Binding of a liver-specific factor to the human albumin gene promoter and enhancer.

A segment of 1,022 base pairs (bp) of the 5'-flanking region of the human albumin gene, fused to a reporter gene, directs hepatoma-specific transcription. Three functionally distinct regions have been defined by deletion analysis: (i) a negative element located between bp -673 and -486, (ii) an enhancer essential for efficient albumin transcription located between bp -486 and -221, and (iii) a promoter spanning a region highly conserved throughout evolution. Protein-binding studies have demonstrated that a liver trans-acting factor which interacts with the enhancer region is the well-characterized transcription factor LF-B1, which binds to promoters of several liver-specific genes. A synthetic oligodeoxynucleotide containing the LF-B1-binding site is sufficient to act as a tissue-specific transcriptional enhancer when placed in front of the albumin promoter. The fact that the same binding site functions in both an enhancer and a promoter suggests that these two elements influence the initiation of transcription through similar mechanisms.

Albumin and alpha-fetoprotein gene expression in various nonhepatic rat tissues.

Albumin and alpha-fetoprotein (AFP), two major serum proteins, are synthesized predominantly in the liver and yolk sac of mammals. In the present paper we report on the developmental expression of the corresponding genes in nonhepatic rat tissues. Significant quantities of mature albumin and AFP mRNAs were revealed in kidney, pancreas, heart, and lung of fetal and/or newborn rats using dot blot and Northern blot assays. Very low levels of these mRNA sequences were also detected in adult kidney and pancreas using sensitive RNA-cDNA solution hybridization assays. In situ hybridization analysis revealed that the albumin and AFP gene transcripts are present in the tubular cells of the 20-day-old fetal kidney. In order to elucidate further the mechanisms governing this expression, we studied the chromatin structure and methylation pattern in the 5'-end of these two genes. A faint band, corresponding to a specific DNase I-hypersensitive site upstream from the albumin gene, was detected in the fetal and neonatal kidney nuclei but not in adult kidney. For both genes, a site CG, demethylation of which is correlated with expression in liver and hepatoma cell lines, is highly methylated in fetal kidney even though AFP and albumin genes are expressed. Taken together, these results show the presence of a cell population in the rat kidney that actively transcribes both the albumin and AFP genes. The expression of these genes may be mediated by mechanisms differing in at least some steps from those exerted in the liver.

Differences in methylation patterns of the alpha-fetoprotein and albumin genes in hepatic and non hepatic developing rat tissues.

By use of different restriction enzymes sensitive to internal cytosine methylation (HpaII, AvaI, HhaI) we have analysed the methylation patterns of albumin and AFP genes in tissues and cell lines with high (liver, yolk sac, hepatoma cell lines), low (fetal and neonatal kidney) or undetectable (spleen, JF1 fibroblasts) expression of either gene. We show that expression of the AFP gene is associated to the demethylation of a whole region or domain extending from -4 to +3 Kb. Moreover, demethylation of a site located at the upstream limit of this domain appears to be correlated with the commitment of the cell type to synthesize AFP. As concerns the albumin gene, we show that the domain in which demethylation is correlated with active gene transcription in hepatoma cell lines has different borders than in tissue. This difference might be related to the different amounts of mRNA synthesized or to an alteration in gene regulation in tumor cells. Finally, we show that low expression of albumin and AFP genes in fetal and neonatal kidney is not correlated with domain demethylation, suggesting that the regulatory mechanisms of expression of these genes are different in kidney as compared with liver.

Albumin and alpha-fetoprotein gene transcription in rat hepatoma cell lines is correlated with specific DNA hypomethylation and altered chromatin structure in the 5' region.

