Alpha 2u-globulin Gene Research Articles

High-resolution FISH mapping of the rat alpha2u-globulin multigene family.

The rat alpha2u-globulins are a group of similar proteins that are encoded by a family of approximately 20 genes located a single locus of < or =880 kbp on Chromosome (Chr) 5q. Individual members of this gene family demonstrate complex tissue, hormonal, and developmental expression patterns despite a high degree of sequence similarity among the members and consequently provide an interesting system for studying the evolution of differential gene expression. Hybridization analysis indicated that gene classes, similar to those identified at the homologous MUP locus in the mouse, do not exist within the rat alpha2u-globulin locus. Furthermore, cross-hybridization analysis revealed the presence of conserved sequences in the 5' and 3' regions flanking the alpha2u-globulin genes, some of which were present in an inverted orientation. We have used high-resolution fiber FISH to examine the structural organization of the alpha2u-globulin locus, and found the genes to be arranged as an array of both direct and inverted repeats. The organization of the rat alpha2u-globulin genes differs from the MUP genes and suggests different evolutionary events have reorganized these homologous sets of genes.

The induction by estrogen of rat alpha 2u-globulin gene expression in mouse L-cells.

Expression of the rat alpha 2u-globulin gene family is regulated in the adult male liver by a number of hormones, including growth hormone, thyroid hormone and several steroids. Upon injection into ovariectomized females, estrogens first induce alpha 2u-globulin expression and then suppress this gene after several days of hormone administration. To study this phenomenon, we developed a mouse L-cell line that expressed the human estrogen receptor. High levels of rat alpha 2u-globulin transcript were induced in stable transfectants of this line carrying a cloned alpha 2u-globulin gene, following exposure to 17 beta-estradiol. Since this induction was inhibited by cycloheximide, the response to estrogen, as to other steroids, appears to be secondary. Using genes with variously deleted 5'-upstream regions, sequences responsible for this induction were located between -730 bp and -223 bp relative to the start of transcription. Examination of the DNA in this region revealed that an estrogen receptor element was located at -590 bp in an area that is highly conserved in most known alpha 2u-globulin genes. Administration of both dexamethasone and estrogen produced a synergistic effect in this system. The induction of alpha 2u-globulin RNA by estrogen in L-cells may re-capitulate the initial response to estrogen in vivo, and therefore represents a good model system to seek the identity of the other factors required to effect full induction.

The analysis of regulatory sequences required for high level induction of rat alpha 2u-globulin gene expression by dexamethasone.

The alpha 2u-globulins are the major urinary proteins in adult male rats. They are encoded by a gene family, the expression of which is under multihormonal control in the liver. Glucocorticoids are positive regulatory hormones and we have analyzed the contribution of 5'-upstream sequences to the induction by dexamethasone of two cloned members of the family transfected into mouse L-cells. The results demonstrate that sequences from -762 bp to -226 bp of clone 91 are required for the 24-fold level of induction that was observed. Addition of 5.5 kb of upstream sequence beyond -762 bp did not alter the level of induction significantly, whereas deletion of the DNA between -762 bp and -226 bp reduced inducibility to about 4-fold. Sequencing of this region revealed that an element, 5'-AGAACAggtTTCAAA-3', similar to the 15 bp consensus glucocorticoid response element 5'-AGAACAnnnTGTACC-3', is situated 513 bp upstream of the transcription start site. We infer that this element or its left half site is necessary for the dexamethasone-induced expression of clone 91 from the observation that a second gene, clone 2, that contained a base substitution at position 5 in the left half site was not inducible. It now appears that at least three distinct cis-acting regulatory regions, all of which bind to the glucocorticoid receptor in vitro, may contribute to the full induction of clone 91 by dexamethasone. These are: the distal upstream region identified by this study, a proximal upstream region that binds not only the receptor but also alpha 2uNF1, a constitutively expressed nuclear protein required for induction and a region within the fourth intron that contains five tandem receptor binding sites.

Three subsets of genes whose tissue specific expression is sex and age-dependent can be identified within the rat alpha 2u-globulin family.

