Hepatocyte differentiation in vitro: initiation of tyrosine aminotransferase expression in cultured fetal rat hepatocytes.

A fetal rat hepatocyte culture system has been used to study the molecular mechanisms of tyrosine aminotransferase (TAT) gene expression during development. It has previously been shown that TAT activity can be detected in 19-d, but not 15-d, gestation hepatocytes on the first day of culture (Yeoh, G. C. T., F. A. Bennett, and I. T. Oliver. 1979. Biochem. J. 180:153-160). In this study enzyme activity, synthesis, and mRNA levels were determined in hepatocytes isolated from 13-, 15-, and 19-d gestation rats maintained in culture for 1, 2, or 3 d and exposed to dexamethasone. TAT expression is barely detectable in 13-d gestation hepatocytes even after 3 d in culture. Hepatocytes isolated from 15-d gestation fetuses have undetectable levels of enzyme activity and synthesis on the first day of culture; both can be assayed by days 2 and 3. TAT mRNA levels in these hepatocytes, measured by hybridization with a specific cDNA, increase substantially during culture. TAT activity, synthesis, and mRNA are evident on the first and subsequent days of culture in 19-d gestation hepatocytes. Transcription measurements in isolated nuclei indicate that the increase in TAT mRNA in 15- and 19-d gestation hepatocytes is associated with an increase in transcription of the gene. Immunocytochemical studies demonstrated that the increase in TAT expression correlated with an increase in the proportion of hepatocytes expressing the enzyme, rather than a simultaneous increase in all hepatocytes. These results support the proposal that a subpopulation of 15-d fetal hepatocytes undergo differentiation in culture with respect to TAT.

Mutations in the glucose-6-phosphatase gene that cause glycogen storage disease type 1a.

Glycogen storage disease (GSD) type 1a is caused by the deficiency of D-glucose-6-phosphatase (G6Pase), the key enzyme in glucose homeostasis. Despite both a high incidence and morbidity, the molecular mechanisms underlying this deficiency have eluded characterization. In the present study, the molecular and biochemical characterization of the human G6Pase complementary DNA, its gene, and the expressed protein, which is indistinguishable from human microsomal G6Pase, are reported. Several mutations in the G6Pase gene of affected individuals that completely inactivate the enzyme have been identified. These results establish the molecular basis of this disease and open the way for future gene therapy.

Identification of mutations in the gene for glucose-6-phosphatase, the enzyme deficient in glycogen storage disease type 1a.

Glycogen storage disease (GSD) type 1a is an autosomal recessive inborn error of metabolism caused by a deficiency in microsomal glucose-6-phosphatase (G6Pase), the key enzyme in glucose homeostasis. Southern blot hybridization analysis using a panel of human-hamster hybrids showed that human G6Pase is a single-copy gene located on chromosome 17. To correlate specific defects with clinical manifestations of this disorder, we identified mutations in the G6Pase gene of GSD type 1a patients. In the G6Pase gene of a compound heterozygous patient (LLP), two mutations in exon 2 of one allele and exon 5 of the other allele were identified. The exon 2 mutation converts an arginine at codon 83 to a cysteine (R83C). This mutation, previously identified by us in another GSD type 1a patient, was shown to have no detectable phosphohydrolase activity. The exon 5 mutation in the G6Pase gene of LLP converts a glutamine codon at 347 to a stop (Q347SP). This Q347SP mutation was also detected in all exon 5 subclones (five for each patient) of two homozygous patients, KB and CB, siblings of the same parents. The predicted Q347SP mutant G6Pase is a truncated protein of 346 amino acids, 11 amino acids shorter than the wild type G6Pase of 357 residues. Site-directed mutagenesis and transient expression assays demonstrated that G6Pase-Q347SP was devoid of G6Pase activity. G6Pase is an endoplasmic reticulum (ER) membrane-associated protein containing an ER retention signal, two lysines (KK), located at residues 354 and 355. We showed that the G6Pase-K355SP mutant containing a lysine-355 to stop codon mutation is enzymatically active. Our data demonstrate that the ER protein retention signal in human G6Pase is not essential for activity. However, residues 347-354 may be required for optimal G6Pase catalysis.

Mutations in the glucose-6-phosphatase gene are associated with glycogen storage disease types 1a and 1aSP but not 1b and 1c.

