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 Edens, A. Articles by Talamantes, F. Search for Related Content PubMed PubMed Citation Articles by Edens, A. Articles by Talamantes, F.

Alternative Processing of Growth Hormone Receptor Transcripts¹

Department of Biology, University of California, Santa Cruz, California 95064

	Abstract

Top Abstract I. Introduction: GH Receptor... II. The GHR Gene III. GHR Transcript... IV. Summary of GHR... References

 

I. Introduction: GH Receptor and GH-Binding Protein

II. The GHR Gene

III. GHR Transcript Heterogeneity

A. 5'-UTR heterogeneity

B. Deletions and truncations of the transmembrane or intracellular domain

C. Deletions in the hormone-binding domain-encoding region

IV. Summary of GHR Alternative Processing and Possible Future Studies

	I. Introduction: GH Receptor and GH-Binding Protein

Top Abstract I. Introduction: GH Receptor... II. The GHR Gene III. GHR Transcript... IV. Summary of GHR... References

 
GH RECEPTOR (GHR) is a transmembrane protein that binds GH with high affinity and specificity. Expression of the receptor is a requirement for cellular responsiveness to GH. The primary transcript that generates the GHR mRNA can undergo alternative processing to produce several related mRNAs, including transcripts that encode the circulating GH-binding protein (GHBP) and a truncated GHR that inhibits normal cellular responses to GH in vitro. In addition, the use of alternative promoters and transcription start sites in the GHR gene-regulatory region generates transcript forms that differ only in the 5'-untranslated region (5'-UTR). This article will review the transcripts originating from the GHR gene. The introduction will address the expression of GHR and GHBP, the major transcripts of the gene. The second section describes the structure of the GHR gene. The third section describes the types of alternative transcripts produced by the gene, the alternative processing mechanisms that produce them, and the biological functions of their protein products. The final section suggests future studies on the expression and function of alternative GHR transcripts.

Normal postnatal growth in mammals requires the presence of GH, a M_r 22,000 polypeptide hormone released primarily from the anterior pituitary gland (reviewed in Ref. 1). GH is bound with high affinity and high specificity by two proteins, the GHR and GHBP. The receptor is associated with the cell membrane of GH target tissues and mediates the somatogenic and metabolic effects of the hormone. The binding protein is found in the serum and forms complexes with circulating GH. Although the exact biological role of the binding protein has not yet been determined, it has been shown to enhance the growth-promoting effects of GH in vivo (2, 3), probably by increasing the half-life of GH in the circulation (4).

GHR in most species has a M_r of 100,000–130,000, although smaller and larger molecular weights have been reported (reviewed in Ref. 5). To date, the amino acid sequence of the GHRs from nine species have been published. These species are the human and rabbit (6), mouse (7), rat (8, 9), bovine (10), sheep (11), pig (12), chicken (13), and monkey (14). In all nine species, the amino acid sequence was deduced by sequencing a cDNA clone of the receptor mRNA.

Based on the deduced amino acid sequences, GHR is predicted to be a single-pass transmembrane protein approximately 620 amino acids in length; the exact number of amino acids varies slightly from species to species. The receptor is initially synthesized as a preprotein roughly 640 amino acids in length, containing a short signal peptide at its amino terminus for direction of the receptor to the cell surface. The mature receptor (after removal of the signal peptide) has an extracellular hormone-binding domain of approximately 245 amino acids (depending on the species) at its N terminus, a hydrophobic transmembrane domain of 24 amino acids, and an intracellular signaling domain of approximately 350 amino acids at its C terminus. Figure 1 compares the sizes of the mature GHRs from the nine species and lists their percent homology to human GHR.

View larger version (28K): [in this window] [in a new window]   Figure 1. Size comparison of GHR and GHBP from different species. GHR and GHBP are shown as boxes. Domains of the proteins are shown as shaded regions inside the boxes. The white region represents the hormone-binding domain of GHR and GHBP, which corresponds to the extracellular domain of GHR. The diagonal striped region represents the GHR hydrophobic transmembrane domain. The black region represents the GHR intracellular signaling domain. The hydrophilic tail of mouse and rat GHBP is shown as a checkered box. Numbers in each region indicate the amino acids present in the corresponding domain. The exact number of amino acids in the rabbit GHBP has not been determined but is known to be at least 197. The percent homology with the human GHR sequence is indicated below each receptor.

Although the substantial amount of GH in serum bound by GHBP, the intracellular presence of GHBP, and cell membrane GHBP all suggest a role of GHBP in the GH axis, investigations of GHBP function have led to conflicting proposals regarding its effects on GH action. Studies have demonstrated that GHBP prolongs the existence of GH in circulation (36, 38, 39, 40), suggesting that GHBP might potentiate GH action by prolonging the availability of GH to target tissues. Consistent with this proposed GHBP function, injection of a GH/GHBP mixture into rats was more effective at promoting growth than injection of GH alone (2). However, GHBP has been shown in vitro to inhibit binding of GH to several cell types and to attenuate the cellular response to GH (41, 42, 43). These inhibitory actions are believed to occur through competition between GHBP and GHR for the hormone. It has recently been proposed that the two seemingly contradictory effects of serum GHBP on GH action may both have biological relevance (44). The authors suggest that the relative concentrations of serum GH and GHBP or the physiological state of the animal might dictate whether the inhibitory or the potentiating effect of serum GHBP will predominate. The function(s) of intracellular and membrane-associated GHBP remain entirely unclear.

Several early studies suggested that the binding protein is either identical, or highly homologous, to the extracellular domain of the receptor. Antibodies raised against the extracellular portion of the receptor cross-reacted with the binding protein (45). Protein sequencing of the N termini of the rabbit receptor and the binding protein showed identical amino acid sequences (6). Recent studies (reviewed in Refs. 4, 46) have explained these findings by demonstrating that GHBP is derived from GHR either by proteolytic cleavage of the receptor’s extracellular domain or by alternative splicing of the GHR primary transcript (GHBP generation will be addressed in greater detail below). Both mechanisms produce GHBP with a hormone-binding domain that is essentially identical to that of GHR. One or the other of these two methods of GHBP generation appear to be used exclusively in some species, while both methods are used in others. The number of amino acids present in circulating GHBP have been best characterized in the rabbit, rat, and mouse. The sizes of the GHBPs in these three species are compared in Fig. 1.

GHR transcripts have been detected in a variety of tissues including liver, muscle, kidney, lung, mammary gland, placenta, and adipose tissue, with the highest level of expression being in the liver (reviewed in Ref. 47). As determined by Northern analysis, the major mRNA that encodes the full-length receptor is approximately 4.6 kb in length, although the exact size varies slightly from species to species. The smallest reported GHR mRNA is the 3.9-kb mouse transcript (7) and the largest is the 5.0-kb monkey transcript (14). Table 1 compares the size of the full-length GHR transcript from nine species. In all nine species the major GHR transcript is more than twice as large as the minimum 1.9 kb necessary to encode the 640-amino acid receptor/signal peptide molecule. The majority of the "excess" size of the GHR mRNA is due to the presence of an approximately 2-kb 3'-untranslated region (3'-UTR) in the transcript (6, 58). Although the specific function of the large 3'-UTR in the GHR transcript has not been investigated, 3'-UTRs in general are thought to affect degradation rates of mRNAs and perhaps also initiation of translation (59, 60).

