An estrogen receptor-alpha splicing variant mediates both positive and negative effects on gene transcription

Analysis of mRNA prepared from a variety of estrogen-responsive cell lines, breast tumor specimens, and normal breast tissue have established that estrogen receptor-alpha (ER alpha) mRNA is typically expressed as a mixture of transcripts. Using PCR amplification, this heterogeneity has been shown to result largely from an imprecise pattern of mRNA splicing that gives rise to a family of correctly processed and exon-skipped ER alpha transcripts. We have reconstructed ER alpha cDNAs representing the single exon-skipped variants ERdeltaE2 through ERdeltaE7 to enable their functional characterization in a well defined cell transfection system. All six of the ER alpha splicing variants support the efficient expression of stable proteins in Cos7 cells, and each shows a characteristic pattern of subcellular distribution. Each of the variants displays a dramatic reduction in DNA-binding activity with a consensus estrogen response element (ERE) in an in vitro gel mobility shift assay. While this DNA-binding defect appears to be complete for ERdeltaE2, ERdeltaE3, ERdeltaE4, and ERdeltaE6, weak DNA binding is observed for ERdeltaE5 and ERdeltaE7. Scatchard analysis of hormone binding demonstrates that among the variants, only ERdeltaE3 binds 17beta-estradiol (E2) and does so with an affinity similar to wild-type ER alpha (wt ER alpha). Individual variants cotransfected with the pERE-TK-CAT reporter plasmid [a consensus ERE-driven chloramphenicol acetyltransferase (CAT) reporter gene that is highly responsive to E2-liganded wt ER alpha] were ineffective at inducing CAT expression in ER-negative HeLa cells. Only ERdeltaE5 showed indications of positive transcriptional activity on the pERE-TK-CAT reporter, but this activity was limited to approximately 5% of the activity of wt ER alpha. When variants were expressed simultaneously with wt ER alpha, ERdeltaE3 and ERdeltaE5 were observed to have a dominant negative effect on wt ER alpha transcriptional activity. Like the wild-type receptor, both ERdeltaE3 and ERdeltaE5 interact with steroid receptor coactivator-1e (SRC-1e) in vitro; however, only ERdeltaE3 retained the ability to dimerize with wt ER alpha. Transcription from a region of the ovalbumin promoter, which contains an ERE half-site and an AP-1 motif, is positively regulated by liganded wt ER alpha and ERdeltaE3 in phorbol ester-treated, transiently transfected HeLa cells. In both cases, this activity was enhanced by cotransfected cJun. These observations suggest that selected ER alpha splicing variants are likely to exert important transcriptional effects, especially on genes that are regulated by nonconsensus EREs and subject to complex hormonal control.

Differential expression and regulation by 20-hydroxyecdysone of mosquito ultraspiracle isoforms

Ultraspiracle (USP), the insect homologue of the vertebrate retinoid X receptor, is an obligatory dimerization partner for the ecdysteroid receptor (EcR). Two USP isoforms, USP-A and USP-B, with distinct N-termini, occur in the mosquito Aedes aegypti. In the fat body and ovary, USP-A mRNA is highly expressed during the pre- and late vitellogenic stages, corresponding to a period of low ecdysteroid titer, while USP-B mRNA exhibits its highest levels during the vitellogenic period, correlating with a high ecdysteroid titer. Remarkably, 20-hydroxyecdysone (20E) has opposite effects on USP isoform transcripts in in vitro fat body culture. This steroid hormone upregulates USP-B transcription and its presence is required to sustain a high level of USP-B expression. In contrast, 20E inhibits activation of USP-A transcription. Although EcR.USP-A recognizes the same ecdysteroid-responsive elements, EcR.USP-B binds them with an affinity twofold higher than that of EcR.USP-A. Likewise, EcR.USP-B transactivates a reporter gene in CV-1 cells twofold more strongly than EcR.USP-A. These results suggest that USP-B functions as a major heterodimerization partner for EcR during the vitellogenic response to 20E in the mosquito.