Glycogen storage disease (GSD) type 1, which is caused by the deficiency of glucose-6-phosphatase (G6Pase), is an autosomal recessive disease with heterogenous symptoms. Two models of G6Pase catalysis have been proposed to explain the observed heterogeneities. The translocase-catalytic unit model proposes that five GSD type 1 subgroups exist which correspond to defects in the G6Pase catalytic unit (1a), a stabilizing protein (1aSP), the glucose-6-P (1b), phosphate/pyrophosphate (1c), and glucose (1d) translocases. Conversely, the conformation-substrate-transport model suggests that G6Pase is a single multifunctional membrane channel protein possessing both catalytic and substrate (or product) transport activities. We have recently demonstrated that mutations in the G6Pase catalytic unit cause GSD type 1a. To elucidate whether mutations in the G6Pase gene are responsible for other GSD type 1 subgroups, we characterized the G6Pase gene of GSD type 1b, 1c, and 1aSP patients. Our results show that the G6Pase gene of GSD type 1b and 1c patients is normal, consistent with the translocase-catalytic unit model of G6Pase catalysis. However, a mutation in exon 2 that converts an Arg at codon 83 to a Cys (R83C) was identified in both G6Pase alleles of the type 1aSP patient. The R83C mutation was also demonstrated in one homozygous and five heterogenous GSD type 1a patients, indicating that type 1aSP is a misclassification of GSD type 1a. We have also analyzed the G6Pase gene of seven additional type 1a patients and uncovered two new mutations that cause GSD type 1a.

Inhibition of tyrosine aminotransferase gene expression by retinoic acid.

Regulation of tyrosine aminotransferase (TAT) gene expression was examined in RALA255-10G, a simian virus-40 tsA mutant-immortalized adult rat hepatocyte line. At the nonpermissive temperature (40 C), these hepatocytes exhibited a differentiated phenotype and actively expressed the TAT gene, but only in the presence of dexamethasone (DEX). The glucocorticoid-mediated TAT expression was inhibited by cycloheximide, a protein synthesis inhibitor, and by RU486, a glucocorticoid antagonist, suggesting that glucocorticoid induction requires protein synthesis and may be mediated through hormone receptors. (Bu)2cAMP (Bt2cAMP) or retinoic acid, individually or in combination, failed to increase TAT mRNA levels. However, Bt2cAMP greatly potentiated the induction by DEX, whereas retinoic acid inhibited the induction by DEX or DEX/Bt2cAMP. Nuclear run-on assays demonstrated that the induction of TAT expression by DEX or DEX/Bt2cAMP in RALA255-10G cells is regulated primarily at the transcriptional level. In contrast, retinoic acid antagonized the DEX- or DEX/Bt2cAMP-mediated induction without affecting the rate of TAT gene transcription. Instead, retinoic acid destabilized TAT mRNA. The half-life values of TAT mRNA in DEX/Bt2cAMP- and DEX/Bt2cAMP/retinoic acid-treated cells were approximately 235-270 min and 90-100 min, respectively. Our results indicate that inhibition of TAT expression by retinoic acid was regulated primarily at the posttranscriptional level.

Effects of dexamethasone and cAMP on tyrosine aminotransferase expression in cultured fetal rat hepatocytes.

Fetal hepatocyte cultures were used to investigate tyrosine aminotransferase (TyrAT) expression during development. Previous studies showed that TyrAT is synthesized by hepatocytes isolated from 15-day-gestation fetuses maintained in culture for two or more day, then exposed to dexamethasone. TyrAT expression was essentially undetectable on the first day of culture of hepatocytes derived from 15-day-gestation, or less mature, fetuses. Dexamethasone and cAMP are potent inducers of TyrAT and they synergistically induce TyrAT to extremely high levels when added simultaneously to cultured fetal hepatocytes. The effects of dibutyryl-cAMP (Bt2cAMP) alone and in combination with dexamethasone on TyrAT expression are investigated. Hepatocytes isolated from 15-day-gestation fetuses exposed to both inducers possessed detectable levels of TyrAT activity and mRNA on day 1 of culture, and this increased by day 3. In contrast, hepatocytes exposed to either inducing agent alone were essentially negative on day 1, but positive on day 3. This was shown to be a consequence of transcription. When 13-day-gestation hepatocytes were maintained in culture under identical conditions detectable levels of TyrAT mRNA were evident on day 1, and this increased by day 3. Immunocytochemical studies revealed that the appearance and subsequent increase in TyrAT production elicited by dexamethasone and Bt2cAMP were due to changes in the proportion of hepatocytes expressing the enzyme. Therefore, in the presence of both dexamethasone and Bt2cAMP, TyrAT expression can be detected in some cells at an earlier stage of liver development than reported previously.