We examined DNA methylation and DNase I hypersensitivity of the alpha-fetoprotein (AFP) and albumin gene region in hepatoma cell lines which showed drastic differences in the level of expression of these genes. We assayed for methylation of the CCGG sequences by using the restriction enzyme isoschizomers HpaII and MspI. We found two methylation sites located in the 5' region of the AFP gene and one in exon 1 of the albumin gene for which hypomethylation is correlated with gene expression. Another such site, located about 4,000 base pairs upstream from the AFP gene, seems to be correlated with the tissue specificity of the cells. DNase I-hypersensitive sites were mapped by using the indirect end-labeling technique with cloned genomic DNA probes. Three tissue-specific DNase I-hypersensitive sites were mapped in the 5' flanking region of the AFP gene when this gene was transcribed. Similarly, three tissue-specific DNase I-hypersensitive sites were detected upstream from the albumin gene in producing cell lines. In both cases, the most distal sites were maintained after cessation of gene activity and appear to be correlated with the potential expression of the gene. Interestingly, specific methylation sites are localized in the same DNA region as DNase I hypersensitive sites. This suggests that specific alterations of chromatin structure and changes in methylation pattern occur in specific critical regulatory regions upstream from the albumin and AFP genes in rat hepatoma cell lines.

Specific sets of DNase I-hypersensitive sites are associated with the potential and overt expression of the rat albumin and alpha-fetoprotein genes.

We have examined the chromatin structure of the 5'-flanking region of the albumin and alpha-fetoprotein (Afp) genes in different developing rat tissues and cloned cell lines that display various functional states of these genes. Nuclease-hypersensitive sites were probed with DNase I, using an indirect end-labeling technique. In albumin-producing rat cells two major DNase I-hypersensitive sites were found near the promoter region and one additional site was located approximately 3 kilobases (kb) upstream. Similarly, in Afp-producing rat tissues and cell lines we mapped one DNase I-hypersensitive region close to the promoter region and two cleavage sites further upstream at approximately 2.2 and approximately 3.8 kb from the cap site. The DNase I-hypersensitive sites of both genes were absent in nonhepatic rat cells and therefore appear to be tissue specific. Loss of specific sets of DNase I-hypersensitive sites accompanies the cessation of transcription for the Afp gene in adult rat liver and in a "dedifferentiated" hepatoma cell line. Likewise, specific sets of DNase I-hypersensitive sites disappear during the inactivation of the albumin gene in hepatoma cells. The distal upstream sites of the Afp and albumin genes display the same DNase I sensitivity in expressing and potentially expressible states. These findings suggest that reversible changes in short chromatin regions may be involved in the actual transcription of the albumin and Afp genes, while more permanent tissue-specific changes at other sites correlate with the capacity of these genes to be expressed during hepatic differentiation and neoplasia.

Activation of alpha-fetoprotein synthesis in rat hepatoma cells with reduced sensitivity to dexamethasone.

The Faza 967 'differentiated', dexamethasone (DEX)-sensitive cell line of Reuber rat hepatoma cells does not synthesize detectable amounts of alpha-fetoprotein (AFP), whereas it does produce albumin. AFP production was activated in 'differentiated' variants of Faza 967 cells with reduced glucocorticoid sensitivity upon culture for several months in the presence of high concentrations of dexamethasone. The stability of AFP production differed among the variants, while albumin synthesis did not change, thus indicating that the regulation of these two genes is not co-ordinated. Using molecular hybridization techniques, we found that the AFP message could not be detected in the non-AFP-producing cells, suggesting that the lack of AFP synthesis most probably originates from a transcriptional block of the AFP gene. AFP-producing and non-AFP-producing variants of Faza 967 cells constitute a valuable model system for studying the regulatory mechanisms involved in the activation and inactivation of the gene coding for the oncodevelopmental protein, AFP.

Organization of the albumin and alpha-fetoprotein genes in fetal and adult rat tissues, and rat hepatomas.

We compared the organization of the albumin and alpha-fetoprotein (AFP) genes in chromosomal DNA from different fetal and adult rat tissues as well as from two rat hepatomas. These two genes are expressed at widely different levels in the tissues and hepatomas analysed. Southern blots of DNAs digested with the restriction endonucleases EcoRI, HindIII or MspI were hybridized to albumin and AFP complementary DNA (cDNA) and genomic probes. No significant difference was observed in the hybridization patterns obtained for the DNAs from the different tissues, except for some interstrain variation between the chromosomal DNAs isolated from Sprague-Dawley and Buffalo rats, which was due to allelic polymorphism. We cannot rule out the possibility of changes in chromosomal gene organization which would result either in small alterations of restriction fragment size or in translocations of large blocks of DNA containing whole sets of restriction enzyme fragments within the chromosome; however, our results indicate that the gross organization of the albumin and AFP genes remains constant throughout the regulatory processes involved in the tissue- and time-specific transcription of these genes.

