Identification of differentially expressed genes in the kidneys of growth hormone transgenic mice.

OBJECTIVE: Bovine growth hormone (bGH) transgenic mice develop severe kidney damage. This damage may be due, at least in part, to changes in gene expression. Identification of genes with altered expression in the bGH kidney may identify mechanisms leading to damage in this system that may also be relevant to other models of kidney damage. DESIGN: cDNA subtraction libraries, northern blot analyses, microarray analyses and real-time reverse transcription polymerase chain reaction (RT/PCR) assays were used to identify and verify specific genes exhibiting differential RNA expression between kidneys of bGH mice and their non-transgenic (NT) littermates. RESULTS: Immunoglobulins were the vast majority of genes identified by the cDNA subtractions and the microarray analyses as being up-regulated in bGH. Several glycoprotein genes and inflammation-related genes also showed increased RNA expression in the bGH kidney. In contrast, only a few genes were identified as being significantly down-regulated in the bGH kidney. The most notable decrease in RNA expression was for the gene encoding kidney androgen-regulated protein. CONCLUSIONS: A number of genes were identified as being differentially expressed in the bGH kidney. Inclusion of two groups, immunoglobulins and inflammation-related genes, suggests a role of the immune system in bGH kidney damage.

Does adiposity status influence femoral cortical strength in rodent models of growth hormone deficiency?

Growth hormone (GH)-deficiency is usually associated with elevated adiposity, hyperleptinemia, and increased fracture risk. Since leptin is thought to enhance cortical bone formation, we have investigated the contribution of elevated adiposity and hyperleptinemia on femoral strength in rodent models of GH deficiency. Quantification of the transpubertal development of femoral strength in the moderately GH-deficient/hyperleptinemic Tgr rat and the profoundly GH-deficient/hypoleptinemic dw/dw rat revealed that the mechanical properties of cortical bone in these two models were similarly compromised, a 25-30% reduction in failure load being entirely due to impairment of geometric variables. In contrast, murine models of partial (GH antagonist transgenic) and complete (GH receptor-null) loss of GH signaling and elevated adiposity showed an impairment of femoral cortical strength proportionate to the reduction of GH signaling. To determine whether impaired femoral strength is exacerbated by obesity/hyperleptinemia, femoral strength was assessed in dw/dw rats following two developmental manipulations that elevate abdominal adiposity and circulating leptin, neonatal monosodium glutamate (MSG) treatment, and maintenance on an elevated fat diet. The additional impairment of femoral strength following MSG treatment is likely to have resulted from a reduction in residual activity of the hypothalamo-pituitary-GH-IGF-I axis, but consumption of elevated dietary fat, which did not reduce circulating IGF-I, failed to exacerbate the compromised femoral strength in dw/dw rats. Taken together, our data indicate that the obesity and hyperleptinemia usually associated with GH deficiency do not exert a significant influence over the strength of cortical bone.

Mouse molar dentin size/shape is dependent on growth hormone status.

Growth hormone (GH) status affects dental development, but how GH influences tooth size/shape is unclear. Since GH affects dental epithelial proliferation, we hypothesized that GH influences the tooth crown and root dimensions. Dentin matrix dimensions were measured in longitudinal sections of decalcified first mandibular molars from 3 genetically modified mice: giant (GH-Excess) mice and dwarf (GH-Antagonist and GH-Receptor-Knockout) mice. GH status was found to influence crown width, root length, and dentin thickness. Analysis of these data suggests that GH influences both tooth crown and root development prior to dentinogenesis as well as during appositional growth of dentin. This is concordant with the expression of paracrine GH and GH receptors during tooth bud morphogenesis, and of GH receptors in the enamel organ, dental papilla, and Hertwig's epithelial root sheath during dentinogenesis. Based on prior studies, these GH morphogenetic actions may be mediated by the induction of both bone morphogenetic protein and insulin-like growth factor-1 expression.