DNA binding and transactivation characteristics of the mosquito ecdysone receptor-Ultraspiracle complex

The steroid hormone 20-hydroxyecdysone is a key regulatory factor, controlling blood-meal triggered egg maturation in mosquitoes. To elucidate the ecdysone hierarchy governing this event, we cloned and characterized the ecdysone receptor (AaEcR) and the nuclear receptor Ultraspiracle (AaUSP), a retinoid X receptor homologue, from the mosquito, Aedes aegypti, which form a functional complex capable of ligand and DNA binding. Here we analyzed the DNA-binding properties of the AaEcR.AaUSP heterodimer with respect to the effects of nucleotide sequence, orientation, and spacing between half-sites in natural Drosophila and synthetic ecdysone response element (EcREs). By using an electrophoretic gel mobility shift assay, we showed that AaEcR.AaUSP exhibits a broad binding specificity, forming complexes with inverted (IR) and direct (DR) repeats of the nuclear receptor response element half-site consensus sequence AGGTCA separated by spacers of variable length. A single nucleotide spacer was optimal for both imperfect (IRhsp-1) and perfect (IRper-1) inverted repeats; adding or removing 1 base pair in an IRhsp-1 spacer practically abolished binding. However, changing the half-site to the consensus sequence AGGTCA (IRper-1) increased binding of AaEcR.AaUSP 10-fold over IRhsp-1 and, at the same time, reduced the stringency of the spacer length requirement, with IRper-0 to IRper-5 showing detectable binding. Spacer length was less important in DRs of AGGTCA (DR-0 to DR-5); although 4 bp was optimal, DR-3 and DR-5 bound AaEcR.AaUSP almost as efficiently as DR-4. Furthermore, AaEcR. AaUSP also bound DRs separated by 11-13 nucleotide spacers. Competition experiments and direct estimation of binding affinity (Kd) indicated that, given identical consensus half-sites and an optimal spacer, the AaEcR.AaUSP heterodimer bound an IR with higher affinity than a DR. Co-transfection assays utilizing CV-1 cells demonstrated that the mosquito EcR.USP heterodimer is capable of transactivating reporter constructs containing either IR-1 or DR-4. The levels of transactivation are correlated with the respective binding affinities of the response elements (IRper-1 > DR-4 > IRhsp-1). Taken together, these analyses predict broad variability in the EcREs of mosquito ecdysone-responsive genes.

Characterization of an intronic hormone response element of the rat liver/skeletal muscle 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase gene

The glucocorticoid response element of the rat liver/skeletal muscle 6- phosphofructo-2-kinase/fructose-2,6-bisphosphatase gene was characterized. The element is composed of two tandem hormone receptor binding sites separated by 12 base pairs. Addition of dexamethasone to HeLa cells transiently transfected with a chloramphenicol acetyl transferase (CAT) reporter plasmid containing the hormone response element and cotransfected with glucocorticoid receptor stimulated transcription 24-fold in an orientation- and position-independent manner. Deletion or mutation of essential G/C pairs of the distal binding site abolished hormone-stimulated CAT activity, whereas deletion or mutation of the proximal binding site decreased the hormone-stimulated response only slightly. Mutation of both distal and proximal binding sites resulted in complete loss of hormone-stimulated CAT activity. Experiments carried out using testosterone and progesterone with their respective receptors revealed qualitatively similar results to those seen with glucocorticoid. Binding of glucocorticoid receptor or androgen receptor DNA binding domains to the hormone response element, visualized by gel mobility shift, was unaffected in the proximal binding site mutant, markedly decreased in the distal binding site mutant, and abolished in the double mutant. In gel mobility shift analysis of separate distal and proximal binding sites, only the native distal site demonstrated high affinity binding to glucocorticoid and androgen receptor DNA binding domains. The results demonstrate that this element is responsible for glucocorticoid, androgen, and progesterone stimulation of transcription of the 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase gene and that the distal receptor binding site is dominant.

Activation of the unliganded estrogen receptor by EGF involves the MAP kinase pathway and direct phosphorylation

The estrogen receptor (ER) can be activated as a transcription factor either by binding of cognate estrogenic ligand or, indirectly, by a variety of other extracellular signals. As a first step towards elucidating the mechanism of 'steroid-independent activation' of the ER by the epidermal growth factor (EGF), we have mapped the ER target domain and determined the signaling pathway. We show that the N-terminal transcriptional activation function AF-1, but not the C-terminal AF-2, is necessary for the EGF response. Both the EGF-induced hyperphosphorylation and the transcriptional activation of the unliganded ER depend on a phosphorylatable serine residue at position 118. However, its phosphorylation is not sufficient and, hence, there must be other target domains or proteins which fulfill an additional requirement for EGF signaling through the ER. Using dominant-negative Ras and MAP kinase kinase (MAPK kinase) and constitutively active MAPK kinase mutants, we show that EGF activates the ER by signaling through the MAPK pathway suggesting that MAPK directly phosphorylates the critical serine 118. Our results also imply that the steroid-independent activation of a variety of ER mutants, which arise during the malignant progression of breast tumors, may contribute to tamoxifen resistance.