Differential DNase I sensitivity of the albumin and alpha-fetoprotein genes in chromatin from rat tissues and cell lines.

We have examined the DNase I sensitivity of the albumin and alpha-fetoprotein (AFP) genes in different rat tissues (adult liver and kidney) and cloned cell lines (hepatoma 7777-C8, JF1 fibroblasts), which show drastic differences in the level of expression of these two genes. This was done by studying the disappearance of defined restriction endonuclease fragments of these genes as a function of limited DNase I digestion. The sensitivity of these genes was compared to that of a gene not expressed in the hepatic cells and to that of a ubiquitously expressed gene. In nuclei from adult rat liver the albumin and AFP genes were preferentially degraded by the nucleolytic action of DNase I, whereas they were not in rat kidney nuclei. In the hepatoma cells the AFP gene was much more sensitive to DNase I digestion than the albumin gene; both genes were very resistant to DNase I action in fibroblastic nuclei. When analyzed in relation to the level of gene expression our results indicate that alterations in the chromatin structure of the albumin and AFP genes might be involved in the early establishment of the tissue-specific potential of overt gene expression; such alterations reflected in an altered DNase I sensitivity do not appear to be responsible for the changes in gene activity occurring during the terminal differentiation of the hepatocyte; and modifications in the chromatin structure of these genes might occur during oncogenic events; these structural modifications could be related to the changes in gene expression observed in hepatocarcinogenic processes.

Detection of rare mRNA species in a complex RNA population by blot hybridization techniques: a comparative survey.

The detection of very rare mRNA species in a complex RNA preparation by current RNA blotting techniques is not straightforward. To be able to determine the size of mRNA molecules representing 10(-6) to 10(-7) of the total mass of an RNA preparation, a quantitative comparison of the level of detection of denatured mRNA species electrophoretically separated on agarose gels, followed by transfer to either nitrocellulose or diazobenzyloxymethyl (DBM) paper and hybridization to specific cDNA probes was carried out. Different transfer procedures were analyzed. Optimal conditions have been found which allowed the detection of RNA bands containing as little as 5 pg of a specific sequence within a few days of autoradiography following hybridization with highly labeled [32P]cDNA probes. Using this procedure it was shown that the low amounts of alpha-fetoprotein (AFP) mRNA sequences present in adult rat liver are mature AFP mRNA molecules.

Changes in methylation pattern of albumin and alpha-fetoprotein genes in developing rat liver and neoplasia.

To determine whether methylation changes in specific DNA sequences of the albumin and AFP genes are implicated in the modulation of transcriptional activity during rat liver development and neoplasia we have analysed the methylation pattern of C-C-G-G sequences within these genes in DNA isolated from fetal and adult hepatocytes, from adult kidney and from a clonal hepatoma cell line which produces AFP but no albumin. We have assayed for methylation of the internal cytosine of this sequence by using the restriction enzyme isoschizomers HpaII and MspI. 32P-labelled cloned cDNA probes were used to reveal the albumin and AFP gene containing fragments. Genomic subclones of the albumin gene were also utilized as molecular probes to measure quantitatively the level of methylation of 6 specific sites within the albumin gene in the different DNA samples. The results indicate that methylation changes at the sites analysed are not responsible for the changes in gene activity during rat liver development. Further they demonstrate that: 1) extensively methylated genes can be actively transcribed; 2) prominent changes in methylation of specific genes during normal development are not necessarily related to alterations in gene activity.

Coordinate secretion of rat and mouse albumin by mouse hepatoma x rat hepatoma hybrid cells directly reflects the intracellular concentration of the corresponding mRNAs.