Chronic changes in peripheral growth hormone levels do not affect ghrelin stomach mRNA expression and serum ghrelin levels in three transgenic mouse models.

Ghrelin is an endogenous ligand for the growth hormone secretagogue (GHS) receptor. Ghrelin is involved in feeding behaviour and is a potent stimulator of GH release. Chronically increased GH concentrations are known to negatively regulate the pituitary GHS receptor. This study tested whether chronic changes in peripheral GH levels/action affect ghrelin mRNA expression and circulating concentrations of ghrelin. Stomach ghrelin mRNA expression and serum concentrations of ghrelin were measured in three groups of transgenic mice and the respective control animals: group 1, GH-receptor gene disrupted mice (GHR/KO); group 2, mice expressing bovine GH (bGH); and group 3, mice expressing GH-antagonist (GHA). Ghrelin mRNA expression in the stomach, pituitary and hypothalamus of young adult male rats were measured using reverse-transcription-polymerase chain reaction. Ghrelin mRNA expression levels were approximately 3000-fold higher in rat stomach than in rat pituitary. Ghrelin mRNA expression in rat hypothalamus was below the detection limits of our assay. Stomach ghrelin mRNA expression, as well as serum concentrations of ghrelin, did not change significantly in any of the three mouse groups compared to the respective control group. These data support previous observations that the stomach is the main source of circulating ghrelin, and also indicate that stomach ghrelin mRNA expression and serum concentrations of ghrelin are not affected by chronic changes in peripheral GH/insulin-like growth factor-I levels/action.

Mouse cellular cementum is highly dependent on growth hormone status.

Cementum is known to be growth-hormone (GH)-responsive, but to what extent is unclear. This study examines the effects of extremes of GH status on cementogenesis in three lines of genetically modified mice; GH excess (giant), GH antagonist excess (dwarf), and GH receptor-deleted (GHR-KO) (dwarf). Age-matched mandibular molar tissues were processed for light microscope histology. Digital images of sections of first molar teeth were captured for morphometric analysis of lingual root cementum. Cross-sectional area of the cellular cementum was a sensitive guide to GH status, being reduced nearly 10-fold in GHR-KO mice, three-fold in GH antagonist mice, and increased almost two-fold in giant mice (p < 0.001). Cellular cementum length was similarly influenced by GH status, but to a lesser extent. Acellular cementum was generally unaffected. This study reveals cellular cementum to be a highly responsive GH target tissue, which may have therapeutic applications in assisting regeneration of the periodontium.

The effect of GHRH on somatotrope hyperplasia and tumor formation in the presence and absence of GH signaling.

Excessive GHRH stimulation leads to somatotrope hyperplasia and, ultimately, pituitary adenoma formation in the metallothionein promoter-driven human GHRH (hGHRH) transgenic mouse. This pituitary phenotype is similar to that observed in humans with ectopic production of GHRH. In both mice and man, GHRH hyperstimulation also results in dramatic increases in circulating GH and IGF-I. To determine whether GH/IGF-I modulates the development and growth rate of GHRH-induced pituitary tumors, pituitary growth and histology were evaluated in mice generated from cross-breeding metallothionein promoter-driven hGHRH transgenic mice with GH receptor binding protein (GHR) gene disrupted mice (GHR(-/-)). Expression of the hGHRH transgene in 2-month-old GHR intact (GHR(+)) mice resulted in the doubling of pituitary weight that was largely attributed to an increase in the number of GH-immunopositive cells. Pituitary weight of GHR(+) hGHRH mice did not significantly change between 2 and 6 months of age, whereas at 12 months, weights increased up to 100-fold those of GHR(+) pituitaries, and 70% of the glands contained grossly visible adenomas. All adenomas stained positively for GH, whereas some showed scattered PRL staining. Pituitaries of GHR(-/-) mice were half the size of those of GHR(+) mice. Although reduced in size, the histological features of GHR(-/-) mouse pituitaries were suggestive of somatotrope hyperplasia. Despite evidence of somatotrope hyperplasia, pituitaries from GHR(-/-) mice as old as 28 months of age were similar in size to those of 2-month-old mice and did not show signs of adenoma formation. Expression of the hGHRH transgene in GHR(-/-) mice did not significantly increase pituitary size between 2 and 6 months of age. However, at 12 months the majority of GHR(-/-), hGHRH pituitaries developed adenomas with mean pituitary weight and histological features similar to those of GHR(+), hGHRH mice. These observations demonstrate that intact GH signaling is not required for GHRH tumor formation. Although the majority of GHR(+), hGHRH and GHR(-/-), hGHRH pituitaries developed tumors by 12 months of age, a small subset remained morphologically indistinct from those at 2 months of age. These observations taken together with the fact that overt tumor formation is preceded by a static pituitary growth phase between 2 and 6 months, indicates that protective mechanisms are in place to maintain pituitary mass despite hGHRH hyperstimulation.

