Differential expression of two linked selection genes (HSVI-tk and Eco.gpt) in transformed teratocarcinoma and in L cells

Upon transfection of (TK-)F9 teratocarcinoma stem-cells and (TK-)L fibroblasts with a plasmid carrying two selection genes, Eco.gpt and HSVI-tk, selection for gpt gene yielded ten times fewer colonies than selection for tk. Only the transformed clones selected for gpt had measurable xanthine guanine phosphoribosyltransferase (XGPRT) activity (Jami et al., 1983). Eco.gpt coding for XGPRT was under the control of simian virus 40 (SV40) early genes' regulating sequences (SV-gpt). In the present study, it was verified that the low efficiency of gpt selection in mouse cells was not due to the eucaryotic controlling sequences added to the bacterial gene. The transformed clones selected for tk that had no XGPRT activity possessed at least one uninterrupted copy of the composite SV-gpt gene and as many copies of the transforming plasmid as the cells selected for gpt expression. In a further test, the gpt gene was placed under the control of tk-regulating sequences and inserted with the tk gene in the same vector. Under these conditions, expression of XGPRT in the transformed clones selected for tk was improved, even though relative selection for gpt remained low.     

Stable transformation of mouse teratocarcinoma stem cells with the dominant selective marker Eco.gpt and retention of their developmental potentialities

Transformation of PCC4 mouse teratocarcinoma stem cells was obtained using a dominant selective marker, the enzyme xanthine-guanine phosphoribosyltransferase (XGPRT), coded by the bacterial Eco.gpt gene placed under the control of the early SV40 genes in the vector pSV2gpt. An average of 20 colonies of transformed cells was obtained, using the calcium phosphate technique, 10 microg DNA vector, no carrier DNA and 1 x 10(6) recipient cells. Five independent Eco.gpt-transformed PCC4 cell lines were propagated in selective medium and assayed for XGPRT activity. All of them had the ability to convert [14C]xanthine to xanthine monophosphate. pSV2gpt sequences were present and associated with high mol. wt. cellular DNA. pSV2gpt sequences and XGPRT activity were both conserved in the three clones that were propagated in non-selective medium for 30 generations. The transformed PCC4 cells retained their ability to produce, in host mice, teratocarcinoma tumors composed of embryonal carcinoma and various differentiated tissues. Thus, pSV2gpt can be used as a dominant marker to select teratocarcinoma stem cells co-transformed with genes that are not selectable by themselves.     

Transformation of teratocarcinoma stem cells and fibroblasts with various vectors containing the Eco.gpt gene as a selection marker

Eco.gpt, which codes for xanthine guanine phosphoribosyltransferase (XGPRT), when placed under the control of SV40 early genes regulating sequences (pSV2gpt) selects transformed teratocarcinoma cells with a low efficiency. The SV40 promoter may not function efficiently in teratocarcinoma stem cells, as suggested by the fact that such cells do not support SV40 T antigen expression. We have tested whether one could change the efficiency of gpt as a dominant selective marker in transformation by several operations. (1) Deletion of 121 base pairs (bp) upstream the bacterial coding sequence gpt (pQS14) did not make any difference. (2) Replacement of the SV40 regulating sequences by the HSV tk regulating sequences (pQS15) resulted in ten times fewer transformants with PCC4 teratocarcinoma cells as well as with L cells. No XGPRT activity was detectable in cultures 48 h after transfection. (3) Reintroduction of the PvuII-HindIII SV40 fragment (which contains an enhancing sequence together with the origin of replication and the early promoter of the virus) into the pQS15 vector, either in 5' or 3' from tk-gpt composite gene (pQS20 and pQS22) allows selection of ten to twenty times more transfected PCC4 or L cells colonies and restores transient XGPRT activity upon transfection. Whatever the vector used, the transformation frequency of PCC4 teratocarcinoma cells remains ten times lower than that of L cells. It appears that the presence of the SV40 PvuII-HindIII fragment in the vector increases cell transformation even with PCC4 cells and that the low frequency obtained with pSV2pgt is likely not due to the use of the SV40 early promoter.(ABSTRACT TRUNCATED AT 250 WORDS)     

Escherichia coli gpt gene provides dominant selection for vaccinia virus open reading frame expression vectors.