Notch-1 inhibits apoptosis in murine erythroleukemia cells and is necessary for differentiation induced by hybrid polar compounds.

Strikingly increased expression of notch-1 has been demonstrated in several human malignancies and pre-neoplastic lesions. However, the functional consequences of notch-1 overexpression in transformed cells remain unclear. We investigated whether endogenously expressed notch-1 controls cell fate determination in mouse erythroleukemia (MEL) cells during pharmacologically induced differentiation. We found that notch-1 expression is modulated during MEL cell differentiation. Premature downregulation of notch-1 during differentiation, by antisense S-oligonucleotides or by enforced expression of antisense notch-1 mRNA, causes MEL cells to abort the differentiation program and undergo apoptosis. Downregulation of notch-1 expression in the absence of differentiation inducer increases the likelihood of spontaneous apoptosis. We conclude that in MEL cells, endogenous notch-1 expression controls the apoptotic threshold during differentiation and growth. In these cells, notch-1 allows differentiation by preventing apoptosis of pre-committed cells. This novel function of notch-1 may play a role in regulating apoptosis susceptibility in notch-1 expressing tumor cells.

Isolation and characterization of mouse hepatocyte lines carrying a lethal albino deletion.

Mice homozygous for chromosomal deletions at or around the albino locus on chromosome 7 express reduced levels of a group of liver genes, including tyrosine aminotransferase (TAT) and phosphoenolpyruvate carboxykinase (PEPCK), and generally die perinatally. Sequences within the deleted region are thought to encode a regulatory factor(s) that affects expression of these genes in trans. To facilitate study of the putative factors, we immortalized hepatocytes derived from newborn cch wild-type and c14CoS deletion homozygous mice as well as cch/c14CoS heterozygous mice using a SV40 temperature-sensitive A255 mutant virus. Three c14CoS deletion homozygous hepatocyte lines were characterized and compared with the homozygous wild-type and heterozygous lines. The SV40 tsA255 mutant-transformed hepatocyte lines were temperature-sensitive for maintenance of transformation and expressed many liver-specific genes. In agreement with in vivo studies, hepatocyte lines derived from mice homozygous for the deletion expressed reduced mRNA levels of a number of liver genes including TAT, PEPCK, X1, X2, and X7 in comparison with heterozygous and wild-type cell lines. Similar mRNA levels of transferrin and albumin, genes whose expression is unaffected by the mutation in vivo, were observed in all cell lines. The expression of two genes, X5 and metallothionein, reported to be reduced in newborn mutant mice, did not differ appreciably among cell lines. TAT and PEPCK have been shown to respond poorly to glucocorticoids and cAMP in newborn mutant mice. Interestingly, all affected liver genes tested were responsive to glucocorticoids and dibutyryl cAMP in deletion homozygous cell lines as well as in wild-type and heterozygote-derived cell lines. This may suggest that effects of the deletion on expression of liver-specific genes do not cause loss of responsiveness to glucocorticoids and cAMP. These immortalized hepatocyte lines, which express most, if not all, liver-specific genes, should provide a useful means for further investigation of the effects of the albino lethal deletion.

Cloning and expression of murine S-adenosylmethionine synthetase.

