Glucocorticoids are insufficient for neonatal gene induction in the liver.

Glucocorticoids and their receptor (GR) play a key role in perinatal gene induction. In the liver, the GR is essential for the neonatal induction of a number of genes, including that coding for tyrosine aminotransferase (TAT). To assess the function of the GR in the perinatal period, we have compared the activity of two types of glucocorticoid responsive elements in transgenic mice; one is the Tat gene glucocorticoid-responsive unit (GRU), an assembly of numerous binding sites for transcription factors, including the GR; the other is a simple dimer of high-affinity GR binding sites (GREs). Both elements confer strong glucocorticoid response in the adult liver. However, only the Tat GRUs are able to promote neonatal induction; the GRE dimer is unresponsive. Because this dimer is responsive to glucocorticoid administration in the neonate, the absence of neonatal induction is not due to the inactivity of the GR at this stage. At birth, the neonate has to withstand a brief period of starvation and hypoglycemia, a nutritional and hormonal situation that resembles fasting in the adult. In transgenic mice, the responses at birth and after fasting in the adult are similar: the Tat GRUs but not the dimeric GREs are activated. Our results show that, in rodents, glucocorticoids are not sufficient for neonatal gene induction in the liver and support the conclusion that the hypoglycemia at birth is the main trigger for expression.

Messenger RNA deadenylylation precedes decapping in mammalian cells.

In yeast, the major mRNA degradation pathway is initiated by poly(A) tail shortening that triggers mRNA decapping. The mRNA is then degraded by 5'-to-3' exonucleolysis. In mammalian cells, even though poly(A) tail shortening also precedes mRNA degradation, the degradation pathway has not been elucidated. We have used a reverse transcription-PCR approach that relies on mRNA circularization to measure the poly(A) tail length of four mammalian mRNAs. This approach allows for the simultaneous analysis of the 5' and 3' ends of the same mRNA molecule. For all four mRNAs analyzed, this strategy permitted us to demonstrate the existence of small amounts of decapped mRNA species which have a shorter poly(A) tail than their capped counterparts. Kinetic analysis of one of these mRNAs indicates that the decapped species with a short poly(A) tail are mRNA degradation products. Therefore, our results indicate that decapping is preceded by a shortening of the poly(A) tail in mammalian cells, as it is in yeast, suggesting that this mRNA degradation pathway is conserved throughout eukaryotic evolution.

Sequence, structure and evolution of the ecdysone-inducible Lsp-2 gene of Drosophila melanogaster.

The Lsp-2 gene encodes a major larval serum protein (hexamerin) of Drosophila melanogaster. Transcription of Lsp-2 is controlled by 20-hydroxyecdysone. Here we report the analysis of the structure of the Lsp-2 gene including the adjacent 5' and 3' sequences. In contrast to all other known hexamerin genes, Lsp-2 does not contain an intron. The Lsp-2 mRNA measures 2312 bases, as deduced from experimental determination of the transcription-start and stop sites and conceptual translation results in a 718 amino acid hexamerin subunit, including a 21-amino-acid signal peptide. While the calculated molecular mass of the native 697-amino-acid subunit is 83.5 kDa, mass spectrometry gave a value of 74.5 kDa. We detected in the Lsp-2 gene a 2052-bp antisense ORF that probably does not code for any protein. An unusual accumulation of rarely used codon triplets was found at the 5' and 3' ends of the Lsp-2 ORF. The calculated secondary structure matches well with that of arthropod hemocyanins. Electron micrographs show for LSP-2 hexamers a cubic shape, which can not be easily reconciled with its hexameric structure. Phylogenetic analysis revealed that LSP-2 diverged from the LSP-1 like hexamerins after separation of the Diptera from other insect orders.

In vivo footprinting of the interaction of proteins with DNA and RNA.

Analysis of the interaction of proteins with either DNA or RNA sequences by in vivo footprinting involves two steps: (i) the in situ modification of nucleic acids by the footprinting reagent and (ii) the visualization of the footprints. Ligation-mediated PCR (LM-PCR) procedures provide a level of sensitivity and specificity that is suitable for visualization of footprints of single-copy genes or low-abundance mRNAs in higher eukaryotes. In this article, we discuss several of the technical aspects of these multistep procedures that contribute to the quality of the results, particularly the parameters that affect the specificity and fidelity of the reactions: (i) the design of the primers, which is important to achieve optimal specificity; (ii) the choice of polymerases so that the amplified material represents faithfully the initial nucleic acid population; and (iii) the impact of the plateau effect within the PCR on the interpretation of the data. We then discuss aspects of in vivo nucleic acid manipulation that may affect the quality of the footprinting image, in particular the choice of the footprinting reagent and its condition of use (e.g., on intact or permeabilized cells or prepared nuclei) and the extent of nucleic acid modification. Finally, we provide detailed experimental procedures corresponding to the techniques we have developed or modified: LM-PCR, reverse ligation-mediated PCR, and nuclease treatment of RNAs in vivo.

