This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Grossmann, M. Articles by Szkudlinski, M. W. Search for Related Content PubMed PubMed Citation Articles by Grossmann, M. Articles by Szkudlinski, M. W.

Novel Insights into the Molecular Mechanisms of Human Thyrotropin Action: Structural, Physiological, and Therapeutic Implications for the Glycoprotein Hormone Family

Laboratory of Molecular Endocrinology, Department of Medicine, University of Maryland School of Medicine and the Institute of Human Virology, Medical Biotechnology Center, Baltimore, Maryland 21201

	Abstract

Top Abstract I. Introduction II. Structure-Function... III. Current Understanding of... IV. Physiological and... V. Evolutionary Considerations VI. Strategies in the... VII. Nonclassic Actions of... VIII. Perspectives on Structure... References

 

I. Introduction A. Historical background

B. TSH and the glycoprotein hormone family

II. Structure-Function Relationships of TSH in Relation to Studies on Gonadotropins

A. Methodological considerations

B. Structure-function studies of protein domains

C. Structure-function studies of carbohydrate chains

III. Current Understanding of TSH/Glycoprotein Hormone Action

A. Structural considerations

B. Hormone-receptor interaction

C. Cooperation of individual hTSH domains in receptor activation

IV. Physiological and Pathophysiological Implications

A. Carbohydrate heterogeneity

B. Naturally occurring glycoprotein hormone mutations

C. "Specificity spillover" syndromes

V. Evolutionary Considerations

A. Glycoprotein hormone specificity

B. Evolutionary changes in TSH activity

VI. Strategies in the Design of Novel TSH Analogs and Therapeutic Implications

A. Clinical use of rhTSH

B. Design of novel glycoprotein hormone analogs

VII. Nonclassic Actions of TSH and Gonadotropins

A. Extrathyroidal/extragonadal glycoprotein hormone actions

B. Relationship to the cystine knot growth factor superfamily

VIII. Perspectives on Structure-Function Studies of TSH

	I. Introduction

Top Abstract I. Introduction II. Structure-Function... III. Current Understanding of... IV. Physiological and... V. Evolutionary Considerations VI. Strategies in the... VII. Nonclassic Actions of... VIII. Perspectives on Structure... References

 
A. Historical background
THE scientific history of thyrotropin (thyroid-stimulating hormone, TSH) began in the 1920s with the discovery of thyroid-stimulating activity in the pituitary gland (reviewed in Refs. 1 and 2). This was followed in the early 1970s by the determination of the primary amino acid sequence of the TSH subunits (3). In the late 1980s, a detailed description of its carbohydrate structures was accomplished (4, 5). The subsequent cloning of the human (h) -subunit (6) and hTSH ß-subunit gene (7, 8, 9) as well as the TSH receptor gene (10, 11, 12, 13) set the stage for the ensuing progress in studies on hTSH structure-function relationships and enabled the production of recombinant (r) hTSH (14), now in clinical trials for the follow-up of patients with differentiated thyroid carcinoma (15, 16). From the standpoint of basic science, another major breakthrough occurred in 1994 with the elucidation of the structure of the closely related human chorionic gonadotropin (hCG) (17, 18), which showed that the glycoprotein hormones belong to the superfamily of cystine knot growth factors. In addition, the crystallization of the ribonuclease inhibitor with specific structural elements termed leucine-rich repeats (LRR) (19) paved the way for the modeling of the extracellular domain of glycoprotein hormone receptors, as these receptors also contain such LRR (10, 20, 21).

Since the last excellent review on TSH in this journal (1, 2), there has been considerable progress in the understanding of the molecular features and the clinical applications of TSH. This review will focus on the structure-function relationships of hTSH in the context of the glycoprotein hormone family and present current views of the molecular mechanisms of glycoprotein hormone action. It will also discuss the physiological, pathophysiological, evolutionary, and therapeutic implications emerging from this research. Novel approaches in structure-function studies and their implications for the rational design of glycoprotein hormone analogs will be summarized. The concomitant progress made in the chromosomal localization, structural organization, and regulation of the TSH - and ß-subunit genes will not be dealt with here, as this topic has recently been covered in detail (22, 23, 24, 25, 26).

B. TSH and the glycoprotein hormone family
TSH is a 28- to 30-kDa glycoprotein produced in the thyrotrophs of the anterior pituitary gland. Its synthesis and secretion are stimulated by TRH and inhibited by thyroid hormone in a classic endocrine negative feedback loop. Differences in the molecular mass of TSH are primarily due to the heterogeneity of carbohydrate chains. In contrast, heterogeneity of its subunit termini as well as the different extent of deamidation of glutamine and asparagine residues are presumably isolation artifacts (27). TSH controls thyroid function upon its interaction with the G protein-coupled TSH receptor (28, 29, 30, 31). TSH binding to its receptor on thyroid cells leads to the stimulation of second messenger pathways involving predominantly cAMP and, in high concentrations, inositol 1,4,5-triphosphate and diacylglycerol, ultimately resulting in the modulation of thyroidal gene expression (32).

Physiological roles of TSH include stimulation of differentiated thyroid functions, such as iodine uptake and organification, the release of thyroid hormone from the gland, and promotion of thyroid growth (27). It also acts as a thyrocyte survival factor and protects the cells from apoptosis (33), perhaps, as has been shown for hCG, via regulation of p53 and the bcl-2 gene family (34, 35). A further interesting finding is that TSH plays a critical role in ontogeny. In a mouse model with targeted disruption of the common -subunit gene and thus devoid of circulating glycoprotein hormones, thyroid development was arrested in late gestation (36).

