Absence of Glucocorticoid Receptor-β in Mice*

Two human glucocorticoid receptor (GR) isoforms, GRα and GRβ, are derived from the same gene by alternative splicing involving exon 9 of the GR locus. The non-ligand binding isoform GRβ was proposed to act as a transdominant negative inhibitor of GRα, thus modulating glucocorticoid responsiveness of target tissues. To study GRβ in mice we characterized the genomic region around exon 9 of the murine GR gene. Sequence analysis revealed that the presumed exon 9β contained an open reading frame of 59 amino acids. In contrast, human exon 9β encoded only 15 amino acids. Using reverse transcriptase polymerase chain reaction the absence of GRβ mRNA was demonstrated in all adult mouse tissues examined. To exclude the possibility that the polymerase chain reaction conditions employed were not suitable for the amplification of GRβ mRNA, we synthesized an artificial template corresponding to the presumed GRβ mRNA spanning exons 7, 8, and 9β. Various amounts of this template were added to brain cDNA preparations and as little as 25 molecules were detectable under the polymerase chain reaction conditions chosen. Since GRβ is not conserved across species its physiological significance in humans appears questionable.

Two human glucocorticoid receptor (GR) isoforms, GR␣ and GR␤, are derived from the same gene by alternative splicing involving exon 9 of the GR locus. The non-ligand binding isoform GR␤ was proposed to act as a transdominant negative inhibitor of GR␣, thus modulating glucocorticoid responsiveness of target tissues. To study GR␤ in mice we characterized the genomic region around exon 9 of the murine GR gene. Sequence analysis revealed that the presumed exon 9␤ contained an open reading frame of 59 amino acids. In contrast, human exon 9␤ encoded only 15 amino acids. Using reverse transcriptase polymerase chain reaction the absence of GR␤ mRNA was demonstrated in all adult mouse tissues examined. To exclude the possibility that the polymerase chain reaction conditions employed were not suitable for the amplification of GR␤ mRNA, we synthesized an artificial template corresponding to the presumed GR␤ mRNA spanning exons 7, 8, and 9␤. Various amounts of this template were added to brain cDNA preparations and as little as 25 molecules were detectable under the polymerase chain reaction conditions chosen. Since GR␤ is not conserved across species its physiological significance in humans appears questionable.
Glucocorticoids are involved in the regulation of a variety of physiological processes such as development, metabolism, maintenance of homeostasis, and regulation of central nervous system functions (1). Their effects are mediated by the glucocorticoid receptor (GR), 1 a ligand-dependent transcription factor that belongs to the superfamily of nuclear receptors (1)(2)(3).
The GR protein is composed of structurally and functionally defined domains. The amino-terminal part of the protein contains a transactivation domain, whereas the central part includes the DNA-binding domain which is crucial for specific interaction of the receptor with DNA sequences containing glucocorticoid receptor responsive elements. The carboxyl terminus includes the ligand-binding domain and sequences which are involved in nuclear translocation, receptor dimeriza-tion, transactivation, and interaction with heat shock proteins (2)(3)(4)(5). Under basal conditions the GR resides in the cytoplasm where it is associated with heat shock proteins and other proteins forming a multiprotein complex (6). Upon hormonal stimulation the GR dissociates from the multiprotein complex and translocates into the nucleus where it can stimulate or suppress transcription of multiple genes (1)(2)(3).
Two human glucocorticoid receptor isoforms termed hGR␣ and hGR␤ have been described (7). Both isoforms are derived from the same gene by differential splicing. Whereas hGR␣ and hGR␤ share the first eight exons of the human GR gene, either of the last two exons, i.e. exon 9␣ or 9␤, is spliced into the respective mRNA (8). This leads to the formation of two protein isoforms having the first 727 amino acids in common. Exon 9␣ encodes 50 amino acids and exon 9␤ 15 amino acids. Whereas hGR␣ can bind hormone and is transcriptionally competent, hGR␤ is not (8).