The rat alpha 2u-globulins are encoded by a multigene family whose 20-25 members are subjected to multihormonal regulation that is dependent upon the sex of the animal, the developmental stage and the tissue being examined. Using RT-PCR and diagnostic restriction analysis of the products, we have examined the specificity of the expression of different members of the gene family. All family members can be classified into three subsets, depending on how the amplified cDNA responds to digestion with ApaLI, SstI and VspI. Subset A contains the restriction sites for both ApaLI and SstI but not VspI and typifies the genes expressed in the salivary glands of both mature and juvenile animals of both sexes, where it is the only subset expressed. This subset of genes also accounts for all the transcripts observed in the kidneys and mammary glands of juvenile males. Although subset A was represented in the transcript populations of all the other tissues examined, its proportion relative to the total varied greatly. The two other subsets were subset V, which contains only the restriction site for VspI, and subset N, which lacks all three restriction sites. In all the other tissues examined, two or all three of the subsets were expressed, usually in a manner that was unique to the sex and age of the tissue in question. The proportion of each of the three alpha 2u-globulin subsets in the alpha 2u-globulin gene family was determined by quantitation of the restriction products of amplified genomic DNA. Interestingly, the most prevalent subset in the genome (N) has the most limited tissue expression pattern, but is found in liver and preputial glands, the tissues expressing the most substantial quantities of alpha 2u-globulin. These results indicate the complexity of the regulation of the alpha 2u-globulins and point to the necessity for gene specific analyses if the expression of the family is to be understood in molecular terms.

Diethylnitrosamine- and partial hepatectomy-induced decrease in alpha 2u-globulin mRNA level in the rat liver.

Changes in gene expression in the rat liver were investigated by analyzing cDNA libraries for liver mRNAs from adult male rats injected with a chemical carcinogen diethylnitrosamine (DEN). Differential screening using normal and DEN-treated liver cDNAs as probes demonstrated that some of the mRNA species had noticeably lower abundance in the DEN-treated liver than in the untreated liver. Surprisingly, most of those clones were found to code for alpha 2u-globulin (A2uG), an abundant protein in the male rat liver. Further analysis by in situ hybridization revealed that the decrease in the A2uG mRNA level occurred in the area where liver cells were proliferating due to DEN treatment and/or partial hepatectomy (PH). The findings indicate coincidence of cell proliferation with a decrease in the A2uG gene expression in the adult male rat liver, implying that the A2uG-related change favors chemical carcinogen-induced cell growth.

A Fos-Jun element in the first intron of an alpha 2u-globulin gene.

The hepatic expression of the alpha-2u-globulin gene family is controlled by a variety of hormones including steroids, growth hormone and insulin. The mechanisms by which these hormones affect alpha 2u-globulin expression are only partially understood. Recently we isolated and characterized clone RAP 01, an alpha 2u-globulin gene expressed in the liver. In preliminary experiments we noted that partial hepatectomy, a procedure which results in a sharp rise in the level of the oncoproteins c-Fos and c-Jun, also causes a transient induction of the messenger RNA corresponding to clone RAP 01. Using the DNAseI footprinting technique we were able to show that this clone contains a TPA (phorbol 12-myristate 13-acetate)-responsive element (TRE) in its first intron. This element (denoted as element X) is identical to the consensus AP-1 binding site (TGACTCAG) and is protected by rat liver nuclear extracts as well as by purified c-Jun. Gel retardation experiments show that an oligonucleotide containing the TRE consensus sequence competes for binding of liver nuclear proteins to element X and that antibodies directed against the M2 peptide of the mouse Fos protein or the PEP-2 peptide of Jun prevent the formation of specific complexes with the same element. Moreover, element X functions as a TRE in transfected BWTG3 hepatoma cells treated with TPA. Co-transfection with fos and jun expression vectors mimics the effects of TPA suggesting that AP-1 is in fact the mediator of the observed response. It is concluded that the first intron of RAP 01 contains a functional Fos-Jun element.

The androgen-dependent C4-Slp gene is driven by a constitutively competent promoter.

