HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 35, 27457-27465, September 1, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/35/27457    most recent M003707200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted 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 Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Leitolf, H. Articles by Szkudlinski, M. W. Search for Related Content PubMed PubMed Citation Articles by Leitolf, H. Articles by Szkudlinski, M. W. Social Bookmarking                      What's this?

Bioengineering of Human Thyrotropin Superactive Analogs by Site-directed "Lysine-scanning" Mutagenesis

COOPERATIVE EFFECTS BETWEEN PERIPHERAL LOOPS^*

Holger Leitolf, Kim Phuong T. Tong, Mathis Grossmann^§, Bruce D. Weintraub, and Mariusz W. Szkudlinski^¶

From the Laboratory of Molecular Endocrinology, Division of Endocrinology, Diabetes, and Nutrition, University of Maryland School of Medicine and Division of Basic Science, Institute of Human Virology, Medical Biotechnology Center, University of Maryland Biotechnology Institute, Baltimore, Maryland 21201

Received for publication, May 2, 2000, and in revised form, June 14, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We have previously engineered the first superactive analogs of human thyrotropin (hTSH) by using a novel design strategy. In this study, we have applied homology comparisons focusing on the L3 loop of the common -subunit of human glycoprotein hormones. Seven highly variable amino acid residues were identified, and charge-scanning mutagenesis revealed three previously unrecognized modification permissive domains and four gain-of-function lysine substitutions. Such gain-of-function mutations were hormone- and receptor-specific and dependent on location and basic charge. Cooperativity of individual substitutions was established in double and triple lysine mutants. In combinations of the most potent L3 loop analog with two previously characterized loop analogs, a higher degree of cooperativity for the L3 loop analog compared with both the L1 loop analog and the hTSH-L3 loop analog was observed. We demonstrated that spatially distinct regions of the common -subunit contribute differentially to the interaction of hTSH with its receptor and that combinations of two modified loops on the same and on opposite sides of the hTSH molecule display similar increases in in vitro biopotency. In addition, combination of all three superactive loops showed cooperativity in receptor binding and activation resulting in the most potent hTSH superactive analog described to date.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Human thyrotropin (hTSH)¹ is a member of a family of structurally related pituitary and placental human glycoprotein hormones that also includes the gonadotropins human chorionic gonadotropin (hCG), human-luteinizing hormone (hLH), and human follicle-stimulating hormone. Each hormone consists of a heterodimer of an -subunit and a noncovalently linked -subunit, and the - heterodimer formation is required for full biological activity. Whereas the primary structure of the -subunit is common to all four glycoprotein hormones, the -subunit is unique and responsible for hormone specificity (1). More recent studies have suggested a potential role for the individual free subunits in pregnancy and tumor growth (2-4).

The recent elucidation of the crystal structure of chemically deglycosylated hCG (5, 6) has revealed that glycoprotein hormones share structural features with a large number of growth factors such as platelet-derived growth factor, neurotrophin (nerve growth factor), and transforming growth factor  (7) and are therefore considered members of the superfamily of cystine knot growth factors. Within each glycoprotein subunit, a central cystine knot motif consisting of three disulfide bonds can be identified together with two peripheral -hairpin loops (L1 and L3) on one side and one long loop (L2) on the other side. Despite a large diversity of membrane receptors within the superfamily of cystine knot growth factors, there are marked similarities in the functional role of the peripheral loop segments despite a limited (approximately 10-15%) amino acid sequence homology between the glycoprotein hormones and other members of the superfamily. Specifically, clusters of positively charged residues within the loops have recently been implicated as important receptor binding domains for members of the neurotrophin and the transforming growth factor  family (8, 9).

Recent structure-function studies of human thyrotropin from our laboratory (for reviews see Refs. 10-12) and others have characterized multiple functionally important domains in both subunits (Fig. 1). We distinguished "specificity determining" domains in the hTSH "seat belt" (hTSH-(88-105) (13)), "modification non-permissive" (33-38 (14), -helix 40-46 (14, 15), glycosylation site 52 (16, 17), -carboxyl terminus 88-92 (18-20)), and "modification permissive" domains in two peripheral -hairpin loops (11-20 (21), hTSH-(58-69) (22)). The mutational analysis of modification permissive domains defined as regions that tolerate the introduction of nonconservative amino acid changes into hTSH without compromising hormone synthesis (10) has been particularly useful in recent structure-function studies.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Schematic drawing of hTSH. Schematic drawing of hTSH depicts the -subunit in gray and the -subunit in black. Functionally critical domains are marked directly within the line drawing (for more details, see text). The peripheral -hairpin loops are indicated as L1 and L3 in the -subunit and as L1 and L3 in the -subunit. Modification permissive domains are the regions between 11 and 20 in the L1 loop and 58 and 69 in the L3 loop. The region between 64 and 81 was the target region for site-directed mutagenesis in this study. Open circles represent the positions of individual amino acid residues that were substituted with basic residues (13, 14, 16, and 20 in L1; 64, 66, 73, and 81 in L3; 58, 63, and 69 in hTSHL3). Modified from Ref. 11.