The expression cassette determines the functional activity of ribozymes in mammalian cells by controlling their intracellular localization.

In order to better understand the influence of RNA transcript context on RNA localization and catalytic RNA efficacy in vivo, we have constructed and characterized several expression cassettes useful for transcribing short RNAs with well defined 5' and 3' appended flanking sequences. These cassettes contain promoter sequences from the human U1 snRNA, U6 snRNA, or tRNA Meti genes, fused to various processing/stabilizing sequences. The levels of expression and the sub-cellular localization of the resulting RNAs were determined and compared with those obtained from Pol II promoters normally linked to mRNA production, which include a cap and polyadenylation signal. The tRNA, Ul, and U6 transcripts were nuclear in localization and expressed at the highest levels, while the standard Pol II promoted transcripts were cytoplasmic and present at lower levels. The ability of these cassettes to confer ribozyme activity in vivo was tested with two assays. First, an SIV-growth hormone reporter gene was transiently transfected into human embryonic kidney cells expressing an anti-SIV ribozyme. Second, cultured T lymphocytes expressing an anti-HIV ribozyme were challenged with HIV. In both cases, we found that the ribozymes were effective only when expressed as capped, polyadenylated RNAs transcribed from Pol II cassettes that generate a cytoplasmically localized ribozyme that facilitates co-localization with its target. We also show that the inability of the other cassettes to support ribozyme-mediated inhibitory activity against their cytoplasmic target is very likely due to the resulting nuclear localization of these ribozymes. These studies demonstrate that the ribozyme expression cassette determines its intracellular localization and, hence, its corresponding functional activity.

Glucocorticoids and protein kinase A coordinately modulate transcription factor recruitment at a glucocorticoid-responsive unit.

The rat tyrosine aminotransferase gene is a model system to study transcriptional regulation by glucocorticoid hormones. We analyzed transcription factor binding to the tyrosine aminotransferase gene glucocorticoid-responsive unit (GRU) at kb -2.5, using in vivo footprinting studies with both dimethyl sulfate and DNase I. At this GRU, glucocorticoid activation triggers a disruption of the nucleosomal structure. We show here that various regulatory pathways affect transcription factor binding to this GRU. The binding differs in two closely related glucocorticoid-responsive hepatoma cell lines. In line H4II, glucocorticoid induction promotes the recruitment of hepatocyte nuclear factor 3 (HNF3), presumably through the nucleosomal disruption. However, the footprint of the glucocorticoid receptor (GR) is not visible, even though a regular but transient interaction of the GR is necessary to maintain HNF3 binding. In contrast, in line FTO2B, HNF3 binds to the GRU in the absence of glucocorticoids and nucleosomal disruption, showing that a "closed" chromatin conformation does not repress the binding of certain transcription factors in a uniform manner. In FTO2B cells, the footprint of the GR is detectable, but this requires the activation of protein kinase A. In addition, protein kinase A stimulation also improves the recruitment of HNF3 independently of glucocorticoids and enhances the glucocorticoid response mediated by this GRU in an HNF3-dependent manner. In conclusion, the differences in the behavior of this regulatory sequence in the two cell lines show that various regulatory pathways are integrated at this GRU through modulation of interrelated events: transcription factor binding to DNA and nucleosomal disruption.

Tissue specificity of a glucocorticoid-dependent enhancer in transgenic mice.

The glucocorticoid-responsive units (GRUs) of the rat tyrosine aminotransferase were associated with the regulatory sequences of a cellular gene expressed ubiquitously--that coding for the largest subunit of RNA polymerase II. In transient expression assays, glucocorticoid responsiveness of the hybrid regulatory regions depends on the spatial relationship and number of regulatory elements. Two parameters affect the ratio of induction by glucocorticoids: the basal level of the hybrid promoter that is affected by the RNA polymerase II regulatory sequences and the glucocorticoid-induced level that depends on the distance between the GRUs and the TATA box. A fully active glucocorticoid-responsive hybrid gene was used to generate transgenic mice. Results show that a composite regulatory pattern is obtained: ubiquitous basal expression characteristic of the RNA polymerase II gene and liver-specific glucocorticoid activation characteristic of the tyrosine aminotransferase GRUs. This result demonstrates that the activity of the tyrosine aminotransferase GRUs is cell-type-specific not only in cultured cells but also in the whole animal.