Studies using fluorescent tetrahydrochrysene estrogens for in situ visualization of the estrogen receptor in living cells

We have analyzed four fluorescent nonsteroidal estrogens for their potential to serve as vital cytological stains to visualize the estrogen receptor (ER) in a model receptor expression system. The novel estrogen fluorophores are based on the rigidified stilbene-like structure of 5,6,11,12-tetrahydrochrysene (THC), and they embody electron-donor (hydroxyl) and electron-acceptor groups (nitrile, amide, ester, or ketone) that afford efficient, long wavelength, and environment-sensitive fluorescence. These probes bind with high affinity to human ER (relative binding affinity, 22-85 vs. estradiol, 100), and they stimulate the transcriptional activity of this receptor. The strong fluorescence of the estrogenic THCs permits visualization, using conventional epifluorescence microscopy, of ER in transfected Cos-7 cells that express elevated levels of receptor. Cell staining by the donor-acceptor THCs characteristically displays a nonuniform pattern of nuclear fluorescence that can be fully inhibited by nonfluorescent estrogens such as estradiol or diethylstilbestrol. Additionally, this staining appears to be specific for ER, since it coincides with the distribution of receptor as determined by indirect immunofluorescence analysis using an ER-specific monoclonal antibody. Using these probes, we have analyzed the intracellular distribution of ER mutants containing a variety of deletions. Evidence is presented to show that removal of amino-terminal sequences within the ER polypeptide results in an altered pattern of intranuclear distribution with preferential accumulation of receptor protein within the nucleolus. These THC fluorophores therefore represent excellent probes for cytological studies involving ER expressed in cultured cells and represent an important advance toward the goal of exploiting fluorescence technology to analyze the expression and distribution of ER within tissue samples.

Leiomyoma primary cultures have elevated transcriptional response to estrogen compared with autologous myometrial cultures

OBJECTIVE

We tested the hypothesis that uterine leiomyomas are hypersensitive to estrogen as compared with autologous human myometrium.

METHODS

The estrogen-induced transcriptional responses of uterine leiomyoma and myometrial primary cultures were determined by transient expression assays. The relative levels of estrogen receptor (ER) in myometrial and leiomyoma tissues were determined by immunoblot.

RESULTS

Myometrial and leiomyoma primary cultures were transcriptionally responsive to the estrogen ethinyl estradiol (eE2). The partial agonist tamoxifen did not elicit a positive transcriptional response and antagonized estrogen-induced transcription in the cultured cells. The responses of hormone-treated leiomyoma cells averaged 4.5-fold higher than those in controls with no hormone (P = .0001). The myometrial cells from women in the follicular phase exhibited little if any transcriptional response to eE2, whereas myometrial cells from women in the luteal phase had a transcriptional response to eE2 averaging threefold higher than that in no-hormone controls (P = .0083). Differences in response between autologous myometrial and leiomyoma cultures were statistically significant by the two-tailed Wilcoxon paired nonparametric signed-rank test (n = 11; P = .0137). These differences were more pronounced in cultures from women in the follicular or early luteal phase. In addition, the levels of ER increased in follicular and early luteal phase myometrial tissues, which correlated well with the number of days from the last menstrual period (n = 8; r = 0.9046; P = .002). Estrogen-receptor levels in myometrial tissues decreased during the late luteal phase. Levels in leiomyoma tissues did not follow the same pattern as in the myometrium and were elevated in tissues taken from women in the follicular phase.

CONCLUSIONS

Autologous leiomyoma cultures have a significantly higher response to estrogen than do matched myometrial cultures, especially if the cultures are derived from the follicular phase. The levels of ER in leiomyoma tissue from women in the follicular phase are significantly elevated.

Estrogenic flavonoids: structural requirements for biological activity

A systematic survey of polycyclic phenols has been performed to identify members of this chemical group with estrogenic activity. Twelve compounds were found to be able to stimulate the transcriptional activity of the human estrogen receptor expressed in cultured cells by transient transfection. These natural estrogens belong to several distinct, but chemically related classes including chalcones, flavanones, flavones, flavonols, and isoflavones. Selected examples of estrogenic flavonoids were further analyzed to determine their biological potencies and their relative affinities for binding to the estrogen receptor. These data are interpreted with respect to the molecular structure of polycyclic phenols required for hormonal activity as nonsteroidal estrogens.

