of M527. Furthermore, its selectivity for Gq, its failure to interact experimentally with ICL3, and its likely cytosolic orientation in modeling provide additional support for a direct interaction of the hydrophobic residue M527 with Gq upon activation. Subsequent studies will provide examinations of these hypotheses by homology modeling of TSHR/G? complexes and related experiments. This will allow additional understanding of the dual G protein coupling of TSHR and particularly of the selectivity of activation.
Acknowledgments
We thank Mrs. Eileen B?senberg for her excellent technical assistance, BRAHMS Diagnostica (Berlin, Germany) for providing [125I]bovine TSH, Dr. Gilbert Vassart for supplying the plasmid TSHR-pSVL, and Dr. Bruce Raaka for critical reading of the manuscript.
References
Wess J 1998 Molecular basis of receptor/G-protein-coupling selectivity. Pharmacol Ther 80:231–264
Gether U 2000 Uncovering molecular mechanisms involved in activation of G protein-coupled receptors. Endocr Rev 21:90–113
Wess J 1997 G-protein-coupled receptors: molecular mechanisms involved in receptor activation and selectivity of G-protein recognition. FASEB J 11:346–354
Farrens DL, Altenbach C, Yang K, Hubbell WL, Khorana HG 1996 Requirement of rigid-body motion of transmembrane helices for light activation of rhodopsin. Science 274:768–770
Gether U, Lin S, Ghanouni P, Ballesteros JA, Weinstein H, Kobilka BK 1997 Agonists induce conformational changes in transmembrane domains III and VI of the ?2 adrenoreceptor. EMBO J 16:6737–6747
Burstein ES, Spalding TA, Brann MR 1998 The second intracellular loop of the m5 muscarinic receptor is the switch which enables G-protein coupling. J Biol Chem 273:24322–24327
Blin N, Yun J, Wess J 1995 Mapping of single amino acid residues required for selective activation of Gq/11 by the m3 muscarinic acetylcholine receptor. J Biol Chem 270:17741–17748
Havlickova M, Blahos J, Brabet I, Liu J, Hruskova B, Prezeau L, Pin J-P 2003 The second intracellular loop of metabotropic glutamate receptors recognizes C termini of G-protein -subunits. J Biol Chem 278:35063–35070
Paschke R, Ludgate M 1997 The thyrotropin receptor in thyroid diseases. N Engl J Med 337:1675–1681
Wonerow P, Schoneberg T, Schultz G, Gudermann T, Paschke R 1998 Deletions in the third intracellular loop of the thyrotropin receptor. A new mechanism for constitutive activation. J Biol Chem 273:7900–7905
Fuhrer D, Lachmund P, Nebel IT, Paschke R 2003 The thyrotropin receptor mutation database: update 2003. Thyroid 13:1123–1126
Wonerow P, Neumann S, Gudermann T, Paschke R 2001 Thyrotropin receptor mutations as a tool to understand thyrotropin receptor action. J Mol Med 79:707–721
Chazenbalk GD, Nagayama Y, Russo D, Wadsworth HL, Rapoport B 1990 Functional analysis of the cytoplasmic domains of the human thyrotropin receptor by site-directed mutagenesis. J Biol Chem 265:20970–20975
Kosugi S, Kohn LD, Akamizu T, Mori T 1994 The middle portion in the second cytoplasmic loop of the thyrotropin receptor plays a crucial role in adenylate cyclase activation. Mol Endocrinol 8:498–509
Fernandez LM, Puett D 1997 Evidence for an important functional role of intracellular loop II of the lutropin receptor. Mol Cell Endocrinol 128:161–169
Timossi C, Maldonado D, Vizcaino A, Lindau-Shepard B, Conn P, Ulloa-Aguirre A 2002 Structural determinants in the second intracellular loop of the human follicle-stimulating hormone receptor are involved in Gs protein activation. Mol Cell Endocrinol 189:157–168
Arora KK, Sakai A, Catt KJ 1995 Effects of second intracellular loop mutations on signal transduction and internalization of the gonadotropin-releasing hormone receptor. J Biol Chem 270:22820–22826
Erlenbach I, Kostenis E, Schmidt C, Serradeil-Le Gal C, Raufaste D, Dumont ME, Pausch MH, Wess J 2001 Single amino acid substitutions and deletions that alter the G protein coupling properties of the V2 vasopressin receptor identified in yeast by receptor random mutagenesis. J Biol Chem 276:29382–29392
Higuchi R 1989 Using PCR to engineer DNA. In: Ehrlich HA, ed. PCR technology. New York: Stockton Press; 61–70
Libert F, Lefort A, Gérard C, Parmentier M, Perret J, Ludgate M, Dumont JE, Vassart G 1989 Cloning, sequencing and expression of the human thyrotropin (TSH) receptor: evidence for binding of autoantibodies. Biochem Biophys Res Commun 165:1250–1255
Swillens S 1995 Interpretation of binding curves obtained with high receptor concentrations: practical aid for computer analysis. Mol Pharmacol 47:1197–1203
Berridge MJ 1983 Rapid accumulation of inositol trisphosphate reveals that agonists hydrolyse polyphosphoinositides instead of phosphatidylinositol. Biochem J 212:849–858
Palczewski K, Kumasaka T, Hori T, Behnke CA, Motoshima H, Fox BA, Le Trong I, Teller DC, Okada T, Stenkamp RE, Yamamoto M, Miyano M 2000 Crystal structure of rhodopsin: a G protein-coupled receptor. Science 289:739–745
Weiner S, Kollmann P, Nguyen D, Case T 1986 An all atom forcefield for simulation of proteins und nucleic acids. J Comp Chem 7:230–252
Laskowski RA, MacArthur MW, Moss DS, Thornton JM 1993 PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Cryst 26:283–291
Biebermann H, Schoneberg T, Schulz A, Krause G, Gruters A, Schultz G, Gudermann T 1998 A conserved tyrosine residue (Y601) in transmembrane domain 5 of the human thyrotropin receptor serves as a molecular switch to determine G-protein coupling. FASEB J 12:1461–1471
Arseven OK, Wilkes WP, Jameson JL, Kopp P 2000 Substitutions of tyrosine 601 in the human thyrotropin receptor result in increase or loss of basal activation of the cyclic adenosine monophosphate pathway and disrupt coupling to Gq/11. Thyroid 10:3–10
Yang M, Wang W, Zhong M, Philippi A, Lichtarge O, Sanborn BM 2002 Lysine 270 in the third intracellular domain of the oxytocin receptor is an important determinant for Gq coupling specificity. Mol Endocrinol 16:814–823
Cetani F, Tonacchera M, Vassart G 1996 Differential effects of NaCl concentration on the constitutive activity of the thyrotropin and the luteinizing hormone/chorionic gonadotropin receptors. FEBS Lett 378:27–31(Susanne Neumann, Gerd Kra)