Mice homozygous for chromosomal deletions at or around the albino locus on chromosome 7 express reduced levels of a group of liver genes. Here, we report the isolation and characterization of cDNA and genomic clones encoding one of the affected genes, the mouse adult liver S-adenosylmethionine (AdoMet) synthetase. This enzyme catalyzes the synthesis of AdoMet, which functions in transmethylation and transsulfuration. Mouse AdoMet synthetase cDNA is 3232 base pairs (bp) in length and contains an open reading frame that encodes an enzymatically active polypeptide of 396 amino acids. The mouse AdoMet synthetase shares 98 and 96% amino acid sequence identity with the adult liver enzyme in the rat and human, respectively. AdoMet synthetases possess the consensus ATP-binding motif Gly-X-Gly-X-X-Gly and a putative ATP-binding Lys residue at conserved locations. As an initial step toward understanding the control of AdoMet synthetase gene expression, we characterized the complete transcription unit of this gene. The AdoMet synthetase gene spans approximately 18 kilobases and consists of nine exons ranging from 78 to 1920 bp. The transcription initiation site was demonstrated by rapid amplification of cDNA ends and confirmed by primer extension studies. A putative TATA box is located at -28 to -23 bp upstream of the transcription start site. The cis-acting DNA elements in the 5'-flanking region of the AdoMet synthetase gene that drive chloramphenicol acetyltransferase gene expression in mouse hepatocytes were identified by transient expression assays. The -365 to -2-bp DNA region upstream of the transcription start site of the AdoMet synthetase gene contains promoter elements, and the -518 to -366-bp DNA region might be involved in negative gene regulation.

Structure-function analysis of human glucose-6-phosphatase, the enzyme deficient in glycogen storage disease type 1a.

Glucose-6-phosphatase (G6Pase) is the enzyme deficient in glycogen storage disease type 1a, an autosomal recessive disorder. We have previously identified six mutations in the G6Pase gene of glycogen storage disease type 1a patients and demonstrated that these mutations abolished or greatly reduced enzymatic activity of G6Pase, a hydrophobic protein of 357 amino acids. Of these, four mutations (R83C, R295C, G222R, and Q347X) are missense and one (Q347X) generates a truncated G6Pase of 346 residues. To further understand the roles and structural requirements of amino acids 83, 222, 295, and those at the carboxyl terminus in G6Pase catalysis, we characterized mutant G6Pases generated by near-saturation mutagenesis of the aforementioned amino acids. Substitution of Arg-83 with amino acids of diverse structures including Lys, a conservative change, yielded mutant G6Pase with no enzymatic activity. On the other hand, substitution of Arg-295 with Lys (R295K) retained high activity, and R295N, R295S, and R295Q exhibited moderate activity. All other substitutions of Arg-295 had no G6Pase activity, suggesting that the role of Arg-295 is to stabilize the protein either by salt bridge or hydrogen-bond formation. Substitution of Gly-222, however, remained functional unless a basic (Arg or Lys), acidic (Asp), or large polar (Gln) residue was introduced, consistent with the hydrophobic requirement of codon 222, which is predicted to be in the fourth membrane-spanning domain. It is possible that Arg-83 is involved in stabilizing the enzyme (His)-phosphate intermediate formed during G6Pase catalysis. There exist 9 conserved His residues in human G6Pase. His-9, His-119, His-252, and His-353 reside on the same side of the endoplasmic reticulum membrane as Arg-83. Whereas H119A mutant G6Pase had no enzymatic activity, H9A, H252A, and H353A mutant G6Pases retained significant activity. Substitution of His-119 with amino acids of diverse structures also yielded mutant G6Pase with no activity, suggesting that His-119 is the phosphate acceptor in G6Pase catalysis. We also present data demonstrating that the carboxyl-terminal 8 residues in human G6Pase are not essential for G6Pase catalysis.

Isolation of the gene for murine glucose-6-phosphatase, the enzyme deficient in glycogen storage disease type 1A.

Glycogen storage disease (GSD) type 1a (von Gierke disease) is caused by a deficiency in glucose-6-phosphatase, the key enzyme in glucose homeostasis catalyzing the terminal step in gluconeogenesis and glycogenolysis. Despite its clinical importance, this membrane-bound enzyme has eluded molecular characterization. Here we report the cloning and characterization of a murine glucose-6-phosphatase cDNA by screening a mouse liver cDNA library differentially with mRNA populations representing the normal and the albino deletion mouse known to express markedly reduced glucose-6-phosphatase activity. Additionally, we identified the gene that consists of 5 exons. Biochemical analyses indicate that the in vitro expressed enzyme is indistinguishable from mouse liver microsomal glucose-6-phosphatase exhibiting essentially identical kinetic constants, latency, thermal lability, and vanadate sensitivity. The characterization of the murine glucose-6-phosphatase gene opens the way for studying the molecular basis of GSD type 1a in humans and its etiology in an animal model.