Based on evolutionary considerations, amino acid structure comparisons between species and between different glycoprotein hormones, as well as computer-assisted homology modeling, we were able to define a design strategy for bioengineering superactive analogs ("superagonists") of human thyrotropin with major increases in both receptor binding affinity as well as signal transduction which were the first to be described for any glycoprotein hormone or any member of the cystine knot growth factor superfamily and which by far exceeded the increases in biopotency obtained for other protein ligands such as growth hormone (23) or interleukin-6 (24) using empirical design approaches. In contrast to the characterization of loss-of-function mutations, which could result in distant conformational effects leading to an indirect alteration of hormone function, the analysis of gain-of-function mutations is expected to be much more specific with respect to their location and the underlying mechanism and therefore more informative in glycoprotein structure-function studies.

In the present study, we have extended our design strategy for bioengineering superactive analogs of human thyrotropin to the second, as yet uncharacterized peripheral -hairpin loop of the common -subunit of human glycoprotein hormones (i.e. the L3 loop, Fig. 1) under the assumption that introduction of positively charged, basic residues into all four peripheral -hairpin loops leads to enhanced electrostatic interactions with the hTSH receptor resulting in an increase in hormone binding affinity and in in vitro biopotency. The aims of the present study were as follows: 1) to identify modification permissive domains within the common L3 loop of human glycoprotein hormones; 2) to identify amino acid residues where substitutions to basic residues lead to the generation of superactive hTSH analogs; 3) to combine single L3 loop mutations to study cooperative effects between individual residues within this loop; and 4) to study cooperative effects between the L3 loop and other peripheral -hairpin loops.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Molecular biology reagents were purchased from Life Technologies, Inc., Roche Molecular Biochemicals, and New England Biolabs, Inc. (Beverly, MA). Cell culture media, media supplements, and the LipofectAMINE reagent were supplied by Life Technologies, Inc. The full-length human -cDNA subcloned into the BamHI-XhoI restriction enzyme sites of the pcDNAI/Neo expression vector (Invitrogen, San Diego) was obtained from T. H. Ji (University of Wyoming, Laramie). The hTSH -minigene in the pLBCMV vector was previously engineered in our laboratory (25). Chinese hamster ovary (CHO) cell clones stably expressing the hTSH wild-type receptor (JP09 and JP26) at different receptor densities were kindly provided by G. Vassart (Free University of Brussels, Brussels, Belgium). In addition, another CHO cell clone (hTSHR-D1) has been established using a construct supplied by B. Rapoport (Veterans Affairs Medical Center, San Francisco, CA) (26), in which amino acid residues 317-366, unique for the hTSH receptor, have been deleted. This construct was used to study the role of TSH receptor-specific structures in the observed increase in hTSH analog bioactivity. cAMP antibody was a gift from J. L. Vaitukaitis (National Institutes of Health, Bethesda). ¹²⁵I-cAMP and ¹²⁵I-hTSH were purchased from Covance Laboratories Inc. (Vienna, VA).

Site-directed Mutagenesis-- Mutagenesis of the human -cDNA was accomplished by the PCR-based megaprimer method of site-directed mutagenesis with two consecutive PCR cycles as described previously (27), using Vent DNA Polymerase (New England Biolabs, Inc., Beverly, MA) and oligonucleotide primers synthesized by Lofstrand Laboratories Ltd. (Gaithersburg, MD) with the wild-type human -cDNA in the pcDNAI/Neo vector as template. For generation of multiple mutations within the L3 loop, various previously mutated -cDNAs in the pcDNAI/Neo or pcDNA 3 expression vector (Invitrogen) with single or double amino acid substitutions were used as templates for PCR. Flanking oligonucleotide primers and amplification conditions had been optimized in the course of our previous studies (21). Mutagenizing primers were designed to introduce the following codon changes: S64K (TCA-AAA), Y65K (TAT-AAA), N66K (AAC-AAG), R67K (AGG-AAG), G73K (GGT-AAA), F74K (TTC-AAA), and A81K (GCG-AAG). Additional primers were used to introduce negatively charged residues as follows: S64E (TCA-GAG), N66E (AAC-GAG), G73E (GGT-GAG), G73D (GGT-GAC), and A81E (GCG-GAG). After simultaneous digestion of the product of the second PCR with BamHI and XhoI and isolation of the resulting fragment, ligation of this fragment into the wild-type human -cDNA in the pcDNA3 vector with the BamHI-XhoI fragment excised was performed. Combination of the L1 loop analog with the L3 loop analog was achieved by single restriction enzyme digest of the L3 loop analog with XbaI and consecutive ligation of the isolated fragment into the L1 loop analog in the pcDNAI/Neo expression vector with the XbaI fragment excised. Chemically competent Escherichia coli cells (DH5 Subcloning Efficient Cells, Life Technologies, Inc., or MC1061/p3, Ultracomp Transformation Kit, Invitrogen) were transformed using a heat-shock protocol. Plasmids were isolated using the Qiaprep 8 Miniprep Kit and Qiaprep 8 Turbo Miniprep Kit for multiple preparations of small DNA quantities and the Qiagen Plasmid Maxi Kit (Qiagen Inc., Santa Clarita, CA) for purifications of large DNA quantities. Correct introduction of single or multiple mutations was verified by bidirectional single-stranded DNA sequencing performed by the Biopolymer Laboratory (University of Maryland School of Medicine, Baltimore).