Mycophenolic acid, an inhibitor of purine metabolism, was shown to block the replication of vaccinia virus in normal cell lines. This observation led to the development of a dominant one-step plaque selection system, based on expression of the Escherichia coli gpt gene, for the isolation of recombinant vaccinia viruses. Synthesis of xanthine-guanine phosphoribosyltransferase enabled only the recombinant viruses to form large plaques in a selective medium containing mycophenolic acid, xanthine, and hypoxanthine. To utilize the selection system efficiently, we constructed a series of plasmids that contain the E. coli gpt gene and allow insertion of foreign genes into multiple unique restriction endonuclease sites in all three reading frames between the translation initiation codon of a strong late promoter and synthetic translation termination sequences. The selection-expression cassette is flanked by vaccinia virus DNA that directs homologous recombination into the virus genome. The new vectors allow high-level expression of complete or partial open reading frames and rapid construction of recombinant viruses by facilitating the cloning steps and by simplifying their isolation. The system was tested by cloning the E. coli beta-galactosidase gene; in 24 h, this enzyme accounted for approximately 3.5% of the total infected-cell protein.     

An assay for transient gene expression in transfected Drosophila cells, using [3H]guanine incorporation.

We have developed an assay for transient gene expression using a dominant-selectable marker previously employed to transform Drosophila cultured cells. Drosophila hydei cells transfected with a functional Escherichia coli xanthine guanine phosphoribosyl transferase gene (gpt), under the control of the long terminal repeats (LTRs) of the copia transposable element, rapidly incorporate guanine into acid-precipitable counts. Autoradiographic analysis in situ shows that approximately 20% of cells take up, and express, the gpt gene. This transient gpt expression depends on the Drosophila promoter sequences since vectors with the gpt gene in reverse orientation to the copia LTRs fail to incorporate guanine. Deletion analysis confirms that the LTRs are essential for gpt gene expression. Similarly, cells transfected with gpt controlled by the Drosophila 70 000 mol. wt. heat-shock (hsp 70) promoter show regulated guanine incorporation when heat shocked. The efficiency of the copia LTRs varies considerably between the cell lines we tested, whereas that of the hsp 70 promoter does not. The heterologous promoters of the Rous sarcoma virus (RSV) and simian virus 40 (SV40) function poorly in these cells.     

Detection of deletion mutations in pSV2gpt-transformed cells.

We have developed a system to study mutations that affect xanthine-guanine phosphoribosyltransferase gene (gpt) expression in hypoxanthine-guanine phosphoribosyltransferase-deficient CHO cells that have been transformed by the plasmid vector pSV2gpt. One isolated transformant, designated AS52, carries a single copy of the Escherichia coli gpt gene stably integrated into the high-molecular-weight DNA and expresses the bacterial gene for the enzyme xanthine-guanine phosphoribosyltransferase. Mutants deficient in this enzyme can be induced in the AS52 cell line by a variety of mutagens, and spontaneous or induced mutants can be selected for resistance to 6-thioguanine (Tgr). Two Tgr clones derived from the AS52 line were analyzed by Southern blot hybridization and were found to contain deletions involving at least a portion of the gpt gene. Because of the small size and stability of the integrated pSV2gpt plasmid, and the well-defined selection protocol for mutant isolation, the AS52 line offers promise as a system suitable for the study of mutation at the molecular level in CHO cells.     