Evidence that age-induced decline in memory retention is delayed in growth hormone resistant GH-R-KO (Laron) mice.

Although the role of growth hormone (GH) in aging is controversial, the recent production of GH-R-KO mice may provide a means for elucidating its importance. Using the inhibitory avoidance learning task as a measure of cognitive aging, the present study compared learning and retention in young and old GH-R-KO mice and their normal siblings. Results for the old normal animals agreed with the current literature, in that the ability of old animals to retain learned information declined over time. However, retention in the old GH-R-KO mice did not decline between the 24-h, 7-day and 28-day retention tests and did not differ from young animals. To determine whether performance differences seen in the old normal vs. old GH-R-KO groups were due to locomotor behavior or emotionality, both groups were tested in the elevated-plus maze. Results showed that the normal and GH-R-KO mice did not differ in number of open or closed arms entered, time spent in closed or open arms or time taken to first enter an open arm. Thus, it was concluded that inhibitory avoidance performance was not affected by differences of locomotor activity or emotionality, and that the absence of GH signaling may be associated with improved long-term memory in aging mice.

The growth hormone (GH)-axis of GH receptor/binding protein gene-disrupted and metallothionein-human GH-releasing hormone transgenic mice: hypothalamic neuropeptide and pituitary receptor expression in the absence and presence of GH feedback.

Elevation of circulating GH acts to feed back at the level of the hypothalamus to decrease GH-releasing hormone (GHRH) and increase somatostatin (SRIF) production. In the rat, GH-induced changes in GHRH and SRIF expression are associated with changes in pituitary GHRH receptor (GHRH-R), GH secretagogue receptor (GHS-R), and SRIF receptor subtype messenger RNA (mRNA) levels. These observations suggest that GH regulates its own synthesis and release not only by altering expression of key hypothalamic neuropeptides but also by modulating the sensitivity of the pituitary to hypothalamic input, by regulating pituitary receptor synthesis. To further explore this possibility, we examined the relationship between the expression of hypothalamic neuropeptides [GHRH, SRIF, and neuropeptide Y (NPY)] and pituitary receptors [GHRH-R, GHS-R, and SRIF receptor subtypes (sst2 and sst5)] in two mouse strains with alterations in the GH-axis; the GH receptor/binding protein gene-disrupted mouse (GHR/BP-/-) and the metallothionein promoter driven human GHRH (MT-hGHRH) transgenic mouse. In GHR/BP-/- mice, serum insulin-like growth factor I levels are low, and circulating GH is elevated because of the lack of GH negative feedback. Hypothalamic GHRH mRNA levels in GHR/BP-/- mice were 232 +/- 20% of GHR/BP+/+ littermates (P < 0.01), whereas SRIF and NPY mRNA levels were reduced to 86 +/- 2% and 52 +/- 3% of controls, respectively (P < 0.05; ribonuclease protection assay). Pituitary GHRH-R and GHS-R mRNA levels of GHR/BP-/- mice were elevated to 275 +/- 55% and 319 +/- 68% of GHR/BP+/+ values (P < 0.05, respectively), whereas the sst2 and sst5 mRNA levels did not differ from GHR/BP intact controls as determined by multiplex RT-PCR. Therefore, in the absence of GH negative feedback, both hypothalamic and pituitary expression is altered to favor stimulation of GH synthesis and release. In MT-hGHRH mice, ectopic hGHRH transgene expression elevates circulating GH and insulin-like growth factor I. In this model of GH excess, endogenous (mouse) hypothalamic GHRH mRNA levels were reduced to 69 +/- 6% of nontransgenic controls, whereas SRIF mRNA levels were increased to 128 +/- 6% (P < 0.01). NPY mRNA levels were not significantly affected by hGHRH transgene expression. Also, MT-hGHRH pituitary GHRH-R and GHS-R mRNA levels did not differ from controls. However, sst2 and sst5 mRNA levels in MT-hGHRH mice were increased to 147 +/- 18% and 143 +/- 16% of normal values, respectively (P < 0.05). Therefore, in the presence of GH negative feedback, both hypothalamic and pituitary expression is altered to favor suppression of GH synthesis and release.