Transient Expression of Recombinant Hormones-- CHO-K1 cells (ATCC, Manassas, VA) were maintained in Ham's F-12 medium supplemented with 10% fetal bovine serum, glutamine (4 mM), penicillin (50 units/ml), and streptomycin (50 µg/ml) at 37 °C and 5% CO₂. Cells were transiently cotransfected in 60-mm dishes after reaching about 80% confluency with wild-type or mutant -cDNA in the pcDNA3 vector for single and multiple mutants within the L3 loop and in the pcDNAI/Neo vector for the L1-L3 analog combination, respectively, and wild-type or mutant hTSH -minigene in the pLBCMV vector using a liposome-mediated approach (LipofectAMINE reagent, Life Technologies, Inc.). After 16 h, the transfected cells were transferred to serum-free medium (CHO-SFM II, Life Technologies, Inc.). In a typical experiment, we used about 10-15 60-mm dishes per mutant and added 6 ml of serum-free medium per dish. After an additional 72 h, cell culture media were harvested, resulting in a total volume of 60-90 ml unconcentrated medium per mutant, which was then first clarified by centrifugation. Second, concentration of this volume by about 50-fold was achieved using Centriprep 10 concentrators (Amicon, Inc., Beverly, MA), resulting in a final volume of about 1.2-1.8 ml, which was then divided into 0.2-ml aliquots, which were stored at 70 °C. These concentrated, unpurified media preparations were then used for all further studies as outlined below. Aliquots were thawed only once before each respective assay.

Immunoassays-- Wild-type recombinant hTSH and mutant hTSH analogs were quantitated with four different immunoassays utilizing different monoclonal and polyclonal antibodies. Three third generation "sandwich" assay systems (Nichols Institute Diagnostics, San Juan Capistrano, CA; ICN Pharmaceuticals, Inc., Costa Mesa, CA; DiaSorin Inc., Stillwater, MN) recognizing hTSH immunochemiluminometrically or immunoradiometrically by forming a bridge between a solid-phase coupled monoclonal antibody and a second polyclonal acridinium ester- or ¹²⁵I-labeled antibody were used. In addition, we performed a polyclonal hTSH radioimmunoassay using antibody (NIDDK anti-hTSH-3) directed against the hTSH -subunit kindly supplied by the National Hormone and Pituitary Program (Torrance, CA) and using recombinant hTSH provided by the Genzyme Corp. (Framingham, MA) as standard (28).

Receptor Binding Assays-- The binding activity of wild-type or mutant hTSH was studied by their ability to displace ¹²⁵I-radiolabeled bovine TSH (¹²⁵I-bTSH) from a solubilized porcine thyroid membrane preparation (Kronus, San Clemente, CA) by serial dilutions of wild-type or mutant hTSH using a buffer containing 0.15% NaCl according to the protocol supplied by the manufacturer.

Biological Activity Assays-- CHO cells stably expressing the human wild-type thyrotropin receptor (JP09 and JP26) or the thyrotropin receptor construct with deletion of amino acid residues 317-366 (hTSHR-D1) were grown in 96-well tissue culture plates in Ham's F-12 medium supplemented with 5% fetal bovine serum, glutamine (4 mM), penicillin (50 units/ml), and streptomycin (50 µg/ml) at 37 °C and 5% CO₂. After reaching confluency, cells were incubated for 2 h with serial dilutions of wild-type or mutant hTSH in a modified Krebs-Ringer buffer under low salt conditions where sucrose was added to maintain isotonicity supplemented with 1 mM 3-isobutyl-1-methylxanthine and 0.1% bovine serum albumin (Sigma). The amount of cAMP released into the buffer was measured by radioimmunoassay as described previously (29).

In vivo bioactivity was assessed by a previously validated bioassay (30). Male albino Swiss Crl:CF-1 mice were pretreated over 4 days with drinking water supplemented with 0.3 µg triiodothyronine/ml ad libitum. This has been shown (30) to stably and reproducibly suppress endogenous thyrotropin secretion, thereby providing indispensable preconditions for the following study by practically eliminating preexisting interanimal variability. Stimulation of total thyroxine (TT4) levels over this suppressed base-line level was chosen as end point. Six hours after an intraperitoneal injection of hTSH wild-type and selected hTSH analogs, blood was collected by retroorbital sinus puncture, and the animals were sacrificed. Total thyroxine (TT4) serum levels were measured by radioimmunoassay (DiaSorin Inc., Stillwater, MN).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Design of Single and Combined Mutants-- Amino acid sequence alignment of the primary amino acid structure of the common L3 loop of human glycoprotein hormones between amino acid residues 64 and 81 was used to select potential target sites for site-directed mutagenesis (Table IA). The 64-81 region appeared to be predominantly surface-exposed and distant from the -subunit based on an hTSH homology model (21) constructed using crystallographic coordinates of hCG (5, 6).