Hepatocyte nuclear factor 3 determines the amplitude of the glucocorticoid response of the rat tyrosine aminotransferase gene.

Hepatocyte nuclear factor 3 (HNF3) recognizes two apparently distinct classes of sequence. However, a detailed mutational analysis of a representative binding site of each class reveals that these sequences display common features. We propose a unified consensus sequence for HNF3-binding sites. The basis of the sequence specificity of the interaction of HNF3 with DNA is analyzed in light of the recently determined structure of an HNF3-DNA complex (Clark et al., Nature 364, 412-420, 1993). Particularly, our study reveals that the DNA site used for this structural analysis is too short to account for all HNF3-DNA interactions. The better knowledge of the sequence determinant recognized by HNF3 has allowed us to analyze its function in the glucocorticoid response of the rat tyrosine aminotransferase (TAT) gene. This response is mediated through a complex array of neighboring and overlapping transcription factor binding sites. Selective inactivation of the HNF3-binding sites in this glucocorticoid response unit (GRU) allows us to demonstrate unambiguously that they play a major role in the amplitude of the glucocorticoid response. Furthermore, HNF3 beta overexpression results in a stimulation of the glucocorticoid response that is dependent on the integrity of its binding sites. We also show that the relative level of HNF3 determines the extent of the contribution of one of the glucocorticoid receptor binding sites. Our results indicate that HNF3 accounts for most of the liver-specific activity of this GRU.

Participation of Ets transcription factors in the glucocorticoid response of the rat tyrosine aminotransferase gene.

We have previously shown that two remote glucocorticoid-responsive units (GRUs) of the rat tyrosine aminotransferase (TAT) gene contain multiple binding sites for several transcription factor families, including the glucocorticoid receptor (GR). We report here the identification of two novel binding sites for members of the Ets family of transcription factors in one of these GRUs. One of these binding sites overlaps the major GR-binding site (GRBS), whereas the other is located in its vicinity. Inactivation of the latter binding site leads to a twofold reduction of the glucocorticoid response, whereas inactivation of the site overlapping the GRBS has no detectable effect. In vivo footprinting analysis reveals that the active site is occupied in a glucocorticoid-independent manner, in a TAT-expressing cell line, even though it is located at a position where there is a glucocorticoid-dependent alteration of the nucleosomal structure. This same site is not occupied in a cell line that does not express TAT but expresses Ets-related DNA-binding activities, suggesting the existence of an inhibitory effect of chromatin structure at a hierarchical level above the nucleosome. The inactive Ets-binding site that overlaps the GRBS is not occupied even in TAT-expressing cells. However, this same overlapping site can confer Ets-dependent stimulation of both basal and glucocorticoid-induced levels when it is isolated from the GRU and duplicated. Ets-1 expression in COS cells mimics the activity of the Ets-related activities present in hepatoma cells. These Ets-binding sites could participate in the integration of the glucocorticoid response of the TAT gene with signal transduction pathways triggered by other nonsteroidal extracellular stimuli.

Expression from the tyrosine aminotransferase promoter (nt -350 to +1) is liver-specific and dependent on the binding of both liver-enriched and ubiquitous trans-acting factors.

The rat tyrosine aminotransferase(TAT) gene promoter (nucleotides -350 to +1; TAT0.35) was able to sustain liver-specific expression both ex vivo in transient transfection (TAT-expressing H411EC3 hepatoma cells vs. TAT non-expressing CCL1.2 fibroblasts) and in in vitro transcription (rat liver vs. spleen crude nuclear extracts). In either case, the index of tissue specificity (6.2 and 6.7 in ex vivo and in vitro experiments, respectively) was close to that obtained with 10 Kb of TAT gene 5'-flanking sequences in transient transfection. Using computer-assisted search of homologies, DNase I footprinting, gel retardation and methylation interference assays, we showed that TAT0.35 sequences spanning nt -156 to -175 and nt -268 to -281 interacted with the liver enriched NF-1Liver (a member of the NF1 gene family) and HNF1 respectively, whereas those encompassing nt -57 to -85 and nt -283 to -288 interacted with the ubiquitous NF-Y and with ubiquitous 'CCAAT'-box binding factor(s), respectively. Competition studies in in vitro transcription carried out with wild type and mutated oligonucleotides, demonstrated that NF-Y cis-elements were crucial for basal TAT promoter activity, both in liver and spleen whereas NF1Liver and HNF1 were only efficient in the liver (supported approximately 60% and 30% of basal TAT0.35 activity respectively). Altogether, these results support the conclusion that TAT0.35 was able to sustain at least part of the liver specificity of TAT gene expression.