The androgen-dependent liver protein Slp, together with its constitutively expressed closely related isoform C4, provides a model to address the question of which minimal alteration in DNA can shut off the expression of a gene in a manner reversible by testosterone or by trans-acting mutations. Previous work indicated that sequences located at -1.9, -0.45, and -0.25 kb from the transcription start site of the C4-Slp gene played a critical role in determining its unusual functional divergence from C4. Now, using quantitatively and qualitatively controlled transfection assays in HepG2 human hepatoma cells and mouse L fibroblasts, we have observed that the C4-Slp promoter is fully effective and unhindered by upstream sequences and that the C4 promoter has a consistent albeit modest superiority. The determinant of this nearly 2-fold difference does not coincide with the sites highlighted in previous studies but lies within the most cap-site-proximal nucleotides, at positions -189 to +48. We have also established conditions for cell-free transcription of C4 and C4-Slp from plasmid and cosmid templates by using nuclear extracts from rat and mouse liver of both sexes as well as from L cells. At variance with the rat alpha 2u-globulin gene, C4-Slp transcription in vitro does not require male factors, for it is expressed as efficiently as C4 by all nuclear extracts. Further, the minimal promoter sequences required to direct accurate initiation extend not farther than the most proximal 19 nucleotides. Because L cells efficiently express transfected cosmids covering the whole C4 gene or C4/C4-Slp recombinants, as well as plasmids carrying the C4-Slp promoter, but fail to express the full C4-Slp gene, we favor a model in which the expression of the gene is modulated intragenically.

Hormone withdrawal triggers a premature and sustained gene activation from delayed secondary glucocorticoid response elements.

Glucocorticoid regulatory elements, denoted GREs and delayed secondary GREs (sGREs), bind the purified glucocorticoid receptors via distinctive sequence motifs and confer a primary and delayed secondary hormone inducibility, respectively, upon a linked reporter construct in stably transfected mammalian cells. The delayed secondary responses, but not the primary responses, are preceded by a time lag of several hours and blocked by protein synthesis inhibitors. In this report, we further characterized and distinguished these hormonal inductions. A 206-base pair DNA fragment from the hepatic rat alpha 2u-globulin (RUG) gene, containing at least two delayed sGREs, was specifically activated by glucocorticoids in a dose-dependent manner via a process which is sensitive to receptor antagonist RU486. Delayed sGRE-stimulated production of correctly initiated transcripts was preceded by a time lag of 2 h, a time when the GRE-mediated induction had reached maximal levels. A pulse of glucocorticoids sustained maximal activation of the delayed secondary response but not the primary response. In fact, hormone withdrawal triggered a premature induction of this delayed secondary response, suggesting that delayed sGREs are under both negative and positive control of the hormone receptor. Two separable elements of the 206-base pair fragment, including the 29-base pair sequence of a single receptor binding site, activated the reporter expression as effectively with transient, pulsatile exposure to hormone as with continuous exposure. Our results suggest that the information content of a hormonal pulse is retained, or "memorized," more persistently by a receptor binding site of delayed sGREs than those of the prototypical GREs.

Interaction of estrogen receptor complexes with the promoter region of genes that are negatively regulated by estrogens: The α 2u-globulins

Since estrogens strongly suppress the expression of alpha 2u-globulin genes in the rat liver, we studied the binding of estrogen-receptor complexes to fragments derived from alpha 2u-globulin gene RAO 01 using a DNA-cellulose competition assay. Rat uterus cytosol labelled with [3H]estradiol was used as a source of the estrogen receptor. As a positive control in these experiments we used an oligonucleotide containing the estrogen response element (ERE) cloned into pUC18. Our experiments indicate that estrogen-receptor complexes bind specifically to the ERE and to a fragment of RAO 01 located in the 5'-upstream region (bp -606 up to -575). This fragment is conserved among other members of this gene family. This is the first time that in vitro estrogen receptor binding is observed to gene fragments derived from a gene that is repressed by this steroid in vivo.

Delayed secondary glucocorticoid response elements. Unusual nucleotide motifs specify glucocorticoid receptor binding to transcribed regions of alpha 2u-globulin DNA.

Glucocorticoids stimulate the transcription of rat alpha 2u-globulin (RUG) genes. Because this induction occurs after a time lag of several hours and is blocked by inhibitors of protein synthesis, it exemplifies a delayed secondary response to steroid hormones. In this report, we show that a region of RUG-transcribed DNA (approximately +1800 to +2174) contains multiple footprint sites for glucocorticoid receptor that are, apparently, organized into at least three independent binding clusters. The DNA sequences bound by the receptor and the location of binding sites were determined. A family of sequences related to half-sites of the consensus primary glucocorticoid response element (GRE) is discernible at each cluster of sites. Compared to the consensus GRE, which contains two pseudo-palindromic hexanucleotides arranged in a tail-to-tail fashion and separated by three bases, the arrangements of hexanucleotides within this segment of RUG DNA are unusual and heterogeneous. Methylation interference of a binding cluster containing three receptor footprints demonstrates that certain guanines of the GRE-like hexanucleotides are essential for efficient receptor binding. A synthetic 29-base pair (bp) RUG element, containing one receptor footprint from this cluster, selectively binds the receptor. Within this 29-bp element, six nucleotides separate two directly repeated copies of GRE-like hexanucleotides. RUG DNA fragments containing all or part of the three binding clusters, including the 29-bp element, confer a delayed secondary hormone responsiveness upon a linked heterologous promoter and reporter gene in stably transfected cell lines. We speculate that the unusual DNA sequence motifs of the receptor-binding sites are crucial for the generation of certain delayed secondary responses.

