Expression of Rabbit Sex Hormone-Binding Globulin during Pregnancy and Prenatal Development and Identification of a Novel Isoform
http://www.100md.com
内分泌学杂志 2005年第4期
Department of Zoology, The University of Hong Kong, Pokfulam, Hong Kong
Address all correspondence and requests for reprints to: Will M. Lee, Department of Zoology, The University of Hong Kong, Pokfulam Road, Hong Kong, China. E-mail: hrszlwm@hku.hk.
Abstract
SHBG is a homodimeric plasma glycoprotein. It functions as a carrier for sex steroids in blood and regulates their access to target cells. In human and rabbit, SHBG is a single-copy gene comprised of eight exons and is expressed primarily in the liver and testis. In the present study, the ontogeny of rabbit SHBG (rbSHBG) gene expression was examined in both fetus and mothers. Trace amounts of rbSHBG mRNA were detected in fetal liver from d 11 to d 29 gestation. These levels increased dramatically at d 30 and remained high until parturition (d 33). In contrast, high levels of rbSHBG mRNA were detected in the maternal liver early during pregnancy, with maximal levels being attained by d 22 and declining markedly thereafter. A rbSHBG transcript lacking the exon 4 sequences was consistently expressed along with the rbSHBG mRNA. When expressed as a glutathione-S-transferase-fusion protein, this alternatively spliced rbSHBG transcript resulted in a product with almost no steroid binding activity, unlike the full-length rbSHBG-glutathione-S-transferase fusion protein, which bound 5-dihydrotestosterone. Antibody specific to the novel rbSHBG isoform lacking the exon 4-encoding domain was raised, and a single immunoreactive protein of 33–35 kDa was detected by Western blot analysis in both fetal and maternal liver, and this indicates that the rbSHBG transcripts lacking exon 4 sequences are translated in vivo. An RT-PCR analysis further revealed that this alternatively spliced SHBG transcript is present in human HepG2 cells as well as human and mouse testes, indicating that exon 4 splicing in SHBG transcription is conserved among mammalian species. To our knowledge, this is the first report of the identification of a SHBG exon 4 splice variant that is translated. Because the SHBG isoform it encodes lacks appreciable steroid-binding activity, it may function beyond that of the widely accepted role of SHBG as a steroid-transport protein.
Introduction
IN MAMMALS, SEX HORMONES play a critical role in controlling phenotypic gender. During fetal development, testosterone produced by the fetal testes is required to initiate the differentiation of Wolffian ducts into male internal genital tract, whereas the development of the female urogenital tract can occur in the absence of gonadal hormones (1). As a result, testosterone is one of the major hormones that influences male sexual differentiation. According to the free hormone hypothesis, only free or unbound hormones are bioavailable. Plasma SHBG is an extracellular protein produced by hepatocytes that binds 5-dihydrotestosterone (DHT), testosterone, and estradiol with high affinity and specificity (2). Thus, SHBG is expected to play an important role in the process of fetal sexual development via its regulation of the bioavailability of sex steroids (for reviews, see Refs. 3, 4, 5).
In humans and rabbits, the SHBG gene is expressed in the testis as well as the liver (6, 7), and the expression of human SHBG in liver and testis is regulated in a tissue-specific manner, as demonstrated by studies of transgenic mice carrying different fragments of the human SHBG (8, 9, 10). Unlike human and rabbit, adult rodents (rats, mice, and hamsters) produce only testicular SHBG, which is otherwise known as the androgen-binding protein, whereas the SHBG in rodents is expressed transiently only in the liver during fetal development (11, 12, 13, 14). Because SHBG is found in the blood plasma of adult humans and rabbits, the rabbit represents a useful model for studies of the production of SHBG and its function in the blood.
The rabbit (rb) SHBG gene is expressed postnatally in both the liver and testis (15, 16, 17). Although SHBG is expressed in the rat brain (18), human placenta (19), and human uterine endometrium (20), the major sites of SHBG synthesis in adult mammals is undoubtedly either liver or testis, depending on the species, and these tissues are responsible for the production and secretion of SHBG into blood and male reproductive tract, respectively (3). It has also been suggested that SHBG may play an important role in the postnatal gonadal development (17). However, the pattern of prenatal expression of SHBG in humans and rabbits and the role of SHBG in prenatal gonadal development remain unknown.
In the present study, we studied the ontogeny of rbSHBG expression in the liver during pregnancy. The levels of SHBG mRNA expression in maternal and fetal rabbits were examined using Northern blotting and RT-PCR, and maternal plasma SHBG levels were measured using a steroid-binding assay. Our results provided insight into the origins of SHBG during sexual development in the fetus. In addition, we discovered that a SHBG transcript lacking exon 4 sequences is present in the liver and testis. The relative amounts of SHBG mRNA and this alternatively spliced SHBG transcript in the liver are approximately 5:1, as estimated by semiquantitative RT-PCR analysis. We also obtained evidence that the SHBG transcripts lacking exon 4 sequences are translated, and we obtained evidence that its translation product does not bind steroid.
Materials and Methods
Animals
New Zealand White rabbits (d 11–32 of pregnancy and at term) were used to study the prenatal and maternal expression of rbSHBG mRNA. New Zealand white rabbits (males and females, aged 7 d and 15 wk) were used to study the tissue distribution of rbSHBG mRNA and protein. BALB/c mice (females, aged 90 d) were used in antiserum production. Rabbits were killed by cervical dislocation after ether anesthesia. Tissues were removed immediately and processed for RNA or protein extraction. The use of animals for this study was approved by the Committee on the Use of Live Animals in Teaching and Research, The University of Hong Kong (approved protocol no. CULATR 499-00).
RNA extraction and Northern blot analysis
Total RNA was extracted from embryo, fetal, and maternal livers using TRIzol reagent (Invitrogen, Carlsbad, CA) following the protocol suggested by the manufacturer. Northern blot analysis for rabbit SHBG mRNA was performed as described previously (21). Samples of RNA (50 μg) were resolved on a 1% formaldehyde-agarose gel and capillary blotted onto nylon membrane (Hybond-N; Amersham Bioscience, Piscataway, NJ). The RNA on the membrane was immobilized by UV cross-linking, followed by hybridization with -32P-labeled rabbit SHBG cDNA or ribosomal protein S16 cDNA probe at 42 C overnight as described previously (16, 22). Radioactivity of hybridized mRNA species was visualized by autoradiography for 2–4 d at –80 C (BioMax films; Eastman Kodak, Rochester, NY).
Steroid-binding analysis
The steroid-binding capacity of rabbit maternal sera and the recombinant SHBG isoforms were determined by saturation analysis as described previously (16). In brief, diluted sera or SHBG in solution were incubated with dextran-coated charcoal (DCC) solution before assay to remove any endogenous steroids. They were then incubated with 10 nM [3H]DHT (SA 42 Ci/mmol; NEN Life Science Products, Boston, MA). Nonspecific binding was assessed in parallel by the addition of a 200-fold excess of unlabeled DHT. Ice-cold DCC (500 μl) was then added to separate bound and free steroid. The dissociation rate of the [3H]DHT-SHBG complex was determined by exposure to DCC for increasing time at 0 C. The steroid-binding capacity of the rabbit maternal sera was calculated and expressed as the binding at zero time exposure to DCC.
RT-PCR and Southern blot analysis
Total RNA was reverse transcribed as described previously (21). Rabbit SHBG and S16 cDNAs were coamplified from 2 μl of the reverse transcription product with 100 pmol/ml of each rbSHBG primer (sense: 5'-GGGATTCAGAGGGAGTGCTTTT; antisense: 5'-GGTCTTGAGCCAGGGGTCAA or sense: 5'-ATGGCTACTCCGCCACTCGTGCCG; antisense: 5'-AATGGCACCGACACCTCCCATTAA) and 7.5 pmol/ml of each ribosomal S16 primer (sense: 5'-TCCGCTGCAGTCCGTTCA-AGTCTT; antisense: 5'-GCCAAACTTCTTGGATTCGCAGCG). The PCR was performed with denaturation at 94 C for 1 min, annealing at 64 C for 1.5 min, and extension at 72 C for 2 min for a total of 22 cycles, after a final extension at 72 C for 15 min. Under these conditions, the amplifications of both rbSHBG and S16 were in linear phase. The authenticity of the RT-PCR products was confirmed by Southern analysis. Southern hybridization was performed with the -32P-labeled rabbit SHBG or S16 cDNA probes used in the Northern blot analysis, as described previously (23). Radioactivity of hybridized products was visualized by autoradiography for 2–4 d at –80 C (BioMax films; Eastman Kodak).
Sequencing analysis of RT-PCR product
The rbSHBG mRNA and alternatively spliced rbSHBG transcript were amplified by RT-PCR using primers flanking exons 1 and 8 of rbSHBG (sense: 5'-ATGGCTACTCCGCCACTCGTGCCG; antisense: 5'-AATGGCACCGACACCTCCCATTAA). The PCR products were purified using a High Pure PCR product purification kit (Roche, Stockholm, Sweden), and sequence analysis was conducted by cycle sequencing using an ABI PRISM BigDye Terminators cycle sequencing kit (version 3.0; Applied Biosystems, Foster City, CA) and the ABI 3100 genetic analyzer (Applied Biosystems). The cDNA sequence of the alternatively spliced rbSHBG transcript was aligned against the sequence of the rbSHBG mRNA sequence with GeneTool software (version 2.0; BioTools Inc., Edmonton, CA).