TSH is a member of the glycoprotein hormone family, which also includes pituitary follitropin (FSH) and lutropin (LH), as well as CG, which is produced predominantly by the placenta. TSH, FSH, and LH are found in all mammalian species as well as in lower vertebrates (3, 37). In contrast, CG is only present in higher primates and in the horse. The CG ß gene had probably only recently evolved from the LH ß gene by a frame-shift mutation with readthrough into the 3'-untranslated region (38). Structurally, the glycoprotein hormones are related heterodimers comprised of a common -subunit and a hormone-specific ß-subunit (3). The common human -subunit contains an apoprotein core of 92 amino acids including 10 half-cystine residues, all of which are in disulfide linkage. It is encoded by a single gene, located on chromosome 6 in humans, and thus identical in amino acid sequence within a given species (39). The ß-subunits can be aligned according to 12 invariant half-cystine residues forming six disulfide bonds. Despite a 30–80% amino acid sequence identity among the hormones, the ß-subunit is sufficiently distinct to direct differential receptor binding with high specificity (less than 0.1% cross-reactivity) (3). The glycoprotein hormone ß-subunit genes differ in length, structural organization, and chromosomal localization (22, 23, 24, 25, 26) (summarized in Table 1). The human TSH ß-subunit gene predicts a mature protein of 118 amino acid residues and is localized on chromosome 1 (27). The fact that human TSH ß-subunit isolated from human pituitaries has an apoprotein core of 112 amino acids is most likely related to carboxyl-terminal truncation during purification. In any case, structure-function studies showed that amino acid residues 113–118 are not required for the activity of hTSH, at least for that in vitro (40).

View larger version (15K): [in this window] [in a new window]   Figure 1. Asparagine (N)-linked oligosaccharides of TSH. The sulfated biantennary structure (A) represents the typical oligosaccharide chain of pituitary bTSH and bLH. The sulfated and sialylated oligosaccharide (B) is more typical of pituitary hTSH and hLH (4, 5). The sialylated nonsulfated structure (C) represents that of rhTSH expressed in CHO cells. Similarly, pituitary hFSH as well as placental hCG are exclusively sialylated (4, 5). Carbohydrate residues are marked as follows: mannose (), N-acetylglucosamine (), N-acetylgalactosamine (•), fucose (), galactose (), sialic acid ().

Traditionally, structure-function relationships of human glycoprotein hormones have been predominantly performed with gonadotropins, particularly hCG (41, 42, 43, 44, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61). This was mostly because hCG purified from urine was readily available and because of the early cloning of the hCG ß-subunit genes reflecting their relative abundance in the placenta (24, 38). Studies on hTSH, in contrast, were hampered by the difficulties in isolating sufficient amounts of hTSH from the pituitary and later by limitations of rhTSH expression after the cloning of the 2-kb hTSH ß-subunit gene fragment (9, 62). Several recent developments have greatly facilitated hTSH structure-function analysis: availability of rhTSH (14), construction of a 981-bp hTSH ß-minigene from the original 2-kb fragment (63), cloning of the hTSH ß-subunit cDNA (M. Grossmann, M. W. Szkudlinski, and B. D. Weintraub, unpublished data), development of suitable hTSH expression systems using eukaryotic cells (14, 50, 63, 64), and the cloning of the TSH receptor cDNA (10, 11, 12, 13). The ensuing progress in understanding hTSH action at the molecular level has highlighted unique features of hTSH, which set this hormone apart from other members of the glycoprotein hormone family. In addition, this progress has helped to understand common principles of glycoprotein hormone action.

	II. Structure-Function Relationships of TSH in Relation to Studies on Gonadotropins

Top Abstract I. Introduction II. Structure-Function... III. Current Understanding of... IV. Physiological and... V. Evolutionary Considerations VI. Strategies in the... VII. Nonclassic Actions of... VIII. Perspectives on Structure... References

 
A. Methodological considerations
Proteins such as TSH are engineered with the goal of better understanding the molecular mechanisms of their function as well as creating novel analogs for practical purposes. Structure-function studies on glycoprotein hormones can be categorized into studies on the carbohydrate moiety as well as on the peptide portion, but the two approaches can also be combined (Table 3). In general, each of these methods has its advantages, which are balanced by inherent limitations. Initially, physicochemical and enzymatic studies have identified amino acids, as well as carbohydrate portions, on both subunits that contribute to receptor binding and signal transduction. These approaches, which are summarized in excellent reviews (41, 42, 52, 53, 54, 56, 57, 59, 60), have been instrumental in gaining an initial understanding of glycoprotein hormone structure-function relationships and continue to provide valuable information to the present time. Other valuable strategies rely primarily on epitope mapping or the use of synthetic peptides (55, 58, 61, 65, 66).

In addition, such combined techniques can lead to the recognition of "modification-permissive domains" that allow introduction of nonconservative changes into hormones, thus enabling modulation of function without compromising protein synthesis (46, 50). Further development of such strategies, including multiple residue replacement, should be helpful to elucidate cooperative effects of individual residues, and this can be extended to the simultaneous mutagenesis of multiple, topically unrelated hormone regions. With such an approach, it should ultimately be possible to individually modulate and dissociate defined biological properties of complex molecules such as hTSH. In fact, this strategy led to the finding that a partial or complete loss of hTSH activity caused by modifications in one domain may in certain instances be completely compensated for by alterations in an unrelated domain (69). Such studies predict that the TSH receptor is capable of tolerating ligands with significant structural modifications, by means of an "analog-induced fit." It may even be possible, therefore, to create alternative contact domains of analog and receptor that are still able to transduce a signal. Such plasticity of ligand-receptor interactions is supported by the observation that the hTSH receptor can be constitutively activated by multiple mutations in various receptor regions (29). Moreover, identification of cooperative, noncooperative, and mutually exclusive hormone domains can provide important leads for further development of therapeutically useful hormone analogs.

It should be pointed out that, as with other approaches, these recombinant techniques are not without limitations. For adequate interpretation of mutagenesis studies, possible effects of a mutation caused by aberrant subunit folding and dimerization should be considered. Such changes could result in distant conformational effects that may alter hormone function in an indirect fashion. This is especially possible if secretion or receptor binding properties of mutated analogs are profoundly impaired. In contrast, "gain of function" changes, such as enhanced receptor binding or switch of hormonal specificity are more likely to be the result of direct residue/domain-specific effects. Nevertheless, it is prudent to ascertain accurate quantification and to rule out the possibility of global conformational changes of analogs with multiple mutations by testing them against a panel of different antibodies or circular dichroism spectroscopy.