View this table: [in this window] [in a new window]   Table I Amino acid sequence alignment of the common L3 loop (Table IA) of human glycoprotein hormones between 64-81; design of single (Table IB) and multiple (Table IC) lysine substitutions

In contrast to the 11-20 region in the common L1 loop of human glycoprotein hormones (21), the 64-81 region in the common L3 loop does not contain clusters of conserved positively charged amino acid residues. We therefore distinguished residues with lower (68-72, 75-80) versus higher (64-67, 73-74, 81) amino acid structure variability during evolution. For their expected higher tolerance toward introduction of nonconservative amino acid changes, we selected the latter as target sites for site-directed mutagenesis. According to our hypothesis that introduction of nonconservative basic residues at these sites may generate "gain-of-function" analogs, we chose to substitute the wild-type amino acid residues at each of the above positions with single lysine residues. This "lysine-scanning" approach resulted in seven single lysine substitutions as follows: S64K, Y65K, N66K, R67K, G73K, F74K, and A81K (Table IB).

In a second step, four out of seven mutations were then combined to generate four double, four triple, and one quadruple mutant (Table IC).

Quantitation and Relative Expression Levels-- Wild-type thyrotropin and mutant human thyrotropin analogs were quantitated with four monoclonal and polyclonal immunoassays. Based on the most sensitive immunochemiluminometric assay, the yield when following the transient transfection procedure outlined above was about 2000 ng of hTSH/ml of concentrated, unpurified medium for the generation of hTSH wild-type. Relative expression levels as compared with wild-type ranged between 28 and 128% for mutants with single substitutions and between 23 and 49% for mutants with multiple substitutions within the L3 loop. A comparison of expression levels for single, double, and triple lysine substitutions is given in Table II, and expression levels of mutant hTSH analogs with multiple substitutions in different loops are shown in Table III. Immunological recognition of the mutants was highly comparable between all four assays used, thereby ensuring accurate quantitation and providing preconditions for reliable calculation of relative increases in biopotency in the following studies. In the course of these (data not shown) as well as our previous studies, we have determined several hTSH amino acid residues critical for binding of the respective monoclonal antibodies used in these assays, and we found them to be distinct and not identical between the assays chosen for this study.

Analogs with Single Mutations-- Introduction of single lysine substitutions at positions 64-67, 73-74, and 81 within the common L3 loop generated seven hTSH analogs all of which were immunologically recognized and biologically active. Therefore, the introduction of nonconservative amino acid residues did not result in a significant alteration of the overall conformation of the hormone or in a significant impairment of its biological activity, thereby defining these regions as modification permissive domains.

Four out of seven lysine substitutions generated hTSH superactive analogs, as they showed a 2-6-fold decrease in the hTSH concentration required for half-maximal stimulation when testing cAMP production in the JP09 (Fig. 2A) and JP26 (data not shown) cell lines. Levels of maximal cAMP stimulation were comparable between hTSH wild-type and hTSH analogs. The increase in in vitro bioactivity was paralleled by a decrease in the hTSH concentration required for a 50 or 75% displacement of bound radiolabeled bTSH in a porcine thyroid membrane preparation (Fig. 2B). The four mutants showed increases in biopotency in the following order: G73K > N66K > S64K = A81K. Lysine substitutions at positions 65, 67, and 74 did not result in an increase in bioactivity (data not shown), thereby establishing site-specific effects for the gain-of-function mutations.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   A and B, single substitutions in the L3 loop. In vitro bioactivity was assessed by cAMP production in the JP09 cell line (A) and receptor binding affinity (B) of L3 loop analogs with single lysine substitutions at positions 64, 66, 73, and 81. Each data point represents the mean ± S.E. of triplicate determinations in a representative experiment repeated two times. WT, wild type.

Analogs with Multiple Mutations-- Fig. 3, A and B, demonstrates the increase in in vitro bioactivity when lysine substitutions were combined to generate analogs with multiple mutations. Fig. 3A shows the gradual decrease in the hTSH concentration required for half-maximal stimulation in the JP09 cell line for the single mutant G73K, the double mutant (S64K/G73K), and the triple mutant (S64K/N66K/G73K), respectively. Fig. 3B exemplifies the same in the JP26 cell line for A81K, the double mutant (G73K/A81K), and the triple mutant (N66K/G73K/A81K). In contrast to the single lysine substitutions, double and triple lysine substitutions in the L3 loop did increase maximal cAMP stimulation levels that ranged between 120 and 130% of those found for the respective wild-type preparation.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   A and B, combined substitutions in the L3 loop. In vitro bioactivity was assessed by cAMP production in the JP09 (A) and JP26 (B) cell lines. The gradual increase in in vitro bioactivity of L3 loop mutants corresponds to an increase in the number of lysine residues from single to double and triple substitutions. Each data point represents the mean ± S.E. of triplicate determinations in a representative experiment repeated two times. WT, wild type.

Interestingly, we could observe maximum increases in bioactivity with two triple lysine substitutions as the addition of A81K to (S64K/N66K/G73K) and S64K to (N66K/G73K/A81K), respectively, did not result in any further gain-of-function as shown in Fig. 4A. In addition, the triple arginine (N66R/G73R/A81R) displayed equal potency compared with the mutant with three lysine residues, whereas the triple histidine mutant (N66H/G73H/A81H) was significantly less bioactive, thereby providing evidence for residue and charge specificity after introduction of nonconservative amino acid residues (Fig. 4B).