Expression, purification, and steroid binding of recombinant rbSHBG isoforms
The reverse transcription product prepared using mature female liver RNA was used as template for PCR using primers flanking exons 2 and 8 of rbSHBG (sense: 5'-GGATCCCTGAGAAGCGTTCTGGCC; antisense: 5'-CCATGGATTAATGGGAGGTGTCGG), to produce cDNA fragments containing a BamHI site (underlined) immediately upstream the codon for Leu1, and an NcoI site (underlined) immediately downstream the stop codon of rbSHBG cDNA. The PCR products obtained were digested with BamHI and NcoI and subcloned into the same sites of the bacterial expression vector pGEX-KGK to produce constructs encoding the rbSHBG isoforms fused in-frame with a glutathione-S-transferase (GST) sequence. Escherichia coli strain BL21 transformed with the constructs were used for fusion protein expression. The recombinant rbSHBG isoforms were expressed, purified, and subjected to SDS-PAGE and Western blot analyses as described previously (16, 24). The binding capacities of the affinity-purified recombinant rbSHBG isoforms for [3H]DHT were determined essentially as described above by using 100 ng of each fusion protein. The relative steroid-binding capacities of the recombinant isoforms were expressed as specifically bound [3H]DHT in counts per minute.
Production of antisera against the rbSHBG isoform encoding by the alternatively spliced rbSHBG transcript
A multiple antigenic peptide (MAP)-conjugated synthetic peptide (Invitrogen) corresponding to the 10-amino acid residues (GSWHQLVPAL) encoded by the nucleotide sequences spanning the exon 3–5 junction of alternative rbSHBG transcript was used to raise antibody against a truncated rbSHBG isoform lacking the exon 4-encoding domain. For immunization, synthetic peptide (150 μg) was administered to BALB/c mice via sc injection, and antiserum was collected 7 d after booster injection.
Immunological detection of the rbSHBG isoform
Rabbit fetal and maternal livers collected at different times during pregnancy were homogenized by sonication in lysis buffer [50 mM Tris, 2 mM EDTA, 0.15 mM NaCl, 2 mM phenylmethylsulfonyl fluoride, 1% Nonidet P-40, and 10% glycerol (pH 7.4)]. Soluble proteins were recovered by centrifugation and quantified using the Bradford protein assay. Samples were resolved by 12% SDS-PAGE and blotted onto nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA). After blocking with PBS containing 5% (wt/vol) skimmed milk, the membranes were incubated in specific antibodies against the MAP-conjugated synthetic peptide (1:2500 dilution of the antiserum) at 4 C overnight. Unbound antibodies were removed by extensive washing with PBS supplemented with 0.05% Tween 20 (vol/vol). The membranes were incubated with alkaline phosphatase-conjugated goat antimouse IgG antibody (1:5000 dilution) (Zymed Laboratories Inc., San Francisco, CA), and immunoreactive proteins on the blots were visualized by chromogenic detection using nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate.
Results
Prenatal and maternal rbSHBG expression
Female rabbits, from 11 to 32 d of pregnancy and 1 d after delivery were studied, and the SHBG mRNA abundance in the embryo, fetal, and maternal livers was examined by Northern blot analysis. A 1.6-kb SHBG mRNA was detected in most of the maternal and fetal liver samples examined (Fig. 1). This SHBG mRNA was not detected in the liver during the early stage of embryonic development and first appeared in the fetal liver at d 25 gestation (Fig. 1A). Its abundance in the fetal liver remained low until d 30 gestation, at which time the levels increased dramatically until parturition (d 33). In the maternal liver, however, high levels of SHBG mRNA are present during early pregnancy from d 11, increasing to a maximum at d 22, and declining thereafter apart from a slight rebound of SHBG mRNA level during d 28 and 29 (Fig. 1B). Serum SHBG levels in pregnant rabbits exhibited a profile similar to that observed for hepatic SHBG mRNA levels in the same animals. The plasma SHBG level reached a maximum at d 22; it then declined rapidly thereafter and remained low despite a slight rebound during the late stage of pregnancy before parturition (Fig. 2).
FIG. 1. Northern blot analysis of rbSHBG mRNA in embryo, fetal, and maternal livers. Total RNA was extracted from rabbit embryos (d 11–20 of pregnancy), fetal livers (d 17–32 of pregnancy), newborn livers (NB, d 1 of birth) (A), and maternal livers from d 11–32 of pregnancy (B). A 1.6-kb SHBG mRNA was detected after 2 d of autoradiographic exposure at –80 C. The blots were washed and reprobed with -32P-labeled S16 cDNA to detect the expression of the ribosomal protein S16 to demonstrate even loading of the samples. The figure shows representative results from one of three independent experiments.
FIG. 2. Serum SHBG levels in pregnant rabbits during pregnancy. Serum SHBG was measured by a steroid-binding capacity assay. Results shown are means ± SD from five independent experiments of pooled animals.
Levels of SHBG mRNA in the fetal and maternal livers during pregnancy, in particular during the early fetal development, were also studied using a more sensitive RT-PCR assay. Results were similar to those obtained by Northern blot analysis (Figs. 3 and 4), but trace amounts of SHBG mRNA were detected in embryo and fetal liver as early as d 14 and 17 gestation, respectively, by RT-PCR (Fig. 3A). In this assay, specific amplification of SHBG mRNA produced a major PCR product of 732 bp. However, a minor PCR product of 570 bp was also identified in all liver samples examined, and this was confirmed to be a rbSHBG transcript by Southern hybridization (Figs. 3B and 4B).
FIG. 3. RT-PCR analysis of rabbit SHBG mRNA levels in embryo and fetal liver. A, RT-PCRs were performed from 10 μg total RNA of rabbit embryo, fetal liver, or newborn liver (NB, d 1 of birth). They were reversely transcribed and subsequently amplified by primers prepared from internal nucleotide sequence to give a major PCR product of 732 bp and a minor PCR product of 570 bp. Coamplification of ribosomal S16 cDNA yielded a 385-bp product. B, The authenticity of the RT-PCR fragments was confirmed by Southern blot hybridization using -32P-labeled rabbit SHBG cDNA. The blots were washed and reprobed with -32P-labeled S16 cDNA. NB, Newborn; –ve, negative control in which reverse transcriptase was omitted from the d 11 sample. The figure shows representative results from one of three independent experiments.
FIG. 4. RT-PCR analysis of rbSHBG mRNA levels in maternal liver. A, RT-PCRs were performed from 10 μg total RNA of pregnant rabbit liver. They were reversely transcribed and subsequently amplified by primers prepared from internal nucleotide sequence to give a major PCR product of 732 bp and a minor PCR product of 570 bp. Coamplification of ribosomal S16 cDNA yielded a 385-bp product. B, The authenticity of the RT-PCR fragments was confirmed by Southern blot hybridization using -32P-labeled rabbit SHBG cDNA. The blots were washed and reprobed with -32P-labeled S16 cDNA. NB, Newborn; –ve, negative control in which reverse transcriptase was omitted from the d 11 sample. The figure shows representative results from one of three independent experiments.
Sequence analysis and tissue distribution of the alternative rbSHBG transcript
When the nucleotide sequence of the PR-PCR product for the alternately spliced rbSHBG transcript was aligned with the rbSHBG mRNA sequence, it was found to contain the same nucleotide sequence as the rbSHBG mRNA, apart from a 162-bp deletion corresponding to the entire exon 4 of rbSHBG (data not shown). This alternative SHBG transcript is most likely produced by in-frame splicing of exon 4 because there is only a single gene for SHBG in rabbit (7). This exon 4-skipped rbSHBG transcript is present not only in the fetal and adult livers but also was found in the testis (Fig. 5), and the abundance of alternatively spliced rbSHBG transcript relative to that of the rbSHBG mRNA is approximately 1:5 in these tissues. By contrast, only trace amounts of SHBG mRNA were detected in the brain and kidney of neonatal and young male rabbit by sequential RT-PCR and Southern blot hybridization (Fig. 5).
FIG. 5. Tissue distribution of rbSHBG transcripts examined by RT-PCR analysis. The rbSHBG expression in various tissues of the fetus (d 30 gestation) and male (young, 7 d old; adult, 15 wk old) and female (young, 7 d old; adult, 15 wk old) rabbits was studied by RT-PCR. Amplification with primers flanking exons 1 and 8 of rbSHBG give a major PCR product of 1197 bp and a minor PCR product of 1035 bp. Authenticities of these products were subsequently confirmed by Southern blot hybridization using -32P-labeled rabbit SHBG cDNA. B, Brain; K, kidney; Li, liver; T, testis. The figure shows representative results from three independent experiments.
Steroid-binding analysis of rbSHBG isoforms
The rbSHBG mRNA and rbSHBG transcripts lacking the exon 4 sequences were expressed as GST-fusion proteins in bacterial cells. After affinity purification, the corresponding major Coomassie-stained products on an SDS-PAGE gel were approximately 67 and 61 kDa, respectively (Fig. 6A). These proteins were identified as GST-fusion proteins by Western blot analysis using anti-GST antibody (Fig. 6A). Because the GST domain of recombinant rbSHBG fusion proteins does not affect its steroid binding activity (16), the steroid-binding capacities of the recombinant fusion proteins were analyzed directly by DHT-binding assay without removal of the GST domain. The results show that the GST-fused recombinant rbSHBG lacking the 54 amino acid residues encoded by the exon 4 sequence has essentially no affinity for DHT, when compared with the recombinant full-length rbSHBG-GST fusion protein (Fig. 6B).