	II. The GHR Gene

Top Abstract I. Introduction: GH Receptor... II. The GHR Gene III. GHR Transcript... IV. Summary of GHR... References

View larger version (19K): [in this window] [in a new window]   Figure 2. Comparison of human and mouse GHR gene size and structure. The black horizontal line represents intron sequence. Diagonal breaks in the lines indicate uncloned portions of the introns. Boxes represent exons (enlarged for clarity). The numerical designation of the exon is listed below each box. Exons with horizontal stripes encode UTRs of the transcripts (5'-UTRs and 3'-UTRs). Vertical striped exons encode signal sequence. White exons encode the hormone-binding domain. Diagonal striped exons encode the transmembrane domain. The black and white checkered exon (exon 8A) encodes the mouse GHBP hydrophilic C-terminal domain and 3'-UTR. Black exons encode the intracellular domain. Human GHR transcripts have 8 known alternative 5'-UTR sequences (V1-V8). The exons encoding four of the 5'-UTRs (V1, V4, V7, and V8) have been cloned and are shown on the human gene map. The human UTR V5 sequence corresponds to the intron sequence immediately upstream of exon 2. Mouse GHR transcripts have two known 5'-UTRs, L1 and L2. The exons encoding both have been cloned and are shown on the mouse gene map.

Exon 4B encodes an 8-amino acid segment (glu-ser-gln-arg-gln-ala-ala-arg) of the mouse GHR extracellular domain. None of the other species for which GHR has been sequenced, including the rat, have amino acids between the exon 4 and exon 5-encoded amino acids. Although the unique nature of the mouse 8-amino acid segment raises questions about its possible effects on GH binding, no investigations have directly addressed this topic. Studies investigating binding of GH to GHR in the human, however, have provided clues regarding the possible function of the exon 4B amino acids. X-ray crystallography of human GHR extracellular domain binding with the GH molecule (67) showed that the amino acids encoded by the exon 4/exon 5 splice junction do not interact with the GH molecule. This suggests that the 8-amino acid segment of mouse GHR would also not directly participate in binding the GH molecule. Mutational analysis of the human GHR (68), however, has shown that conversion of the last 2 amino acids of exon 4 (Arg⁷⁰ and Arg⁷¹) or the fifth amino acid of exon 5 (Trp⁷⁶) to alanine reduced GH binding by 2-fold, 2.7-fold, and 2.5-fold, respectively. Together, these results suggest that the region encoded by the exon 4/exon 5 splice junction in the human, while not directly involved in binding GH, is important for proper folding of the GHR extracellular domain. Whether the same results apply to the mouse exon 4B amino acids remains to be clarified through mutagenesis and x-ray crystallography studies of the mouse extracellular domain.

Exon 8A also has no homolog in the human GHR gene. It is located in the 2-kb segment separating exons 7 and 8. Exon 8A encodes the C-terminal hydrophilic domain of mouse GHBP and the 3'-UTR of the GHBP transcript. Exons 7, 8A, and 8 can undergo alternative splicing to produce the GHR and the GHBP transcripts (see Section II.B below). Splicing of exon 7 to exon 8A generates the 1.2-kb GHBP transcript, while splicing of exon 7 to exon 8 generates the 4.0-kb GHR transcript.

The chicken GHR gene has been partially characterized by genomic Southern blotting and sequencing of selected exons (58). Although smaller in overall size due to shorter introns, the exon structure appears to be similar to that of the human GHR gene (Fig. 3). The coding exons of the chicken GHR gene were numbered to correspond to the homologous exons in the human GHR gene. Exon 3 in the human GHR gene, however, does not have a homolog in the chicken GHR gene. Consequently, exon 4 follows exon 2 in the map of the chicken GHR gene. Studies of human GHR have shown that the amino acids encoded by exon 3 are not critical for GH binding (69, 70, 71). There have been two reported 5'-UTRs found in chicken GHR transcripts, but the exon(s) encoding them have not yet been cloned.

View larger version (12K): [in this window] [in a new window]   Figure 3. The chicken GHR gene. See the legend to Fig. 2 for an explanation of exon and intron symbols. The exons were given numerical designations to indicate homologies with coding exons of the human GHR gene. Exon 4 follows exon 2 in the diagram to indicate that the chicken GHR gene has no homolog of human exon 3. The exact locations of exons 2 and 7 were not determined. The location of these exons on the diagram is therefore an approximation. Although chicken GHR transcripts have two known alternative 5'-UTR sequences, the exon(s) encoding these 5'-UTRs have not been cloned.

	III. GHR Transcript Heterogeneity

Top Abstract I. Introduction: GH Receptor... II. The GHR Gene III. GHR Transcript... IV. Summary of GHR... References

In cases where the transcript size is not significantly altered, restriction analysis and cDNA sequencing have been used to detect alternative GHR transcripts. For example, sequencing of bovine GHR cDNAs, which appear as a single band by Northern analysis, has revealed GHR transcript heterogeneity in this species as well (74). Regardless of the actual size of the transcript, the terms "full length" and "major" GHR transcript will be used to indicate transcripts that encode the entire GHR, including complete extracellular, transmembrane, and intracellular domains. For the purposes of this review, GHR transcripts generated by alternative processing will be divided into three categories: 1) transcripts with 5'-UTR heterogeneity, 2) transcripts that encode deletions or truncations of the receptor transmembrane or intracellular domains, and 3) transcripts that encode deletions in the receptor hormone-binding domain.

A. 5'-UTR heterogeneity
Heterogeneity in the 5'-UTR of GHR transcripts was observed in the first GHR cDNAs to be cloned and sequenced, those of the rabbit and human (6). Of the six human clones that were sequenced, five had distinct 5'-UTRs. All three of the rabbit clones had distinct 5'-UTR sequences. Within each species, the 5'-UTRs were identical for the first 11 bases upstream of the initiating AUG codon but diverged in sequence beginning at the 12th base. Human GHR exon 2 encodes the final 11 bases of the 5'-UTR as well as the first 23 amino acids of the coding region. The point of divergence of the human and rabbit 5'-UTR sequences was therefore exactly coincident with the 5'-boundary of exon 2, strongly suggesting that the sequence divergence is produced by splicing of alternative 5' UTR-encoding exons to exon 2.