View larger version (16K): [in this window] [in a new window]   Fig. 4.   A-D, combined substitutions in the L3 loop. In vitro bioactivity of triple and quadruple lysine substitutions within the L3 loop (A) and comparison of triple lysine, arginine, or histidine (B) substitutions are assessed by cAMP production in the JP09 cell line. Each data point represents the mean ± S.E. of triplicate determinations. In vitro bioactivity of selected single and triple acidic (glutamic acid, aspartic acid) amino acid substitutions within the L3 loop are assessed by cAMP production in the JP09 cell line (C). In vitro bioactivity of single and triple lysine substitutions within the L3 loop are assessed by cAMP stimulation in the hTSHR-D1 cell line (D). WT, wild type.

In addition to the assessment of substitutions to basic residues within the L3 loop as outlined above, we also characterized the effects of selected single substitutions to acidic, negatively charged amino acid residues (glutamic acid, aspartic acid) at positions 64, 66, 73, and 81 as well as one combined mutant with three glutamic acid residues (Fig. 4C). We observed either no change or a decrease in bioactivity as compared with the cAMP stimulation elicited by the wild-type hormone for the single substitutions S64E, N66E, G73D, and A81E as well as for the triple mutant (S64E/N66E/G73E), thereby supporting our concept of a basic charge-dependent enhancement of TSH biopotency. However, a G73E substitution, in contrast to our above findings at positions 64, 66, and 81 for glutamic acid and at position 73 for aspartic acid, resulted in a 2-fold lower increase in biopotency than the G73K mutation.

When testing the four single superactive lysine substitutions in the L3 loop in a CHO cell line stably expressing a hTSH receptor construct (26) in which amino acid residues 317-366, unique for the hTSH receptor and not found in the hLH/CG receptor, had been deleted, we observed no loss of superactive potency (Fig. 4D), thereby indicating that the 317-366 region of the thyrotropin receptor does not seem to be involved in mediating superactive effects of the four single lysine substitutions within the L3 loop. Assessment of cAMP stimulation by the two most potent L3 loop analogs with triple lysine substitutions also failed to reveal a "loss-of-superagonism" (Fig. 4D).

Cooperative Effects between Peripheral Loops-- We chose one of the two previously characterized most potent triple lysine substitutions ((N66K/G73K/A81K), i.e. "L3(3K)") for further comparison and combination with two other superactive peripheral -hairpin loop analogs that had been bioengineered in our laboratory (Fig. 1). We had previously generated an L1 loop analog with four lysine substitutions at positions 13, 14, 16, and 20 (i.e. " L1(4K)" (21)) and an hTSH-L3 loop analog with three lysine substitutions at positions hTSH-58, hTSH-63, and hTSH-69 (i.e. " L3(3R)" (22)).

Fig. 5A shows a comparison of the in vitro bioactivity for the L1(4K) and the L3(3K) analog after co-expression with either hTSH- wild-type or the hTSH-L3(3R) analog. We observed a higher degree of cooperative effects for the L3 loop analog versus the L1 loop analog with the hTSH-L3 loop analog, because L3(3K) alone increased cAMP production in the JP09 cell line by about 10-20-fold and was therefore significantly less potent than L1(4K) which showed an increase in in vitro bioactivity of about 30-40-fold. However, after combination with the hTSH-L3 loop analog, the increase in biopotency observed for the L1-hTSH-L3 loop combination as well as for the L3-hTSH-L3 loop combination was comparable and ranged between 100- and 150-fold. This was also reflected by the receptor binding studies (data not shown).

View larger version (24K): [in this window] [in a new window]   Fig. 5.   A and B, combination of two superactive loops. A, comparison of in vitro bioactivity after combination of superactive L1 and L3 loop analogs with either hTSH- wild-type (WT) or the hTSH-L3 loop analog showing different degrees of cooperativity between loops. In analogy, combination of the L3 loop analog or the hTSH-L3 loop analog with the L1 loop analog results in combined mutants with identical increases in in vitro bioactivity, although the L3 loop analog alone is significantly less potent than the hTSH-L3 loop analog (B). Increases in in vitro bioactivity were assessed by cAMP production in the JP09 cell line. Each data point represents the mean ± S.E. of triplicate determinations in a representative experiment repeated two times.

In analogy to these observations, the combination of the L3(3K) analog or the hTSH-L3(3R) analog with the L1(4K) analog resulted in surprisingly similar increases in in vitro bioactivity for the two resulting combinations of superactive loops on either the same (i.e. "unipolar" combination) or on opposite (i.e. "bipolar" combination) sides of the hTSH molecule (Fig. 1), although the L3 loop analog alone was significantly less potent than the hTSH-L3 loop analog alone (Fig. 5B), so that a higher degree of cooperative effects could again be established for the L3 loop analog versus the hTSH-L3 loop analog with the L1 loop analog.