Restoration of the activity of a mutant hormone analog by appropriate modifications of the receptor can also demonstrate that a mutation causes a site-specific decrease of hormone activity. Such parallel mutagenesis of ligand and receptor is a promising approach that is more complex and has so far received only scant attention (71). This combined strategy should allow identification of cooperative interactions of specific domains of ligand and receptor and therefore be highly informative in understanding mechanistic aspects of glycoprotein hormone signal transduction.

B. Structure-function studies of protein domains
Multiple domains of both the - and ß-subunits have been shown to be important for heterodimer assembly, secretion, and bioactivity of the glycoprotein hormones. Among these regions, several segments that are highly conserved among different species have been confirmed to be particularly important for receptor binding and bioactivity of hTSH by a variety of different approaches. Whereas Fig. 2 summarizes the results from site-directed mutagenesis studies in the linear subunit gene sequences, Fig. 3 shows the topical relations of identified domains in a hTSH ribbon model based on the structure of hCG (17, 18).

View larger version (16K): [in this window] [in a new window]   Figure 2. Site-directed mutagenesis of hTSH. This figure shows the primary structure of the human -subunit (A) and the hTSH ß-subunit (B), using the one-letter code for amino acids. Functionally important residues are numbered and highlighted in boldface in the wild type sequence. Specific mutations are shown either above the corresponding wild-type residue (if the mutation resulted in an increase of hTSH activity), or below (if mutation caused a decrease of activity). Mutations N78Q and ßN23Q, which disrupt a glycosylation recognition sequence (highlighted by the outlined font) decrease in vivo, but not in vitro activity (see text), whereas disruption of the glycosylation recognition sequence at N52 (N52Q, N52D) increases in vitro, but slightly decreases in vivo, activity (74). Multiple residue mutations are underlined. Also depicted is the location of selected structural features. The ß-hairpin loops correspond to the continuous lines above the primary structure flanked by arrows. The bold part of these lines indicates ß-sheet and the thin part of the line indicates the actual loops. -Helix (between 16–18 and 40–46) is marked with a dashed line above the sequence. The seat belt between ßC88 and ßC105 is marked by a dashed line flanked by arrows. Chimeric substitutions in this not only decreased hTSH activity but also changed receptor specificity (see text). The extent of truncations at the carboxyl termini of both subunits is shown by the dashed lines below the primary structure. The following mutations not depicted in the figure did not change hTSH activity: Q13A, P16A, Q20A, P40A, L41A, L41P, Q50P, R67K, H90L, K91A, K91D, K91T, ßR60A, ßR60E. * P38D, The activity of this construct could not be tested due to failure of this mutant -subunit to heterodimerize with the TSH ß-subunit.

View larger version (23K): [in this window] [in a new window]   Figure 3. hTSH ribbon homology model showing domains important for hTSH activity. The schematic drawing of hTSH is based on a molecular homology model of hTSH (50), built on the template of an hCG model derived from crystallographic coordinates obtained from the Brookhaven Data Bank (17). The two ß-hairpin loops (L1, L3) in each subunit are marked. Each subunit also has a long loop (L2), which extends from the opposite side of the central cystine knot. The ßL2 loop corresponds to the "Keutmann loop" of the hCG ß-subunit, whereas the L2 loop has a peripheral -helical structure (residues 40–46). Depicted are the topical relationships of regions important for hTSH receptor binding and activity. The -subunit is shown as a checkered, and the ß-subunit as a solid, line. The functionally important -subunit domains are boxed. Important domains of the ß-subunit are marked directly within the line drawing: the TSH ß-subunit 58–69 region in the ßL3-loop is depicted by the crossed line. The seat belt region between the 10th (C88) and 12th (C105) cysteine is highlighted as follows: the N-terminal part of the seat belt, the "determinant loop" (C88-C95 in the hTSH ß-subunit), corresponds to the beaded line, and the carboxyl-terminal segment (C95-C105 in the hTSH ß-subunit) corresponds to the dashed line. Because the carboxyl terminus beyond hCGß 111 was not traceable in the original electron density map (17), the hTSH ß-subunit is only drawn to the corresponding residue 106. Because of deglycosylation of hCG before crystallization, the large and flexible oligosaccharide side chains are not shown. The -asparagine52 carbohydrate (origin marked) is predicted to project away from the protein backbone into the proposed receptor-binding domain (116), which also includes the 40–46 helix and the -carboxyl-terminus 88–92 (17, 18). In contrast, the 11–20 domain is not located in proximity to the hTSH ß-subunit seat belt.

1. Common -subunit domains.
Despite the general importance of these -subunit domains (Figs. 2 and 3) in glycoprotein hormone activity, recent studies on hTSH have revealed important differences in the role of certain domains for hTSH compared with hCG and hFSH. Coexpression of selected mutant -subunits with the ß-subunits of hTSH, hFSH, and hCG showed that specific residues within the 33–38 domain played strikingly different roles for glycoprotein heterodimer secretion. In light of the high degree of structural and functional homology, these differences were surprising: for example, an -subunit in which -alanine³⁶ was replaced by glutamic acid was not able form a dimer with the hCG ß-subunit, whereas this mutated -subunit combined efficiently with the hTSH ß-subunit to give rise to a bioactive heterodimer (72). Alanine scanning showed that residues -phenylalanine³³ and -arginine³⁵ were critical for hCG, but not hTSH, receptor binding (72). Conversely, en bloc alanine replacement of the surface exposed positively charged -helical fragment -arginine⁴²-serine⁴³-lysine⁴⁴ reduced hTSH, but not hCG, activity (72, 86, 92). Similarly, the -asparagine⁵² oligosaccharide played opposite roles for hCG and hTSH signal transduction, as outlined below (74, 81). In addition, a single amino acid, the ultimate -serine⁹², was identified to play an important role for heterodimer secretion, receptor binding, and bioactivity of hTSH, but not for that of hCG or hFSH (64, 85, 93). This observation explains the evolutionary constraint to preserve this residue in CG, LH, and FSH, because the -subunit is encoded by a single gene (39). A study using overlapping -subunit peptides also showed that 26–46 and the -carboxyl-terminus 81–92 were important receptor-binding domains of hTSH (65), illustrating the validity of both complementary approaches. However, a comprehensive study using alanine-substituted peptides encompassing the 26–46 region identified specific residues important in receptor binding (73), only some of which were confirmed by creation of the corresponding hTSH mutants with site-directed mutagenesis (72). Thus, such comparisons indicate that the effect of a substitution of an amino acid within a linear, structurally not constrained peptide may not always be comparable to the same substitution within the context of the heterodimeric hormone.