Prolactin, growth hormone, and epidermal growth factor activate Stat5 in different compartments of mammary tissue and exert different and overlapping developmental effects.

Prolactin (Prl)-induced phosphorylation of Stat (signal transducer and activator of transcription) 5 is considered a key event in functional mammary development and differentiation. We now demonstrate that not only Prl, but also growth hormone (GH) and epidermal growth factor (EGF), can activate Stat5 in mammary tissue. We investigated the roles of these hormones in mammary development using mice in which the respective receptors had been inactivated. Although Prl receptor (PrlR)-null mice are infertile, we were able to maintain pregnancies in a few mice by treatment with progesterone. Mammary tissue in these mice was severely underdeveloped and exhibited limited differentiation as assessed by the phosphorylation status of Stat5 and the expression of milk protein genes. PrlR +/- mice showed impaired mammary development and alveolar differentiation during pregnancy, which corresponded with reduced phosphorylation levels of Stat5a and 5b, and impaired expression of milk protein genes. Development of the glands in these mice was arrested at around day 13 of pregnancy. While Prl activated Stat5 only in the epithelium, GH and EGF activated Stat5 preferentially in the stroma. To assess the relevance of the GH receptor (GHR) in the mammary gland, we transplanted GHR-null epithelium into cleared fat pads of wild-type mice. These experiments demonstrated that the GHR in the epithelium is not required for functional mammary development. Similarly, the EGFR in the epithelium is not required for alveolar development. In contrast, epithelial PrlR is required for mammary development and milk protein gene expression during pregnancy. Although GH is not required for alveolar development, we were able to demonstrate its lactogenic function in cultured mammary epithelium from PrlR-null mice. However, ductal development in GHR-null mice was impaired, supporting the notion that GH signals through the stromal compartment. Our findings demonstrate that GH, Prl, and EGF activate Stat5 in separate compartments, which in turn reflects their specific roles in ductal and alveolar development and differentiation.

Secretagogues and the somatotrope: signaling and proliferation.