Finally, we combined all three as yet characterized peripheral -hairpin loops (Fig. 6, A-D). We observed an increase in in vitro bioactivity from "single loop" (i.e. L3(3K)) to "double loop" (i.e. L3(3K) + L3(3R)) and "triple loop" (i.e. L1(4K)/L3(3K) + L3(3R)) analogs (Fig. 6A) and a parallel increase in binding affinity (Fig. 6B). Increases in maximal cAMP levels for combinations of superactive peripheral loops ranged between 150 and 180% of the wild-type preparation. The addition of the L3(3K) analog to our previously most potent L1-hTSH-L3 loop combination was even further enhancing in vitro bioactivity (Fig. 6C) and receptor binding affinity (Fig. 6D), thereby generating the most potent hTSH superactive analog described to date.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   A-D, combination of three superactive loops. The gradual increase in in vitro bioactivity and receptor binding for mutants with one, two, and three superactive peripheral loops are shown (A and B). Combination of the L3 loop analog with our previously most potent (22) L1-L3 loop combination (C and D) results in a further gain in biopotency. Increases in in vitro bioactivity were assessed by cAMP production in the JP09 cell line. Each data point represents the mean ± S.E. of triplicate determinations. WT, wild type.

Fig. 7 compares the in vivo bioactivity for hTSH wild-type, the L3(3K), and the L1(4K)/L3(3K) superactive analog. Both preparations were highly potent in vivo and displayed fold increases in in vivo bioactivity comparable to those found in vitro. Although more extensive in vivo studies were beyond the scope of this paper, the levels of hTSH analogs 6 h after intraperitoneal injection suggested no major change in the clearance rate for these analogs. This supported a general applicability of such analogs for in vivo stimulation of radioiodine uptake, which will include future tests in rats and primates.

View larger version (10K): [in this window] [in a new window]   Fig. 7.   In vivo bioactivity. The comparison of in vivo bioactivity for hTSH wild-type (WT) and the L3(3K) and the L1(4K)/L3(3K) superactive analogs is shown. Increases in in vivo bioactivity were assessed by TT4 serum levels 6 h after an intraperitoneal injection of wild-type (WT) hTSH or mutant hTSH analogs. Each data point represents the mean ± S.E. of groups of 4-5 animals. The experiment was repeated once, and representative results are shown.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In the present study, we have designed mutations in the common L3 loop of human glycoprotein hormones resulting in gain-of-function thyrotropin analogs. Whereas studies using peptide approaches have suggested a role of this loop in receptor binding (31-33), previous mutagenesis studies have shown no effects or limited "loss-of-function" (34-36). We have modified a recently developed rational design strategy for bioengineering glycoprotein hormone superactive analogs for mutational analysis of the common L3 loop. Studies from our laboratory and others had suggested a major role of the peripheral -hairpin loops of cystine knot growth factors in receptor binding (8, 9, 21). We have previously designed hTSH and hCG superactive analogs with significant increases in both receptor binding and signal transduction based on the introduction of positively charged lysine and arginine residues into the common L1 and the hTSH-L3 loop (21, 22). The amino acid residues chosen as target sites for site-directed mutagenesis in these studies were delineated from sequence comparisons either between species reflecting different stages of hormone evolution or between different hormones in a given species. Whereas this evolutionary approach proved to be extremely efficient and specific to guide selection of target sites for mutagenesis, it was based on the presence of conserved clusters of positively charged lysine or arginine residues, which could not be seen in amino acid sequence comparisons within the common L3 loop of human glycoprotein hormones. However, highly variable residues undergoing multiple evolutionary changes could be identified. We hypothesized that three regions with particularly high variability between species (64-67, 73-74, and 81) may tolerate the introduction of nonconservative amino acid changes, and these were therefore subjected to a lysine-scanning approach. In contrast to alanine-scanning mutagenesis that had been previously utilized in functional studies (37-39), the current strategy might be particularly useful for bioengineering superagonists of glycoprotein hormones because it allowed us to identify three modification permissive domains, and about 60% (i.e. four out of seven) of the single lysine substitutions tested in this study proved to generate gain-of-function mutants.

In preliminary studies we have additionally co-expressed the analogs with single lysine substitutions in the L3 loop with the hCG -subunit (data not shown). We observed a differential pattern in the modulation of hTSH and hCG bioactivity by identical lysine substitutions in the common L3 loop of human glycoprotein hormones, in particular for the A81K substitution, which increased hTSH bioactivity but decreased hCG bioactivity. In addition, biopotency was greater for the G73K than for the N66K substitution when expressed in the context of hTSH, whereas N66K was more potent than G73K when expressed as hCG. Thus, certain mutations in the -subunit are increasing hTSH bioactivity in a hormone-specific fashion, which can be related to subunit interaction and/or specific hormone-receptor interaction. However, the peripheral location of such mutations suggests that an effect mediated via subunit interaction is less likely. Therefore, it is possible to envision that a specific basic residue in the ligand and an acidic residue in the TSH receptor, not present in the homologous position of the LH receptor, participate in the formation of a new local electrostatic interaction.