In addition to these differences in the importance of such common -subunit regions for TSH activity compared with the gonadotropins, there are also similar roles of these domains for the activity of all members of the glycoprotein hormone family. Thus, truncation of three or more residues from the -carboxyl terminus eliminates the activity of hTSH, hCG, and hFSH almost entirely (64, 76, 77, 84, 85). Moreover, a combination of alanine/proline scanning revealed that several residues of the 40–51 region were critical for both hTSH and hCG (-proline³⁸, -lysine⁵¹), although the role of some residues appeared to be hormone-dependent (-phenylalanine³³, -arginine³⁵, -alanine³⁶, -arginine⁴²-serine⁴³-lysine⁴⁴, -leucine⁴⁸) (69, 72, 86).

The 11–20 region contains a cluster of basic residues in all vertebrates except hominoids and forms a previously unrecognized domain with the ability to potentiate receptor binding and signal transduction, as well as an important motif in the evolution of glycoprotein hormone bioactivity (Table 4 and 50 . In contrast to the above domains, 11–20 is not highly conserved among the species and is a modification-permissive site. Hence, this region allows amino acid substitutions with no or minimal effect on hormone production, but substantial increases of bioactivity. In contrast, tightly conserved regions are usually "modification nonpermissive sites" and cannot be altered without a perturbation in hormone structure resulting in major decreases of hormone production and concomitant loss of function. Based on evolutionary considerations detailed below, positively charged lysine residues were inserted into the -cysteine¹⁰-proline²¹ region of the human -subunit, as well as a single nonconservative ß-leucine⁶⁹-arginine mutation in the TSH ß-subunit. Such changes, individually as well as in various combinations, increased the potency and efficacy of hTSH and hCG mutants. Most notably, each mutation to a lysine residue in the 11–20 region caused a substantial increase in activity, but alanine mutagenesis of these residues in the hTSH did not significantly alter hormone activity, indicating that only the selective reconstitution of basic amino acids was functionally significant (50). Moreover, the substitution of -serine⁴³ to arginine (69) and replacements of -histidine⁹⁰ and -lysine⁹¹ (64) either decreased or did not change TSH activity. Thus, introduction of basic residues does not uniformly lead to an increase of hormone activity, but the importance of such basic residues varies depending on their location within the molecule.

View this table: [in this window] [in a new window]   Table 4. Amino acid sequence alignment of various vertebrate -subunits in the L1 ß-hairpin loop

Another functionally important domain in the hTSH ß-subunit was recently identified by focusing on regions of nonhomology between the different human ß-subunits. In this respect, targeting of residues with charge differences is of particular interest, as basic residues have been implicated to play a role in receptor binding and activation of TSH, as described above (50, 96). Such nonconserved regions of the ß-subunits could be involved in regulating glycoprotein hormone specificity or may represent modification-permissive domains generally important for signal transduction, which diverged during evolution of the different ß-subunits. If the latter was true, these would constitute regions in which site-directed mutagenesis may be useful to specifically alter hormone activity and therefore would be of primary interest for the generation of hormone analogs. Using this approach, a novel domain within the ß-hairpin ßL3 loop of the hTSH-ß subunit was identified that appears to modulate hTSH receptor binding and signal transduction (77a). Sequence comparison of hCG and hTSH ß-subunits showed a region (residues 58–69 of the TSH-ß subunit) that contains a cluster of basic residues in hCG, but not in hTSH (net charge +2 in hCG vs. 0 in hTSH). This domain is located peripherally within the ß-hairpin ßL3-loop and appears surface-exposed in the crystal structure in hCG. Interestingly, epitope-mapping studies of hCG/hCG receptor complexes had suggested that this region may be in direct contact with the hCG receptor (97, 98). Analogous to previous studies of the -subunit 11–20 domain, introduction of single and multiple basic residues into this hTSH ß-subunit domain led to additive, substantial increases of TSH receptor binding affinity as well as intrinsic activity.

C. Structure-function studies of carbohydrate chains
The oligosaccharide moieties assume importance in every aspect of the life span of TSH, from early translational events during biosynthesis to its removal from the circulation and degradation. The specific functions of the oligosaccharides change as the hormone travels through distinct intracellular compartments during its synthesis, as well as after secretion. Overall, the carbohydrates serve comparable functions among the members of the glycoprotein hormone family (5, 41, 42, 99). However, more recent work has shown that, in certain cases, the oligosaccharides have unique side-chain and residue-dependent roles for hTSH, which are different from those for the gonadotropins. Studies on oligosaccharides of individual hormones are therefore, by analogy to those of the protein component, important to recognize the hormone-specific roles of these structures. Moreover, they can have substantial implications for the design and production of clinically useful glycoprotein hormone analogs. This is especially relevant because an understanding of their function offers the possibility to modify them in a rational fashion using recombinant DNA methodology and heterologous cell expression (100, 101). Indeed, several studies have demonstrated that bioreactor conditions or cell culture techniques can affect the carbohydrate structures of cell culture-derived glycoproteins including hTSH (100, 101, 102).

1. Postranslational modifications and intracellular processing.
Various methods have been used to study the functional role of the oligosaccharides for TSH and the other glycoprotein hormones in experimental settings, including physicochemical, enzymatic, and molecular methods (Table 3). Similar to findings for other members of the glycoprotein hormone family, the cotranslational attachment of the oligosaccharides which protects the nascent polypeptide from intracellular degradation is essential for the subunit folding and combination of TSH and is necessary for the secretion of the mature hormone from the cell.