FIG. 6. A, Expression of GST fusion SHBG isoforms. Affinity-purified GST-fusion rbSHBG isoforms were analyzed by SDS-PAGE, followed by Coomassie blue staining and Western blot with anti-GST antibody (Ab). Lanes 1 and 3 correspond to GST-fused full-length rbSHBG; whereas lanes 2 and 4 correspond to GST-fused truncated rbSHBG lacking the exon 4-encoding domain. B, Steroid binding analysis of the affinity-purified GST-fusion proteins. The relative affinities of the recombinant rbSHBG isoforms for DHT were expressed as the specific [3H]DHT bound after exposure to DCC for 10 min at 0 C. Results shown are mean ± SD from three individual experiments. M.W., Molecular weight.
Immunological detection of the truncated SHBG protein
The antiserum against the truncated rbSHBG polypeptide encoded by the exon 4-skipped rbSHBG transcript was able to differentiate between the full-length rbSHBG polypeptide and the novel SHBG isoform (Fig. 7A). This antiserum was then used to detect this SHBG isoform in fetal and maternal liver lysates prepared at different stages of pregnancy. A single immunoreactive protein of about 33–35 kDa was detected in maternal and fetal liver samples by Western blot analysis (Fig. 7B). This immunoreactive band was present in the maternal liver lysates from d 11 to d 32 of pregnancy. The strongest signal was observed on d 11, which then gradually decreased until parturition. These results correlate quite well with the RT-PCR analysis of the exon 4-skipped rbSHBG transcript levels in the maternal liver (Fig. 3). For the fetal liver lysates, the 33- to 35-kDa immunoreactive protein was detected on d 31 and 32 gestation (Fig. 7B), and no signal could be detected before d 29 of pregnancy (data not shown), despite the fact that as much as 100 μg of total soluble proteins (compared with 40 μg for the maternal liver) were analyzed for these samples. This may be due to lower levels of the alternatively spliced rbSHBG transcripts present in fetal liver, compared with the maternal liver.
FIG. 7. Western blot analysis of the truncated rbSHBG protein lacking the exon 4-encoding domain. Purified GST-fused recombinant rbSHBG proteins (A), total lysates prepared from maternal and fetal livers, respectively (B), and total soluble protein from various tissues of male and female rabbits (C) were resolved by 12% SDS-PAGE and transferred onto nitrocellulose membrane. Western blot was then performed using mouse antiserum against the novel rbSHBG isoform as primary antibody. Immunoreactive protein-antibody complex was then detected using alkaline phosphatase-conjugated goat antimouse IgG antibody and visualized by chromogenic detection using nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate as substrate. B, Brain; K, kidney; Li, liver; S, spleen; T, testis. The figure shows representative results from three independent experiments.
To examine the tissue distribution of the 33- to 35-kDa immunoreactive protein, other tissues from male and female rabbits were analyzed by Western blot. As shown in Fig. 7C, an immunoreactive protein of the same molecular size was observed consistently in the liver lysates of mature male and female rabbits as well as in testis lysates. The immunoreactive signal of the testis lysates was weaker than those observed in the liver lysates of the same animal. This matched the RT-PCR results in that the alternatively spliced rbSHBG transcript is less abundant in testis than liver (Fig. 5). On the other hand, no immunoreactive signal could be detected in the brain, kidney, and spleen lysates of male or female rabbits. This reflects the fact that SHBG expression in these tissues is either very low or nonexistent.
Detection of exon 4-skipped SHBG transcripts in mouse and human tissue and cells
The presence of the exon 4-skipped SHBG transcripts in mouse and human tissue and cells were examined by RT-PCR and sequence analysis using species-specific primers. As in the case of rabbit, an appropriately sized product was detected in the mouse testis (500 bp), human testes, and human HepG2 cells (537 bp) (Fig. 8). This slight difference in the sizes of the PCR products amplified between the species is due to the differences in the best annealing positions of corresponding primers. Sequence analysis and pairwise alignment confirmed that these RT-PCR products were derived from SHBG transcripts lacking exon 4 sequences (data not shown). To our knowledge, this is the first demonstration that exon 4-skipped SHBG transcripts exist in human and mouse tissues (for summary, see Table 1).
FIG. 8. Detection of the exon 4-skipped SHBG transcripts in mammalian species. Total RNA isolated the liver of mature rabbit and testes of mature rabbit, mouse, and human as well as human liver cell-line HepG2 were used as template, and RT-PCRs were performed using universal sense primer annealing to the exon 3 and species-specific antisense primers annealing to the specific regions with the exon 7 of the corresponding mature mRNA (Table 2). The major bands of about 700 bp represented the PCR product amplified from the SHBG mRNA, whereas the minor bands of about 550 bp were confirmed by automated nucleotide sequencing to be the PCR product amplified from the exon 4-skipped SHBG transcripts of the corresponding species. The figure shows representative results from three independent experiments.
TABLE 1. Summary of alternative SHBG transcripts in human, rabbit, mouse, and rat
Discussion
In this report, we presented the results of a detailed ontogenic study of rbSHBG expression in fetal and maternal livers. In the rabbit fetus, a high level of SHBG expression was observed during late stages of gestation. This suggests a marked increase in fetal liver SHBG biosynthesis during late gestation and supports the concept that changes in plasma SHBG levels could influence fetal and neonatal sexual development through modulating the amounts of testosterone available to developing male reproductive tissues (for review, see Ref. 26). In contrast, high SHBG mRNA levels were observed in the maternal liver during early pregnancy and declined markedly thereafter with a slight rebound observed at d 28 and 29. This pattern of maternal hepatic rbSHBG expression mirrored that of maternal serum SHBG levels, which is in line with the fact that the majority of plasma SHBG is made by the liver. This high level of SHBG expression in the maternal rabbit liver during early pregnancy may be due to an increased production of estrogens during this time (27) because estrogens stimulate the production and secretion of SHBG by HepG2 cells (28). The observed increase of maternal SHBG production at early pregnancy suggests that SHBG may play a role related to the development of the uterus during pregnancy. Although it is generally assumed that increasing plasma SHBG levels restricts the action of its sex steroid ligand, SHBG might act directly by interacting with membrane receptors that have been identified on endometrial cells (29, 30) and mediate a nongenomic action of estrogens (31).
To our surprise, an alternative rbSHBG transcript was detected by RT-PCR alongside rbSHBG mRNA in both the fetal and maternal livers. Sequence analysis revealed that this alternative transcript contained an in-frame deletion of 162 bp corresponding to the entire exon 4 of the rbSHBG. This exon 4-skipped rbSHBG transcript was also expressed concurrently with the rbSHBG mRNA in testis and liver of young and mature male and female rabbits. The ratios of the alternative rbSHBG transcript to the rbSHBG mRNA are similar (1:5) in all tissues, suggesting there is no tissue specificity in the production of the exon 4-skipped rbSHBG transcript.
In human, the 54 amino acid residues (Val103-Pro156) encoded by the 162 bp of the exon 4 are critical structural elements of the steroid-binding pocket of the N-terminal laminin G-like (LG) domain (for reviews, see Refs. 32, 33, 34, 35). Thus, deletion of these 54 amino acid residues was expected to disrupt the steroid-binding site, and this was confirmed by expressing and studying the steroid-binding properties of a GST-fused truncated rbSHBG protein lacking the exon 4-encoding domain. Unlike the full-length rbSHBG-GST fusion protein, the truncated rbSHBG-GST fusion protein does not bind steroid. This suggests that if the rbSHBG transcript lacking exon 4 sequences is translated in vivo, it probably functions in ways that are not directly related to steroid transport.
To determine whether the exon 4-skipped SHBG transcript is translated in vivo, we generated an antiserum capable of discriminating between the rbSHBG isoform lacking the exon 4-encoding domain and the mature rbSHBG polypeptide. This antiserum detected an immunoreactive rbSHBG isoform in the cell lysates of maternal and fetal liver as well as the testis of adult rabbit, in which the alternatively spliced rbSHBG transcript was most abundant. The size of this immunoreactive protein was about 34 kDa, which matches its deduced size of 33.7 kDa. The size of the full-length rbSHBG in the circulation that has been reported ranged from 41–45 kDa, dependent on the degree of glycosylation (36). Because the deletion of 54 amino acid residues in exon 4 would lead to a reduction of about 6 kDa, the truncated rbSHBG detected in the Western blot is most likely either not glycosylated or incompletely glycosylated. Differential glycosylation of SHBG has been reported in human, rat, rabbit, and zebrafish (10, 11, 36, 37), and the actual glycosylation status of rbSHBG isoform in tissues therefore remains to be determined.
Another question we sought to address is the localization of the truncated rbSHBG protein. In human and rabbit, hepatic SHBG is normally rapidly secreted into the blood, in which it functions as a steroid-transporting protein. RT-PCR and sequence analysis indicated that the exon 4-skipped rbSHBG transcript contains all the other seven exons of the full-length transcript, including exon 1, which encodes the signal peptide of SHBG. Therefore, the truncated rbSHBG made in the liver was expected to be secreted. Attempts have been made to detect the truncated SHBG in crude rabbit serum, but these were not successful. However, our Western blot results suggest that the truncated rbSHBG protein accumulates in the tissues in which it is made.