Subsequent studies identified two distinct 5'-UTRs in GHR transcripts of the sheep (11, 52), chicken (58), cow (10, 74), and mouse (75), five 5'-UTRs in the rat (76), and eight 5'-UTRs in the human (77). One of the eight human 5'-UTRs (designated 5'-UTR V5) was identical in sequence to the intron immediately upstream of exon 2 in the human GHR gene (61) and therefore might represent a partially spliced form of the GHR transcript instead of a true alternative first exon. Another of the human 5'-UTRs (V3) was found in three distinct forms, designated V3a, V3b, and V3c. The V3c form had a 93-bp internal deletion compared with the other two forms, while V3b contained a divergent 5'-sequence compared with the other two V3 sequences. Most of the GHR 5'-UTRs contain one or more upstream AUG codons. AUG codons occur in less than 10% of all eukaryotic 5'-UTRs but are found at much higher frequencies in the 5'-UTRs of protooncogenes, growth factors, and cell surface receptors (reviewed in Ref. 78). Upstream AUG codons occurring as part of the strong translation-initiating consensus sequence A/GNNAUGG (79) have been shown to suppress translation at the downstream translation initiating AUG (reviewed in Ref. 80). Although the majority of GHR 5'-UTR AUG codons are not found in strong start sequences and therefore are not predicted to influence translation efficiency, their frequency remains a curious feature of GHR 5'-UTRs. Table 2 lists the known GHR 5' UTRs by species and indicates their name designations, their homology to the 5'-UTRs of other species, and the number of upstream AUG codons.

1. The sheep. Using the two 5'-UTRs (1A and 1B) of the sheep GHR cDNAs as probes, sheep genomic clones containing the two alternative first exons were isolated and sequenced (52, 83). The upstream-most of the two exons encodes 5'-UTR 1B and is located at least 27 kb upstream of exon 2. The region immediately upstream of exon 1B contains a CCAAT box and a GC box, the binding consensus sequences for the common transcription factors CCAAT transcription factor (CTF) and promoter specific transcription factor 1 (Sp1), as well as several less common binding consensus sequences, such as E boxes (a binding site for the muscle-specific transcription factor MyoD) and a C/EBP site, which is bound by members of the CCAAT/enhancer binding protein family of transcription factors. C/EBP sites are associated with genes expressed in the liver (reviewed in Ref. 85). No TATA boxes were found upstream of the 1B exon. Northern analysis indicated that 5'-UTR 1B is expressed in liver, skeletal muscle, testes, intestine, and adrenal GHR transcripts. Lower levels of expression were detected in heart, pancreas, and bone. When ligated upstream of a luciferase reporter gene, the promoter was able to drive in vitro expression of the reporter gene in cell lines derived from human liver (HuH7 cells), human cervix (HeLa cells), and hamster ovary (CHO cells). Together, the Northern analysis and transfection studies indicated that the 1B promoter had little tissue specificity and therefore might function as the general promoter for the GHR gene in sheep.

The downstream exon encodes 5' UTR-1A and is located 17 kb from exon 2. In contrast to the general tissue expression of the exon 1B, Northern analysis and ribonuclease protection assay (RPA) showed that GHR transcripts with 5'-UTR 1A are found only in the liver (11). Several putative binding sites for factors associated with liver-specific genes were found upstream of the exon, including four C/EBP sites, two GR sites (to which the activated glucocorticoid receptor binds), and binding site for hepatic nuclear factor 5. A TATA box is present near the major transcription start site. Reporter gene constructs demonstrated that this promoter region was active in vitro in two hepatocyte cell lines. Figure 4 shows the genomic structure and splicing pattern of the two alternative first exons and summarizes the characteristics of the two sheep GHR promoters.

View larger version (23K): [in this window] [in a new window]   Figure 4. The regulatory region of the sheep and mouse GHR genes. Exon and intron symbols are explained in the legend to Fig. 2. Dotted diagonal lines connecting the exons represent alternative splicing pathways. Arrows indicate the promoter regions upstream of the alternative first exons. A brief description of the promoter’s tissue specificity and transcription factor-binding sites is shown below the arrows.

2. Bovine. The exon encoding bovine 5'-UTR 1A has been cloned and sequenced (96). The region upstream of the exon was not analyzed for promoter function. However, the first 500 bases of this region were 94% identical in sequence to the sheep 1A promoter, leading the authors to hypothesize that it functions as the promoter for the cow GHR exon 1A. Although the exon(s) encoding bovine 5'-UTR 1B have not yet been cloned, the 1B sequence shows significant sequence homology with sheep exon 1B (74), suggesting that the associated promoters might also be similar in sequence. Consistent with the hypothesis that the bovine 1A and 1B promoters are functionally similar to the sheep 1A and 1B promoters, bovine 1A is expressed almost exclusively in the liver while bovine 1B is expressed in the liver as well as in several extrahepatic tissues.

3. Mouse. Mouse GHR transcripts contain two distinct 5'-UTRs, designated L1 and L2 (75). The exon encoding mouse 5'-UTRs L1 has been cloned and the adjacent upstream regions sequenced (63, 97). The region upstream of exon L1 contained two TATA boxes and an C/EBP binding site, elements that the sheep liver-specific 1A promoter also contained. Binding sites for the common transcription factors, activator protein-2 (AP-2) and CTF, were also identified. Functional analysis of the putative promoter region was carried out by ligating a 3.5-kb fragment of the DNA upstream of the exon to a luciferase reporter gene and transfecting into a hepatoma cell line. Luciferase activity increased 10-fold above that of the promoterless luciferase gene, demonstrating that the sequence upstream of the exon acts as functional promoter in vitro. Deletion of portions of the 3.5-kb fragment identified two transcription enhancer elements, located at -3.4 and -3.0 kb relative to the exon, which were necessary for maximal promoter activity. The enhancer element at -3.0 kb contained a binding site for a member of the CTF/NF-1 family of transcription factors (97). The sequence of the enhancer element at -3.4 kb did not match the binding consensus sequence of any known transcription factor. A reporter gene construct containing the -3.4 enhancer element was more than twice as active in adult rat hepatocytes as in fetal hepatocytes, while a similar reporter gene construct lacking the -3.4 enhancer element showed equivalent activity in adult and fetal hepatocytes. These results suggest the -3.4 enhancer element might play a role in the postnatal up-regulation of GHR expression (9, 11).

Sequencing mouse exon L2 and the adjacent upstream region revealed two GC boxes and a CAAT box, both of which occur in the ovine 1B promoter region (66). Also in common with the ovine 1B promoter, no TATA box was found near the transcription start site cluster. Although reporter gene analysis will be necessary to fully characterize the mouse L2 promoter, the similarities between it and the sheep 1B promoter suggest that both function as general tissue promoters for the GHR gene. The regulatory region of the mouse GHR gene is diagrammed in Fig. 4.