Androgen Receptor Messenger Ribonucleic Acid (mRNA) in the Rat Liver: Changes in mRNA Levels during Maturation, Aging, and Calorie Restriction*

By means of RNAase protection assay with an antisense cRNA probe, we have shown that the liver of the young adult male rat contains androgen receptor (AR) mRNA to a level of 4% compared to the prostate. Steady state levels of AR mRNA in the liver show both sex and age specificity. Compared to that of the male, the female liver contains a markedly reduced amount of AR mRNA. AR mRNA is almost undetectable in livers of prepubertal male (less than 35 days old) and senescent male (greater than 750 days old) rats. Both prepubertal and senescent animals are relatively insensitive to the androgenic induction of alpha 2u-globulin, a hepatic secretory protein. The age-dependent decline in hepatic androgen sensitivity and AR mRNA level can be delayed considerably by a 40% reduction in the dietary calorie intake. Analysis of poly(A)-containing RNA from two liver cell populations, hepatocytes and nonhepatocytes, revealed that only the hepatocytes that express alpha 2u-globulin gene contain AR mRNA. From these results and our earlier observation of in vitro induction of alpha 2u-globulin in isolated rat liver, we conclude 1) that androgen can act directly on hepatocytes to promote alpha 2u-globulin synthesis; 2) that changes in the hepatic androgen sensitivity during maturation and aging are reflections of the age-dependent expression of the receptor gene; and 3) that retardation of the age-dependent loss of androgen sensitivity by calorie restriction is due to a concomitant delay in the decline of the hepatic AR mRNA level.

Androgen regulated expression of the alpha 2u-globulin gene in pancreatic hepatocytes of rat.

Under a copper-deficient regimen, pancreatic cells in the adult rat can be found to undergo differentiation into hepatocytes. Pancreatic hepatocytes induced in male and female rats were examined for the expression of the androgen-inducible hepatic protein, alpha 2u-globulin. Alpha 2u-Globulin protein was demonstrable by immunoperoxidase method in all the pancreatic hepatocytes of male rats. Northern blot analysis confirmed the presence of 1.3 kb alpha 2u-globulin mRNA transcript in the pancreas of male rats with hepatocytes. Orchiectomy resulted in marked decrease of alpha 2u-globulin protein and its mRNA. Administration of dihydrotestosterone to castrated rats resulted in increased levels of alpha 2u-globulin mRNA and the amount of alpha 2u-globulin protein in the pancreatic hepatocytes. Unlike normal males, in intact and ovariectomized females alpha 2u-globulin was not detectable in pancreatic hepatocytes. These results indicate that similar to hepatic parenchymal cells pancreatic hepatocytes synthesize alpha 2u-globulin under androgenic regulation. Furthermore, unlike in liver where it is expressed predominantly in perivenular and midlobular hepatocytes, there is no localized difference in the expression of this gene in the transdifferentiated pancreatic hepatocytes.

Identification of Nuclear Proteins That Bind to the Glucocorticoid Regulatory Region of a Rat α2u-Globulin Gene

Abstract The induction of rat alpha 2u-globulin by glucocorticoids is a secondary response to the hormone, that is protein synthesis is absolutely required for induction. Using the DNase I protection assay, we have identified three proteins present in rat liver nuclei that bind in or near the regulatory region of a cloned alpha 2u-globulin gene. One protein which we term alpha 2u-globulin nuclear factor 1 (alpha 2uNF1), binds to precisely the sequence we have previously shown to be required for hormonal induction. Genes containing linker-scanning mutations in this region show diminished binding to this nuclear factor and display greatly reduced or abolished glucocorticoid response. alpha 2uNF1 was detected in nuclei from several sources, and its level is apparently unaffected by glucocorticoids. Its recognition sequence is unlike those of previously reported transcription factors. We detect two other proteins, alpha 2uNF2 and alpha 2uNF3, that bind near the alpha 2u-globulin regulatory region. A mutant alpha 2u-globulin promoter which does not bind alpha 2uNF2 shows increased inducibility by glucocorticoids in transfected mouse L-cells. The binding of alpha 2uNF3 is not required for alpha 2u-globulin induction by the hormone.