In the endoplasmatic reticulum, high mannose type oligosaccharides are transferred onto an asparagine residue with the recognition sequence asparagine-x-serine/threonine (where x is any amino acid except for proline, and other local structural restrictions that determine enzyme accessibility may apply). Subsequently, the oligosaccharides are partially trimmed by glycosidases, such as Mannosidase I and II (103). In the endoplasmatic reticulum, oligosaccharides are believed to stabilize a conformation that facilitates disulfide bond formation and are hence important for proper subunit folding. Moreover, the carbohydrates are part of a quality control program that ensures correct posttranslational processing. Thus, molecular chaperones have been identified that retain glycoproteins in the endoplasmatic reticulum until proper trimming of the carbohydrates has been accomplished. Only then are the nascent glycoproteins released to the next compartment/chaperone in the postranslational cascade (104). Incubation of mouse pituitary cells with tunicamycin, an inhibitor of oligosaccharide attachment during translation, led to aggregation and intracellular degradation of TSH (105). Similarly, folding kinetics and disulfide bond formation of the hCG-ß subunit lacking carbohydrate consensus sequences were delayed, leading to slow secretion and partial intracellular retention and degradation of the hCG ß-subunit (106, 107). Even the selective disruption of single glycosylation sites using site-directed mutagenesis caused significant decreases of hTSH secretion from transiently transfected Chinese hamster ovary (CHO) cells (62, 74).

In the Golgi apparatus, the carbohydrates are further trimmed and subsequently processed to mature complex oligosaccharides by sequential addition of carbohydrate residues catalyzed by various specific glycosyltransferases (5, 103). In this compartment, the oligosaccharides assume a critical role for intracellular translocation and direct the transport of the glycoproteins to specific cell compartments.

2. Intrinsic activity.
After secretion from the cell, the carbohydrates become important for the intrinsic activity, plasma half-life, and final in vivo activity of TSH. Earlier studies on gonadotropins and bovine TSH using chemical and enzymatic deglycosylation as well as hybrid studies had shown that the oligosaccharides, and predominantly those of the -subunit, are necessary for full in vitro activity of these hormones (108, 109, 110, 111, 112). In contrast to their critical role in receptor activation, they play a much less important role for high-affinity receptor binding. Thus there is a consensus that carbohydrates affect signal transduction predominantly at a post receptor-binding step. In fact, deglycosylated hCG acted as a competitive antagonist in certain in vitro assays (111, 112). By comparison, hTSH was shown to retain higher residual intrinsic activity upon deglycosylation (113, 114, 115).

In the absence of structural information on ligand-receptor complexes, the precise molecular basis of how carbohydrates contribute to TSH activity remains unclear. In this respect, it should be emphasized that because of the difficulty in obtaining high quality crystals of intact glycoproteins due to the microheterogeneity and relative flexibility of the oligosaccharide conformations, hCG was partially deglycosylated with hydrogen fluoride before crystallization (17). It is important to bear in mind that deglycosylated hCG acts as a competitive receptor antagonist, and the carbohydrates may be important to stabilize the active conformation of the hormone (see below). Therefore, it is not known how the structure of a fully agonistic hormone compares with the reported crystal structure. Recent structural analysis of the oligosaccharides of ¹³C, ¹⁵N-enriched recombinant hCG by nuclear magnetic resonance suggested that the -subunit carbohydrates do not interact with the protein backbone, but project outward into solution. Furthermore, the carbohydrates exist in an extended conformation with significant internal motion and have considerable conformational freedom (116).

Whereas one of the more recent models suggests that the carbohydrates appear to affect signal transduction primarily by their bulk (97, 98), other studies indicate that additional features, including specific carbohydrate-receptor interactions, may also be important. For example, sequential enzymatic deglycosylation of hTSH and its expression in glycosylation mutant cell lines, combined with site-directed mutagenesis, suggested that the terminal sugar residues, especially negatively charged sialic acid residues, critically affect the role of a carbohydrate side chain (74, 110, 117, 118, 119). Arguments in favor of a direct interaction of the carbohydrates with the receptor stem from the demonstration of oligosaccharide or glycopeptide binding to corpus luteum slices expressing the CG/LH receptor (120). In this respect, it has been pointed out that a segment of the extracellular domain of the LH/CG receptor shares considerable sequence identity with a domain of the Dolichos biflorus seed lectin as well as the soybean agglutinin (20). However, at least for TSH, an indirect mechanism involving a conformational change and/or aberrant ligand binding appears more likely as this lectin-like component identified in the hCG receptor is not present in the hTSH receptor (42). A possible role of the carbohydrates in maintaining glycoprotein hormones in a conformation able to activate the receptor is supported by the observation that certain antibodies can convert receptor-bound deglycosylated CG from an antagonist to an agonist (121). Several studies suggested that deglycosylated hormone does not elicit a signal because it binds to the receptor in an aberrant fashion. Thus, it was observed that deglycosylated hCG binds to different domains of the CG/LH receptor from native hCG (122). Further, there is evidence of differences in antibody accessibility of receptor-bound native and deglycosylated hCG (123).

The use of site-directed mutagenesis in combination with expression in glycosylation mutant cell lines (74), as well as the expression of hTSH in insect cells using a baculovirus system (124), have emphasized unique roles of individual side chains for hTSH activity. From these and studies using dimerization of heterologous subunits (110) and sequential enzymatic digestion (119), it appears that the roles of the terminal sialic acids as well as of individual oligosaccharides are different for the in vitro activity of hTSH compared with hCG and hFSH. This indicates that conserved structures within the context of a given ligand-receptor complex may contribute to signal transduction in different ways. In hCG, which is exclusively sialylated, sialic acid is required for full expression of in vitro activity (111, 125). In hFSH, which is predominantly sialylated, removal of sialic acid residues does not change in vitro activity (126). However, if hTSH or LH, which contain significant amounts of sulfated GalNAc termini (4) when produced in the pituitary thyrotroph, are expressed in CHO cells that produce exclusively sialylated termini, in vitro activity is attenuated (127, 128). Studies using site-directed mutagenesis of individual glycosylation recognition sites showed that the oligosaccharide at -asparagine⁵², but not the one at -asparagine⁷⁸, was necessary for hCG and hFSH action (81, 88, 89, 129). In contrast, the -asparagine⁵² chain and specifically its terminal sialic acid residues markedly attenuated TSH receptor binding and activation (74). As posttranslational modifications of carbohydrates regulate glycoprotein hormone activity in normal physiology (1, 5, 42, 43), modulation of terminal sialylation of the -asparagine⁵² oligosaccharide, which appears more heterogeneous than other side chains (56), may thus be important in regulating activity in a hormone-specific manner. Interestingly, deletion of this -asparagine⁵² side chain increased the weak inherent thyrotropic activity of hCG, opposite to the effect at its native receptor (74). Thus, as shown in Fig. 4 (and see below), the differential role of this oligosaccharide chain suggests that its composition, sialic acid/sulfate-dependent negative charge and possibly spatial orientation are critically important not only for signal transduction, but also for the specificity of ligand-receptor interaction, at least for that of hTSH.