4. Rat. Rat GHR transcripts have five 5'-UTRs, designated V1-V5 (76). Northern blot studies show that UTR V2 is expressed exclusively in the liver, and the other four are expressed in various tissues, including the liver, but at differing levels. The exon encoding 5'-UTR V2 and the surrounding genomic region have been cloned (82). Sequencing upstream of the exon showed no TATA box or other binding sites for common transcription factors. However, two GR sequences, a consensus sequence that can be bound by activated glucocortocoid receptor, were found upstream of the exon. GR sequences are found upstream of the sheep GHR liver-specific promoter as well. Although no reporter gene studies were undertaken, the authors investigated the in vivo activity of the exon’s promoter by using Northern analysis and RPAs to quantitate the expression of the liver-specific 5'-UTR as a percentage of total GHR transcripts. They found that 30% of the liver GHR transcripts in females contain the liver-specific 5'-UTR V2, while in males, only 2% of the liver GHR transcripts contain 5'-UTR V2. Removal of the gonads caused the percentage of GHR transcripts containing 5'-UTR V2 to rise to 7% in males and to fall to 16% in females. Infusion of exogenous GH to males in a manner that mimics the female GH secretion pattern caused a rise in expression to 17%. Together, these data indicate that the 5'-UTR V2 is expressed in the liver in a sexually dimorphic manner and suggest that the sex hormones and GH secretion pattern play a major role in determining the level of its expression.

The regions of the rat GHR gene encoding 5'-UTRs V1, V3, V4, and V5 have not yet been cloned. However, a model of the rat GHR gene regulatory region consistent with RPA, cDNA cloning, and genomic Southern blotting data has been proposed (76). The cDNA cloning studies showed that 5'-UTRs V3, V4, and V5 all begin with an identical 25-bp sequence. Downstream of this 25-bp segment the sequences diverge and become unique for 22 bp (V3), 101 bp (V4), and 175 bp (V5). Exactly 9 bp before the initiating AUG of the coding region, the 5' UTR sequences reconverge. These results suggested to the authors that the 25-bp segment common to 5'-UTRs V3, V4, and V5 might be encoded by a separate upstream exon. According to this model, the "common 25-bp" exon is spliced to downstream exons that encode the unique segments of 5'-UTRs V3, V4, and V5. These exons are, in turn, spliced to the 5'-end of exon 2, which includes the final 9 bp of the rat 5'-UTR sequence (Fig. 5). Although in the simplest version of this model the common 25-bp sequence and the unique segments of 5'-UTRs V3, V4, and V5 are each encoded by single exons, the cloning studies did not rule out that any or all of these segments might be encoded by multiple exons. In support of a more complex version of the model, RPA studies using a probe containing the 5' UTR V4 sequence produced a 135-bp band corresponding to full-length 5'-UTR V4 and a smaller band corresponding to 71 bp of 5'-UTR V4 sequence. The smaller band could be generated by any one of at least three forms of alternative processing: 1) The unique portion of the V4 sequence could be composed of 2 exons (V4A and V4B in Fig. 5). Inclusion of both V4-unique exons generates full-length 5'-UTR V4. Skipping either exon V4A or V4B deletes a portion of the 5'-UTR V4 sequence. 2) The single exon encoding the V4-unique sequence could contain an internal splice donor or splice acceptor site. Use of the internal site generates the smaller 5'-UTR V4. Use of the terminal splice donor/splice acceptor site generates full-length 5'-UTR V4. 3) Exon V4 could contain a functional transcription start site. Use of this start site deletes the 25-bp common sequence and the 5'-portion of exon V4 from the 5'-UTR sequence. Sequencing of all V4 variants and the regions of the rat GHR gene encoding them will be necessary to determine the mechanisms that generate the observed heterogeneity. Human 5'-UTR V3 was also found to exist in multiple forms (77), indicating that heterogeneity within an individual GHR 5'-UTR is not limited to the rat.

View larger version (17K): [in this window] [in a new window]   Figure 5. Hypothetical models of the rat GHR gene-regulatory region. Exon and intron symbols are explained in the legend of Fig. 2. Dotted diagonal lines connecting the exons represent alternative splicing pathways. The exon labeled 25 bp encodes the 25-bp sequence common to 5'-UTRs V3, V4, and V5. The exons labeled V5, V4, and V3 encode the unique sequences of these three 5'-UTRs. Exons labeled V2 and V1 encode 5'-UTRs V1 and V2. Exon 2 encodes the final 9 bp of the 5'-UTR and the beginning of the GHR-coding region. All introns are of unknown length. Arrows represent promoter regions. A, Simple model of the GHR-regulatory region. The 25-bp exon can be spliced to exons V5, V4, and V3, which are spliced to exon 2. Exons V2 and V1 are spliced to exon 2 without splicing to the 25-bp common exon. In this model exons V1, V2, and 25 bp are each associated with a distinct promoter. B, Three models to account for the short form of the 5'-UTR V4 observed by RPA studies. The short form of the 5'-UTR may be generated by 1) exon skipping if the unique portion of the V4 sequence is composed of 2 or more exons (shown as V4A and V4B in the diagram), 2) use of an an internal splice donor or splice acceptor site (splice donor shown in diagram) in exon V4, or 3) use of a transcription start site in exon V4.

Probes corresponding to 5'-UTRs V3, V4, and V5 hybridized to EcoRI restriction fragments distinct in size from the 4.5-kb fragment proposed to contain exons V1, V2, and 2. These results indicate that exons V3, V4, and V5 are located further upstream than the V1 and V2 exons. Although there is no EcoRI restriction site within the V4 sequence, the probe corresponding to 5'-UTR V4 hybridized with two EcoRI fragments, whereas a V4 probe from which the 25-bp common sequence had been removed (leaving only the V4-unique sequence) hybridized to a single EcoRI fragment. These results are consistent with the common 25-bp segment being encoded by a separate exon than the V4-unique segment.

5. Human. Human GHR transcripts have eight alternative GHR 5'-UTRs, designated V1–V8 (77). Exons encoding UTRs V1, V4, V7, and V8 have been cloned (62). The four exons are clustered within a 2-kb section of genomic DNA. The distance between these exons and exon 2 was not established. Northern analysis showed that 5'-UTR V1 was expressed exclusively in the liver. The sequence 200 bases upstream of exon V1 contained a TATA box but no other binding sites of known transcription factors. A reporter gene containing 2 kb of DNA upstream of the exon V1 was able to increase luciferase expression 20-fold above that of the promoterless luciferase gene when transfected into a hepatocyte cell line, demonstrating that the region upstream of the liver-specific exon is a functional promoter in vitro. Expression of 5'-UTRs V2–V8 was not detected by Northern analysis in the liver or any of 19 other tissues examined. This was an unexpected result for two reasons. All eight of the human 5'-UTRs were originally cloned from a liver cDNA library (6, 77) and therefore detection of all eight was expected in the liver. Second, GHR transcripts had been detected by RT-PCR in several tissues (98), including kidney, pancreas, heart, testis, ovary, lung, and the intestine, tissues that did not express any of the eight known GHR 5'-UTRs as detected by Northern analysis. The apparent contradiction in these studies might indicate that extrahepatic GHR transcripts contain novel (uncharacterized) 5'-UTRs, or it might simply reflect the different sensitivities of RT-PCR and Northern analysis.