With respect to its physiological role, the novel rbSHBG isoform is unlikely to provide a steroid transportation function. Nevertheless, this rbSHBG isoform contains a complete C-terminal LG domain, and this may be functionally important. This is of interest because the C-terminal LG domain of SHBG resembles the LG 5 domain structure within the C-terminal SHBG-like regions of protein S and Gas 6. The SHBG-like regions of protein S and Gas 6 have been reported to interact with the C4b-binding protein (38) and the Tyro3 family receptor tyrosine kinases (39), respectively, and this may be mediated via protein-protein interactions involving this C-terminal LG domain.
Although the novel rbSHBG isoform lacks steroid-binding activity, its expression profile during prenatal development, and the occurrences of exon 4-skipped SHBG transcripts in mouse and human tissues and cells, suggests that this novel SHBG isoform might have a biological significance beyond a steroid-transporting protein.
TABLE 2. Primers used in the detection of SHBG transcripts in mammalian species
Acknowledgments
The authors thank Dr. Ping-Fai Ki for the special technical contributions in animal handling and antibody production and Ms. Connie Wong for the general technical comments and suggestions.
References
Porterfield SP 2001 Male reproductive system. In: Schmitt W, ed. Endocrine physiology: the Mosby physiology monograph series. St. Louis: Mosby, Inc.; 153–176
Westphal U 1986 Steroid-protein interactions II. In: Monographs on endocrinology. Berlin: Springer-Verlag; 204–209
Hammond GL 1993 Extracellular steroid-binding proteins In: Parker M, ed. Frontiers in molecular biology, steroid hormone action. Oxford: IRL; 12–15
Joseph DR 1994 Structure, function and regulation of androgen-binding protein/sex hormone-binding globulin. Vitam Horm 49:197–280
Hammond GL 2002 Access of reproductive steroids to target tissues. Obstet Gynecol Clin North Am 29:411–423
Hammond GL, Underhill DA, Rykse HM, Smith CL 1989 The human sex hormone-binding globulin gene contains exons for androgen-binding protein and two other testicular messenger RNAs. Mol Endocrinol 3:1869–1876
Ip YC, Lee WM, Hammond GL 2000 The rabbit sex hormone-binding globulin gene: structural organization and characterization of its 5' flanking region. Endocrinology 141:1356–1365
Janne M, Deol HK, Power SG, Yee SP, Hammond GL 1998 Human sex hormone-binding globulin gene expression in transgenic mice. Mol Endocrinol 12:123–136
Janne M, Hogeveen KN, Deol HK, Hammond GL 1999 Expression and regulation of human sex hormone-binding globulin transgenes in mice during development. Endocrinology 140:4166–4174
Selva DM, Hogeveen KN, Seguchi K, Tekpetey F, Hammond GL 2002 A human sex hormone-binding globulin isoform accumulates in the acrosome during spermatogenesis. J Biol Chem 277:45291–45298
Sullivan PM, Petrusz P, Szpirer C, Joseph DR 1991 Alternative processing of androgen-binding protein RNA transcripts in fetal rat liver. J Biol Chem 266:143–154
Wang YM, Sullivan PM, Petrusz P, Yarbrough W, Joseph DR 1989 The androgen binding protein gene is expressed in CD1 mouse testis. Mol Cell Endocrinol 63:85–92
Cates JM, Damassa DA, Gagin GA, Dempsey RV 1995 Hepatic expression of sex hormone-binding globulin associated with the postnatal surge of serum androgen-binding activity in the Djungarian hamster. J Steroid Biochem Mol Biol 55:147–158
Cates JM, Damassa DA 1997 Characterization and developmental expression patterns of testicular androgen-binding protein in the Djungarian hamster (Phodopus sungorus). J Reprod Fertil 111:291–298
Khan MS, Knowles BB, Aden DP, Rosner W 1982 Secretion of testosterone-estradiol-binding globulin by a human hepatoma-derived cell line. J Clin Endocrinol Metab 53:448–452
Lee WM, Wong AST, Tu AWK, Cheung C-H, Li JCH, Hammond GL 1997 Rabbit sex hormone-binding globulin: primary structure, tissue expression, and structure/function analyses by expression in Escherichia coli. J Endocrinol 153:373–384
Wong AST, Lui WY, Hui ITC, Lee WM 2001 Rabbit sex hormone-binding globulin: expression in the liver and testis during postnatal development and structural characterization by truncated proteins. Int J Androl 24:165–174
Wang YM, Bayliss DA, Millhorn DE, Petrus Z, Joseph DR 1990 The androgen-binding protein gene is expressed in male and female rat brain. Endocrinology 127:3124–3130
Larrea F, Diaz L, Carino C 1993 Evidence that human placenta is a site of sex hormone-binding globulin gene expression. J Steroid Biochem Mol Biol 46:497–505
Misao R, Itoh N, Mori H, Fujimoto J, Tatnaya T 1994 Sex hormone-binding globulin mRNA levels in human uterine endometrium. Eur J Endocrinol 131:623–629
Wong CC, Lee WM 2002 The proximal cis-acting elements Sp1, Sp3 and E2F regulate mouse mer gene transcription in Sertoli cells. Eur J Biochem 269:3789–3800
Chan YL, Paz V, Olvera J, Wool IG 1990 The primary structure of rat ribosomal protein S16. FEBS Lett 263:85–88
Cheng CK, Cheung CH, Lee WM 2003 Mouse testin: complementary DNA cloning, genomic organization, and characterization of its proximal promoter region. Biol Reprod 68:1376–1386
Hildebrand C, Bocchinfuso WP, Dales D, Hammond GL 1995 Resolution of the steroid-binding and dimerization domains of human sex hormone-binding globulin by expression in Escherichia coli. Biochemistry 34:3231–3238
Misao R, Iwagaki S, Fujimoto J, Nakanishi Y, Sun WS, Tamaya T 1999 Expression of sex hormone-binding globulin exon 7 splicing variant mRNA in secondary spreading lesions of gynecological cancers. J Steroid Biochem Mol Biol 68:103–109
Renfree MB, Wilson JD, Shaw G 2002 The hormonal control of sexual development. Novartis Found Symp 244:136–152
Challis JR, Davies J, Ryan KJ, Parker CR 1973 The concentrations of progesterone, estrone and estradiol-17? in the plasma of pregnant rabbits. Endocrinology 93:971–976
Rosner W, Aden DP, Khan MS 1984 Hormonal influences on the secretion of steroid-binding proteins by a human hepatoma-derived cell line. J Clin Endocrinol Metab 59:806–808
Nakhla AM, Khan MS, Rosner W 1990 Biologically active steroids activate receptors-bound human sex hormone binding globulin to cause LNCaP cells to accumulate adenosine 3', 5'-monophosphate. J Clin Endocrinol Metab 71:398–404
Rosner W, Hryb DJ, Khan MS, Nakhla AM, Romas NA 1999 Sex hormone-binding globulin mediates steroid hormone signal transduction at the plasma membrane. J Steroid Biochem Mol Biol 69:481–485
Avvakumov GV, Zhuk NI, Strel’chyonok OA 1986 Subcellular distribution and selectivity of the protein-binding component of the recognition system for sex-hormone-binding protein-estradiol complex in human decidual endometrium. Biochim Biophys Acta 881:489–498
Grishkovskaya I, Avvakumov GV, Sklenar G, Dales D, Hammond GL, Muller YA 2000 Crystal structure of human sex hormone-binding globulin: steroid transport by a laminin G-like domain. EMBO J 19:504–512
Avvakumov GV, Grishkovskaya I, Muller YA, Hammond GL 2002 Crystal structure of human sex hormone-binding globulin in complex with 2-methoxyestradiol reveals the molecular basis of high affinity interactions with C-2 derivatives of estradiol. J Biol Chem 277:45219–45225
Grishkovskaya I, Avvakumov GV, Hammond GL, Catalano MG, Muller YA 2002 Steroid ligands bind human sex hormone-binding globulin in specific orientations and produce distinct changes in protein conformation. J Biol Chem 277:32068–32093
Hammond GL, Avvakumov GV, Muller YA 2003 Structure/function analyses of human sex hormone-binding globulin: effects of zinc on steroid-binding specificity. J Steroid Biochem Mol Biol 85:195–200
Danzo BJ, Black JH, Bell BW 1989 The microheterogeneity of rabbit testosterone-binding globulin is due to differential glycosylation of its single protomer. Biol Reprod 41:957–965
Miguel-Queralt S, Knowlton M, Avvakumov GV, Al-Nouno R, Kelly GM, Hammond GL 2004 Molecular and functional characterization of sex hormone binding globulin in zebrafish. Endocrinology 145:5221–5230
Saposnik B, Borgel D, Aiach M, Gandrille S2003 Functional properties of the sex-hormone-binding globulin (SHBG)-like domain of the anticoagulant protein S. Eur J Biochem 270:545–555
Lu Q, Gore M, Zhang Q, Camenisch T, Boast S, Casagranda F, Lai C, Skinner MK, Klein R, Matsushima GK, Earp HS, Goff SP, Lemke G 1999 Tyro-3 family receptors are essential regulators of mammalian spermatogenesis. Nature 398:723–728(Kwong-Man Ng, Man-Ting So)
Address all correspondence and requests for reprints to: Will M. Lee, Department of Zoology, The University of Hong Kong, Pokfulam Road, Hong Kong, China. E-mail: hrszlwm@hku.hk.