Steroid receptor variants and their potential role in cancer

Variant forms of steroid receptors that possess altered structure have been described in a number of human cancers. This is particularly true for tumors of the breast, which frequently contain a variety of defective estrogen receptors. Evidence for the existence of steroid receptor variants in human tumors is summarized along with what is known regarding the functional behavior of these aberrant receptors. This article also discusses the implications that steroid receptor variants have for the diagnosis and treatment of tumors of the breast and other human malignancies.

Mutational analysis of cysteine residues within the hormone-binding domain of the human estrogen receptor identifies mutants that are defective in both DNA-binding and subcellular distribution

To further our understanding of the role played by cysteine residues in ligand recognition by the human estrogen receptor (hER), we have individually mutated residues 381, 417, 447, and 530 within the hormone-binding domain from cysteine to serine. We have also examined 10 additional nonconservative amino acid changes at C530. Four mutants were identified with defects in their ability to stimulate transcription in response to hormone: C447S, C530P, C530W, and C530G. Accumulation of the mutant proteins to comparable steady state levels after transient expression in COS-7 cells leads us to conclude that none of these substitutions results in increased turnover of receptor protein. None of these mutations had a reduced affinity for estradiol when compared with wild-type hER [dissociation constant (Kd) = 0.29 nM]. Although C447 and C530 reside outside of motifs previously defined as being necessary for DNA binding or receptor dimerization, all four of these mutants unexpectedly displayed altered DNA-binding properties when analyzed using a gel mobility shift assay. Their loss of DNA binding could be overcome, at least in part, by hormone treatment or by the addition of antibody. Additionally, mutations C447S, C530S, and C530P displayed patterns of subcellular localization in the absence of hormone that differed from wild-type hER. These results argue that cysteine residues 447 and 530 within the hormone-binding domain play no direct role in the ability of this receptor to bind estradiol. These mutations, however, can effect DNA-binding activity and the distribution of hER within cells.

Interaction of naturally occurring nonsteroidal estrogens with expressed recombinant human estrogen receptor

The interaction between the recombinant human estrogen receptor and a variety of nonsteroidal estrogens was studied using a transient transfection assay in mammalian cells. Eight naturally occurring compounds were confirmed to stimulate the transcriptional activity of the human estrogen receptor and to compete for the binding of radiolabeled 17 beta-estradiol to this protein. In order of biological potency, these were zearalenone, beta-zearalenol, coumestrol, genistein, daidzein, phloretin, formononetin, and biochanin A. As with steroidal estrogens, the hormonal activity of these compounds was specific for the estrogen receptor and sensitive to inhibition by 4-hydroxytamoxifen and ICI-164,384. Evidence is also presented to indicate that the stimulatory activity of genistein is unrelated to the protein tyrosine kinase inhibitory activity of this isoflavone. These results demonstrate that a significant number of structurally diverse plant and fungal secondary metabolites exist in nature that may contribute to the total estrogen exposure of the human population.

Characterization of estrogen receptor cDNAs from human uterus: identification of a novel PvuII polymorphism

Using reverse transcription and the polymerase chain reaction, we have cloned estrogen receptor complementary DNAs from normal human uterine tissue. Restriction endonuclease analysis identified a polymorphic PvuII recognition site within half of the receptor cDNAs. Sequence analysis revealed a number of differences with the sequence previously reported for the ER cDNA isolated from MCF7 cells and confirmed that the codon for amino acid 400 was erroneously assigned as valine (GTG) rather than glycine (GGG). Sequencing also defined the nature of the PvuII polymorphism, with allele A coding for Glu22 and allele B (with an additional PvuII site) coding for Gln22. We demonstrate that both alleles of this receptor activate transcription of an estrogen-responsive gene to the same extent. This selective cloning method should have wide application in the investigation of naturally occurring cDNA variants from diseased tissues, such as breast cancer cell lines and primary tumor specimens.