6. Conclusion. In summary, GHR transcripts in all species studied except the pig and the monkey are known to contain alternative 5'-UTRs. The exons encoding 10 of the 5'-UTRs (2 from sheep, 1 from bovine, 2 from mouse, 4 from human, and 1 from rat) have been cloned from genomic DNA, and in each case each 5'-UTR was shown to be encoded by a separate exon. Reporter gene studies and Northern analysis suggest that the 5'-UTR exons are associated with unique promoters. In each species in which 5'-UTR expression has been studied, one of the 5'-UTRs displayed exclusive or elevated activity in the liver. The purpose of multiple 5'-UTRs in the GHR transcripts is not clear at this time. One possibility is that they serve a regulatory role by differentially affecting the translational efficiency of the mRNAs. Another possibility is that the complexity of the tissue and physiological regulation of GHR transcription necessitates multiple promoters and, as a consequence, the expression of multiple 5'-UTR sequences.

B. Deletions and truncations of the transmembrane or intracellular domain
Sequence heterogeneity in the GHR transcript’s coding region, unlike the heterogeneity in the 5'-UTRs, has the potential to directly alter the receptor’s amino acid sequence and, therefore, its biological properties and function. Transcripts with altered or deleted transmembrane and intracellular domain-encoding regions are of particular interest because these domains are associated with two central functions of the receptor, integration into target tissue cell membranes and signal transduction. In this section, deletions of the transmembrane domain will be reviewed first, followed by truncations of the intracellular domain.

1. Deletions of the transmembrane domain.
a. Mouse.
The first evidence for a change in the function of GHR due to alternative processing of the transcript was the finding that in the mouse, the circulating GHBP is produced by alternative splicing of the primary GHR transcript (7). Early studies had shown that the M_r of the circulating mouse GHBP and GHR were 42,000 and 103,000, respectively (99, 100), and that the two proteins were immunologically similar (99). Northern blots of mouse liver RNA revealed two transcripts of 3.9 kb and 1.2 kb, which hybridized to a GHR probe (50). Cloning of the cDNAs showed the transcripts were identical in sequence for the first 899 bp, after which point they diverged (7). Based on the predicted amino acid sequence, both transcripts encoded an identical GH-binding domain but diverged in sequence beyond this point. The 3.9-kb transcript’s GH-binding domain is followed by a 24-amino acid hydrophobic transmembrane domain and a 350-amino acid intracellular domain, indicating that it encodes the full-length GHR. The 1.2-kb transcript’s GH-binding domain is followed by a 27-amino acid hydrophilic C-terminal domain. The smaller size of the 1.2-kb transcript’s predicted protein product and its lack of a transmembrane domain suggested that it encoded the mouse serum GHBP. Transfection of a cDNA clone encoding the 1.2-kb transcript into the Chinese hamster ovary (CHO) cell line and the monkey kidney COS-7 cell line resulted in high levels of GH-binding activity in the media (5). An antibody was raised against a synthetic peptide containing the 27-amino acid C-terminal domain encoded by the 1.2-kb transcript (101). The antibody was able to immunoprecipitate GHBP from serum, confirming that mouse GHBP is the translation product of the 1.2-kb transcript. Although the absolute identity of the first 899 bp of the transcripts strongly suggested the 3.9 kb and the 1.2 kb mRNAs arise by alternative splicing of a single primary transcript, the possibility that the two transcripts originated from two highly related genes was not ruled out.

Experimental evidence confirming the alternative splicing hypothesis was obtained when the region of the mouse GHR gene containing exons 7 and 8 was cloned (64, 65). An exon encoding the 27-amino acid hydrophilic C terminus and the 3'-UTR of the GHBP transcript was discovered in the 2-kb region between exons 7 and 8. The GHBP-specific exon has no homolog in the human GHR gene. To indicate that the exon follows exon 7 while differentiating it from the transmembrane domain-encoding exon 8, the GHBP-specific exon has been designated exon 8A. Figure 6 diagrams the portion of the mouse GHR gene containing exons 7, 8A, and 8. The 1.2-kb GHBP transcript contains coding exons 2–7 spliced to exon 8A. The GHR-specific exons 8, 9, and 10 are absent in the GHBP transcript. The 4.0-kb GHR transcript contains coding exons 2–7 spliced to exons 8, 9, and 10. Exon 8A does not appear in the GHR transcript.

View larger version (6K): [in this window] [in a new window]   Figure 6. The section of the mouse GHR gene containing exons 7, 8A, and 8. Exon and intron symbols are explained in the legend to Fig. 2. Dotted diagonal lines connecting the exons represent alternative splicing pathways. The mutually exclusive splicing of exon 7 to either exon 8A or to exon 8 generates the GHBP and GHR transcripts, respectively. Exon 8A encodes the 27-amino acid hydrophilic domain of the binding protein and the 3'-UTR of the GHBP transcript. A portion of the exon 8A 3'-UTR sequence is shown below the gene diagram. The 3'-UTR contains two overlapping polyadenylation consensus sequences (underlined) located 12 bp upstream of the potential cleavage/polyadenylation locations (indicated by arrows). Exon 8 encodes the 24-amino acid transmembrane domain of the receptor and a small number of extracellular and intracellular domain amino acids.

View larger version (17K): [in this window] [in a new window]   Figure 7. Alternative splicing pathways for generation of GHR and GHBP transcripts in mice. Exon and intron symbols are explained in the legend to Fig. 2. The primary transcript, GHR transcript, and GHBP transcript are boxed. Exons 1–6 (which are common to GHR and GHBP transcripts) are shown as a single unit for clarity. Primary transcripts can become committed to the GHR splicing pathway (upper diagram) by splicing of exon 8 to exon 7. This removes the GHBP-specific exon 8A from the transcript. Primary transcripts can become committed to the GHBP-splicing pathway (lower diagram) by splicing of exon 8A to exon 7 followed by cleavage of the transcript in the 3'-UTR of exon 8A (lower left), or by cleavage of the transcript in the 3'-UTR of exon 8A followed by splicing of exon 8A to exon 7 (lower right).

View larger version (13K): [in this window] [in a new window]   Figure 8. GHR-GHBP minigene constructs and the processing of their nascent transcripts in mouse L cells. Exon and intron symbols are explained in the legend to Fig. 2. The columns to the right of the minigene diagrams indicate the processing steps that the nascent transcript was able to undergo, as determined by RT-PCR and Northern analysis. The intact minigene (top diagram) was able to splice exon 7 to exons 8A or 8, and to undergo cleavage in exon 8A. Deletion of the exon 8A polyadenylation consensus sequences and the flanking AT-rich region (middle diagram) eliminated cleavage in exon 8A but did not abolish splicing of exons 8A and 8 to exon 7. Mutation of the exon 8A splice acceptor site (lower diagram) abolished splicing of exon 7 to exon 8A but did not eliminate cleavage in exon 8A or splicing of exon 8 to exon 7.