Abstract
SHBG is a homodimeric plasma glycoprotein. It functions as a carrier for sex steroids in blood and regulates their access to target cells. In human and rabbit, SHBG is a single-copy gene comprised of eight exons and is expressed primarily in the liver and testis. In the present study, the ontogeny of rabbit SHBG (rbSHBG) gene expression was examined in both fetus and mothers. Trace amounts of rbSHBG mRNA were detected in fetal liver from d 11 to d 29 gestation. These levels increased dramatically at d 30 and remained high until parturition (d 33). In contrast, high levels of rbSHBG mRNA were detected in the maternal liver early during pregnancy, with maximal levels being attained by d 22 and declining markedly thereafter. A rbSHBG transcript lacking the exon 4 sequences was consistently expressed along with the rbSHBG mRNA. When expressed as a glutathione-S-transferase-fusion protein, this alternatively spliced rbSHBG transcript resulted in a product with almost no steroid binding activity, unlike the full-length rbSHBG-glutathione-S-transferase fusion protein, which bound 5-dihydrotestosterone. Antibody specific to the novel rbSHBG isoform lacking the exon 4-encoding domain was raised, and a single immunoreactive protein of 33–35 kDa was detected by Western blot analysis in both fetal and maternal liver, and this indicates that the rbSHBG transcripts lacking exon 4 sequences are translated in vivo. An RT-PCR analysis further revealed that this alternatively spliced SHBG transcript is present in human HepG2 cells as well as human and mouse testes, indicating that exon 4 splicing in SHBG transcription is conserved among mammalian species. To our knowledge, this is the first report of the identification of a SHBG exon 4 splice variant that is translated. Because the SHBG isoform it encodes lacks appreciable steroid-binding activity, it may function beyond that of the widely accepted role of SHBG as a steroid-transport protein.
Introduction
IN MAMMALS, SEX HORMONES play a critical role in controlling phenotypic gender. During fetal development, testosterone produced by the fetal testes is required to initiate the differentiation of Wolffian ducts into male internal genital tract, whereas the development of the female urogenital tract can occur in the absence of gonadal hormones (1). As a result, testosterone is one of the major hormones that influences male sexual differentiation. According to the free hormone hypothesis, only free or unbound hormones are bioavailable. Plasma SHBG is an extracellular protein produced by hepatocytes that binds 5-dihydrotestosterone (DHT), testosterone, and estradiol with high affinity and specificity (2). Thus, SHBG is expected to play an important role in the process of fetal sexual development via its regulation of the bioavailability of sex steroids (for reviews, see Refs. 3, 4, 5).
In humans and rabbits, the SHBG gene is expressed in the testis as well as the liver (6, 7), and the expression of human SHBG in liver and testis is regulated in a tissue-specific manner, as demonstrated by studies of transgenic mice carrying different fragments of the human SHBG (8, 9, 10). Unlike human and rabbit, adult rodents (rats, mice, and hamsters) produce only testicular SHBG, which is otherwise known as the androgen-binding protein, whereas the SHBG in rodents is expressed transiently only in the liver during fetal development (11, 12, 13, 14). Because SHBG is found in the blood plasma of adult humans and rabbits, the rabbit represents a useful model for studies of the production of SHBG and its function in the blood.
The rabbit (rb) SHBG gene is expressed postnatally in both the liver and testis (15, 16, 17). Although SHBG is expressed in the rat brain (18), human placenta (19), and human uterine endometrium (20), the major sites of SHBG synthesis in adult mammals is undoubtedly either liver or testis, depending on the species, and these tissues are responsible for the production and secretion of SHBG into blood and male reproductive tract, respectively (3). It has also been suggested that SHBG may play an important role in the postnatal gonadal development (17). However, the pattern of prenatal expression of SHBG in humans and rabbits and the role of SHBG in prenatal gonadal development remain unknown.
In the present study, we studied the ontogeny of rbSHBG expression in the liver during pregnancy. The levels of SHBG mRNA expression in maternal and fetal rabbits were examined using Northern blotting and RT-PCR, and maternal plasma SHBG levels were measured using a steroid-binding assay. Our results provided insight into the origins of SHBG during sexual development in the fetus. In addition, we discovered that a SHBG transcript lacking exon 4 sequences is present in the liver and testis. The relative amounts of SHBG mRNA and this alternatively spliced SHBG transcript in the liver are approximately 5:1, as estimated by semiquantitative RT-PCR analysis. We also obtained evidence that the SHBG transcripts lacking exon 4 sequences are translated, and we obtained evidence that its translation product does not bind steroid.
Materials and Methods
Animals
New Zealand White rabbits (d 11–32 of pregnancy and at term) were used to study the prenatal and maternal expression of rbSHBG mRNA. New Zealand white rabbits (males and females, aged 7 d and 15 wk) were used to study the tissue distribution of rbSHBG mRNA and protein. BALB/c mice (females, aged 90 d) were used in antiserum production. Rabbits were killed by cervical dislocation after ether anesthesia. Tissues were removed immediately and processed for RNA or protein extraction. The use of animals for this study was approved by the Committee on the Use of Live Animals in Teaching and Research, The University of Hong Kong (approved protocol no. CULATR 499-00).
RNA extraction and Northern blot analysis
Total RNA was extracted from embryo, fetal, and maternal livers using TRIzol reagent (Invitrogen, Carlsbad, CA) following the protocol suggested by the manufacturer. Northern blot analysis for rabbit SHBG mRNA was performed as described previously (21). Samples of RNA (50 μg) were resolved on a 1% formaldehyde-agarose gel and capillary blotted onto nylon membrane (Hybond-N; Amersham Bioscience, Piscataway, NJ). The RNA on the membrane was immobilized by UV cross-linking, followed by hybridization with -32P-labeled rabbit SHBG cDNA or ribosomal protein S16 cDNA probe at 42 C overnight as described previously (16, 22). Radioactivity of hybridized mRNA species was visualized by autoradiography for 2–4 d at –80 C (BioMax films; Eastman Kodak, Rochester, NY).
Steroid-binding analysis
The steroid-binding capacity of rabbit maternal sera and the recombinant SHBG isoforms were determined by saturation analysis as described previously (16). In brief, diluted sera or SHBG in solution were incubated with dextran-coated charcoal (DCC) solution before assay to remove any endogenous steroids. They were then incubated with 10 nM [3H]DHT (SA 42 Ci/mmol; NEN Life Science Products, Boston, MA). Nonspecific binding was assessed in parallel by the addition of a 200-fold excess of unlabeled DHT. Ice-cold DCC (500 μl) was then added to separate bound and free steroid. The dissociation rate of the [3H]DHT-SHBG complex was determined by exposure to DCC for increasing time at 0 C. The steroid-binding capacity of the rabbit maternal sera was calculated and expressed as the binding at zero time exposure to DCC.
RT-PCR and Southern blot analysis
Total RNA was reverse transcribed as described previously (21). Rabbit SHBG and S16 cDNAs were coamplified from 2 μl of the reverse transcription product with 100 pmol/ml of each rbSHBG primer (sense: 5'-GGGATTCAGAGGGAGTGCTTTT; antisense: 5'-GGTCTTGAGCCAGGGGTCAA or sense: 5'-ATGGCTACTCCGCCACTCGTGCCG; antisense: 5'-AATGGCACCGACACCTCCCATTAA) and 7.5 pmol/ml of each ribosomal S16 primer (sense: 5'-TCCGCTGCAGTCCGTTCA-AGTCTT; antisense: 5'-GCCAAACTTCTTGGATTCGCAGCG). The PCR was performed with denaturation at 94 C for 1 min, annealing at 64 C for 1.5 min, and extension at 72 C for 2 min for a total of 22 cycles, after a final extension at 72 C for 15 min. Under these conditions, the amplifications of both rbSHBG and S16 were in linear phase. The authenticity of the RT-PCR products was confirmed by Southern analysis. Southern hybridization was performed with the -32P-labeled rabbit SHBG or S16 cDNA probes used in the Northern blot analysis, as described previously (23). Radioactivity of hybridized products was visualized by autoradiography for 2–4 d at –80 C (BioMax films; Eastman Kodak).
Sequencing analysis of RT-PCR product
The rbSHBG mRNA and alternatively spliced rbSHBG transcript were amplified by RT-PCR using primers flanking exons 1 and 8 of rbSHBG (sense: 5'-ATGGCTACTCCGCCACTCGTGCCG; antisense: 5'-AATGGCACCGACACCTCCCATTAA). The PCR products were purified using a High Pure PCR product purification kit (Roche, Stockholm, Sweden), and sequence analysis was conducted by cycle sequencing using an ABI PRISM BigDye Terminators cycle sequencing kit (version 3.0; Applied Biosystems, Foster City, CA) and the ABI 3100 genetic analyzer (Applied Biosystems). The cDNA sequence of the alternatively spliced rbSHBG transcript was aligned against the sequence of the rbSHBG mRNA sequence with GeneTool software (version 2.0; BioTools Inc., Edmonton, CA).