The ratio of mouse to rat albumin secreted by mouse hepatoma x rat hepatoma hybrid cells is constant (of the order of 5.0) irrespective of the total amounts produced. The present results establish for seven independent hybrid clones that the coordination in the ratio of mouse to rat product applies also at the level of accumulation of albumin mRNAs of the two species. The interpretation that coordinate synthesis reflects coordinate transcription of the relevant genes is thus reinforced.

No evidence for post-transcriptional control of albumin and alpha-fetoprotein gene expression in developing rat liver neoplasia.

Rot analysis of hybridization data using highly labeled alpha-fetoprotein (AFP) and albumin (32P)cDNA probes has been used to quantitate AFP and albumin mRNA sequences in RNA preparations from different subcellular fractions of developing rat liver and Morris hepatoma 7777. In addition, size analysis of these mRNA sequences has been carried out by electrophoretic fractionation on agarose gels containing methylmercury hydroxyde and hybridization to radioactive cloned albumin and AFP cDNA probes. In all the tissues examined (fetal, newborn and adult rat liver, and hepatoma 7777) most of the albumin and AFP mRNA sequences were found associated with the polysomes as mature mRNA molecules; less than 2% of these sequences were present in the nuclear or the non polysomal cytoplasmic compartments. The number of AFP mRNA molecules was found to decrease in parallel in all the cellular compartments during rat liver development. In Morris hepatoma 7777 the content of albumin mRNA was considerably decreased in all the cellular fractions as compared to normal liver. These results demonstrate that post-transcriptional control mechanisms leading to an accumulation of non-functional mRNA molecules are not implicated in the changes of expression of albumin and AFP genes during rat liver development and neoplasia.

Expression of the albumin gene in rat hepatoma cells and their dedifferentiated variants.

Rat hepatoma clones whose cells do and do not produce albumin, as well as somatic hybrid between the two types of cells, have been examined for albumin mRNA. A direct proportionality between the rate of albumin production and the concentration of albumin mRNA sequences was found for all albumin-producing hepatoma and hybrid clones, indicating that rate of synthesis of the protein is determined by the concentration of its mRNA. Albumin-negative dedifferentiated variant and somatic hybrid cells contain fewer than one to five molecules of albumin mRNA per cell; the block in expression of the gene appears to be at the same (probably transcriptional) level in variants and their somatic hybrids.

Molecular mechanisms of control of albumin and alphafetoprotein production: a system to study the early effects of chemical hepatocarcinogens.

The cellular mechanism of control of alphafetoprotein (AFP) and albumin (ALB) production during normal and pathologic permissive states is being examined using complementary DNA probes. AFP is the major serum protein for most of fetal life, but normally is produced in very small amounts in the adult. AFP production recurs in the adult during restitutive proliferation of the liver following partial hepatectomy or chemical injury, early after the exposure to chemical carcinogens and in animals with hepatocellular or yolk sac carcinomas. AFP production is roughly proportional in each case examined so far to the amount of mRNA available. On the other hand, there appears to be no difference in the gene number or gene organization in permissive or non-permissive states and there is no evidence of selective degradation of AFP or AFP (and ALB) production is probably at the level of gene transcription. Understanding of how carcinogens act to permit expression of the AFP gene may lead to important insights into carcinogenic mechanisms.

Changes in expression of albumin and alpha-fetoprotein genes during rat liver development and neoplasia.