In a parallel experiment, the mouse cell line was transfected with a minigene designed to eliminate splicing of exon 7 to exon 8A through a mutation of the exon 8A splice acceptor site. No transcripts containing exons 7 and 8A spliced together were detected by RT-PCR, indicating the mutation had successfully abolished GHBP splicing. However, in addition to GHR transcripts, cells transfected with the mutated minigene produced a transcript retaining intron 7/8A and exon 8A. Northern analysis indicated this transcript was cleaved in the 3'-UTR of exon 8A, demonstrating that the first step in the C/S GHBP pathway (cleavage in exon 8A without prior splicing exon 7 to 8A) can also occur in vitro.

Together, the minigene data indicated that neither exon 8A splicing nor cleavage in the 3'-UTR of exon 8A is absolutely dependent on the other process having occurred previously. Although this result is consistent with both the S/C and C/S pathways being able to contribute to the cellular GHBP mRNA pool, it is important to note the experiments do not indicate that both pathways occur in vivo or that both are used to equal degrees to produce GHBP mRNA. Additional experiments, designed to quantitate the transcript levels produced by the intact minigene and the mutated forms, will be necessary to estimate the contribution of each pathway in vivo.

b. Rat.
Rat GHBP is also produced by alternative splicing of the GHR primary transcript (104). The sizes of rat GHR and GHBP are similar to those of the mouse (21, 105). Two transcripts of 4.75 kb and 1.2 kb hybridized with a GHR probe in Northern blots of rat tissues (8). A cDNA clone of the 1.2-kb transcript was identical in sequence to the receptor transcript until nine bases before the hydrophobic transmembrane domain-encoding region, at which point the sequence diverged completely. Beyond the divergence point, the transmembrane and intracellular domain-encoding regions of the receptor transcript were replaced in the smaller transcript by a 17-amino acid hydrophilic domain-encoding region. Transfection of the 1.2-kb cDNA into COS-7 cells resulted in 96% of the assayable GH-binding activity being located in the medium, indicating that the smaller transcript encoded a secreted GHBP. An antibody raised against the 17-amino acid hydrophilic domain encoded by the 1.2-kb transcript immunopreciptated almost all GHBP from rat serum (106, 107).The portion of the rat GHR gene between exon 7 and exon 8 was cloned and analyzed (104). The GHR gene structure in this region was essentially identical with that of the mouse. The intron between exons 7 and 8 contained an exon encoding the hydrophilic domain and 3'-UTR of the GHBP transcript. These findings confirmed the alternative splicing hypothesis of rat GHBP generation. As with the mouse, the pathways by which the GHR and GHBP transcripts (i.e., S/C or C/S GHBP mRNA generation) are produced in vivo remain to be determined.

c. Rabbit.
Studies suggest that the serum GHBP is generated in some species by proteolytic cleavage of the receptor’s hormone-binding domain rather than through the alternative splicing mechanism found in mice and rats (reviewed in Ref. 28). Despite the presence of GHBP in the circulation, there have been no reports of transcripts lacking the hydrophobic transmembrane-encoding region in the human, rabbit, sheep, bovine, or pig, suggesting that in these species GHBP is produced posttranslationally. Posttranslational generation of GHBP from GHR has been best characterized in the rabbit. Previous studies demonstrated that rabbit GHR has an M_r of 100,000–130,000 (6) and the GHBP has an M_r of 51,000 (108). Transfection of the full-length rabbit GHR cDNA into CHO cells, the human hepatoma HepG2 cell line, and COS-7 cells (109, 110) resulted in expression of GHR in the plasma membrane and secretion of GHBP into the culture medium. Because the transfected GHR cDNA did not contain intron sequences, it was not believed to undergo any splicing. The GHBP in the medium was therefore thought to originate from the intact receptor by proteolytic cleavage. Although no specific protease(s) involved in the cleaving reaction were identified, the fact that three cell lines of diverse tissue and species origin were able to carry out the cleavage reaction suggested that the protease(s) responsible for the cleavage have a wide pattern of expression. The exact point of receptor cleavage has not been precisely determined. Protein sequencing, however, has demonstrated that amino acids 190–197 of the 246-amino acid rabbit GHR extracellular region are definitely present in the binding protein (6). Additional protein sequencing in the same study tentatively identified extracellular amino acids 237–239 as also being present in the binding protein, but due to weak signals their identification was not certain. Together, these data indicate that the rabbit-binding protein is produced by proteolytic cleavage of the receptor hormone binding-domain and that the cleavage site is located C-terminal to amino acid 197, possibly within a few residues of the transmembrane domain. The sequence specificity of the cleavage reaction was demonstrated by the finding that COS-7 cells transfected with rat GHR cDNA do not secrete GHBP into the medium (110), implying that the rat GHR does not contain the proper amino acid sequence for cleavage by the protease(s) responsible for rabbit GHR cleavage in COS-7 cells.

d. Monkey.
Transfection of the human kidney fibroblast 293 cell line with monkey GHR cDNA resulted in release of soluble GHBP into the medium as well as expression of the full-length receptor in the cell membrane, suggesting that generation of GHBP in the monkey can take place through proteolytic cleavage of the receptor (14). Like the rat and the mouse, however, the monkey also expresses an alternatively spliced GHR transcript that encodes a functional GHBP. Analysis of monkey cDNA revealed a transcript that was identical in sequence to the GHR transcript for the first 784 bases of the hormone-binding domain-encoding region but which diverged in sequence 8 bp before the beginning of the transmembrane domain-encoding region of the receptor mRNA. Downstream of the divergence point, the GHBP transcript contained 236 bp of novel sequence, which included a polyadenylation signal and terminated in a polyadenosine tail. Comparison of the monkey GHR and GHBP sequences with the human GHR genomic sequence revealed that the point at which the monkey GHBP and GHR transcripts diverged is exactly coincident with the 3'-end of human exon 7 and, furthermore, that downstream of the divergence point, the monkey GHBP transcript is identical in sequence to the 5' end of human intron 7/8. The transcript’s predicted translation product would contain the first 243 amino acids of the receptor’s 246-amino acid extracellular domain. The final 3 amino acids of the extracellular domain and the entire transmembrane and intracellular domains of the receptor would be replaced by 8 novel amino acids before an in-frame stop codon occurred. The similarity of the transcript’s novel sequence with human intron 7/8 and the lack of a transmembrane domain in the translation product suggested to the authors that the transcript is a GHBP-encoding mRNA produced by retention of intron 7/8 (Fig. 9). PCR analysis showed that GHBP transcript expression is highest in the stomach, intermediate in the kidney and heart, and lowest in the liver. No expression was detected in the intestine and pancreas.