Recently, it was demonstrated that hGR␤ transiently expressed in COS-7 cells bound specifically to glucocorticoidresponsive elements and acted as a dominant negative inhibitor of hGR␣ activity (9). Since hGR␤ mRNA was expressed in all human tissues examined it was assumed that this isoform might play an important role in modulating tissue sensitivity to glucocorticoids (9). However, the low abundance of hGR␤ mRNA is incompatible with such a function. Quantitative RT-PCR experiments revealed that in all human tissues and cell lines analyzed the hGR␤ message was expressed 200 -500-fold less than the hGR␣ message (10). In contrast, it was shown by Western blots that the ratio of hGR␤ to hGR␣ varied from 1.0 to 5.0 (11). Moreover, some conflicting results concerning the subcellular localization of the hGR␤ protein were obtained. Using hGR␤ specific antibodies it was demonstrated that after dexamethasone treatment hGR␤ translocated from the cytosol into the nucleus (11). On the other hand, it had been reported that hGR␤ resided primarily in the cell nucleus independently of hormonal stimulation (10). To date all these discrepancies have not been conclusively explained.
Since GR␤ has been studied so far exclusively in human tissues we were interested to investigate whether this GR isoform would be of physiological importance in other species. Therefore we characterized the genomic region around the presumed exon 9␤ of the murine GR locus and studied the tissue distribution of GR␤ in mice.

Materials
DNA modifying enzymes, Taq DNA polymerase, deoxynucleotide triphosphates, MgCl 2 , and PCR buffer were purchased from Boehringer Mannheim. Reverse transcriptase and first strand buffer were obtained from Life Technologies, Inc., Pfu polymerase from Stratagene, and RNasin from Promega.

Methods
Library Screen and DNA Sequencing-A murine genomic stem cell library (12) was screened with a mouse cDNA probe spanning exons * This work was supported in part by Deutsche Forschungsgemeinschaft Grant SFB 229, Volkswagen-Stiftung project I 172526, the Fonds der Chemischen Industrie, BMFT Project 0310681, and European Community Grants BI02-CT93-0319 and PL 960179. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) Z97061.
6 -9 of the GR gene using high stringency hybridization and washing conditions (13). Three hybridizing genomic phages were purified and characterized by restriction mapping. One of the phages containing exon 9 was subcloned into Bluescript KSII (Stratagene). A 2.8-kilobase HindIII fragment containing the 3Ј-untranslated region of exon 9␣ and the putative exon 9␤ was sequenced on both strands (14).
RNA Isolation and RT-PCR Analysis-Total RNA from a variety of adult mouse tissues was isolated after homogenization in guanidinium thiocyanate (15). The quality of the RNA preparations was controlled by ethidium bromide staining of the 18 S and 28 S rRNA after electrophoretic separation in denaturing agarose gels. 2 g of total RNA were subjected to cDNA synthesis. After DNase I digestion, cDNA synthesis was performed in first strand buffer containing 10 mM dithiothreitol, 0.5 mM dNTPs, 10 units of RNasin, 200 units of reverse transcriptase, 1 g of random hexamers in a final volume of 20 l. To control for DNA contamination, addition of reverse transcriptase was omitted in control samples of each RNA preparation. The cDNA was precipitated and dissolved in 20 l. For the amplification of GR␣ and GR␤ mRNA by PCR, 2 l of cDNA solution were used. PCR was performed in a final volume of 50 l containing 1.5 mM MgCl 2 , 1 unit of Taq DNA polymerase, 0.2 mM dNTPs, and 0.2 M upstream and downstream primers. After an initial incubation for 3 min at 95°C, samples were subjected to 35 cycles of 1 min at 95°C, 1 min at 52°C, and 1 min at 72°C. This was followed by a final extension step at 72°C for 10 min. 35 cycles of PCR were sufficient for the amplification of GR␣ mRNA, whereas in the case of control cDNA experiments and amplification of GR␤ mRNA a second round of 35 cycles was done under identical conditions using 2 l of the first amplification round as template. Primers used for the amplification of the GR␣ message were as follows: 5Ј-AGCAGAGAATGACTC-TAC-3Ј (upstream, corresponding to nucleotides 1928 -1945 of the murine GR cDNA (16)) and 5Ј-GAATTCAATACTCATGGAC-3Ј (downstream, corresponding to nucleotides 2265-2283 of the murine GR cDNA (16)). The primers used for the amplification of the GR␤ message were as follows: 5Ј-AGCAGAGAATGACTCTAC-3Ј (upstream, corresponding to nucleotides 1928 -1945 of the murine GR cDNA (16)) and 5Ј-AGCTCTTTATACATCTAATG-3Ј (downstream, corresponding to the putative murine exon 9␤). Amplified cDNA fragments were analyzed on a 2% agarose gel, subcloned, and sequenced.