Commonly occurring plant flavonoids have estrogenic activity

A remarkable diversity of naturally occurring and synthetic compounds have been shown to mimic the biological effects of 17 beta-estradiol by virtue of their ability to bind to and activate the nuclear estrogen receptor. This report extends the family of nonsteroidal estrogens to include several multiply hydroxylated chalcones, flavanones, and flavones. The hormone-like activity of these natural plant products is indicated by their ability to stimulate an estrogen receptor-dependent transcriptional response and to promote growth of estrogen-dependent MCF7 cells in culture. The transcriptional response can be inhibited by the steroidal estrogen antagonist ICI-164,384 and is specific for the estrogen receptor. Evidence is presented to show that selected hydroxylated flavonoids interact directly with the estrogen receptor, based on their ability to compete for the binding of 17 beta-[3H]estradiol to the receptor in cell-free extracts. These compounds are less active, on a molar basis, than 17 beta-estradiol or the synthetic dihydroxystilbene estrogens, but they have potencies comparable to those of other known phytoestrogens. Together, these findings broaden our understanding of the structure-activity relationships for nonsteroidal estrogens and present a series of new chemical prototypes for the future development of potentially useful agonists and antagonists for this nuclear receptor. The wide distribution of weakly estrogenic flavonoid pigments in food crops and medicinal plants raises additional questions about the possible health risks and benefits of these compounds, meriting closer examination of their presence in the human diet.

In situ localization of the estrogen receptor in living cells with the fluorescent phytoestrogen coumestrol

Coumestrol is a naturally occurring plant coumarin that displays high affinity for the hormone-binding site of the human estrogen receptor (hER), for which it serves as a potent non-steroidal agonist. Coumestrol emits intense blue fluorescence when bound to this protein, making it ideally suited for use as a cytological stain to detect ER in fixed and intact cells. Conditions are reported for the efficient detection of recombinant ER protein artificially expressed in cultured cells by calcium phosphate-mediated transfection. Coumestrol fluorescence co-localizes with hER protein detected by indirect immunofluorescence with an hER-specific anti-peptide antiserum, confirming the specificity of this reagent. Fluorescence from ER occupied by coumestrol shows a predominantly nuclear distribution, even when the receptor protein is cross-linked in situ by fixation with glutaraldehyde plus paraformaldehyde before coumestrol exposure. This corroborates previous conclusions, based on indirect immunofluorescence analysis, that the ER remains nuclear in the absence of hormone. Examination of the pattern of coumestrol staining in mitotic cells further indicates that hER remains associated with condensed chromatin. Such observations illustrate the potential for using coumestrol to investigate real-time effects of a variety of physiological stimuli on the subcellular distribution of hER in living cells.

Exon skipping gives rise to alternatively spliced forms of the estrogen receptor in breast tumor cells

We have previously described three messenger RNA variants coding for the human estrogen receptor (ER) [1]. These variants were identified using the polymerase chain reaction to perform directed cloning of ER cDNAs synthesized from polyadenylated RNA extracted from the human breast cancer cell line T47D. Each of the variants is characterized by the precise deletion of a single exon within the protein coding region of this message and was presumably derived by inaccurate or promiscuous splicing of primary estrogen receptor transcripts. We report here the results of RNAse protection experiments which independently confirm the existence of these splicing variants in T47D cells. Similar analysis of RNA from MCF-7 cells also revealed the presence of variant ER transcripts, suggesting that they may be a common finding in tumor cell lines which express the estrogen receptor. However, attempts to identify splicing variants in a number of nominally ER-negative cell lines using either RNAse protection or PCR amplification were without success.

Regulation of gene expression of rat skeletal muscle/liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase. Isolation and characterization of a glucocorticoid response element in the first intron of the gene

At least two genes encode isoenzymes of rat 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase. Alternative splicing of one of these genes generates a skeletal muscle-specific transcript from an upstream promoter and a liver-specific transcript from a downstream promoter. A potent glucocorticoid response element was identified in the first intron of the gene, i.e. between liver exon I and exon II. The element is approximately 3.5 kilobase pairs (kb) downstream of the liver isoenzyme transcription start site and 13 kb upstream of exon II of the gene and confers dexamethasone-sensitive expression of chloramphenicol acetyltransferase (CAT) activity from a heterologous thymidine kinase promoter and from both homologous 5'-flanking regions of the gene. This glucocorticoid response element also exhibits androgen- but not estrogen-sensitive expression of CAT activity in HeLa cells cotransfected with the appropriate receptor expression vector. DNase footprint and sequence analysis revealed that the element is comprised minimally of two adjacent 15-mer glucocorticoid receptor dimer binding sites situated in opposite orientations. Glucocortcoid regulation of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase gene expression in liver and skeletal muscle is mediated by a single complex glucocorticoid response element located in the first intron of the skeletal muscle/liver gene.