View larger version (16K): [in this window] [in a new window]   Figure 9. Generation of monkey GHR and GHBP transcripts. A, The section of the monkey GHR gene containing exons 7 and 8 and intron 8A. Exon and intron symbols are explained in the legend to Fig. 2. Dotted diagonal lines represent alternative splicing pathways. Splicing of exon 7 to exon 8 (upper dotted line) generates the GHR mRNA. Retention of intron 7/8A and cleavage/polyadenylation in the intron generates the GHBP mRNA, which contains the 5' 236 bases of intron 7/8 (lower dotted line) and a polyadenosine tail (polyadenylation site marked by arrow). B, The monkey GHR and GHBP nucleotide sequences surrounding the exon 7/exon 8 splice junction and the corresponding protein sequences. Nucleotide sequences are arranged by codon. Exon sequences are shown in uppercase and are boxed in thick lines. Human intron 7/8 sequence appears in lowercase and is boxed by thin lines. Diagonal breaks in the boxes indicate parts of the exon and intron sequences not shown in the diagram. Amino acids are shown connected to their respective codons by vertical lines. Uppercase amino acids are encoded by exons. Amino acids encoded by intron 7/8 (the eight novel amino acids of GHBP) are in lowercase. The 14 C-terminal amino acids (amino acids 238–251) of the binding protein are shown in the GHBP diagram. Receptor amino acids 238–258 are shown in the GHR diagram. Transmembrane domain amino acids are marked with asterisks.

A second alternatively spliced GHR transcript was also detected by PCR analysis of monkey RNA. Sequencing of the PCR product revealed that the segments encoded by exons 4, 5, and 6 in the human GHR gene were deleted in the monkey transcript. Translation would result in a protein containing the signal sequence and the first 28 amino acids of the hormone-binding domain (encoded by exons 2 and 3), followed by 16 novel amino acids (encoded by a frame shifted reading of exon 7) before a stop codon is encountered. The function of the protein encoded by this transcript is not clear. The absence of a hydrophobic transmembrane domain suggests the protein would not function as membrane-bound receptor, while the absence of the amino acids critical for GH binding (reviewed in Ref. 47) suggests that it would not function as circulating GHBP. Further studies of the expression and translation of this transcript will be necessary to determine its biological significance.

e. Chicken.
Western analysis has identified two major GHBPs in chicken serum. The M_r of these proteins is 69,500 and 27,500 (24). The source of GHBP in the chicken, however, is not clearly understood. Although it might arise from proteolytic cleavage of GHR in a fashion analogous to that of the rabbit, there are two chicken GHR transcript variants that potentially encode GHBPs. One transcript variant is produced by the use of an alternative polyadenylation site located within the coding region of the transcript. The second transcript variant is produced by the use of an alternative 3'- splice acceptor site upstream of exon 7. Northern analysis of liver mRNA showed that the chicken has two mRNAs of 4.7 kb and 0.7 kb and several minor transcripts of intermediate size that hybridize to GHR probe (13, 56, 57). Sequencing of cDNA clones showed that the larger mRNA corresponded to the full-length receptor. The cDNA corresponding to the 0.7-kb mRNA contained the first 325 nucleotides of the full-length receptor sequence followed by a polyadenosine tail. In most mRNAs, polyadenylation occurs in the 3'-UTR and requires two sequence motifs, a polyadenylation consensus sequence (AAUAAA) and a GU-rich region located slightly downstream of the RNA cleavage site (reviewed in Refs. 111, 112). The coding region of the full-length chicken GHR transcript contains both of these motifs. An AAUAAA sequence occurs at nucleotide positions 304–309, and a GU-rich sequence occurs between nucleotides 330 and 346. The authors hypothesized that the 0.7-kb transcript was produced by RNA cleavage and polyadenylation of the GHR transcript between these two sites (Fig. 10). RT-PCR showed that turkey, duck, and quail, which also express a 0.7-kb GHR transcript, all have a polyadenylation consensus sequence and the GU-rich region at this location. Mammalian GHR-coding regions, on the other hand, do not contain a polyadenylation consensus sequence at this location. These findings suggest that truncation of the GHR transcript by polyadenylation is restricted to certain avian species.

View larger version (10K): [in this window] [in a new window]   Figure 10. Generation of GHR/GHBP protein isoforms in the chicken by alternative RNA cleavage/polyadenylation and use of an alternative splice acceptor site in intron 6/7. The top portion of the diagram shows the section of the chicken GHR gene containing exons 5, 6, and 7. Exon and intron symbols are explained in the legend of Fig. 2. Dotted diagonal lines indicate alternative splicing pathways. The striped box inside exon 5 shows the location of transcript cleavage when the alternative RNA cleavage/polyadenylation site within the coding region is used. Cleavage and subsequent polyadenylation at this site result in production of the 0.7-kb transcript. Transcripts not cleaved in exon 5 can undergo alternative splicing of exon 6 to either a splice acceptor site within intron 6/7 or to exon 7. Both of the alternatively spliced transcripts are 4.6 kb. The lower portion of the diagram shows the predicted protein products of the alternative processing pathways. Protein symbols are explained in the legend of Fig. 1. The C-terminal domain of the 0.7-kb transcript’s protein product is predicted to consist entirely of the amino acid lysine due to translation of the transcript’s polyadenosine tail. The number of lysine residues in this domain is not known.

Although the 0.7-kb transcript was shown to be associated with polyribosomes in liver cells, indicating that it is translated, the biological properties of the truncated protein product were not investigated in the above studies. Certain predictions, however, based on the amino acid sequence encoded by the 0.7-kb mRNA, can be made. Because the truncated GHR would lack the hydrophobic transmembrane and intracellular domains of the full-length receptor, it is very unlikely that the protein is an integral membrane protein or that it could activate known GHR signal transduction pathways. Whether the translated 0.7-kb transcript encodes a functional GHBP is not clear. Only the first 95 amino acids of the 221-amino acid hormone-binding domain of the receptor are encoded. The 95 amino acids that would be present in the protein include roughly half of the amino acids that have been implicated by mutagenesis (reviewed in Ref. 47) and crystallography studies (67) as being important for GH binding. Because there is no stop codon in the truncated transcript, the polyadenosine tail would be translated into a polylysine domain at the C terminus of the protein, perhaps forming a structure analogous to the hydrophilic C-terminal domain of the rat and mouse GHBPs. Purification and GH-binding studies of the truncated receptor must be undertaken to determine whether it does encode a functional GHBP.