Synthesis of an Artificial GR␤ Template-An artificial GR␤ template was generated by overlap extension using PCR (17). Briefly, two different sets of PCR experiments using Pfu polymerase were performed. In the first round two different DNA fragments were amplified: a 290-bp fragment (amplified from brain cDNA) containing exons 7 and 8 of the murine GR gene and having a 3Ј-overhang homologous to the beginning of the presumed exon 9␤ and a 194-bp fragment (amplified from the genomic 2.8-kilobase HindIII fragment) containing the putative exon 9␤ having a 5Ј-overhang homologous to the end of exon 8. Primers used for the amplification of the 290-bp fragment were as follows: 5Ј-AGCA-GAGAATGACTCTAC-3Ј (upstream, outer primer 1) and 5Ј-GGGTTTA-ACCATATAACATTATCATGCATGGAGTCCAA-3Ј (downstream, inner primer 1). Primers used for the amplification of the 194-bp fragment were as follows: 5Ј-TTGGACTCCATGCATGATAATGTTATATGGTTA-AACCC-3Ј (upstream, inner primer 2) and 5Ј-AGCTCTTTATACATCT-AATG-3Ј (downstream, outer primer 2). Both PCR fragments were gel purified and fused in a second PCR experiment using the two outer primers. PCR conditions were the same as described above. The 484-bp fusion product spanning exons 7, 8, and 9␤ was purified by gel electrophoresis, subcloned into Bluescript KSII, and sequenced on both strands (14).

RESULTS
To investigate GR␤ in mice we isolated and characterized a genomic phage containing exons 8 and 9 of the murine GR locus. A 2.8-kb genomic HindIII fragment including exon 9 was sequenced on both strands to locate exons 9␣ and 9␤. The nucleotide sequence of the presumed exon 9␤ and its preceeding intron is shown in Fig. 1A. This sequence shows strong similarity to the sequence of human exon 9␤. However, as depicted in Fig. 1B the murine and rat sequences (19) homologous to human exon 9␤ are not preceeded by a splice site conforming to the G(T/A)G rule (18). The putative murine exon 9␤ would contain an open reading frame of 59 amino acids (Fig.  1A) since the stop codon of human exon 9␤ is not conserved in mice and rats (Fig. 1B). The first 15 amino acids of the presumed murine exon 9␤ are highly conserved, the remaining contain a remarkable stretch of basic amino acids (Fig. 1A). These sequence data make the existence of GR␤ in mice unlikely.
To obtain evidence for or against the existence of GR␤ in mice we performed RT-PCR analysis of GR␣ and GR␤ mRNA in  ). B, comparison of the nucleotide sequences from mouse and rat (19) homologous to the nucleotide sequence around human exon 9␤ (depicted in uppercase letters). In mouse and rat, the acceptor splice site preceeding the first nucleotide (indicated with an arrow) of the putative exon 9␤ does not conform to the GT/AG rule (18) and the stop codon of human exon 9␤ (underlined) is not conserved. various adult mouse tissues using an upstream primer specific for murine exon 7 and either one of the downstream primers specific for exon 9␣ and the presumptive exon 9␤, respectively. The primers were chosen in close analogy to those employed in RT-PCR experiments examining the tissue distribution of GR␤ in humans (10). GR␣ cDNA producing a 356-bp band was easily detectable after 35 cycles of PCR in liver, brain, kidney, heart, and pituitary ( Fig. 2A). In contrast, amplification of the presumptive GR␤ cDNA which would produce a 484-bp band was not successful even after two rounds of 35 PCR cycles. The absence of GR␤ in the examined tissues was confirmed by radioactive hybridization of the blotted PCR gels (data not shown). To ensure that the PCR conditions employed were suitable for the amplification of GR␤ cDNA, we generated, by use of the overlap extension technique (17), an artificial murine GR␤ template. This template spanned exons 7, 8, and 9␤ thus resembling the proposed structure of the carboxyl-terminal part of human GR␤ (10). The concentration of the stock solution of the artificial GR␤ template was determined spectrophotometrically and serial dilutions were prepared. Various amounts of this template were added to brain cDNA preparations and subjected to two rounds of 35 PCR cycles (Fig. 2B). Whereas native GR␤ was not detected under the PCR conditions chosen as little as about 25 molecules of