Expression, purification, and steroid binding of recombinant rbSHBG isoforms
The reverse transcription product prepared using mature female liver RNA was used as template for PCR using primers flanking exons 2 and 8 of rbSHBG (sense: 5'-GGATCCCTGAGAAGCGTTCTGGCC; antisense: 5'-CCATGGATTAATGGGAGGTGTCGG), to produce cDNA fragments containing a BamHI site (underlined) immediately upstream the codon for Leu1, and an NcoI site (underlined) immediately downstream the stop codon of rbSHBG cDNA. The PCR products obtained were digested with BamHI and NcoI and subcloned into the same sites of the bacterial expression vector pGEX-KGK to produce constructs encoding the rbSHBG isoforms fused in-frame with a glutathione-S-transferase (GST) sequence. Escherichia coli strain BL21 transformed with the constructs were used for fusion protein expression. The recombinant rbSHBG isoforms were expressed, purified, and subjected to SDS-PAGE and Western blot analyses as described previously (16, 24). The binding capacities of the affinity-purified recombinant rbSHBG isoforms for [3H]DHT were determined essentially as described above by using 100 ng of each fusion protein. The relative steroid-binding capacities of the recombinant isoforms were expressed as specifically bound [3H]DHT in counts per minute.
Production of antisera against the rbSHBG isoform encoding by the alternatively spliced rbSHBG transcript
A multiple antigenic peptide (MAP)-conjugated synthetic peptide (Invitrogen) corresponding to the 10-amino acid residues (GSWHQLVPAL) encoded by the nucleotide sequences spanning the exon 3–5 junction of alternative rbSHBG transcript was used to raise antibody against a truncated rbSHBG isoform lacking the exon 4-encoding domain. For immunization, synthetic peptide (150 μg) was administered to BALB/c mice via sc injection, and antiserum was collected 7 d after booster injection.
Immunological detection of the rbSHBG isoform
Rabbit fetal and maternal livers collected at different times during pregnancy were homogenized by sonication in lysis buffer [50 mM Tris, 2 mM EDTA, 0.15 mM NaCl, 2 mM phenylmethylsulfonyl fluoride, 1% Nonidet P-40, and 10% glycerol (pH 7.4)]. Soluble proteins were recovered by centrifugation and quantified using the Bradford protein assay. Samples were resolved by 12% SDS-PAGE and blotted onto nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA). After blocking with PBS containing 5% (wt/vol) skimmed milk, the membranes were incubated in specific antibodies against the MAP-conjugated synthetic peptide (1:2500 dilution of the antiserum) at 4 C overnight. Unbound antibodies were removed by extensive washing with PBS supplemented with 0.05% Tween 20 (vol/vol). The membranes were incubated with alkaline phosphatase-conjugated goat antimouse IgG antibody (1:5000 dilution) (Zymed Laboratories Inc., San Francisco, CA), and immunoreactive proteins on the blots were visualized by chromogenic detection using nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate.
Results
Prenatal and maternal rbSHBG expression
Female rabbits, from 11 to 32 d of pregnancy and 1 d after delivery were studied, and the SHBG mRNA abundance in the embryo, fetal, and maternal livers was examined by Northern blot analysis. A 1.6-kb SHBG mRNA was detected in most of the maternal and fetal liver samples examined (Fig. 1). This SHBG mRNA was not detected in the liver during the early stage of embryonic development and first appeared in the fetal liver at d 25 gestation (Fig. 1A). Its abundance in the fetal liver remained low until d 30 gestation, at which time the levels increased dramatically until parturition (d 33). In the maternal liver, however, high levels of SHBG mRNA are present during early pregnancy from d 11, increasing to a maximum at d 22, and declining thereafter apart from a slight rebound of SHBG mRNA level during d 28 and 29 (Fig. 1B). Serum SHBG levels in pregnant rabbits exhibited a profile similar to that observed for hepatic SHBG mRNA levels in the same animals. The plasma SHBG level reached a maximum at d 22; it then declined rapidly thereafter and remained low despite a slight rebound during the late stage of pregnancy before parturition (Fig. 2).
FIG. 1. Northern blot analysis of rbSHBG mRNA in embryo, fetal, and maternal livers. Total RNA was extracted from rabbit embryos (d 11–20 of pregnancy), fetal livers (d 17–32 of pregnancy), newborn livers (NB, d 1 of birth) (A), and maternal livers from d 11–32 of pregnancy (B). A 1.6-kb SHBG mRNA was detected after 2 d of autoradiographic exposure at –80 C. The blots were washed and reprobed with -32P-labeled S16 cDNA to detect the expression of the ribosomal protein S16 to demonstrate even loading of the samples. The figure shows representative results from one of three independent experiments.
FIG. 2. Serum SHBG levels in pregnant rabbits during pregnancy. Serum SHBG was measured by a steroid-binding capacity assay. Results shown are means ± SD from five independent experiments of pooled animals.
Levels of SHBG mRNA in the fetal and maternal livers during pregnancy, in particular during the early fetal development, were also studied using a more sensitive RT-PCR assay. Results were similar to those obtained by Northern blot analysis (Figs. 3 and 4), but trace amounts of SHBG mRNA were detected in embryo and fetal liver as early as d 14 and 17 gestation, respectively, by RT-PCR (Fig. 3A). In this assay, specific amplification of SHBG mRNA produced a major PCR product of 732 bp. However, a minor PCR product of 570 bp was also identified in all liver samples examined, and this was confirmed to be a rbSHBG transcript by Southern hybridization (Figs. 3B and 4B).
FIG. 3. RT-PCR analysis of rabbit SHBG mRNA levels in embryo and fetal liver. A, RT-PCRs were performed from 10 μg total RNA of rabbit embryo, fetal liver, or newborn liver (NB, d 1 of birth). They were reversely transcribed and subsequently amplified by primers prepared from internal nucleotide sequence to give a major PCR product of 732 bp and a minor PCR product of 570 bp. Coamplification of ribosomal S16 cDNA yielded a 385-bp product. B, The authenticity of the RT-PCR fragments was confirmed by Southern blot hybridization using -32P-labeled rabbit SHBG cDNA. The blots were washed and reprobed with -32P-labeled S16 cDNA. NB, Newborn; –ve, negative control in which reverse transcriptase was omitted from the d 11 sample. The figure shows representative results from one of three independent experiments.
FIG. 4. RT-PCR analysis of rbSHBG mRNA levels in maternal liver. A, RT-PCRs were performed from 10 μg total RNA of pregnant rabbit liver. They were reversely transcribed and subsequently amplified by primers prepared from internal nucleotide sequence to give a major PCR product of 732 bp and a minor PCR product of 570 bp. Coamplification of ribosomal S16 cDNA yielded a 385-bp product. B, The authenticity of the RT-PCR fragments was confirmed by Southern blot hybridization using -32P-labeled rabbit SHBG cDNA. The blots were washed and reprobed with -32P-labeled S16 cDNA. NB, Newborn; –ve, negative control in which reverse transcriptase was omitted from the d 11 sample. The figure shows representative results from one of three independent experiments.
Sequence analysis and tissue distribution of the alternative rbSHBG transcript
When the nucleotide sequence of the PR-PCR product for the alternately spliced rbSHBG transcript was aligned with the rbSHBG mRNA sequence, it was found to contain the same nucleotide sequence as the rbSHBG mRNA, apart from a 162-bp deletion corresponding to the entire exon 4 of rbSHBG (data not shown). This alternative SHBG transcript is most likely produced by in-frame splicing of exon 4 because there is only a single gene for SHBG in rabbit (7). This exon 4-skipped rbSHBG transcript is present not only in the fetal and adult livers but also was found in the testis (Fig. 5), and the abundance of alternatively spliced rbSHBG transcript relative to that of the rbSHBG mRNA is approximately 1:5 in these tissues. By contrast, only trace amounts of SHBG mRNA were detected in the brain and kidney of neonatal and young male rabbit by sequential RT-PCR and Southern blot hybridization (Fig. 5).
FIG. 5. Tissue distribution of rbSHBG transcripts examined by RT-PCR analysis. The rbSHBG expression in various tissues of the fetus (d 30 gestation) and male (young, 7 d old; adult, 15 wk old) and female (young, 7 d old; adult, 15 wk old) rabbits was studied by RT-PCR. Amplification with primers flanking exons 1 and 8 of rbSHBG give a major PCR product of 1197 bp and a minor PCR product of 1035 bp. Authenticities of these products were subsequently confirmed by Southern blot hybridization using -32P-labeled rabbit SHBG cDNA. B, Brain; K, kidney; Li, liver; T, testis. The figure shows representative results from three independent experiments.
Steroid-binding analysis of rbSHBG isoforms
The rbSHBG mRNA and rbSHBG transcripts lacking the exon 4 sequences were expressed as GST-fusion proteins in bacterial cells. After affinity purification, the corresponding major Coomassie-stained products on an SDS-PAGE gel were approximately 67 and 61 kDa, respectively (Fig. 6A). These proteins were identified as GST-fusion proteins by Western blot analysis using anti-GST antibody (Fig. 6A). Because the GST domain of recombinant rbSHBG fusion proteins does not affect its steroid binding activity (16), the steroid-binding capacities of the recombinant fusion proteins were analyzed directly by DHT-binding assay without removal of the GST domain. The results show that the GST-fused recombinant rbSHBG lacking the 54 amino acid residues encoded by the exon 4 sequence has essentially no affinity for DHT, when compared with the recombinant full-length rbSHBG-GST fusion protein (Fig. 6B).
