Urea improves efficiency of bisulphite-mediated sequencing of 5'-methylcytosine in genomic DNA.

The detection of 5'-methylcytosine by the bisulphite-mediated genomic sequencing method has considerably aided study of the role of methylation in areas such as X chromosome inactivation, genomic imprinting and cancer research. However on occasion difficulty has been experienced in obtaining complete conversion of cytosine to uracil in regions of the target DNA. We report here a simple improvement to the method involving addition of urea to the bisulphite reaction, a step which greatly improves the reaction efficiency, presumably by maintaining the target DNA in single stranded form, thereby allowing complete and reliable conversion.

Identification of a novel tissue-specific processed HPRT gene and comparison with X-linked gene transcription in the Australian marsupial Macropus robustus.

The genome of the Australian marsupial Macropus robustus contains a highly conserved processed hypoxanthine phosphoribosyltransferase homologue, HPRT-2. Using the techniques of reverse transcriptase-polymerase chain reaction (RT-PCR) and protein isoelectric focusing (IEF) we have shown this processed gene to be fully functional, but liver specific. In contrast, the unprocessed X-linked parent gene HPRT-1 was expressed in all somatic tissues. Expression of the HPRT-2 gene effectively doubles the total HPRT enzyme activity in liver compared to other tissues. Analysis of the 5'-flanking sequence of HPRT-2 revealed regions with homology to the liver-specific regulatory motifs C/EBP, NF-IL6, LF-A1 and LF-B1, although the functional significance of these regions remains unknown. Consistent with X chromosome inactivation in female mammals, transcript levels of the unprocessed X-linked gene HPRT-1 were similar in males and females in all tissues examined. No HPRT-2 activity was detected in testes, indicating that this gene does not compensate for sex chromosome inactivation during spermatogenesis. Moreover, the demonstration of very high HPRT-1 enzyme levels in testes indicated that such a compensatory mechanism may not be required. Phylogenetic analyses attribute considerable antiquity to the processed gene and PCR with conserved primers spanning exons 4-8 of genomic DNA from several different kangaroo species inferring the existence of a conserved processed HPRT-2 homologue in these marsupial species. However, no such conserved PCR product was obtained with DNA from eutherian species, suggesting that integration of HPRT-2 occurred after the separation of the metatherian and eutherian lineages.

Demethylation of the human MDR1 5' region accompanies activation of P-glycoprotein expression in a HL60 multidrug resistant subline.

Chemotherapy is frequently limited by the development of multidrug resistance, a major cause of which is activation of the P-glycoprotein-encoding MDR1 gene. We have previously developed a P-glycoprotein-expressing multidrug resistant subline (HL60/E8) from the non-P-glycoprotein-expressing human HL60 promyelocytic leukemia cell line. A possible cause of MDR1 silencing in HL60 cells is methylation of the promoter proximal region, thus demethylation occurring as a result of drug treatment may be responsible for MDR1 activation in the multidrug resistant subline. Using the bisulphite genomic sequencing technique we demonstrated that HL60 DNA is methylated at multiple sites within two distinct areas, one upstream and one downstream of the transcription start point. Only a single site in each area was methylated in all strands examined, with the remaining adjacent sites showing partial methylation. In contrast, DNA from the multidrug resistant HL60/E8 subline was unmethylated at essentially all sites in both areas. Thus the development of the P-glycoprotein-expressing multidrug resistant subline was associated with demethylation of the MDR1 5' region.

Methylation sensitive protein binding to an intragenic active X-specific methylated region in the M. robustus Hprt gene.

An intragenic region of the wallaroo (Macropus robustus) hypoxanthine phosphoribosyltransferase gene which contains three active X-specific methylated cytosines was examined for protein(s) binding. In vitro DNase I footprinting of unmethylated DNA identified three footprints, one of which (footprint I) contained two of the known differentially methylated sites (a HpaII and a HhaI site). Methylation of the footprint I HpaII site only, abolished the formation of several, specific DNA-protein complexes in mobility shift assays. UV cross-linking experiments indicated that polypeptides involved in the methylation sensitive interactions with footprint I had molecular weights ranging from 72 to 48 kDa. Analogous results were obtained with nuclear extracts from both eutherian and metatherian cells, indicating that these proteins are conserved. We suggest that the binding of these proteins to the inactive X may play some role in gene silencing, with the active gene being protected from this effect by methylation of the binding site.

Intragenic matrix attachment and DNA-protein interactions in the human X-linked Hprt gene.

To investigate the possible contribution of intragenic differentially methylated cytosines to X-linked gene expression, we examined DNA-protein interactions in a region in intron 3 of the human hypoxanthine phosphoribosyltransferase gene which contains at least one HpaII site methylated specifically on the active X. In vitro DNase I footprinting experiments using unmethylated DNA and HeLa nuclear extract identified three footprints (I-III). Footprints I and III flank an Alu repeat containing the HpaII site(s), one of which is contained within footprint II. Although methylation of the HpaII site had no effect on footprint II binding interactions, methylation of nearby CpGs substantially reduced the formation of three of the specific DNA-protein complexes binding to footprint II in mobility shift assays. Additionally, an A+T rich region immediately 5' to the HpaII-containing Alu repeat was found to bind specifically to nuclear matrices in vitro. We suggest that differential methylation of CpGs may affect the binding of regulatory proteins in vivo, and that interactions between the footprint proteins and those binding to the matrix attachment region may be involved in controlling X-linked Hprt expression.

Isolation of a potentially functional HPRT processed pseudogene from the hill kangaroo Macropus robustus.