Somatotrope function requires consideration of both growth hormone (GH) secretion and cellular proliferation. The regulation of these processes is, to a large extent, controlled by three hypothalamic hormones: GH-releasing hormone (GHRH), somatostatin (SRIF), and an as-yet-unidentified GH secretagogue (GHS). Each binds to G protein-linked membrane receptors through which signaling occurs. Our laboratory has used a series of genetic and transgenic models with perturbations of individual components of the GH regulatory system to study both somatotrope signaling and proliferation. Impaired GHRH signaling is present in the lit mouse, which has a GHRH receptor (R) mutation, and the dw rat, which has a post-receptor signaling defect. Both models also have impaired responses to a GHS, implying an interaction between the two signaling systems. The spontaneous dwarf rat (SDR), in which a mutation of the GH gene results in total absence of the hormone, shows characteristic changes in the hypothalamic regulatory hormones due to an absence of GH feedback and alterations in the expression of each of their pituitary receptors. Treatment of SDRs with GHRH and a GHS has allowed demonstration of a stimulatory effect of GHRH on GHRH-R, GHS-R, and SRIF type 2 receptor (SSTR-2) expression and an inhibitory effect on SSTR-5 expression. GH also modifies the expression of these receptors, though its effects are seen at later time periods and appear to be indirect. Overall, the results indicate a complex regulation of GH secretion in which somatotrope receptor, as well as ligand expression, exerts an important physiological role. Both the SDR and the GH-R knockout (ko) mouse have small pituitaries and decreased somatotropes, despite elevated GHRH secretion and intact GHRH-R signaling. Introduction of the hGHRH transgene into GH-R ko mice confirmed that the proliferative effects of GHRH require GH/insulin-like growth factor-I (IGF-I) action. The results offer new insights into factors participating in somatotrope proliferation.

Assessment of growth parameters and life span of GHR/BP gene-disrupted mice.

GH has many biological roles, including promotion of growth. Most, if not all, of its roles are achieved through interaction with its receptor. We chose to study the effects of loss of GH signaling on growth and aging in a mouse model for Laron Syndrome (LS) in which the GHR/BP gene has been disrupted. We observed that mice homozygous for the disruption (-/-) were significantly smaller than normal wild-type (+/+) mice as well as mice heterozygous for the disruption, even at 1.5 yr of age. IGF-I levels were also significantly lower in the -/- mice and remained low as the mice aged. IGFBP-3 levels were severely reduced in the -/- mice, whereas IGFBP-1, -2, and -4 levels remained unchanged. Finally, the -/- mice lived significantly longer than +/+ and +/- mice. The latter result contradicts the anti-aging GH data and suggests the need for further analysis of GH and aging.

Evidence for growth hormone (GH) autoregulation in pituitary somatotrophs in GH antagonist-transgenic mice and GH receptor-deficient mice.

Growth hormone (GH) modulates the hypothalamic release of somatostatin and GH-releasing hormone; however, there has been no evidence of GH autoregulation on the pituitary somatotroph. To determine the effects of GH on its own regulation, we examined the pituitaries of giant transgenic mice expressing a GH agonist (E117L), dwarf transgenic mice expressing a GH antagonist (G119K), and dwarf mice devoid of the GH receptor/binding protein (GHR/BP). In the E117L transgenic mice, the number and distribution of pituitary GH-immunoreactive cells were unchanged from nontransgenic littermate controls; an ultrastructural examination revealed typical, densely granulated somatotrophs. In contrast, the pituitaries of the G119K mice contained both moderately granulated somatotrophs and a sparsely granulated (SG) population with well-developed synthetic organelles and a distinct juxtanuclear globular GH-staining pattern. GHR/BP-deficient mice exhibited a marked reduction in the intensity of cytoplasmic GH immunoreactivity; however, prominent GH staining in the juxtanuclear Golgi was seen. GH-immunoreactive cells were increased in number, and the reticulin network pattern was distorted; stains for proliferating cell nuclear antigen confirmed mild hyperplasia. Electron microscopy showed that the somatotrophs were hyperactive SG cells with prominent endoplasmic reticulum membranes, large Golgi complexes, and numerous mitochondria. These findings are consistent with synthetic and secretory hyperactivity in pituitary somatotrophs due to the reduced GH feedback regulation. The changes are most striking in animals that are devoid of GHR/BP and less marked in animals expressing a GH antagonist; both models had reduced insulin-like growth factor-I levels, but the more dramatic change in the GHR/BP animals can be explained by abrogated GH signaling. This represents the first evidence of direct GH feedback inhibition on pituitary somatotrophs, which may have implications for the use of GH analogs in different clinical settings.