FIG. 6. A, Expression of GST fusion SHBG isoforms. Affinity-purified GST-fusion rbSHBG isoforms were analyzed by SDS-PAGE, followed by Coomassie blue staining and Western blot with anti-GST antibody (Ab). Lanes 1 and 3 correspond to GST-fused full-length rbSHBG; whereas lanes 2 and 4 correspond to GST-fused truncated rbSHBG lacking the exon 4-encoding domain. B, Steroid binding analysis of the affinity-purified GST-fusion proteins. The relative affinities of the recombinant rbSHBG isoforms for DHT were expressed as the specific [3H]DHT bound after exposure to DCC for 10 min at 0 C. Results shown are mean ± SD from three individual experiments. M.W., Molecular weight.
Immunological detection of the truncated SHBG protein
The antiserum against the truncated rbSHBG polypeptide encoded by the exon 4-skipped rbSHBG transcript was able to differentiate between the full-length rbSHBG polypeptide and the novel SHBG isoform (Fig. 7A). This antiserum was then used to detect this SHBG isoform in fetal and maternal liver lysates prepared at different stages of pregnancy. A single immunoreactive protein of about 33–35 kDa was detected in maternal and fetal liver samples by Western blot analysis (Fig. 7B). This immunoreactive band was present in the maternal liver lysates from d 11 to d 32 of pregnancy. The strongest signal was observed on d 11, which then gradually decreased until parturition. These results correlate quite well with the RT-PCR analysis of the exon 4-skipped rbSHBG transcript levels in the maternal liver (Fig. 3). For the fetal liver lysates, the 33- to 35-kDa immunoreactive protein was detected on d 31 and 32 gestation (Fig. 7B), and no signal could be detected before d 29 of pregnancy (data not shown), despite the fact that as much as 100 μg of total soluble proteins (compared with 40 μg for the maternal liver) were analyzed for these samples. This may be due to lower levels of the alternatively spliced rbSHBG transcripts present in fetal liver, compared with the maternal liver.
FIG. 7. Western blot analysis of the truncated rbSHBG protein lacking the exon 4-encoding domain. Purified GST-fused recombinant rbSHBG proteins (A), total lysates prepared from maternal and fetal livers, respectively (B), and total soluble protein from various tissues of male and female rabbits (C) were resolved by 12% SDS-PAGE and transferred onto nitrocellulose membrane. Western blot was then performed using mouse antiserum against the novel rbSHBG isoform as primary antibody. Immunoreactive protein-antibody complex was then detected using alkaline phosphatase-conjugated goat antimouse IgG antibody and visualized by chromogenic detection using nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate as substrate. B, Brain; K, kidney; Li, liver; S, spleen; T, testis. The figure shows representative results from three independent experiments.
To examine the tissue distribution of the 33- to 35-kDa immunoreactive protein, other tissues from male and female rabbits were analyzed by Western blot. As shown in Fig. 7C, an immunoreactive protein of the same molecular size was observed consistently in the liver lysates of mature male and female rabbits as well as in testis lysates. The immunoreactive signal of the testis lysates was weaker than those observed in the liver lysates of the same animal. This matched the RT-PCR results in that the alternatively spliced rbSHBG transcript is less abundant in testis than liver (Fig. 5). On the other hand, no immunoreactive signal could be detected in the brain, kidney, and spleen lysates of male or female rabbits. This reflects the fact that SHBG expression in these tissues is either very low or nonexistent.
Detection of exon 4-skipped SHBG transcripts in mouse and human tissue and cells
The presence of the exon 4-skipped SHBG transcripts in mouse and human tissue and cells were examined by RT-PCR and sequence analysis using species-specific primers. As in the case of rabbit, an appropriately sized product was detected in the mouse testis (500 bp), human testes, and human HepG2 cells (537 bp) (Fig. 8). This slight difference in the sizes of the PCR products amplified between the species is due to the differences in the best annealing positions of corresponding primers. Sequence analysis and pairwise alignment confirmed that these RT-PCR products were derived from SHBG transcripts lacking exon 4 sequences (data not shown). To our knowledge, this is the first demonstration that exon 4-skipped SHBG transcripts exist in human and mouse tissues (for summary, see Table 1).
FIG. 8. Detection of the exon 4-skipped SHBG transcripts in mammalian species. Total RNA isolated the liver of mature rabbit and testes of mature rabbit, mouse, and human as well as human liver cell-line HepG2 were used as template, and RT-PCRs were performed using universal sense primer annealing to the exon 3 and species-specific antisense primers annealing to the specific regions with the exon 7 of the corresponding mature mRNA (Table 2). The major bands of about 700 bp represented the PCR product amplified from the SHBG mRNA, whereas the minor bands of about 550 bp were confirmed by automated nucleotide sequencing to be the PCR product amplified from the exon 4-skipped SHBG transcripts of the corresponding species. The figure shows representative results from three independent experiments.
TABLE 1. Summary of alternative SHBG transcripts in human, rabbit, mouse, and rat
Discussion
In this report, we presented the results of a detailed ontogenic study of rbSHBG expression in fetal and maternal livers. In the rabbit fetus, a high level of SHBG expression was observed during late stages of gestation. This suggests a marked increase in fetal liver SHBG biosynthesis during late gestation and supports the concept that changes in plasma SHBG levels could influence fetal and neonatal sexual development through modulating the amounts of testosterone available to developing male reproductive tissues (for review, see Ref. 26). In contrast, high SHBG mRNA levels were observed in the maternal liver during early pregnancy and declined markedly thereafter with a slight rebound observed at d 28 and 29. This pattern of maternal hepatic rbSHBG expression mirrored that of maternal serum SHBG levels, which is in line with the fact that the majority of plasma SHBG is made by the liver. This high level of SHBG expression in the maternal rabbit liver during early pregnancy may be due to an increased production of estrogens during this time (27) because estrogens stimulate the production and secretion of SHBG by HepG2 cells (28). The observed increase of maternal SHBG production at early pregnancy suggests that SHBG may play a role related to the development of the uterus during pregnancy. Although it is generally assumed that increasing plasma SHBG levels restricts the action of its sex steroid ligand, SHBG might act directly by interacting with membrane receptors that have been identified on endometrial cells (29, 30) and mediate a nongenomic action of estrogens (31).
To our surprise, an alternative rbSHBG transcript was detected by RT-PCR alongside rbSHBG mRNA in both the fetal and maternal livers. Sequence analysis revealed that this alternative transcript contained an in-frame deletion of 162 bp corresponding to the entire exon 4 of the rbSHBG. This exon 4-skipped rbSHBG transcript was also expressed concurrently with the rbSHBG mRNA in testis and liver of young and mature male and female rabbits. The ratios of the alternative rbSHBG transcript to the rbSHBG mRNA are similar (1:5) in all tissues, suggesting there is no tissue specificity in the production of the exon 4-skipped rbSHBG transcript.
In human, the 54 amino acid residues (Val103-Pro156) encoded by the 162 bp of the exon 4 are critical structural elements of the steroid-binding pocket of the N-terminal laminin G-like (LG) domain (for reviews, see Refs. 32, 33, 34, 35). Thus, deletion of these 54 amino acid residues was expected to disrupt the steroid-binding site, and this was confirmed by expressing and studying the steroid-binding properties of a GST-fused truncated rbSHBG protein lacking the exon 4-encoding domain. Unlike the full-length rbSHBG-GST fusion protein, the truncated rbSHBG-GST fusion protein does not bind steroid. This suggests that if the rbSHBG transcript lacking exon 4 sequences is translated in vivo, it probably functions in ways that are not directly related to steroid transport.
To determine whether the exon 4-skipped SHBG transcript is translated in vivo, we generated an antiserum capable of discriminating between the rbSHBG isoform lacking the exon 4-encoding domain and the mature rbSHBG polypeptide. This antiserum detected an immunoreactive rbSHBG isoform in the cell lysates of maternal and fetal liver as well as the testis of adult rabbit, in which the alternatively spliced rbSHBG transcript was most abundant. The size of this immunoreactive protein was about 34 kDa, which matches its deduced size of 33.7 kDa. The size of the full-length rbSHBG in the circulation that has been reported ranged from 41–45 kDa, dependent on the degree of glycosylation (36). Because the deletion of 54 amino acid residues in exon 4 would lead to a reduction of about 6 kDa, the truncated rbSHBG detected in the Western blot is most likely either not glycosylated or incompletely glycosylated. Differential glycosylation of SHBG has been reported in human, rat, rabbit, and zebrafish (10, 11, 36, 37), and the actual glycosylation status of rbSHBG isoform in tissues therefore remains to be determined.
Another question we sought to address is the localization of the truncated rbSHBG protein. In human and rabbit, hepatic SHBG is normally rapidly secreted into the blood, in which it functions as a steroid-transporting protein. RT-PCR and sequence analysis indicated that the exon 4-skipped rbSHBG transcript contains all the other seven exons of the full-length transcript, including exon 1, which encodes the signal peptide of SHBG. Therefore, the truncated rbSHBG made in the liver was expected to be secreted. Attempts have been made to detect the truncated SHBG in crude rabbit serum, but these were not successful. However, our Western blot results suggest that the truncated rbSHBG protein accumulates in the tissues in which it is made.