The histone deacetylase inhibitor trichostatin A reduces nickel-induced gene silencing in yeast and mammalian cells

We have previously reported that nickel (Ni)-silenced expression of the URA3 gene in yeast (Saccharomyces cerevisiae) and gpt transgene in G12 Chinese hamster cells. In both cases, close proximity to a heterochromatic region was required for gene silencing. Yeast exposed to Ni exhibited reduced acetylation of the lysine residues in the N-terminal tail of histone H4. Ni-induced silencing of the gpt gene in mammalian cells involved hypermethylation of promoter region DNA. Yeast do not employ DNA methylation to silence gene expression. To determine if histone deacetylation participates in Ni-induced silencing of the URA3 and gpt genes, we exposed yeast and G12 hamster cells to the histone deacetylase inhibitor trichostatin A (TSA) prior to and concurrently with Ni. Treatment of yeast cells with 0.2-0.6mM NiCl(2) resulted in reduced expression of the URA3 gene as assessed by increased resistance to 1g/l 5-fluorotic acid (5-FOA). This effect was lessened when yeast were pre-treated with 50 microg TSA/ml. Similarly, treatment of G12 cells with 5 ng/ml TSA during and after exposure to 0.3 microg Ni(3)S(2)/cm(2) reduced silencing of the gpt gene as gauged by resistance to 10 microg/ml 6-thioguanine (6-TG). The ability of TSA alone and in combination with the DNA-demethylating agent (5-AzaC) to reactivate the gpt gene in Ni-silenced variants was also assessed. Although treatment with 100 ng/ml TSA for 48 h was partially effective in reactivating the gpt gene, treatment with 5 microM 5-AzaC was more efficacious. The greatest gpt gene reversion frequencies were observed following a sequential 5-AzaC/TSA treatment. Taken all together, our data from mammalian cells suggests that both DNA methylation and histone deacetylation participate in Ni-induced silencing of the gpt gene with DNA hypermethylation playing the more dominant role in maintaining the silenced state.     

Angiotensinogen production by rat hepatoma cells in culture and analysis of its regulation by techniques of somatic cell genetics

Angiotensinogen was synthesized by cells derived from the Reuber H35 rat hepatoma. Independent clones produced similar amounts of angiotensinogen, which corresponded to about four times more than expected for normal hepatocytes. The protein was secreted rapidly but could be visualized within cells using immunofluorescence. For one clone, it is shown that maximal angiotensinogen synthesis occurred during mid-exponential growth. Somatic cell genetics techniques have been used to investigate the regulation of angiotensinogen expression. Eleven clones of dedifferentiated variant hepatoma cells that failed to produce most or all of the liver specific proteins analyzed including albumin fell into two groups: Seven clones produced only 1-3% as much angiotensinogen as the differentiated clones, and four showed a reduction to 10-30%. Clones of the latter class were the only ones among the eleven analyzed that retained the potential to give rise to revertants, showing restoration of the differentiated state. All revertants fully restored angiotensinogen production, but only some of them re-expressed albumin. Somatic hybrids between differentiated hepatoma cells and one of the variants showed a substantial reduction in angiotensinogen production, whereas for some clones, albumin synthesis was fully maintained. These results show that regulation of the expression of angiotensinogen and of a second serum protein, albumin, was independent and that angiotensinogen synthesis was a faithful indicator of the general differentiation profile of all classes of clones.     

Selection against expression of the Escherichia coli gene gpt in hprt+ mouse teratocarcinoma and hybrid cells.

Thioxanthine is toxic for mammalian cells transformed by the dominant selectable marker gpt. It allowed us to select, in the presence of the endogenous hypoxanthine-guanine phosphoribosyltransferase gene, mutants that did not express gpt any more and also hybrid cells that had lost the chromosome carrying it. The gpt marker is thus dominant in negative as well as in positive selection, which makes it potentially very useful for genetic studies of mammalian cells.     

Inactivation of a transfected gene in human fibroblasts can occur by deletion, amplification, phenotypic switching, or methylation.