Transgenic models of growth hormone action.

A growth-promoting principle of the pituitary gland was discovered in 1921, and bovine growth hormone (GH) was isolated in 1944. Since then, the structure of GH as it relates to its biological activities has been an exciting research topic. Equally fascinating is the relationship between GH structure and its metabolic activities. In attempts to define some of these activities, several investigators have used GH transgenic mice as models. In this review we summarize what is known about the molecular mechanisms of GH action. We then describe some of the GH transgenic models and point out potential targets for nutrition research.

Deficits in female reproductive function in GH-R-KO mice; role of IGF-I.

Mice homozygous for targeted disruption of the GH receptor/GH binding protein gene (GH-R-KO mice; -/-) exhibit reduced plasma IGF-I levels, elevated plasma GH levels, and dwarf phenotype. Although most GH-R-KO mice are fertile, age at first conception is greatly delayed in -/- x -/- matings. Here we report that the age of vaginal opening is significantly delayed in GH-R-KO vs. normal mice, but it can be advanced by treatment with recombinant human (rh)IGF-I. In pregnant GH-R-KO females, fetal size is reduced and pregnancy is prolonged while placental weight is, unexpectedly, increased. Alterations in fetal and placental weight are related to maternal rather than fetal genotype. Moreover, litter size and body weight of newborn pups are significantly reduced in GH-R-KO vs. normal females. Reduction in litter size reflects both dam and sire effects. We conclude that GH resistance and consequent reduction in peripheral IGF-I levels is associated with delay of female puberty, alterations in fetal and placental growth, delay of parturition, and reduced litter size.

Pituitary and testicular function in growth hormone receptor gene knockout mice.

The role of GH in the control of pituitary and testicular function is poorly understood. GH receptor gene knockout (GHR-KO) mice were recently produced. As these mice are good experimental animals to assess the influence of the effects of GH and insulin-like growth factor-I (IGF-I), the present studies were undertaken. Young adult male GHR-KO mice and their normal siblings were tested for fertility and subsequently injected (i.p.) with saline or GnRH (1 ng/g BW) in saline. Fifteen minutes later, blood was obtained via heart puncture. Plasma IGF-I, PRL, LH, and testosterone concentrations were measured by RIAs. In addition, the testicular testosterone response to LH treatment was evaluated in vitro. The results indicate that the absence of GH receptors (GHRs) was associated with an increase (P < 0.005) in plasma PRL levels, and circulating IGF-I was not detectable. Although the basal plasma LH levels were similar in GHR-KO mice relative to those in their normal siblings, the circulating LH response to GnRH treatment was significantly (P < 0.001) attenuated. Plasma testosterone levels were unaffected by disruption of the GHR gene. However, basal (P < 0.01) and LH-stimulated (P < 0.001) testosterone release from the isolated testes of GHR-KO mice were decreased. The rate of fertility in GHR-KO male mice was also reduced. These results indicate that the lack of GHRs (with GH resistance and lack of IGF-I secretion) induces hyperprolactinemia and alters the effect of GnRH on LH secretion as well as testicular function. Thus, GH and IGF-I influence pituitary and gonadal functions in male mice.

Arabidopsis gls mutants and distinct Fd-GOGAT genes. Implications for photorespiration and primary nitrogen assimilation.