Albumin mRNA was isolated and purified from rat liver polysomes by a combination of immunoprecipitation of specific polysomes, poly(U)-Sepharose 4B chromatography, and fractionation of the resulting poly(A)-containing RNA on a sucrose gradient. alpha-Fetoprotein (AFP) mRNA was isolated from Morris hepatoma 7777 by a similar procedure. The purity of the mRNA preparations was determined by analytical gel electrophoresis under denaturing conditions, analysis of sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the polypeptides synthesized in a wheat germ cell-free system, and the kinetics of hybridization to cDNA transcribed from albumin mRNA and AFP mRNA. The albumin mRNA possessed a chain length of approximately 2265 nucleotides and the AFP mRNA possesed a length of approximately 2235 nucleotides when examined under stringent denaturing conditions on agarose gels containing 10 mM methylmercury hydroxide. Analysis of poly(A) content by a hybridization assay with [3H]poly(U) revealed the presence in albumin mRNA of a poly(A) region containing approximately 100 adenosine residues. The AFP mRNA preparation was found to contain an average poly(A) tract of approximately 190 bases. Thus, albumin mRNA appears to contain approximately 330 untranslated nucleotides, and AFP mRNA appears to contain a similar number (approximately 285) of noncoding, nonpoly(A) bases. The purified albumin and AFP mRNA's were used as templates for synthesis of full-length cDNA hybridization probes. Both of the probes selectively hybridized to their templates with kinetics expected for single RNA species the sizes of albumin and AFP mRNA. ROt analysis was used to quantitate albumin and AFP mRNA sequences during normal liver postnatal development and liver oncogenesis. The number of polysomal AFP mRNA molecules per liver was found to drastically decrease during the first weeks of postnatal life, concomitant with a decline in the AFP synthetic capacity of the livers and in the serum concentrations of AFP. During this period, the concentration of albumin mRNA molecules per cell in the liver remained at high, approximately constant levels. In Morris hepatoma 7777, the concentration of AFP-specifying sequences was at least 10(3)-fold higher than that found in normal adult liver, whereas the content of albumin nRNA was four- to five-fold lower. These changes in concentration of albumin and AFP mRNA sequences closely correlated with a parallel variation in the specific protein synthetic capacity of the tissues.

alpha-Fetoprotein and albumin genes of rats: no evidence for amplification-deletion or rearrangement in rat liver carcinogenesis.

Full-length radiolabeled albumin and alpha-fetoprotein (AFP) cDNAs were synthesized from pure albumin and AMP mRNA preparations by using avian myeloblastosis virus reverse transcriptase (RNA-dependent DNA polymerase). The cDNAs have been used to quantitate the number of albumin and AFP genes in different rat tissues by two independent methods, both of which yielded similar results. First, the kinetics of the association of these cDNAs with nuclear DNA from rat liver, rat kidney, and Morris hepatoma 7777 under conditions of vast DNA excess indicated that the albumin and AFP mRNA's are transcribed from "nonrepetitive DNA." Second, saturation hybridization experiments in which a constant amount of rat liver DNA or Morris hepatoma 7777 was hybridized with increasing amounts of cDNA to albumin mRNA have shown the presence of 1--2 albumin genes per rat haploid genome. The number of AFP genes obtained in similar titration experiments was approximately 2--3. This was true whether rat liver DNA or hepatoma 7777 DNA was used in the reassociation experiments. When high molecular weight DNA preparations from both these tissues were digested with the restriction endonuclease EcoRI and the fragments were transferred to a nitrocellulose filter, the albumin and AFP [32P]cDNA probes hybridized to different sets of DNA fragments. However, each probe gave the same hybridization pattern whether Buffalo rat liver DNA or hepatoma 7777 DNA was utilized.

Isolation and characterization of poly(adenylic acid)-containing messenger ribonucleic acid from rat liver polysomes.