Plasmids containing the bacterial gpt gene under control of the simian virus 40 promoter were transfected into a simian virus 40-transformed human fibroblast line. Two transfectants, E2 and C10, which contain stably integrated single copies of the gpt gene, were isolated. These two lines produce Gpt- variants spontaneously with a frequency of about 10(-4). We carried out a detailed molecular analysis of the spectrum of alterations which gave rise to the Gpt- phenotype in these variants. DNA from 14 of 19 Gpt- derivatives of one of the cell lines (E2) contains deletions or rearrangements of gpt-containing sequences. In four of the remaining five lines, the Gpt- phenotype was correlated with reduced levels of expression rather than with changes in the gross structure of the gpt gene, and it was possible to reactivate the gpt gene. In one Gpt- line, gpt mRNA was present at normal levels, but no active enzyme was produced. Spontaneous Gpt- derivatives of the other cell line (C10) produced a completely different spectrum of alterations. Very few deletions were found, but several derivatives contained additional extrachromosomal gpt sequences, and, remarkably, in two other Gpt- lines, gpt-containing sequences were amplified more than 100-fold. The phenotypes of the majority of the Gpt- derivatives of C10 could be attributed to alterations in gene expression caused by methylation.     

Amplification and hormone-regulated expression of a mouse mammary tumor virus-Eco gpt fusion plasmid in mouse 3T6 cells.

Mouse 3T6 cells were transformed with a chimeric DNA plasmid, pSVMgpt, in which the mouse mammary tumor virus (MMTV) promoter was fused to the Escherichia coli gene encoding xanthine-guanine phosphoribosyl transferase (Eco gpt). The transformants exhibited glucocorticoid-inducible expression of Eco gpt. With limiting xanthine concentrations, conditions were established in which cell growth became hormone dependent. Cells selected for their ability to grow in limiting concentrations of both xanthine and glucocorticoids contained amplified levels of Eco gpt DNA, and expression of Eco gpt remained glucocorticoid inducible in these amplified cells. Thus, amplification of the MMTV promoter region in itself did not abolish hormonal responsiveness of a gene. In addition to increased levels of Eco gpt DNA, some of the selected cells also exhibited increased levels (two- to threefold) of glucocorticoid receptors. Lastly, we found that excessive expression of Eco gpt is toxic to 3T6 cells; by maintaining low hormone levels and, therefore, low levels of expression, we were able to select cells with amplified Eco gpt. Thus, the MMTV promoter may be of general utility in expressing genes whose products may be lethal if they are produced in excessive quantities.     

Molecular analysis of ethyl methanesulfonate-induced reversion of a chromosomally integrated mutant shuttle vector gene in mammalian cells.

The molecular mechanisms of ethyl methanesulfonate-induced reversion in mammalian cells were studied by using as a target a gpt gene that was integrated chromosomally as part of a shuttle vector. Murine cells containing mutant gpt genes with single base changes were mutagenized with ethyl methanesulfonate, and revertant colonies were isolated. Ethyl methanesulfonate failed to increase the frequency of revertants for cell lines with mutant gpt genes carrying GC----AT transitions or AT----TA transversions, whereas it increased the frequency 50-fold to greater than 800-fold for cell lines with mutant gpt genes carrying AT----GC transitions and for one cell line with a GC----CG transversion. The gpt genes of 15 independent revertants derived from the ethyl methanesulfonate-revertible cell lines were recovered and sequenced. All revertants derived from cell lines with AT----GC transitions had mutated back to the wild-type gpt sequence via GC----AT transitions at their original sites of mutation. Five of six revertants derived from the cell line carrying a gpt gene with a GC----CG transversion had mutated via GC----AT transition at the site of the original mutation or at the adjacent base in the same triplet; these changes generated non-wild-type DNA sequences that code for non-wild-type amino acids that are apparently compatible with xanthine-guanine phosphoribosyltransferase activity. The sixth revertant had mutated via CG----GC transversion back to the wild-type sequence. The results of this study define certain amino acid substitutions in the xanthine-guanine phosphoribosyltransferase polypeptide that are compatible with enzyme activity. These results also establish mutagen-induced reversion analysis as a sensitive and specific assay for mutagenesis in mammalian cells.     