Can hammerhead ribozymes be efficient tools to inactivate gene function?

In order to improve hammerhead ribozyme efficiency and specificity, we have analyzed, both in vitro and in vivo, the activity of a series of ribozyme/substrate combinations that have the same target sequence but differ in the length of the ribozyme/substrate duplex or in their structure, i.e., the total length of the RNA. In vitro, we have found that optimal kcat/Km (at 37 degrees C) is obtained when the ribozyme/substrate duplex has a length of 12 bases, which according to the base composition represents a calculated free energy of binding of -16 kcal/mol. We discuss the importance of this value for ribozyme specificity and present strategies that may improve it. Increasing the length of the duplex from 14 to 17 bases (from -19 to -26 kcal/mol) produces a reduced ribozyme activity which is probably due to a slower rate of product dissociation. In addition, inclusion of either the substrate or the ribozyme in a long transcript produces a reduction (10 fold) of the kcat/Km, probably because of a different accessibility of the target sequence. In vivo, the activity of the trans-acting ribozyme was extremely low and detected in only one case: with a ribozyme/substrate duplex length of 13 bases and with both ribozyme and substrate embedded in short RNAs expressed at a very high level. The similarity of the results obtained in vitro and in vivo indicates that it is possible to use an in vitro system to optimize ribozymes which are to be used in vivo. Satisfactory results were obtained in vivo only with cisacting ribozymes. Altogether these results suggest that the ribozyme/substrate hybridization step is the limiting step in vivo and therefore it is not clear if ribozymes represent an improvement over antisense RNAs.

Visualization of the interaction of a regulatory protein with RNA in vivo.

We have adapted to RNA molecules the ligation-mediated polymerase chain reaction (LMPCR) procedure of genomic sequencing [Mueller, P. R. & Wold, B. (1989) Science 246, 780-786]. This new procedure, the reverse ligation-mediated PCR (RLPCR), is sufficiently sensitive to allow "in vivo" footprinting of minor RNA species. It is based on the ligation of an RNA linker of known sequence to every 5' end resulting from the cleavage of total cellular RNA. Target RNA molecules are specifically reverse-transcribed and the resulting products are amplified by PCR. The localization of the initial 5' ends is ultimately determined on a sequencing gel. To demonstrate the validity of this strategy, we have used RNase T1 treatment of permeabilized cells and RLPCR and have detected in vivo iron-depletion-dependent footprints on two iron-responsive elements of the transferrin receptor mRNA.

Co-expression of insulin and somatostatin genes in pancreatic endocrine cells selected for their high level of insulin gene transcription.

RW cells are pancreatic endocrine RIN cells that have been stably transfected with a chimeric gene that places the expression of the dominant selection gpt gene under the control of the insulin gene regulatory sequences. These RW cells were examined for hormone content using immunocytochemistry. This analysis shows that: first, there are cells that are negative for insulin although they were cultured under selective pressure. Second, there is a higher proportion of somatostatin-producing cells than in the parental RIN cells; these somatostatin cells form two populations: one of cells containing only somatostatin and, surprisingly, one made of cells containing both insulin and somatostatin. Thus: (1) expression of the transfected and endogenous insulin regulatory sequences is not regulated in a coordinate fashion; (2) the presence of both hormones in the same cell suggests that the regulation of the expression of insulin and somatostatin genes and the differentiation pathway of the two respective cell types may be closely related.

Two liver-enriched trans-acting factors support the tissue-specific basal transcription from the rat tyrosine aminotransferase promoter.