Ferredoxin-dependent glutamate synthase (Fd-GOGAT) plays a major role in photorespiration in Arabidopsis, as has been determined by the characterization of mutants deficient in Fd-GOGAT enzyme activity (gls). Despite genetic evidence for a single Fd-GOGAT locus and gene, we discovered that Arabidopsis contains two expressed genes for Fd-GOGAT (GLU1 and GLU2). Physical and genetic mapping of the gls1 locus and GLU genes indicates that GLU1 is linked to the gls1 locus, whereas GLU2 maps to a different chromosome. Contrasting patterns of GLU1 and GLU2 expression explain why a mutation in only one of the two genes for Fd-GOGAT leads to a photorespiratory phenotype in the gls1 mutants. GLU1 mRNA was expressed at the highest levels in leaves, and its mRNA levels were specifically induced by light or sucrose. In contrast, GLU2 mRNA was expressed at lower constitutive levels in leaves and preferentially accumulated in roots. Although these results suggest a major role for GLU1 in photorespiration, the sucrose induction of GLU1 mRNA in leaves also suggests a role in primary nitrogen assimilation. This possibility is supported by the finding that chlorophyll levels of a gls mutant are significantly lower than those of the wild type when grown under conditions that suppress photorespiration. Both the mutant analysis and gene regulation studies suggest that GLU1 plays a major role in photorespiration and also plays a role in primary nitrogen assimilation in leaves, whereas the GLU2 gene may play a major role in primary nitrogen assimilation in roots.

THE MOLECULAR-GENETICS OF NITROGEN ASSIMILATION INTO AMINO ACIDS IN HIGHER PLANTS.

Nitrogen assimilation is a vital process controlling plant growth and development. Inorganic nitrogen is assimilated into the amino acids glutamine, glutamate, asparagine, and aspartate, which serve as important nitrogen carriers in plants. The enzymes glutamine synthetase (GS), glutamate synthase (GOGAT), glutamate dehydrogenase (GDH), aspartate aminotransferase (AspAT), and asparagine synthetase (AS) are responsible for the biosynthesis of these nitrogen-carrying amino acids. Biochemical studies have revealed the existence of multiple isoenzymes for each of these enzymes. Recent molecular analyses demonstrate that each enzyme is encoded by a gene family wherein individual members encode distinct isoenzymes that are differentially regulated by environmental stimuli, metabolic control, developmental control, and tissue/cell-type specificity. We review the recent progress in using molecular-genetic approaches to delineate the regulatory mechanisms controlling nitrogen assimilation into amino acids and to define the physiological role of each isoenzyme involved in this metabolic pathway.

Sex-specific transcriptional regulation by the male and female doublesex proteins of Drosophila.

The somatic sexual phenotype of Drosophila is regulated by the sexual differentiation pathway. Male (DSXM) and female (DSXF) proteins encoded by doublesex (dsx), a gene at the end of this pathway, bind to three sites within a 127-bp enhancer that directs sex- and tissue-specific transcription of Yolk protein genes. We describe mutagenesis of these binding sites and the resulting effects on DSXM and DSXF binding in vitro and on gene regulation in wild-type and dsx mutant flies. The results demonstrate that DSXM represses and DSXF activates transcription from the two strongest binding sites. Thus, the pathway regulates sex-specific transcription through the male and female dsx proteins that act directly on the target gene, but with opposite effects.

The doublesex proteins of Drosophila melanogaster bind directly to a sex-specific yolk protein gene enhancer.

The doublesex (dsx) gene of Drosophila melanogaster encodes both male-specific and female-specific polypeptides, whose synthesis is regulated by alternative sex-specific splicing of the primary dsx transcript. The alternative splicing of the dsx mRNA is the last known step in a cascade of regulatory gene interactions that involves both transcriptional and post-transcriptional mechanisms. Genetic studies have shown that the products of the dsx locus are required for correct somatic sexual differentiation of both sexes, and have suggested that each dsx product functions by repressing expression of terminal differentiation genes specific to the opposite sex. However, these studies have not shown whether the dsx gene products function directly to regulate the expression of target genes, or indirectly through another regulatory gene. We report here that the male- and female-specific DSX proteins, expressed in E.coli, bind directly and specifically in vitro to three DNA sequences located in an enhancer region that regulates female-specific expression of two target genes, the yolk protein genes 1 and 2. This result suggests strongly that dsx is a final regulatory gene in the hierarchy of regulatory genes controlling somatic sexual differentiation.