the added artificial GR␤ template were easily detectable. From these results we conclude that GR␤ mRNA comparable to that described in humans (10) is not synthesized in adult mice. DISCUSSION Recently, first attempts were made to unravel if GR␤ plays a physiological role in humans. Using transient transfection experiments it was shown that GR␤ could act as a transdominant negative inhibitor of GR␣ (9). From these results it was hypothesized that a possible in vivo function of GR␤ could be the modulation of glucocorticoid sensitivity of target tissues (9). It was speculated that abnormally high expression levels of GR␤ might contribute to the pathogenesis of glucocorticoid resistance, whereas low expression could cause syndromes of glu-cocorticoid hypersensitivity (9). However, the in vitro experiments revealed that GR␤ was only a weak transdominant negative inhibitor of GR␣ since 5-fold overexpression was required to get a 50% reduction of GR␣ activity (9). Moreover, there was a large discrepancy observed concerning the mRNA and protein levels of GR␤ in humans. While the GR␤ mRNA was 200 -500-fold less expressed than the GR␣ mRNA (10) the ratio of GR␤ to GR␣ protein varied between 1 and 5 (11). These discrepancies which have never been explained conclusively, as well as conflicting results about the subcellular localization of GR␤ protein (10,11), raised some doubts concerning the existence and hence the physiological importance of this GR isoform in humans. We therefore decided to study as an additional approach GR␤ in mice and to examine whether it was conserved across species. Using quantitative RT-PCR experiments we could not find any evidence for the existence of GR␤ mRNA in all adult mouse tissues examined, among them liver, brain, pituitary, and kidney which have been reported to contain considerable amounts of GR␤ protein in humans (11). We conclude from these experiments that GR␤ mRNA structurally comparable to that in humans (10) is not synthesized in mice. Whereas GR␤ mRNA was ubiquitously expressed in humans and was amplified under comparable PCR conditions after 45 cycles (10) we failed to demonstrate GR␤ in mice even after 2 ϫ 35 cycles of PCR. To control the suitability of our PCR conditions we synthesized an artificial GR␤ template structurally corresponding to the human GR␤ mRNA. The amplification product of this template was easily detectable even if very low amounts of template were added to the cDNA preparation. In addition, the obvious absence of GR␤ in mice is supported by two other findings. First, splicing of the putative murine exon 9␤ would give rise to a protein distinctly different from human GR␤. Apart from the first 15 amino acid residues of the presumptive exon 9␤ which indeed show high homology to the 15 amino acids encoded by human exon 9␤, the additional 44 amino acids of the open reading frame have no counterpart in human GR␤. They contain a remarkable stretch of basic residues. Therefore one can conclude that, if GR␤ would exist in mice it might have properties distinctly different from its human variant. In rat, a putative exon 9␤ would encode 48 amino acids and only the first 7 amino acids would show high homology to human exon 9␤.
Second and even more important, the splice site in front of the putative exon 9␤ is not in accordance with the consensus splice sequences at all (18). In sharp contrast, however, all other splice junctions in the murine GR gene (20) and the intron preceeding human exon 9␤ (8) conformed to the GT/AG rule (18). Therefore it seems reasonable to assume that exon 9␤ is not spliced and belongs to the 3Ј-untranslated region of the murine GR gene.
From these results we conclude that GR␤ is not conserved across species. It is tempting to speculate that during evolution a point mutation in the splice acceptor site is responsible for the absence of GR␤ mRNA in rodents. The obvious lack of GR␤ protein in mice argues against an important role of this isoform in humans. This assumption is supported by recent findings (21) demonstrating that the expression level of GR␤ protein in humans is much lower than previously reported (11) and providing strong evidence against a transdominant negative activity of GR␤.