Identification of a dominant negative form of the human estrogen receptor

Experiments were undertaken to characterize mRNAs coding for the estrogen receptor (ER) in the human breast cancer cell line T47D. We report here the isolation of cDNAs corresponding to three isoforms of this receptor in addition to a majority of wild-type clones. Sequence analysis showed that these isoforms are generated through alternative splicing. None of the alternatively spliced isoforms of ER is able to bind to an estrogen-responsive element (ERE) in a gel mobility shift assay in vitro or to activate transcription of a reporter gene containing an ERE in vivo. One isoform, ER delta E3, which harbors a deletion of exon 3 encoding the second zinc finger of the DNA-binding domain, inhibits estrogen-dependent transcription activation in a dominant negative fashion when it is cotransfected with the wild-type ER and reporter plasmid. It also inhibits DNA binding of wild-type ER in a gel mobility shift assay in vitro. Since ER delta E3 is not able to bind to its response element, the observed inhibitory effect probably occurs through protein-protein interactions. This could involve the formation of a heterodimer between mutant and wild-type receptors, competition for a limiting transcription factor, or both. These results may have implications for understanding the loss of estrogen responsiveness that frequently occurs in breast cancer.

Use of a cloned cDNA sequence to measure changes in 6-phosphogluconate dehydrogenase mRNA levels caused by thyroid hormone and dietary carbohydrate

A cDNA clone containing sequences complementary to the mRNA coding for rat hepatic 6-phosphogluconate dehydrogenase has been isolated and used to measure changes in specific mRNA levels during dietary and hormonal regulation of this enzyme. Hepatic mRNA was fractionated by sucrose gradient centrifugation to enrich for 6-phosphogluconate dehydrogenase mRNA sequences. A cDNA library was prepared from the fraction with maximal activity and then screened by differential colony hybridization using probes synthesized either from 6-phosphogluconate dehydrogenase mRNA enriched by polysome immunoadsorption or from unenriched hepatic mRNA. A single colony giving an appropriate differential signal was confirmed to contain sequences encoding 6-phosphogluconate dehydrogenase by specific immunoprecipitation of hybrid-selected translational products. 6-Phosphogluconate dehydrogenase mRNA contains about 2400 bases. The cloned cDNA comprises about 880 bases, or 35% of the mRNA. Southern analysis of restriction endonuclease digests of genomic DNA suggests that the major 6-phosphogluconate dehydrogenase gene is probably present in a single copy in the rat genome. Feeding a fat-free, high carbohydrate diet and administration of thyroid hormone increased the concentration of hybridizable 6-phosphogluconate dehydrogenase mRNA in liver. Thus, both dietary and hormonal regulation of 6-phosphogluconate dehydrogenase synthesis occurs at a pretranslational level.

Changes in the rates of synthesis and messenger RNA levels of hepatic glucose-6-phosphate and 6-phosphogluconate dehydrogenases following induction by diet or thyroid hormone

The lipogenic capacity of rat liver is increased in animals fed a high carbohydrate, fat-free diet or by the administration of 2,2',5'-triiodo-L-thyronine. Underlying this change is a generalized induction of the enzymes involved in lipogenesis, including glucose-6-phosphate dehydrogenase, 6-phosphogluconate dehydrogenase, and malic enzyme, which together serve to generate the additional NADPH required for increased fatty acid synthesis. This report presents evidence indicating that induction of the hexose-shunt dehydrogenases involves increased enzyme synthesis secondary to elevated enzyme specific mRNA levels, as has previously been shown for malic enzyme. Activities of specific mRNAs, estimated by cell-free translation of hepatic poly(A)-containing RNA in the mRNA dependent rabbit reticulocyte lysate, were compared with enzyme specific activities and relative rates of specific enzyme synthesis. The 2-fold increase in glucose-6-phosphate dehydrogenase specific activity in hyperthyroid rats and the 13-fold increase in rats fed a high carbohydrate, fat-free diet, relative to euthyroid, chow-fed controls were paralleled by comparable increases in the synthetic rates and mRNA levels of this enzyme. Similarly, consonant changes in the rate of enzyme synthesis and concentration of 6-phosphogluconate dehydrogenase mRNA accompanied the 2.5- and 3-fold increases in specific activity of this enzyme observed in response to hormonal and dietary induction, respectively. Thus, both thyroid hormone and carbohydrate feeding appear to induce glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase primarily by increasing the effective cellular concentrations of their respective mRNAs and, consequently, their rates of synthesis.