The other potential GHBP-encoding transcript in the chicken is an alternatively spliced product of the GHR gene. A clone containing a 17-bp segment inserted in between the sequences encoded by exons 6 and 7 (53) was isolated from a cDNA library. Expression of transcripts containing the insertion was confirmed by RT-PCR. Sequencing of the chicken genomic DNA showed that the inserted segment matched the sequence of the final 17 bases of intron 6/7, suggesting that the transcript was produced by use of an alternative 3'-splice acceptor site 17 bp upstream of exon 7 (Fig. 10). Although the transcript contains transmembrane and intracellular domain-encoding sequences, the insertion causes a frame shift that would produce a truncated GHR protein containing the first 177 of the receptor’s 221 extracellular amino acids followed by 6 novel amino acids before a stop codon is reached. The presence of 80% of the hormone-binding domain and the lack of the transmembrane and intracellular domains led the authors to hypothesize that the protein encoded by this mRNA might function as a secreted GHBP.

2. Truncations of the intracellular domain. Several species express GHR transcripts with deletions or truncations in the intracellular domain-encoding region. Because the intracellular domain of the receptor is responsible for initiating signal transduction, these alterations could potentially cause an increase or decrease in the normal cellular response to GH or cause a nonstandard cellular response to GH. The first report of a GHR with alternations in the intracellular domain was made when the rabbit GHR cDNA was cloned (6). In addition to the full-length GHR cDNA, cDNAs encoding a receptor with a truncated intracellular domain were also cloned. These cDNAs encoded the full-length receptor’s extracellular and transmembrane domains and the first 4 amino acids of the intracellular domain. Beyond the fourth intracellular amino acid, they encoded 4 novel amino acids (not found in full-length receptor) before a stop codon was encountered, making the intracellular domain in these transcripts a total of 8 amino acids in length (compared with the 350-amino acid intracellular domain of the full-length rabbit GHR). Exon 8 of the human GHR gene encodes the transmembrane domain and the first four amino acids of the intracellular domain. The point of divergence of the truncated rabbit clones is therefore exactly coincident with the 3'-end of exon 8 in the human gene, suggesting that alternative splicing involving the splice donor site of exon 8 of the rabbit gene is the mechanism by which these clones were generated.

The presence of GHR transcripts with divergent sequences in the intracellular domain was confirmed in the rabbit and detected also in humans, rats, and mice by PCR (113). Primers complementary to sequences in human exon 7 and exon 10 were synthesized and used to amplify the region of the human GHR cDNA corresponding to exons 8 and 9. The predicted 295-bp amplification product (i.e., full-length receptor containing exons 8 and 9) was observed. In addition, a smaller 269-bp product was also observed. Cloning and sequencing the smaller product showed that it contained a deletion of the first 26 bp of exon 9 sequence. This resulted in a frame shift that would give rise to a truncated receptor containing the first 273 amino acids of GHR (the 246-amino acid extracellular hormone-binding domain, the 24-amino acid transmembrane domain, and the first 3 amino acids of the intracellular domain), followed by 6 novel intracellular amino acids before a stop codon is read. Because the total number of amino acids in the truncated GHR protein would be 279, it will be referred to as the human GHR₂₇₉. The proposed alternative splicing mechanism by which the GHR₂₇₉ transcript is produced is diagrammed in Fig. 11. The splice donor site of exon 8 can be joined either to the canonical splice acceptor site at the 5'-end of exon 9 to generate the full-length transcript or to an alternative splice acceptor site located 26 bp within exon 9 (generating the GHR₂₇₉ transcript). The internal exon 9 splice site sequence resembles the mammalian 3'-splice acceptor consensus sequence (5'-12(T/C)NCAG-3') found at the 3'-end of most introns (114), further supporting the proposed alternative splicing model. The GHR₂₇₉ transcript was found to be expressed at varying levels in liver, mammary gland, adipose, placenta, lung, kidney, stomach, and muscle. In the mammary gland, adipose tissue, and placenta it was comparable in level to the full-length transcript. In liver and kidney its levels were lower, the hepatic levels being less than 10% of that of the full-length transcript (115). In lung, stomach, and muscle the GHR₂₇₉ transcript was barely detectable. Transcripts analogous to the human GHR₂₇₉ transcript were detected in the rabbit, mouse, and rat using the same RT-PCR method described above. The GHR₂₇₉ transcript in all three species had a deletion of the first 26 bases of exon 9, suggesting that the same alternative splicing mechanism proposed for the human transcript also operates in these species.

View larger version (16K): [in this window] [in a new window]   Figure 11. Generation of the human GHR279 by use of an alternative splice acceptor site located within exon 9. A, The portion of the human GHR gene containing exons 8 and 9. Exon and intron symbols are explained in the legend of Fig. 2. Dotted lines represent alternative splicing pathways. Splicing of exon 8 to the 5'-end of exon 9 (upper dotted lines) generates the full-length GHR transcript. Splicing of exon 8 to the splice-acceptor site 26 bp within exon 9 (lower dotted lines) generates the GHR279 transcript. B, The nucleotide sequence surrounding the exon 8/exon 9 splice site and the corresponding protein sequence. Diagram symbols are explained in the legend of Fig. 9. Striped lines joining the GHR and GHR279 diagrams indicate the GHR sequence that is deleted from the GHR279 sequence.

The affinity of GH-GHR₂₇₉ binding was measured by Scatchard analysis. Interestingly, the study in 293 cells showed that full-length GHR had 2-fold higher association constant (K_a) values than GHR₂₇₉, while in the COS-7 cell study, GHR₂₇₉ had a 2-fold higher K_a than full-length GHR. The difference in relative K_a values might simply reflect variations in analysis technique or sample purity, or it might indicate actual differences in K_a, perhaps due to dissimilar posttranslational modification of GHR in the two cell types.

The functional properties of the truncated receptor were tested in the 293 cell line. A luciferase reporter gene ligated to a promoter containing the binding consensus sequence for the transcription factor STAT5 was cotransfected into the 293 cells with GHR₂₇₉ or full-length GHR cDNAs. Cells transfected with full-length GHR showed a 12-fold induction in luciferase activity when stimulated by GH, indicating that the full-length receptor expressed in 293 cells was functional with regard to activating STAT5. Cells transfected with GHR₂₇₉ cDNA were unable to stimulate any increase in luciferase activity, consistent with experiments showing that the intracellular domain amino acids missing in GHR₂₇₉ are necessary for STAT5 activation (119, 120, 121). Simultaneous transfection with GHR₂₇₉ and full-length GHR cDNAs resulted in a lower level of luciferase activity than by transfection with full-length GHR alone, indicating that GHR₂₇₉ acts in vitro as an inhibitor of STAT5 activation by the full-length receptor. This finding raises the possibility that GHRs with truncated intracellular domains might function in vivo as attenuators of GH signal transduction. Normal GHR signal transduction requires GH-mediated dimerization of plasma membrane GHR (reviewed in Ref. 122). The authors hypothesized that the inhibition might be due to either GHR₂₇₉ competing with the full-length receptor for GH or to GH-mediated formation of GHR₂₇₉/full-length GHR heterodimers that were not able to initiate signal transduction. In support of the latter mechanism of inhibition, immunoprecipitation and Western blotting revealed that the two GHR types did form heterodi