Undegraded rat liver polysomes were obtained after homogenizing the tissue in a medium containing NH4Cl, heparine, and yeast tRNA. Purification of poly(A)-containing RNA from polysomal RNA was accomplished by affinity chromatography on oligo(dT)-cellulose columns. Poly(A)-containing RNA molecules were monitored by the formation of ribonuclease-resistant hybrids with [3H]poly(U). To improve the separation of messenger RNA and ribosomal RNA by oligo(dT)-cellulose it was found essential to dissociate the aggregates formed between both molecular species by heat treatment in the presence of dimethylsulfoxide (Me2SO) prior to chromatography. Sucrose gradient analysis under denaturing conditions showed that the preparations obtained were virtually free of ribosomal RNA. Poly(A)-containing RNA constituted approx. 2.2% of the total polysomal RNA and the number average size was 1500--1800 nucleotides, as judged by sedimentation analysis on sucrose density gradients containing Me2SO. Approximately 8.2% of the purified preparation obtained was able to anneal with [3H]poly(U); the number average nucleotide length of the poly(A) segment of the RNA population was calculated to be 133 adenylate residues. Based on these values, our preparations appear to be greater than 90% pure. The RNA fractions obtained after oligo(dT)-cellulose chromatography were used to direct the synthesis of liver polypeptides in a heterologous cell-free system derived from wheat-germ. The system was optimized with respect to monovalent and divalent cations, and presence of polyamines (spermine). More than 65% of the translational activity present in the unfractionated polysomal RNA was recovered in the final poly(A)-containing RNA fraction. However, about 25% of the activity was found to be associated with the unbound fraction which was essentially free of poly(A)-containing RNA. Immunoprecipitation analysis with a specific antiserum to rat serum albumin demonstrated that about 6--8% of the labeled synthetic products translated from the poly(A)-containing RNA sample corresponded to serum albumin. Analysis of the translation products by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed a heterogeneous distribution of molecular sizes ranging from 15 000 to greater than 70 000 daltons. Spermine not only increased the overall yield and extent of protein synthesis, but also resulted in higher yields of large protein products. Under optimal translation conditions a discrete peak representing about 7% of the total radioactivity was observed to migrate with rat serum albumin.

Isolation of a chromatin fraction from calf endometrium highly enriched in estradiol binding sites.

The intranuclear distribution of [3H]-estradiol binding sites was studied in highly purified nuclei isolated from calf endometrial tissue pre-incubated with the labeled hormone. The major part (approximately 85%) of the receptor bound estradiol was found associated with the extranucleolar chromatin; only a negligible amount of [3H]-estradiol (approximately 8%) sedimented with the nucleolar fraction. [3H]-estradiol labeled chromatin was then fragmented by sonication and fractionated by sucrose density gradient sedimentation under different conditions of centrifugation. The vast majority of the [3H]-estradiol was invariably found to be associated with a fast sedimenting fraction which contained only 5 to 10% of the nuclear DNA. The concentration of estradiol receptors (per weight of DNA) in this fraction was 25- to 50-fold higher than that found in the slow sedimenting major chromatin component. Chemical analysis showed this fraction to have a high protein/DNA ratio but no phospholipids were detected.

Limited proteolysis of cytoplasmic and nuclear uterine estradiol receptors yields identical estradiol-binding fragments.

Limited tryptic hydrolysis of the estradiol cytoplasmic receptor from calf uterus has been demonstrated to yield in a high-salt buffer a stable estradiol-binding molecule with the following characteristics: sedimentation coefficient 4.0 +/- 0.1 S; Stokes radius 3.5 +/- 0.05 nm; molecular weight 60000 (for an assumed v value of 0.73 ml g-1) and frictional ratio 1.36. Nuclear KCl extracts, prepared from uteri preincubated at 37 degrees C with labeled estradiol, were analysed by Sephadex G-200 chromatography and sucrose density gradient centrifugation. The following molecular parameters were found for the estradiol-receptor complex: sedimentation coefficient 4.4 +/- 0.1 S; Stokes radius 4.12 +/- 0.02 nm; molecular weight 77000 and frictional ratio 1.47 (v = 0.73 ml g-1). Limited tryptic proteolysis of this extract gave an estradiol-binding fragment with molecular characteristics identical to the trypsin-modified cytoplasmic receptor. In addition, mild tryptic digestion of whole labeled nuclei allowed us to solubilize almost quantitatively the nuclear [3H]estradiol in a macromolecular bound form. The molecule thus obtained showed molecular parameters very similar to the 60000-dalton trypsin fragments obtained from high-salt cytoplasmic and nuclear extracts. These molecules were undistinguishable by gel electrophoresis analysis at six different acrylamide concentrations. These results in conjunction with those derived from dissociation kinetics experiments and ligand specificity studies indicate the cytosolic protein is a functional part of the nuclear receptor. Based upon these and other studies we suggest that proteolytic cleavage of the estradiol-receptor complex, which results in the removal of the estradiol-binding sites from the nuclear recognition sites of the molecule, could play a role in the inactivation of the estradiol receptor in vivo.