The rat tyrosine aminotransferase gene (TAT) is a glucocorticoid-inducible gene, specifically expressed in liver. Using gel retardation assays, we have shown that its promoter (nt + 1 to -350; TAT.35) binds a combination of both ubiquitous and liver-specific trans-acting factors. Cis-acting sequences spanning: (i) nt -65 to -85 bound NF-Y, an ubiquitous "AACCAAT" box binding factor; (ii) nt -157 to -171 bound a liver-enriched member of the NF1 gene family [NF1Liver (NF1L hereafter)]; (iii) nt -266 to -281 bound the liver specific factor HNF1; and (iv) nt -283 to -288 bound ubiquitous "CCAAT" box binding factor(s). Moreover, the TAT gene promoter was able to drive liver-specific basal transcription, even in an in vitro assay using TAT-expressing (liver) vs non-expressing (spleen) crude nuclear extracts (NEs). Competition studies in transcription with both unmutated and mutated ds-oligonucleotides (ds-oligos) demonstrated that NF1L and HNF1 supported approx. 60 and 25% of the basal transcriptional activity sustained by TAT.35 in the liver, respectively. Neither of these oligos affected the very low level of transcription sustained by spleen NEs. This suggests a minor role for HNF1 in liver-specific basal TAT gene expression, consistent with previous observations with dedifferentiated C2 hepatoma cells (which does not express HNF1) [Deschatrette and Weiss. Biochimie 56 (1974) 1603-1611 and Cereghini et al. EMBO Jl9 (1990) 2257-2263]. Competition studies in liver-specific in vitro transcription with ds-oligo -265/-290 yielded a 90% inhibition, suggesting either that sequences spanning nt -283 to -288 sequester "CCAAT-box" binding factor(s) that may be relevant elsewhere for TAT promoter function (e.g. NF-Y which interacts with nt -65 to -85), or that such a factor interacts functionally with HNF1.

In vivo footprinting of rat TAT gene: dynamic interplay between the glucocorticoid receptor and a liver-specific factor.

HNF5, a liver-specific DNA-binding protein, interacts with DNA in a manner that allows DNAase I cleavage in the middle of its recognition sequence. Using this property we have identified in vivo HNF5 bound to its sites within two glucocorticoid-responsive units of the rat tyrosine aminotransferase (TAT) gene. One HNF5-binding site is also a glucocorticoid receptor-binding site; glucocorticoid-dependent HNF5 binding could be detected at this site even though it is incompatible with glucocorticoid receptor binding. HNF5 binds within 10 min of hormone addition, indicating that it participates in transcriptional activation. In the TAT gene glucocorticoid-dependent HNF5 binding occurs where there is glucocorticoid-dependent disruption of nucleosomal structure; constitutive binding occurs in constitutively disrupted regions. These results suggest a hit-and-run mechanism of transcriptional activation by glucocorticoid receptor: the activated receptor binds its target sequence, modifies local chromatin structure, then leaves its site accessible to another factor.

Trimethoprim binds in a bacterial mode to the wild-type and E30D mutant of mouse dihydrofolate reductase.

Previous crystallographic studies of the antibacterial trimethoprim in complexes with bacterial and avian dihydrofolate reductases have shown substantial differences in the mode of binding, providing plausible explanations for the origin of the remarkable species selectivity of this inhibitor (Matthews, D. A., Bolin, J. T., Burridge, J. M., Filman, D. J., Volz, K. W., Kaufman, B. T., Beddell, C. R., Champness, J. N., Stammers, D. K., and Kraut, J. (1985) J. Biol. Chem. 260, 381-391; Matthews, D. A., Bolin, J. T., Burridge, J. M., Filman, D. J., Volz, K. W., and Kraut, J. (1985) J. Biol. Chem. 260, 392-399). A major species difference between the active sites is that the only carboxylate present is always Glu in vertebrates and Asp in bacteria. Crystallographic studies of the wild-type and E30D mutant of the enzyme from mouse now reveal that in both cases trimethoprim is bound in an identical fashion to that observed with the bacterial enzyme, and there is no obvious single explanation for the origin of the 10(5)-fold selectivity of trimethoprim binding. In an earlier study of a mouse wild-type enzyme using more limited data it was proposed that trimethoprim bound in the avian mode (Stammers, D. K., Champness, J. N., Beddell, C. R., Dann, J. G., Eliopoulos, E. E., Geddes, A. J., Ogg, D., and North, A. C. T. (1987) FEBS Lett. 218, 178-184), but a re-examination indicates that the occupancy of the active site by trimethoprim is less than had been thought, and we are currently unable to make an unambiguous interpretation of the electron density maps and cannot confirm the avian mode of binding in those crystals.