A highly conserved hypoxanthine phosphoribosyltransferase processed pseudogene (KPH) has been isolated from a female kangaroo (Macropus robustus) lambda EMBL3 genomic library. The pseudogene contains only transcribed material with all of the introns precisely removed and has possible direct repeats at either end of the message. It has a 654-nucleotide open reading frame (ORF) from the Met start codon to the stop codon that contains no additions, deletions or premature stops relative to expressed HPRT genes and, therefore, the possibility exists that it is expressed in vivo. Possible CAAT and GC boxes are present in the region 5' to the ORF and a polyadenylation signal is present in the region 3' to the ORF. If not expressed, the age of the pseudogene is estimated to be 10.7 million years. We propose that integration into the genome occurred specifically in a homocopolymeric region within a highly repeated region unique to the kangaroo genome.

Isolation of a clone partially encoding hill kangaroo X-linked hypoxanthine phosphoribosyltransferase: sex differences in methylation in the body of the gene.

An X-linked clone encoding exons 4-9 of the hypoxanthine phosphoribosyltransferase (HPRT) gene was isolated from a kangaroo (Macropus robustus: Marsupialia) lambda EMBL4 genomic library. Sequence similarity between the kangaroo and eutherian HPRT coding sequences was high; however, intron sizes varied significantly between the kangaroo and other eutherian species. HpaII and HhaI sites in the body of the gene were generally hypermethylated in vivo on the active, relative to the inactive X, with sites within intron 3 showing essentially complete correspondence of activity with methylation and inactivity with unmethylation. At approximately 5 kb downstream from the gene, a switch to unmethylation of active X-linked sites occurred. This switch occurred within a cluster of HpaII and HhaI sites that may represent a CG island associated with a subsequent gene.

Critical modulation by thymidine and hypoxanthine of sequential methotrexate-5-fluorouracil synergism in murine L1210 cells.

Treatment of murine L1210 cells with methotrexate (MTX) followed by 5-fluorouracil (FUra) produced synergistic cytotoxicity, but only in media containing serum with low concentrations of hypoxanthine, such as horse serum and dialyzed fetal calf serum. Addition of hypoxanthine (1 to 10 microM) during drug exposure reduced the synergism of sequential MTX (1 to 100 microM)-FUra (30 to 300 microM) treatment. The reduction of synergy by hypoxanthine varied with the MTX concentration, higher hypoxanthine concentrations being required to prevent synergy at higher MTX concentrations. The cytotoxicity produced by sequential MTX (10 microM)-FUra (30 to 300 microM) treatment was also reduced if thymidine was added to the regrowth media following drug exposure. The rescue by thymidine was concentration dependent, but as little as 0.5 microM thymidine was sufficient to substantially reduce the synergistic cytotoxicity. These results indicate that both hypoxanthine and thymidine are critical determinants of sequential MTX-FUra synergy and call into question the relevance of experiments in low-thymidine- and low-hypoxanthine-containing media to the clinical situation, where plasma hypoxanthine and thymidine concentrations are often greater than 1 and 0.5 microM, respectively.

Allopurinol modulation of fluorouracil toxicity.

Considerable interest has developed in the modulation of fluorouracil activity by nucleosides. The toxicity of fluorouracil in mice is known to be reduced by concurrent administration of allopurinol, presumably because biochemical pathways activating fluorouracil in normal tissues are blocked. We have given allopurinol (300 mg t.d.s. PO) concurrently with continuous infusions of fluorouracil (2.0--2.25 g/m2/day X 5) to 34 patients with colorectal cancer and 11 patients with various adenocarcinomas. There were 41 patients assessable for toxicity. Stomatitis was the predominant dose-limiting toxicity (22% grade 1, 19% grade 2, and 27% grade 3 toxicity). Neutropenia (Less Than 1,000/mu l) occurred in 17% patients. Among 26 colorectal cancer patients assessable for response there was a 15.4% response rate. We conclude that allopurinol modulates fluorouracil toxicity in man, allowing a two-fold increase in dose. However, at least in colorectal cancer no greater frequency of tumour response is seen than with lower doses of fluorouracil given by standard schedules of administration without allopurinol.

The activities of thymidine metabolising enzymes during the cell cycle of a human lymphocyte cell line LAZ-007 synchronised by centrifugal elutriation.

The activities throughout the cell cycle of thymidine kinase (EC 2.7.1.21), dihydrothymine dehydrogenase (EC 1.3.1.2), thymidine phosphorylase (EC 2.4.2.4) and dTMP phosphatase (EC 3.3.3.35) were measured in the Epstein-Barr virally transformed human B lymphocyte line LAZ-007. Cells were synchronised at different stages of the cell cycle using the technique of centrifugal elutriation. The degree of synchrony in each cycle-stage cell population was determined by flow microfluorimetric analysis of DNA content and by measurement of thymidine incorporation into DNA. The activity of the anabolic enzyme thymidine kinase was low in the G1 phase cells, but increased manyfold during the S and G2 phases, reaching a maximum after the peak of DNA synthesis, then decreasing in late G2 + M phase. By contrast, the specific activities of the enzymes involved in thymidine and thymidylate catabolism, dihydrothymine dehydrogenase, thymidine phosphorylase and dTMP phosphatase remained essentially constant throughout the cell cycle, indicating that the fate of thymidine at different stages of the cell cycle is governed primarily by regulation of the level of the anabolic enzyme thymidine kinase and not by regulation of the levels of thymidine catabolising enzymes.