Research Resource: Novel Structural Insights Bridge Gaps in Glycoprotein Hormone Receptor Analyses

Abstract

The first version of a glycoprotein hormone receptor (GPHR) information resource was designed to link functional with structural GPHR information, in order to support sequence-structure-function analysis of the LH, FSH, and TSH receptors (http://ssfa-gphr.de). However, structural information on a binding- and signaling-sensitive extracellular fragment (∼100 residues), the hinge region, had been lacking. A new FSHR crystal structure of the hormone-bound extracellular domain has recently been solved. The structure comprises the leucine-rich repeat domain and most parts of the hinge region. We have not only integrated the new FSHR/FSH structure and the derived homology models of TSHR/TSH, LHCGR/CG, and LHCGR/LH into our web-based information resource, but have additionally provided novel tools to analyze the advanced structural features, with the common characteristics and distinctions between GPHRs, in a more precise manner. The hinge region with its second hormone-binding site allows us to assign functional data to the new structural features between hormone and receptor, such as binding details of a sulfated tyrosine (conserved throughout the GPHRs) extending into a pocket of the hormone. We have also implemented a protein interface analysis tool that enables the identification and visualization of extracellular contact points between interaction partners. This provides a starting point for comparing the binding patterns of GPHRs. Together with the mutagenesis data stored in the database, this will help to decipher the essential residues for ligand recognition and the molecular mechanisms of signal transduction, extending from the extracellular hormone-binding site toward the intracellular G protein-binding sites.

The family of glycoprotein hormone receptors (GPHRs), a subgroup of the family A G protein-coupled receptors (GPCRs), consists of the TSH receptor (TSHR), the FSH receptor (FSHR), and the lutropin/choriogonadotropin receptor (LHCGR).

The general structural topology of the homologous GPHRs is identical with that of other GPCRs and is characterized by an N-terminal extracellular region (N-ECR) and an intracellular C-terminal section, with 7-transmembrane helices (TMHs) connected by 3 intracellular loops and 3 extracellular loops. The TMHs and loops constitute the serpentine domain or 7-transmembrane domain (7-TMD), which spans the membrane from the extra- to the intracellular site. A special structural feature of all GPHRs, in contrast to other family A GPCRs, is a large N-ECR (Figure 1A), which is made up of more than 320 amino acids (reviewed in Reference 1).

Figure 1.

Schematic Illustration of GPHR Topology and a Homology Model of hLHCGR N-ECR with Bound Lutropin. A, GPHRs possess a 7-TMH topology (serpentine domain consists of 7-TMHs connected by 3 extracellular loops [ECLs] and 3 intercellular loops [ICLs]), as do all GPCRs. In contrast to most family A GPCRs, they have a very large N-ECR. This N-ECR is composed of the LRRD and the hinge region. The LRRD is made up of 11 full repeats. Cb-1 comprises a cysteine from the N terminus, 2 cysteines from an antiparallel β-strand upstream and one from the first regular repeat. The hinge region, now structurally defined by the new FSHR crystal structure, starts after a short helix-element of the 11th repeat and contains the first 2 cysteines of Cb-2, which are linked via disulfide bridges to Cb-3. After the helix, an unstructured region harbors the cysteine box 2/3 linker and the cleavable peptide (C-peptide), in the case of TSHR. The sulfotyrosine is known from mutagenesis studies and contributes to hormone binding in all 3 GPHR subtypes. The C-terminal section of the hinge region contains a β-strand that is located parallel to the last repeat of the LRRD and is connected via a disulfide bridge. B, Homology model of the N-ECR of hLHCGR in complex with hLH. The gray ribbon presentation visualizes the Cα atom backbone of the LRRD with 9 repeats, according to previous information concerning the LRRD dimension. The new structural information for the LRRD (33) is colored magenta and includes an extension of the LRRD by repeat 11 (β-strand [light green] and a short helix [light blue]), a coiled region with the sulfated tyrosine 331 (blue sticks) and finally a β-strand orientated parallel to the last strand of the LRRD. The 2 cysteines of the short helix are bridged via disulfide bonds between the β-strand and the C terminus of the hinge region. The hLH is visualized as a surface (α-subunit in pale orange, β-subunit in pale green).