With respect to its physiological role, the novel rbSHBG isoform is unlikely to provide a steroid transportation function. Nevertheless, this rbSHBG isoform contains a complete C-terminal LG domain, and this may be functionally important. This is of interest because the C-terminal LG domain of SHBG resembles the LG 5 domain structure within the C-terminal SHBG-like regions of protein S and Gas 6. The SHBG-like regions of protein S and Gas 6 have been reported to interact with the C4b-binding protein (38) and the Tyro3 family receptor tyrosine kinases (39), respectively, and this may be mediated via protein-protein interactions involving this C-terminal LG domain.
Although the novel rbSHBG isoform lacks steroid-binding activity, its expression profile during prenatal development, and the occurrences of exon 4-skipped SHBG transcripts in mouse and human tissues and cells, suggests that this novel SHBG isoform might have a biological significance beyond a steroid-transporting protein.
TABLE 2. Primers used in the detection of SHBG transcripts in mammalian species
Acknowledgments
The authors thank Dr. Ping-Fai Ki for the special technical contributions in animal handling and antibody production and Ms. Connie Wong for the general technical comments and suggestions.
References
Porterfield SP 2001 Male reproductive system. In: Schmitt W, ed. Endocrine physiology: the Mosby physiology monograph series. St. Louis: Mosby, Inc.; 153–176
Westphal U 1986 Steroid-protein interactions II. In: Monographs on endocrinology. Berlin: Springer-Verlag; 204–209
Hammond GL 1993 Extracellular steroid-binding proteins In: Parker M, ed. Frontiers in molecular biology, steroid hormone action. Oxford: IRL; 12–15
Joseph DR 1994 Structure, function and regulation of androgen-binding protein/sex hormone-binding globulin. Vitam Horm 49:197–280
Hammond GL 2002 Access of reproductive steroids to target tissues. Obstet Gynecol Clin North Am 29:411–423
Hammond GL, Underhill DA, Rykse HM, Smith CL 1989 The human sex hormone-binding globulin gene contains exons for androgen-binding protein and two other testicular messenger RNAs. Mol Endocrinol 3:1869–1876
Ip YC, Lee WM, Hammond GL 2000 The rabbit sex hormone-binding globulin gene: structural organization and characterization of its 5' flanking region. Endocrinology 141:1356–1365
Janne M, Deol HK, Power SG, Yee SP, Hammond GL 1998 Human sex hormone-binding globulin gene expression in transgenic mice. Mol Endocrinol 12:123–136
Janne M, Hogeveen KN, Deol HK, Hammond GL 1999 Expression and regulation of human sex hormone-binding globulin transgenes in mice during development. Endocrinology 140:4166–4174
Selva DM, Hogeveen KN, Seguchi K, Tekpetey F, Hammond GL 2002 A human sex hormone-binding globulin isoform accumulates in the acrosome during spermatogenesis. J Biol Chem 277:45291–45298
Sullivan PM, Petrusz P, Szpirer C, Joseph DR 1991 Alternative processing of androgen-binding protein RNA transcripts in fetal rat liver. J Biol Chem 266:143–154
Wang YM, Sullivan PM, Petrusz P, Yarbrough W, Joseph DR 1989 The androgen binding protein gene is expressed in CD1 mouse testis. Mol Cell Endocrinol 63:85–92
Cates JM, Damassa DA, Gagin GA, Dempsey RV 1995 Hepatic expression of sex hormone-binding globulin associated with the postnatal surge of serum androgen-binding activity in the Djungarian hamster. J Steroid Biochem Mol Biol 55:147–158
Cates JM, Damassa DA 1997 Characterization and developmental expression patterns of testicular androgen-binding protein in the Djungarian hamster (Phodopus sungorus). J Reprod Fertil 111:291–298
Khan MS, Knowles BB, Aden DP, Rosner W 1982 Secretion of testosterone-estradiol-binding globulin by a human hepatoma-derived cell line. J Clin Endocrinol Metab 53:448–452
Lee WM, Wong AST, Tu AWK, Cheung C-H, Li JCH, Hammond GL 1997 Rabbit sex hormone-binding globulin: primary structure, tissue expression, and structure/function analyses by expression in Escherichia coli. J Endocrinol 153:373–384
Wong AST, Lui WY, Hui ITC, Lee WM 2001 Rabbit sex hormone-binding globulin: expression in the liver and testis during postnatal development and structural characterization by truncated proteins. Int J Androl 24:165–174
Wang YM, Bayliss DA, Millhorn DE, Petrus Z, Joseph DR 1990 The androgen-binding protein gene is expressed in male and female rat brain. Endocrinology 127:3124–3130
Larrea F, Diaz L, Carino C 1993 Evidence that human placenta is a site of sex hormone-binding globulin gene expression. J Steroid Biochem Mol Biol 46:497–505
Misao R, Itoh N, Mori H, Fujimoto J, Tatnaya T 1994 Sex hormone-binding globulin mRNA levels in human uterine endometrium. Eur J Endocrinol 131:623–629
Wong CC, Lee WM 2002 The proximal cis-acting elements Sp1, Sp3 and E2F regulate mouse mer gene transcription in Sertoli cells. Eur J Biochem 269:3789–3800
Chan YL, Paz V, Olvera J, Wool IG 1990 The primary structure of rat ribosomal protein S16. FEBS Lett 263:85–88
Cheng CK, Cheung CH, Lee WM 2003 Mouse testin: complementary DNA cloning, genomic organization, and characterization of its proximal promoter region. Biol Reprod 68:1376–1386
Hildebrand C, Bocchinfuso WP, Dales D, Hammond GL 1995 Resolution of the steroid-binding and dimerization domains of human sex hormone-binding globulin by expression in Escherichia coli. Biochemistry 34:3231–3238
Misao R, Iwagaki S, Fujimoto J, Nakanishi Y, Sun WS, Tamaya T 1999 Expression of sex hormone-binding globulin exon 7 splicing variant mRNA in secondary spreading lesions of gynecological cancers. J Steroid Biochem Mol Biol 68:103–109
Renfree MB, Wilson JD, Shaw G 2002 The hormonal control of sexual development. Novartis Found Symp 244:136–152
Challis JR, Davies J, Ryan KJ, Parker CR 1973 The concentrations of progesterone, estrone and estradiol-17? in the plasma of pregnant rabbits. Endocrinology 93:971–976
Rosner W, Aden DP, Khan MS 1984 Hormonal influences on the secretion of steroid-binding proteins by a human hepatoma-derived cell line. J Clin Endocrinol Metab 59:806–808
Nakhla AM, Khan MS, Rosner W 1990 Biologically active steroids activate receptors-bound human sex hormone binding globulin to cause LNCaP cells to accumulate adenosine 3', 5'-monophosphate. J Clin Endocrinol Metab 71:398–404
Rosner W, Hryb DJ, Khan MS, Nakhla AM, Romas NA 1999 Sex hormone-binding globulin mediates steroid hormone signal transduction at the plasma membrane. J Steroid Biochem Mol Biol 69:481–485
Avvakumov GV, Zhuk NI, Strel’chyonok OA 1986 Subcellular distribution and selectivity of the protein-binding component of the recognition system for sex-hormone-binding protein-estradiol complex in human decidual endometrium. Biochim Biophys Acta 881:489–498
Grishkovskaya I, Avvakumov GV, Sklenar G, Dales D, Hammond GL, Muller YA 2000 Crystal structure of human sex hormone-binding globulin: steroid transport by a laminin G-like domain. EMBO J 19:504–512
Avvakumov GV, Grishkovskaya I, Muller YA, Hammond GL 2002 Crystal structure of human sex hormone-binding globulin in complex with 2-methoxyestradiol reveals the molecular basis of high affinity interactions with C-2 derivatives of estradiol. J Biol Chem 277:45219–45225
Grishkovskaya I, Avvakumov GV, Hammond GL, Catalano MG, Muller YA 2002 Steroid ligands bind human sex hormone-binding globulin in specific orientations and produce distinct changes in protein conformation. J Biol Chem 277:32068–32093
Hammond GL, Avvakumov GV, Muller YA 2003 Structure/function analyses of human sex hormone-binding globulin: effects of zinc on steroid-binding specificity. J Steroid Biochem Mol Biol 85:195–200
Danzo BJ, Black JH, Bell BW 1989 The microheterogeneity of rabbit testosterone-binding globulin is due to differential glycosylation of its single protomer. Biol Reprod 41:957–965
Miguel-Queralt S, Knowlton M, Avvakumov GV, Al-Nouno R, Kelly GM, Hammond GL 2004 Molecular and functional characterization of sex hormone binding globulin in zebrafish. Endocrinology 145:5221–5230
Saposnik B, Borgel D, Aiach M, Gandrille S2003 Functional properties of the sex-hormone-binding globulin (SHBG)-like domain of the anticoagulant protein S. Eur J Biochem 270:545–555
Lu Q, Gore M, Zhang Q, Camenisch T, Boast S, Casagranda F, Lai C, Skinner MK, Klein R, Matsushima GK, Earp HS, Goff SP, Lemke G 1999 Tyro-3 family receptors are essential regulators of mammalian spermatogenesis. Nature 398:723–728(Kwong-Man Ng, Man-Ting So)