We have therefore tested various single and combined lysine substitutions in the L3 loop for their ability to exceed the cAMP stimulation elicited by the wild-type hormone in the context of a hTSH receptor construct (26) where amino acid residues 317-366, unique for the hTSH receptor and not present in the corresponding structure of the hLH/CG receptor, had been deleted. Interestingly, superactive effects of the various hTSH analogs observed with the wild-type receptor could be reproduced with this receptor mutant, thereby indicating that the deleted hTSH receptor region is not involved in mediating the superactive effects of the L3 loop superactive analogs. This strategy of assessing the presence or absence of superactive effects originally characterized for the wild-type receptor with different receptor chimeras or receptor mutants constitutes a separate tool in the analysis of ligand-receptor interaction of glycoprotein hormones. By first screening for possible regions of interaction with overlapping libraries of receptor chimeras, it should be possible to exclude significant portions of the receptor as contact sites for a given mutated ligand. In a second step, certain regions of interest can then be subjected to a more detailed analysis by characterizing defined receptor mutants where a limited number of amino acid residues have been changed. Finally, a charge reversal technique switching individual oppositely charged residues between the ligand and the receptor can be used assuming that this should reestablish the original superactive interaction. Therefore, it should be possible to pinpoint specific interaction sites between hTSH and the hTSH receptor with the help of superactive thyrotropin analogs.

As in our previous studies, gain-of-function effects were highly site-specific, as only introduction of lysine residues at positions 64, 66, 73, and 81, but not at positions 65, 67 and 74, resulted in the generation of superactive analogs. They were also residue- and charge-specific, as a mutant with three lysine or arginine (i.e. strongly basic) residues at positions 66, 73, and 81 showed similar increase in biopotency and was significantly more potent than a triple histidine (i.e. weakly basic) analog. A maximal increase in biopotency was reached by combination of three basic residues within the L3 loop, in contrast to our previous observation where the most potent L1 loop analog had been engineered by introduction of four basic substitutions (21).

Two published crystal structures of deglycosylated hCG, the homology model of hTSH, and epitope mapping studies indicated that amino acid residues 64, 66, 73, and 81 in the 64-81 region are surface-oriented and not involved in hydrogen bonding with the TSH -subunit (5, 6, 21, 33, 40). An introduction of basic residues in such locations enhanced the hormone-receptor interaction. However, 68, 70, 71, 74, and 76 contribute to the formation of a hydrophobic patch between two -subunit loops. Accordingly, an F74K mutation did not show any effect on TSH binding to its receptor.

The observation that the introduction of negatively charged glutamic acid or aspartic acid residues at these positions resulted in all but one mutant characterized in either no change or a decrease in biopotency is further supporting our hypothesis of a modulation of in vitro hTSH analog bioactivity by alteration of electrostatic interactions between ligand and receptor. The slight increase of hTSH bioactivity demonstrated in a single mutant with an amino acid substitution to glutamic acid at position 73, however, may be attributed to other effects than a change in electrostatic interactions with the receptor, i.e. that the effects of introducing negative charges in the ligand may not always depend on electrostatic repulsion. Considering the fact that an amino acid substitution to aspartic acid at the same position did result in a decrease in biopotency, the different lengths of the amino acid side chain may also play a role in ligand-receptor interaction. In addition, position 73 is in close proximity to the tip of the L3 loop, with the most exposed location of all amino acid residues investigated in this study which might account for considerable conformational flexibility after introduction of nonconservative amino acid changes. Finally, from this study, hydrophobic interactions or a modification of hydrogen bonding cannot be excluded as additional events modulating the overall change in bioactivity observed in the course of our experiments. More detailed studies in the future will be needed to specifically address these issues.

The assessment of in vitro bioactivity of the various L3 loop mutants by measurement of cAMP production as well as binding studies were performed under isotonic low salt conditions. Although we did not study the effect of low salt versus normal salt conditions on the in vitro bioactivity of analogs with the L3 loop mutations, we have previously compared the effect of low salt versus physiological salt concentration in the assessment of in vitro bioactivity of superactive TSH analogs with mutations in the L1 and L3 loop (22). These studies revealed only small differences in fold increases in in vitro bioactivity under different (0-0.9% NaCl) ionic conditions, which could be demonstrated to be comparatively minor when considering the fold increases elicited between wild-type and superactive mutant hormone. Furthermore, we have previously tested the effect of L1 and L3 analogs on proliferation of FRTL-5 cells and thyroid hormone production in cultured human thyroid follicles using physiological salt conditions (21). Again, only small differences in fold increases were noted in comparison to the standard isotonic low salt conditions of the cAMP bioassay.

Parallel to our previous studies, we observed cooperative effects when single basic substitutions within one peripheral -hairpin loop were combined. These findings confirmed that additional electrostatic interactions between the ligand and the receptor resulting from the introduction of additional lysine residues into the ligand are not only site-specific but also directly cooperative indicating that they are not likely resulting from distant conformational changes.

The combination of the three as yet characterized superactive peripheral -hairpin loops of human thyrotropin allowed us to extend our studies to the analysis of cooperative effects between loops. The higher degree of cooperativity between the L3 versus the L1 loop analog with the hTSH-L3 loop analog and the surprisingly similar increases in biopotency between unipolar and bipolar combinations suggest that two modified superactive peripheral loops in closer proximity may exhibit higher cooperativity than two more spatially distant loops, further indicating considerable conformational flexibility of both the ligand and the hTSH receptor at the ligand-receptor interface. In addition, the fact that the combination of the superactive L1, L3, and the hTSH-L3 loop analogs was more potent than any combination of two loops supports the concept that all three engineered loops are involved in receptor binding and signal transduction.