The major binding region for the hormones TSH, LH, chorionic gonadotropin (CG), and FSH (2) has been identified by experimental studies on the LRRDs (3–7). A complementary pattern of amino acid side-chain properties is responsible for specific hormone recognition (8). Furthermore, the hinge region (comprising about 90–130 residues) has also been shown to be important for hormone binding and signal induction (examples in References 9–15). On the basis of site-directed mutagenesis, it has been concluded that the extracellular hinge region of GPHRs is necessary for stabilization of a signaling-competent basal receptor conformation (12, 15–23), and it has been proposed that the hinge region plays an important role in the regulation and amplification of receptor activity (11, 16, 21, 24).

The first crystal structure of the FSHR leucine-rich repeat domain (LRRD) complex with the hormone FSH (PDB [Protein Database] entry code: 1XWD) was published in 2005 (26). Moreover, the structural features of the FSHR LRRD were subsequently confirmed by crystal structure complexes between the TSHR LRRD and an activating (27) (PDB entry code: 3G04) or inactivating (28) autoantibody (PDB entry code: 2XWT), respectively.

However, the detailed molecular mechanism of receptor activation and the regulation of signaling activity at the extracellular site are not fully understood. One major reason for this has been the lack of structural information for the hinge region containing the second hormone-binding site, which would be a prerequisite to clarify the exact arrangement of receptor components to each other or to describe the entire hormone-binding process. This missing structural section was also absent in the freely available GPHR information resource (web-address: http//:ssfa-gphr.de). This database and linked web applications were designed to collect relevant GPHR information, with the aim of supporting sequence-structure-function analysis (SSFA) of this GPCR subfamily (29). The SSFA-GPHR information resource allows a focused analysis of semiquantitative mutant data from GPHR subtypes and diverse experimental approaches (30, 31). A second and complementary GPHR information resource with mutant data and several online applications has been provided by others (Reference 32; http://gris.ulb.ac.be/).

A new GPHR crystal structure (hFSHR in complex with hFSH) that is a milestone in GPHR research has recently been published (33). This structure presents structural information on almost the entire extracellular region, including the LRRD and most parts of the hinge region that had not yet been solved. Binding site analysis revealed that both the LRRD and the hinge region interact with the hormone. In consequence, combination of this novel structural information with available functional information allows us to answer important questions and to interpret experimental findings on GPHRs. For example, a section of the hinge region interacts directly with the hormone, and structural details of this second binding site can now be elucidated (33).

This advanced information is also relevant to the homologous TSH and LH receptor and demands a revision and upgrade of our GPHR research resource. Therefore, we have now implemented not only the new extracellular human FSHR structure, but have also provided the derived new homologous structural models for GPHR subtypes (hTSHR, hLHCGR, rLHCGR, rFSHR) and linked functional information with these new structural insights. This enables more detailed extraction of data for these receptors and analyses of differences or similarities among the GPHRs. This upgrade includes, for the first time, not only the extracellular receptor region, but also the bound hormones. To ensure optimal exploitation of the complete hormone-binding interface at the extracellular region, a new Interface Analysis Tool has been added that offers visualization and analysis options to illuminate the contacts and interactions between the hormone and receptor complex.

Materials and Methods

Homology modeling of different GPHRs bound with hormone subtypes

Structural homology models for the extracellular domains of the TSHR and LHCGR were built on the basis of the newly available FSHR crystal structure (33). The procedure of hTSHR modeling has been recently published (34), and this model (TSHR extracellular region with bound bTSH) was also used for the database. In brief, the incomplete LRRD of hTSHR with 9 repeats was extracted from the LRRD-antibody complex (27) (PDB entry code: 3G04) and extended by adding amino acids from the superimposed full FSHR N-ECR crystal structure with 11 repeats (33). The 11-repeat is followed by a helical element comprising cysteine box 2 (Cb-2).

Consecutively to Cb-2, the new FSHR N-ECR crystal structure includes the N-terminal fragment of the hinge region (hFSHR positions, Trp291–Ser295; hTSHR positions, Gln289–Ser304). This portion was also added to the hTSHR LRRD model, and side chains of TSHR were substituted. Disulfide bridges between cysteines of Cb-2 and Cb-3 were built as suggested by the FSHR crystal structure. Furthermore, the corresponding region of the hFSHR crystal structure was used as a structural template (hFSHR positions, Thr331–Ile359) in the design of the hTSHR C-terminal hinge region fragment (hTSHR positions, Ser383-Ile411).