View larger version (44K): [in this window] [in a new window]   Figure 4. Mechanistic model depicting cooperation of individual hTSH domains in receptor activation: role of the sialylated -asparagine52 oligosaccharide chain in glycoprotein hormone receptor activation. This mechanistic model of glycoprotein hormone receptor interaction highlights possible explanations for the differential role of the -asparagine52 oligosaccharide and other domains in hTSH receptor binding and activation. This simplified scheme does not reflect the dynamic, multistep interaction between ligand-receptor interaction in three-dimensional space, nor does it predict the nature of subsequent events after ligand receptor interaction that lead to G protein coupling. The model allows for reconciliation of the following findings from site-directed mutagenesis studies: deletion of the -asparagine52 chain decreases hCG activity at the CG/LH receptor (81), increases hTSH activity at the TSH receptor (74), decreases hCG activity at the TSH receptor (74), and decreases the activity of a hTSH/hCG seat belt chimera at the CG/LH receptor (78). Panel A 1 shows interaction of hCG with the CG/LH receptor. The carbohydrate at -asparagine52, which is important for hCG action, interacts with the receptor either directly or indirectly, e.g., by favorably influencing the spatially related -helix and the -carboxyl-terminus, also important for receptor activation. Also shown is a second receptor-binding domain, 11–20. Upon deletion of the -asparagine52 chain (A 2), activity is reduced (81) either because of the lack of direct interaction of this chain with the CG/LH receptor or by a conformational effect. Panel B 1 shows the interaction of hTSH with its receptor. In hTSH, the seat belt has a different conformation than in hCGß, which in turn leads to a different spatial orientation of the -asparagine52 chain. In contrast, the orientation of the distant 11–20 domain, which is equally important for the receptor binding of hCG and hTSH (50), is not affected by the seat belt. Note that the carbohydrate does not allow for optimal interaction of the hormone with the receptor. Rather, it restricts interaction of other activation domains, such as of the -helix as well as the -carboxyl-terminus with the receptor. Thus, deletion of -asparagine52 oligosaccharide (B 2) increases receptor activation (74). Panel C 1 shows how the hTSH/hCG seat belt chimera is able to activate the CG/LH receptor similar to hCG. Since the seat belt residues are identical to those of hCG, the orientation of the -asparagine52 chain is thus similar to its orientation in hCG and allows for efficient receptor interaction. Similar to hCG itself, the deletion of the chain reduces receptor activation (C 2; Ref. 74). Panel D 1 shows that the -asparagine52 oligosaccharide interferes with the ability of hCG to activate the TSH receptor. Deletion of the chain increases the ability of hCG to bind to the TSH receptor, similar to hTSH (D2, 74). Deletion of the -asparagine52 chain allows other functionally important domains of the composite receptor interaction site, e.g., -helix and/or the -carboxyl terminus to better interact with the receptor. This could be due to a direct role (lack of repulsion of negatively charged sialic residues from negatively charged receptor interface) or an indirect role of the carbohydrate (change of conformation of the composite domain). This repulsion may be more prominent at the TSH receptor, as the hCG receptor possesses an overall higher number of positive charges in its extracellular domain. Thus, the -asparagine52 chain, in addition to attenuating activity of hTSH, acts as a negative specificity determinant, restricting the inappropriate interaction of hCG with the TSH receptor. H, -helix, C, a carboxyl terminus; 52, -asparagine52 oligosaccharide chain; S, sialic acid; SB, seat belt. The relative degree of receptor activation is indicated by the arrows.

In analogy to what was observed for intrinsic activity, the specific carbohydrate structure at different glycosylation sites may affect hormonal clearance to a different degree. It was shown that the peripherally located single carbohydrate chain of the TSH ß-subunit appears to be the most important in determining the MCR of hTSH (110), whereas the -asparagine⁷⁸ chain is more critical than the -asparagine⁵² chain in this respect (74). Similar findings for the relative roles of individual carbohydrates for clearance have also been reported for hFSH (129). An important lesson to be learned from such findings is the lack of direct correlation between the effects of carbohydrates on in vitro and in vivo activities of glycoproteins. This fact is a consequence of the fundamental difference between a hormone-specific interaction with the target organ receptor and carbohydrate-dependent clearance mechanisms determining the circulatory half-life of a given glycoprotein. Such studies highlight the difficulties of translating results obtained using in vitro systems into whole organism physiology and illustrate the importance of determining the activity of glycoprotein hormone analogs in adequate animal models.

	III. Current Understanding of TSH/Glycoprotein Hormone Action

Top Abstract I. Introduction II. Structure-Function... III. Current Understanding of... IV. Physiological and... V. Evolutionary Considerations VI. Strategies in the... VII. Nonclassic Actions of... VIII. Perspectives on Structure... References

 
A. Structural considerations
Proposed models of glycoprotein hormone action have greatly benefitted from the structure of hCG (17, 18), which is also suitable for testing these models prospectively. However, due to the absence of structural data on glycoprotein receptors, the precise molecular mechanisms of glycoprotein hormone signal transduction remain largely unknown. Despite the recent availability of relatively large amounts of pure protein using systems such as baculovirus-based insect cell expression, high resolution structure of any G protein-coupled receptor has not yet been obtained. This is in part due to the difficulties in obtaining crystals that include the transmembrane domain. Furthermore, the large carbohydrate component of the extracellular domain of the glycoprotein hormone receptors makes this task even more difficult. Ultimately, crystal structures of receptor-ligand complexes should directly identify domains of the hormone that contact the receptor. Understanding in molecular detail how glycoprotein hormones interact with their receptor and how the signal is subsequently transmitted to the G proteins will require three-dimensional structures of all the components as well as identification of their conformational changes upon activation.