The lysine-scanning approach first used for this study as an extension of our previous evolutionary approach in the rational design of superactive analogs of human thyrotropin constitutes a new, separate tool in the design of glycoprotein hormone superagonists. Moreover, the results of this highly efficient approach support the concept that it is possible to increase the electrostatic component of hormone-receptor interaction by introduction of positively charged amino acid substitutions into the ligand either resulting in a recruitment of de novo binding sites not used by the natural hormone or leading to higher binding affinity in preexisting interaction domains.

Superactive analogs of hTSH are not only providing new insights into hTSH structure-function relationships but also have a wide range of potential diagnostic and therapeutic clinical applications. Wild-type recombinant human thyrotropin has been under extensive basic and clinical study (41, 42) as diagnostic preparation for the follow-up of patients after thyroidectomy for thyroid carcinoma. Our design strategy for superactive analogs of human thyrotropin combining evolutionary and lysine-scanning approaches provides the bioengineering tools for developing a second generation recombinant thyrotropin with major improvements in clinical efficacy and therefore enhanced diagnostic and therapeutic utility. The principles derived from these studies are applicable not only to other glycoprotein hormones, but may also be useful for bioengineering other members of the cystine knot growth factor superfamily.

	ACKNOWLEDGEMENTS

We are indebted to Dr. Basil Rapoport for supplying us with the hTSHR-D1 construct. We thank Dr. Christine Leitolf for valuable help and indispensable support during all stages of this study, especially during the preparation and review of the manuscript. We also thank Dr. Rasa Kazlauskaite for technical assistance and general support.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Recipient of Grant Le 1037/1-1 from the Deutsche Forschungsgemeinschaft, Kennedyallee 40, 53715 Bonn, Germany. Present address: Dept. of Clinical Endocrinology, Center of Internal Medicine and Dermatology, Hannover Medical School, 30625 Hannover, Germany.

^§ Present address: The Walter and Eliza Hall Institute of Medical Research, P. O. Royal Melbourne Hospital, Victoria 3050, Australia.

^¶ To whom all correspondence and requests for reprints should be addressed: Section of Protein Engineering, Laboratory of Molecular Endocrinology, Division of Basic Science, Institute of Human Virology, MBC, UMBI, Rm. N457, 725 West Lombard St., Baltimore, MD 21201. Tel.: 410-706-1946; Fax: 410-706-4574; E-mail: szkudlin@umbi.umd.edu.

Published, JBC Papers in Press, June 19, 2000, DOI 10.1074/jbc.M003707200

	ABBREVIATIONS

The abbreviations used are: hTSH, human thyrotropin; bTSH, bovine thyrotropin; hCG, human chorionic gonadotropin; hLH, human luteinizing hormone; TT4, total thyroxine; PCR, polymerase chain reaction; CHO, Chinese hamster ovary.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				G. Kleinau and G. Krause Thyrotropin and Homologous Glycoprotein Hormone Receptors: Structural and Functional Aspects of Extracellular Signaling Mechanisms Endocr. Rev., April 1, 2009; 30(2): 133 - 151. [Abstract] [Full Text] [PDF]

				G. Kleinau, H. Jaeschke, S. Mueller, B. M. Raaka, S. Neumann, R. Paschke, and G. Krause Evidence for cooperative signal triggering at the extracellular loops of the TSH receptor FASEB J, August 1, 2008; 22(8): 2798 - 2808. [Abstract] [Full Text] [PDF]

				S. Mueller, G. Kleinau, H. Jaeschke, R. Paschke, and G. Krause Extended Hormone Binding Site of the Human Thyroid Stimulating Hormone Receptor: DISTINCTIVE ACIDIC RESIDUES IN THE HINGE REGION ARE INVOLVED IN BOVINE THYROID STIMULATING HORMONE BINDING AND RECEPTOR ACTIVATION J. Biol. Chem., June 27, 2008; 283(26): 18048 - 18055. [Abstract] [Full Text] [PDF]

				S. M. Sine, H.-L. Wang, and N. Bren Lysine Scanning Mutagenesis Delineates Structural Model of the Nicotinic Receptor Ligand Binding Domain J. Biol. Chem., August 2, 2002; 277(32): 29210 - 29223. [Abstract] [Full Text] [PDF]

				M. W. Szkudlinski, V. Fremont, C. Ronin, and B. D. Weintraub Thyroid-Stimulating Hormone and Thyroid-Stimulating Hormone Receptor Structure-Function Relationships Physiol Rev, April 1, 2002; 82(2): 473 - 502. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 275/35/27457    most recent M003707200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted 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 Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Leitolf, H. Articles by Szkudlinski, M. W. Search for Related Content PubMed PubMed Citation Articles by Leitolf, H. Articles by Szkudlinski, M. W. Social Bookmarking                      What's this?