The sulfated hFSHR Tyr335 was already known as mandatory for hormone binding and signaling and interacts tightly with amino acids of the hormone subunits between the β-subunit loop 2 and the α-loop 1 (33). Therefore, if it is assumed that this sulfated tyrosine in TSHR also interacts with the same hormone-binding site as that observed for the FSHR-FSH complex in the TSHR model Tyr385 (termed Tys385), it then substitutes for the sulfated FSHR Tyr335.

The hormone bTSH was modeled on the basis of hFSH, so that the general mode of orientation between the hormone and receptor was maintained as suggested by the human FSHR-FSH complex.

The same procedure was repeated for the complex structures of human and rat LHCGR. For hLHCGR, both interaction models, hLHCGR/hLH and hLHCGR/hCG, are provided. For the complex with bound hCG, hFSH in the template structure of the hFSH-hFSHR complex was substituted with the crystallized structure of hCG (35). In the hLH model, this hormone structure was used as a template. The CG interaction model is currently only preliminary, due to the lack of any structural information about its C-terminal tail (24 amino acids), which might influence the binding mode of hCG (36). In analogy, the rLHCGR model was created. In rodents, only LH is responsible for rLHCGR activation (37); therefore, the respective interaction model, rLHCGR/CG, is not provided.

The rFSHR model was designed by substituting subtype-specific residues, while keeping the hFSHR/hFSH at conserved positions.

Glycosylation was not considered in any structures or structural models. Side chains of the homology models were subjected to conjugate gradient minimizations (until they converged at a termination gradient of 0.05 kcal/[mol*Å]). The AMBER F99 force field was used. Finally, the models were minimized without constraints. Structural modifications and homology modeling procedures were performed with Sybyl X2.0 (Tripos, Inc, St Louis, Missouri).

General modifications of the database according to the new information from the FSHR crystal structure

The new FSHR structure revealed new structural definitions of specific receptor sections, eg, the length of the LRRD or the hinge region (33). We, therefore, refined our previous annotations in the database for specific sections and amino acids, in accordance with the new set of structural descriptions, and also revised our structural annotations in the alignments, numbering scheme, and snake-plot designer.

In the analysis of the hormone receptor-binding interface and for localization of functional information in proximity to the binding area, and in contrast to former database versions, we added the whole complex and the derived homology models. The 3-dimensional interactive windows to visualize functional information of the hormone-receptor complex were adapted and optimized. For example, in the 3D-Structural Search, one can easily switch between solely the N-ECR, the complex of N-ECR and hormone, or only the 7-transmembrane domain.

Furthermore, in the previous version of our web application, amino acids in the hinge region with functional insights (mutations) were not linked to any structural information, even though the hinge region is known to be important for signaling and ligand interaction. We have now added available structural information for the hinge regions of GPHR subtypes, in order to visualize specifically wild-type amino acids with known functional data from mutagenesis studies.

New tool for the interface analysis of hormone-receptor complex

The new FSHR structure and homologous models for TSHR and LHCGR provide additional insights into hormone recognition, as well as binding and receptor activation. For the identification and analysis of essential amino acids, a new tool was implemented to analyze opposing binding patterns by selectively visualizing the contact surface, either for the hormone or for the receptor, or by highlighting interaction partners.

For displaying intermolecular interactions via a surface, a new Jmol (38) function for contact maps is applied (Figure 2A). The color of the surface is based on the contact proximity of the van der Waals surface between potentially interacting partners. Red spots are the closest contacts and usually correspond to hydrogen bonds. On the other hand, yellow and green spots highlight more distant side chains and often correspond to hydrophobic interactions. The user can visualize contacting residues with distinct properties on opposing sides of the complex, eg, charge-charge interactions − positive vs negative charges, with the tool to highlight them separately (Figure 2B).

Figure 2.

Protein-Protein Interface Analysis Tool. The display shows the homology model of hTSHR (LRRD in gray; hinge region in magenta) in complex with the hormone bTSH (α-subunit in orange, β-subunit in green) as visualized at the SSFA-GPHR website. The surface of the hormone is colored according to the receptor-hormone contact proximity of the van der Waals radius between the interaction partners. Red spots are the closest contacts and usually correspond to hydrogen bonds. Yellow and green spots highlight more distant side chains and often correspond to hydrophobic interactions. As an example, in panel A, all negatively charged N-ECR residues in close contact to the hormone are visualized together with the contact surface displayed on bTSH. The inset shows a close-up of particular negatively charged side chains of the hinge region interacting with the hormone. In panel B the interaction partners (positively charged side chains) on the interface are highlighted (labeled manually for representation).