Loss of Androgenic Induction of α2u-Globulin Gene Family in the Liver of NIH Black Rats*

Unlike all known strains of rat, the androgen-inducible alpha 2u-globulin gene family is totally silent in the liver of NIH black (NB) rats. No endocrinological or reproductive abnormalities are apparent, and the mRNA for the androgen-repressible hepatic protein SMP-2 is normally regulated in these animals. Furthermore, immunoblot analysis shows a normal level of the male-specific cytoplasmic androgen-binding protein. Cross-breeding of the NB male and Sprague-Dawley female shows that the hybrid male in the F-1 generation regains the androgen-dependent expression of alpha 2u-globulin in the liver. These results along with the observation of high constitutive level of alpha 2u-globulin mRNA in the preputial gland of NB rats indicate a tissue- and gene-specific regulatory defect which prevents androgenic induction of alpha 2u-globulin in the liver.

Developmentally Regulated Male-Specific Transfactor(s) Enablein VitroTranscription of a Cloned α2U-Globulin Gene

Selected members of the rat alpha 2u-globulin gene family are expressed in several tissues, manifesting characteristic developmental and endocrine transcriptional control. Studies are now underway to identify the responsible cis sequences and transacting factors. We recently reported that the cloned rat alpha 2u-globulin 207 gene manifests tissue-specific androgen-dependent expression; it is expressed in livers of male, but not female, transgenic mice. In the present study a portion of this gene (-639 to +1395) is used as an in vitro template in the presence of nuclear extracts derived from hepatic nuclei of prepubescent and mature male and female rats. alpha-Amanitin-sensitive in vitro transcription of the alpha 2u 207 gene by extracts derived from mature male rats is highly active and is at least 13 times as rapid as that with preparations from mature female or immature animals. In contrast, the rate of transcription of a control template, the adenoviral late promoter, is the same with all of these nuclear extracts. S1 nuclease analysis and the size of the transcript indicate that transcription is initiated in vitro at the same site as it is in vivo and that it continues to the 3' terminus of the alpha 2u-globulin template. Thus, cis sequences are present in this gene fragment which are controlled by developmentally regulated male transacting factors, enabling selective transcription of this alpha 2u-globulin gene. Mixing experiments indicate that this transcriptional control is positive.

Structure and negative transcriptional regulation by glucocorticoids of the acute-phase rat .alpha.1-inhibitor III gene

DNA clones representing the negative acute-phase gene coding for the plasma proteinase inhibitor alpha 1-inhibitor III were isolated from a rat genomic library. Structural analysis established the existence of at least four different members of the alpha 1-inhibitor III gene family. Partial DNA sequence analysis of the 5'-terminal regions was performed for the alpha 1-inhibitor III gene and the related alpha 1-inhibitor IV gene. The transcription start site of the alpha 1-inhibitor III gene was located by S1 mapping and primer extension. No stable alpha 1-inhibitor IV mRNA was detected in rat liver. In an experimentally induced acute-phase reaction, the transcription rate of the alpha 1-inhibitor III gene was reduced 12.7-fold at 6 h after stimulation. Four hours after injection of a high dose of dexamethasone into rats, the transcription rate of this gene was reduced 9-fold. Thus, glucocorticoids alone are capable of causing a strong transient down-regulation of the transcription of this gene in rats, independent of other inflammatory mediators. An inverted consensus glucocorticoid responsive element (5'GGACACAATAT3') shared with the glucocorticoid-regulated alpha 1-fetoprotein, alpha 2u-globulin, and alpha 1-acid glycoprotein genes was detected by computer-assisted sequence analysis in the promoter proximal 5'-flanking region of the alpha 1-inhibitor III gene.