Even though cocrystallization can map the topography of complementary surfaces of ligand and receptor, functional analysis of such contacts will be necessary. For example, crystallization of the GH-receptor complex showed a large ligand receptor interface. However, solely systematic site-directed mutagenesis of GH revealed that of the residues that contacted the receptor, only a small set was actually important in maintaining high-affinity interaction. The functional relevance of individual residues did not correlate with the extent to which their side chains were buried at the interface of the crystal complex and was therefore not predictable from the structure (67, 68). Conversely, due to the dynamic nature of ligand-receptor interactions, deletion of protein domains that do not contact the target in cocrystallization studies can, in certain instances, still be important for signal induction (134). In fact, there are many examples in the literature showing that protein functions are influenced by residues far from active sites (135).

In this context, it should be emphasized that glycoprotein hormones range among the largest (28–34 kDa) and most complex naturally occurring ligands. In addition, their receptors are notable for a large extracellular domain that is unusual for G protein-coupled receptors. This extracellular part is encoded by several exons (21, 28, 30, 31, 136). The presence of LRR in their extracellular domains, which appears unique among G protein-coupled receptors, has led to the realization that the glycoprotein hormone receptors belong to the superfamily of LRR proteins (137). This family encompasses a vast variety of molecules with diverse functions and cellular localizations, the common characteristic being that they are involved in protein-protein interaction. Despite their diverse functions, conservation of the LRR indicates similar roles of such modules for these proteins (137). Cocrystallization of the LRR-containing ribonuclease inhibitor complexed with its ligand (19) has inspired recent modeling of the hTSH (138) and hCG receptor (98, 139), and these models have recently begun to be tested using site-directed mutagenesis of the receptor (140). Crystallization of the ribonuclease inhibitor revealed a nonglobular shape with solvent-exposed parallel ß-sheets and flexibility of the module, allowing elastic alteration of the entire structure. These aspects support the suitability of LRR for protein-protein interactions. Moreover, the concave surface formed by the repeats allowed for a large interface with the ribonuclease (19, 137). Interestingly, this is compatible with results from epitope mapping, showing that most of the surface of glycoprotein hormones is masked upon interaction with their receptors (98, 123, 141). Thus, these findings predict a similarly large ligand-receptor interface for glycoprotein hormones, which is also supported by the identification of multiple functionally important regions on both subunits. In fact, it was speculated earlier that the extracellular domain of glycoprotein hormone receptors represents a rather flexible entity that wraps around the ligand in a "process-like adaptive manner" (123).

B. Hormone-receptor interaction
There is no general consensus of the specific mechanisms by which the glycoprotein hormone docks into its receptor. It is generally accepted that the ß-heterodimer is required for glycoprotein hormone activity, and individual subunits do not possess significant activity at the glycoprotein hormone receptors (3). In fact, multiple contact points of both -subunit and ß-subunit with the receptor, perhaps in a stepwise fashion, appear necessary to induce a conformational change of the receptor, favoring receptor G-protein coupling and subsequent second messenger generation (60). It appears likely that the initial interaction involves specific high-affinity binding of the hormone to the LRR-containing extracellular domain of the receptor. This initial binding event may control specificity by negative determinants that restrict heterologous ligand-receptor interaction (57, 95). Whether the extracellular domain of the TSH receptor by itself is sufficient for high-affinity ligand binding has not been unequivocally established (28, 30, 31, 142). In addition to interactions with the extracellular domain, secondary contacts between common, possibly -subunit, domains with the transmembrane portion of the receptor may initiate the signal by analogy to G protein-coupled receptor activation by small ligands, such as for the adrenergic receptors (143). However, it is not known how even parts of the bulky glycoprotein hormones could be accommodated in such a hypothetical pocket. In this respect, modeling of the transmembrane domain of the glycoprotein hormone receptor indicated that, in contrast to the tight hydrophobic pocket of adrenergic receptors, the glycoprotein hormone receptor domain may form a deeper, yet broader, hydrophilic groove (144).

Binding of glycoprotein hormones to additional receptor domains was supported by the identification of a direct interaction between a counterionic pair of residues of the -carboxyl terminus of hCG and the first exoloop of the CG/LH receptor (71). Subsequently, specific binding of an -carboxyl-terminal peptide to the CG/LH receptor was demonstrated (145). Further, binding of hCG to the extracellular domain of the receptor unmasked an immunoreactive site on the -subunit, which was not accessible if the hormone bound to the full-length receptor (141), supporting the notion that some -subunit regions may contact the carboxyl-terminal half of the receptor. Moreover, coexpression of the extracellular domain of the CG/LH receptor with the transmembrane domain restored efficient hCG-mediated signal transduction (146). However, in a recently proposed model of hCG action, binding to the extracellular domain alone could account for G protein activation without the need for secondary contact points (98). Even though it was reported that hCG can bind with low affinity to, and activate a truncated form of the CG/LH receptor lacking the extracellular domain (147), this was not observed with similar studies of the TSH receptor (148). In fact, several N-terminally truncated TSH receptor constructs were not stimulated by either TSH or hCG (148). These findings again underscore the need for structural data on hormone-receptor complexes to understand potential causes for such discrepancies. In any case, ligand binding is believed to modulate interactions between the transmembrane helices, effecting conformational changes in the intracellular loops and thus altering G protein coupling (149), ultimately activating the second messenger systems (150). Figure 5 shows a potential orientation of hTSH within the hormone-receptor complex, by analogy to models that have been proposed by others (30, 98, 138, 139, 140).