Temporal, spatial and tissue-specific expression of a myogenin-lacZ transgene targeted to the Hprt locus in mice.

A lacZ transgene, expressed by the myogenin promoter, was introduced into the mouse hypoxanthine phosphoribosyltransferase (Hprt) locus by gene targeting in embryonic stem cells. Embryos between E10.5-E18.5 days were analyzed for expression of the transgene after staining for beta-galactosidase activity. Transgene expression was restricted to the skeletal muscle lineages reflecting a similar temporal and spatial pattern previously demonstrated for the endogenous myogenin gene. Additionally, a second transgene, MC1tk, showed expression in 87% of the clones when targeted to Hprt. This strategy, called targeted transgenesis, provides control for analyzing promoter sequences and for comparing various transgenes expressed by the same promoter.     

Albumin is encoded by 2 messenger RNAs in Xenopus laevis

A cDNA clone library was prepared from liver poly(A) RNA pf non-estrogenized Xenopus laevis. Albumin coding sequences were screened by hybridization to a cDNA prepared from poly(A) RNA enriched by sucrose density gradient centrifugation, and by a sensitive solid-phase radioimmunoassay to detect clones that contain templates for albumin antigenic determinants. Nine clones were obtained by this approach, and all but one have the cDNA inserted in phase with the beta-lactamase gene of pBR322. Mapping of these clones with restriction endonucleases yielded 2 distinct patterns, suggestive of heterogeneity in the coding sequences. This was confirmed by heteroduplex analyses of hybrids formed between clones representative of each of the 2 classes. Both classes of albumin cDNA clones were used to select mRNAs of the same size (2.3kb) that code for peptides that are indistinguishable by SDS gel electrophoresis. Examination of the organization of the albumin genes by blot hybridization of the cDNA clones to restriction fragments of Xenopus DNA failed to detect any differences at the genomic level. The considerable diversity of the albumin cDNAs is suggestive of a multiplicity of albumin genes, rather than differential processing of a common precursor RNA.     

Hybrid Cytomegalovirus-U6 Promoter-based Plasmid Vectors Improve Efficiency of RNA Interference in Zebrafish

Short hairpin RNA (shRNA) directed by RNA polymerase III (Pol III) or Pol II promoter was shown to be capable of silencing gene expression, which should permit analyses of gene functions or as a potential therapeutic tool. However, the inhibitory effect of shRNA remains problematic in fish. We demonstrated that silencing efficiency by shRNA produced from the hybrid construct composed of the CMV enhancer or entire CMV promoter placed immediately upstream of a U6 promoter. When tested the exogenous gene, silencing of an enhanced green fluorescent protein (EGFP) target gene was 89.18 +/- 5.06% for CMVE-U6 promoter group and 88.26 +/- 6.46% for CMV-U6 promoter group. To test the hybrid promoters driving shRNA efficiency against an endogenous gene, we used shRNA against no tail (NTL) gene. When vectorized in the zebrafish, the hybrid constructs strongly repressed NTL gene expression. The NTL phenotype occupied 52.09 +/- 3.06% and 51.56 +/- 3.68% for CMVE-U6 promoter and CMV-U6 promoter groups, respectively. The NTL gene expression reduced 82.17 +/- 2.96% for CMVE-U6 promoter group and 83.06 +/- 2.38% for CMV-U6 promoter group. We concluded that the CMV enhancer or entire CMV promoter locating upstream of the U6-promoter could significantly improve inhibitory effect induced by the shRNA for both exogenous and endogenous genes compared with the CMV promoter or U6 promoter alone. In contrast, the two hybrid promoter constructs had similar effects on driving shRNA.