Nuclear Factors from Expressing Tissues Interact in Vitro with a Rat α-2u Globulin Gene Intron

Members of the rat alpha 2u globulin gene family are expressed in a tissue-specific manner. Using a DNase I footprinting assay, we find that expressing tissues (liver, lachrymal, and salivary gland) contain nuclear proteins that interact specifically with two sites in the third intron of a cloned gene. These sequences are similar to a CCAAT-like sequence in the 5'-flanking region of the gene and the binding site for the transcription factor Sp1. Extracts of nonexpressing tissues (kidney, testis, brain, and spleen) do not show footprinting activity at these sites. Whereas hepatic expression of alpha 2u globulin is restricted to normal adult males, nuclear extracts of liver from all animals (adult and juvenile, male and female, intact and hypophysectomized) produce similar footprints. Footprints by extracts of lachrymal and salivary glands, tissues that express a subset of alpha 2u globulin genes distinct from the hepatic set, are also similar to the liver activity. The findings suggest that the third intron of an alpha 2u globulin gene contains sites that interact with tissue-specific factors to establish the potential for hormonal and developmental control of expression.

Glucocorticoid receptor binding to defined regions of α2 u-globulin genes

A DNA-cellulose competition assay was used to study binding of glucocorticoid receptor complexes to two alpha 2u-globulin genes, RAP 01 and RAO 01. Two binding regions were found in RAP 01, one localized between bp -642 and -452, the other between -252 and -118 from the transcriptional initiation site. Only the second region was found in RAO 01. The binding affinity was comparable to that observed using a long terminal repeat fragment of mouse mammary tumor virus. Both regions contain sequences homologous to a 17-bp consensus proposed for the glucocorticoid receptor binding. Therefore, despite evidence that glucocorticoid induction of alpha 2u-globulin may be indirect, a direct action of the receptor in the 5'-upstream region of the genes cannot be excluded.

Tissue-Specific and Hormonally Regulated Expression of a Rat α2u Globulin Gene in Transgenic Mice

To investigate the tissue-specific and hormonal regulation of the rat alpha 2u globulin gene family, we introduced one cloned member of the gene family into the mouse germ line and studied its expression in the resulting transgenic mice. Alpha 2u globulingene 207 was microinjected on a 7-kilobase DNA fragment, and four transgenic lines were analyzed. The transgene was expressed at very high levels, specifically in the liver and the preputial gland of adult male mice. The expression in male liver was first detected at puberty, and no expression was detected in female transgenic mice. This pattern of expression is similar to the expression of endogenous alpha 2u globulin genes in the rat but differs from the expression of the homologous mouse major urinary protein (MUP) gene family in that MUPs are synthesized in female liver and not in the male preputial gland. We conclude that these differences between rat alpha 2u globulin and mouse MUP gene expression are due to evolutionary differences in cis-acting regulatory elements. The expression of the alpha 2u globulin transgene in the liver was abolished by castration and fully restored after testosterone replacement. The expression could also be induced in the livers of female mice by treatment with either testosterone or dexamethasone, following ovariectomy and adrenalectomy. Therefore, the cis-acting elements responsible for regulation by these two hormones, as well as those responsible for tissue-specific expression, are closely linked to the alpha 2u globulin gene.

Tissue-specific and hormonally regulated expression of a rat alpha 2u globulin gene in transgenic mice.

To investigate the tissue-specific and hormonal regulation of the rat alpha 2u globulin gene family, we introduced one cloned member of the gene family into the mouse germ line and studied its expression in the resulting transgenic mice. Alpha 2u globulingene 207 was microinjected on a 7-kilobase DNA fragment, and four transgenic lines were analyzed. The transgene was expressed at very high levels, specifically in the liver and the preputial gland of adult male mice. The expression in male liver was first detected at puberty, and no expression was detected in female transgenic mice. This pattern of expression is similar to the expression of endogenous alpha 2u globulin genes in the rat but differs from the expression of the homologous mouse major urinary protein (MUP) gene family in that MUPs are synthesized in female liver and not in the male preputial gland. We conclude that these differences between rat alpha 2u globulin and mouse MUP gene expression are due to evolutionary differences in cis-acting regulatory elements. The expression of the alpha 2u globulin transgene in the liver was abolished by castration and fully restored after testosterone replacement. The expression could also be induced in the livers of female mice by treatment with either testosterone or dexamethasone, following ovariectomy and adrenalectomy. Therefore, the cis-acting elements responsible for regulation by these two hormones, as well as those responsible for tissue-specific expression, are closely linked to the alpha 2u globulin gene.