View larger version (33K): [in this window] [in a new window]   Figure 5. Structural model of the hTSH-hTSH receptor complex. This schematic drawing is intended to reflect a potential initial interaction of hTSH with its receptor. The receptor is depicted in accordance to models of the hTSH receptor (138) based on the LRR-containing ribonuclease inhibitor (19). Actual spatial relationships will require cocrystallization of hormone receptor complexes. The two parallel ß-hairpin loops of the -subunit (L1, L3) face toward the receptor and may participate in the interaction of the hormone with the extracellular loops of the receptor transmembrane domain (139). The equivalent loops of the TSH ß-subunit (ßL1, ßL3) are shown projecting to a proposed binding site within the concave surface of the nine LRR of the extracellular domain of the TSH receptor (138). However, site-directed mutagenesis studies have implicated additional domains to contact the receptor. It is conceivable that conformational changes in receptor and/or ligand upon high-affinity binding would allow for secondary interactions.

C. Cooperation of individual hTSH domains in receptor activation
This paragraph attempts to integrate the results of individual site-directed mutagenesis studies into a model highlighting several aspects of hTSH action. A hypothesis of how individual hTSH domains may interplay in receptor activation is summarized in Fig. 4. As stated earlier, the importance of several highly conserved domains in the common -subunit for the signal transduction of all glycoprotein hormones emphasizes that these hormones elicit their biological responses in a similar fashion. Yet, as described above, the -asparagine⁵² oligosaccharide, and in particular its negatively charged sialic acid moieties, play an opposite role for hTSH activity compared with hCG or hFSH (74). Further, as discussed above, the relative contributions of the -helix and the -carboxyl-terminus to signal transduction are, at least in part, different for each glycoprotein hormone (64, 69, 72, 85, 86, 93). This implies that these -subunit activity domains may, to a certain degree, function in a ß-subunit-dependent fashion.

As mentioned earlier, chimeric studies have shown that the ß-subunit seat belt appears to direct, at least in part, glycoprotein hormone specificity (78, 83, 90, 95). Accordingly, the seat belt may achieve this by influencing common -subunit domains important for signal transduction, such as the -asparagine⁵² oligosaccharide, to function in a hormone-dependent fashion. This was shown by deleting the -asparagine⁵² oligosaccharide in a hTSH chimera in which the native seat belt sequence had been replaced with the corresponding residues of hCG (M. Grossmann, M. W. Szkudlinski, and B. D. Weintraub, unpublished data). This oligosaccharide was chosen because of its differential effect on glycoprotein hormone activity: absence of this oligosaccharide, if sialylated, decreased hCG activity (81), but increased hTSH activity (74). Remarkably, the hCG-like activity of this hTSH/hCG seat belt chimera decreased upon deletion of the asparagine⁵² oligosaccharide. Thus, the function of this domain in the chimera was similar to its function in hCG, but different from that in hTSH. This suggests that the seat belt may indirectly modulate hormonal specificity by orienting -subunit domains that are in close proximity (see Fig. 4). This is consistent with the hormone-dependent differences in the contribution of these domains for receptor activation.

In contrast, an 11- to 20--subunit domain engineered for increased binding, located within the ß-hairpin L1 loop, appears to be important for all glycoprotein hormones (50). This relative absence of specificity of the engineered 11–20 domain may be associated with its distance from the seat belt and other regions of the ß-subunit. In a recent model of hCG bound to its receptor, the 11–20 region may contact the transmembrane portion of the receptor, further supporting its possible direct involvement in receptor binding (139). Accordingly, the potential orientation of hTSH within the hormone-receptor complex is depicted in Fig. 5.

	IV. Physiological and Pathophysiological Implications

Top Abstract I. Introduction II. Structure-Function... III. Current Understanding of... IV. Physiological and... V. Evolutionary Considerations VI. Strategies in the... VII. Nonclassic Actions of... VIII. Perspectives on Structure... References

 
A. Carbohydrate heterogeneity
1. Regulation of physiological microheterogeneity.
Different degrees in the terminal processing of oligosaccharides give rise to a mixture of circulating glycoforms, which in turn are responsible for the physiological microheterogeneity of the glycoprotein hormones. The processed hormone N-linked oligosaccharides are typically bi- or multiantennary structures displaying notable hormone-dependent differences in their terminal residues (Fig. 1). Pituitary TSH and LH are unique in that they contain predominantly sulfated oligosaccharides, due to the presence of GalNAc-transferase and GalNAc-4-sulfotransferase in the pituitary thyrotrophs and luteotrophs, and terminate mostly in SO₄-4GalNAcß1–4GlcNAcß1–2Man. On the other hand, CG and FSH, like almost all other serum glycoproteins, terminate in Sia2–3(6)Galß1–4GlcNAcß1–2Man (4, 5). Such selective glycosylation may have primarily evolved as a means to preserve the pulsatile pattern of TSH and LH levels in the circulation and thus avoid receptor desensitization of the target organ. In fact, a separate hepatic receptor specific for oligosaccharides terminating with sulfated GalNAc residues has been implicated in the rapid clearance of LH and TSH (156). By contrast, terminal sialylation enables the glycoprotein hormone to escape such specific receptor-mediated hepatic clearance mechanisms, and the kidney becomes the major organ of (less efficient) clearance. For example, rhTSH is produced in Chinese hamster ovary cells that lack GalNAc-transferase and GalNAc-4-sulfotransferase and TSH produced in these cells terminates exclusively in sialic acids (14, 102, 127). This appears to be the main reason why its circulatory half-life is prolonged compared with its predominantly sulfated pituitary counterpart (110, 127, 130).

The concept of how the carbohydrates affect clearance and hence in vivo bioactivity is also exemplified by hTSH produced in insect cells using a baculovirus system. Insect cell-expressed hTSH, which lacks sialic acids but contains predominantly high-mannose residues, was cleared very rapidly compared with rhTSH, presumably via the hepatic mannose receptor (157, 158), and had a lower bioactivity than sialylated rhTSH (124). Such observations emphasize that the main physiological role of carbohydrate moieties and their terminal residues lies in the differential targeting and clearance of the hormones. Therefore, the glycosylation state and specifically the degree of terminal sialylation have a powerful im