Results and Discussion

We now report an updated and revised version of a Glycoprotein-Hormone Receptor Database (29), which has been improved significantly by implementation of new structural information. This mainly comprises missing sections (90–130 residues) of the N-ECR, namely the extracellular hinge region that was recently solved for the FSHR by x-ray crystallography in a complex with FSH (33). In addition, extended insights into the structure of the LRRD were observed, namely by enlarging this domain (already predicted in Reference 11). Finally, the second hormone-binding site at the hinge region is deciphered in detail around a sulfated tyrosine and can be analyzed in complex with the hormone for all GPHR subtypes.

In addition to the participation of Cb-1 in the N-terminal fold of the LRRD, the new structure revealed that the LRRD continues through specific amino acids of Cb-2 (Figure 1). Thus, the LRRD is constituted by 11 LRRs, instead of 9 as assumed previously. The short helical structure at repeat 11 contains the 2 consecutive cysteines that are bridged to the last 2 extracellular cysteines of Cb-3. These are located on one side in a short β-strand parallel to the LRRD β-strand 11 and on the other side are adjacent to TMH-1 (Figure 1). In conclusion, the previous prediction of tight spatial proximity between Cb-2 and Cb-3 (10) is confirmed, as well as the suggested cysteine bridges (17, 36, 39). Regarding the short helix linking the extended LRRD and the hinge region toward the transmembrane domain via 2 disulfide bridges, it is conceivable that this may act as a pivot and center, where the conformational effects of 2 hormone-binding sites (LRRD and hinge) converge toward the transmembrane domain (25).

Moreover, a specific feature appearing at the second hormone-binding site of the new crystal structure is a sulfated tyrosine at the C terminus of the hinge region that binds in a pocket located between the α- and β-subunit of the hormone. All GPHRs exhibit such a sulfotyrosine (FSHR Tyr335, LHR Tyr331, TSHR Tyr385). Several functional studies had already confirmed the general importance of this sulfated residue for hormone binding and signal transduction at GPHRs (36, 40). Tyrosylprotein sulfotransferases (such as tyrosylprotein sulfotransferase 2) are enzymes that catalyze the posttranslational modification of specific tyrosines by adding an SO₃ group. This addition enhances the negative electrostatic potential of the side chain and therefore increases the number of possible hydrogen bonds formed. Their relevance for TSHR might also be indicated in vivo by the absence of tyrosylprotein sulfotransferase 2 in knockout mice, which leads to hypothyroidism (41).

Interestingly, distinct sulfated tyrosines at extracellular N-terminal domains of several chemokine-receptors have also been shown to interact with the corresponding endogenous ligands: CXCR4, CCR2B, CCR5, CXCR3, CX3CR1, C5a receptor, and CCR8 (eg, Refs. 42–44). As observed in the FSHR crystal structure, the crystal structure of the N-terminal tail of the CXCR4 (PDB entry code 2K03) contains a sulfated tyrosine, with strong hydrogen bond interactions between arginine and asparagine side chains in the chemokine (45). Finally, tyrosine sulfation is here of pathophysiologic relevance too, as indicated by interrelations with HIV infection mechanisms via chemokine receptors (46).

Moreover, several examples of important functions for sulfotyrosines are also known in non-GPCR proteins: 1) direct contacts in selectins, 2) proteolytic processing (eg, gastrin processing), and 3) proteolytic activation of extracellular proteins (eg, factor V and VIII activation) (47).

The more extensive implications for ligand/protein recognition reflect the general importance of this issue and are relevant to the understanding of the hormone (ligand)-receptor interaction in more detail. In the new version of our database, we have therefore implemented a visualization tool of interactions between the receptor and hormone. Particular amino acids with specific biophysical properties can be depicted and highlighted (Figure 2); on the other hand, surface rendering combined with distance mapping can be used for contact analyses. In particular, this might guide comparative studies between the GPHRs in the now almost complete extracellular complexes. Additionally, it might also be helpful in explaining the diverse functional properties of the GPHRs (48, 49).

Finally, this current report and work provide a progressive step toward a comprehensive understanding of GPHRs, including important aspects such as signal transduction and structural organization.

Funding Statement

This work was supported by research grants KL2334/2–2 and KR1273/4–1 from the Deutsche Forschungsgemeinschaft.