Demethylation and expression of methylated plasmid DNA stably transfected into HeLa cells.

In vitro methylation at CG dinucleotides (CpGs) in a transfecting plasmid usually greatly inhibits gene expression in mammalian cells. However, we found that in vitro methylation of all CpGs in episomal or non-episomal plasmids containing the SV40 early promoter/enhancer (SV40 Pr/E) driving expression of an antibiotic-resistance gene decreased the formation of antibiotic-resistant colonies by only approximately 30-45% upon stable transfection of HeLa cells. In contrast, when expression of the antibiotic-resistance gene was driven by the Rous sarcoma virus long terminal repeat or the herpes simplex virus thymidine kinase promoter, this methylation decreased the yield of antibiotic-resistant HeLa transfectant colonies approximately 100-fold. The low sensitivity of the SV40 Pr/E to silencing by in vitro methylation was probably due to demethylation upon stable transfection. This demethylation may be targeted to the promoter and extend into the gene. By genomic sequencing, we showed that four out of six of the transfected SV40 Pr/E's adjacent Sp1 sites were hotspots for demethylation in the HeLa transfectants. High frequency demethylation at Sp1 sites was unexpected for a non-embryonal cell line and suggests that DNA demethylation targeted to certain aberrantly methylated regions may function as a repair system for epigenetic mistakes.

Frequent hypomethylation in Wilms tumors of pericentromeric DNA in chromosomes 1 and 16.

Rearrangements in the pericentromeric heterochromatin of chromosome 1 or 16 are often found in many types of cancers, including Wilms tumors, and have been suggested to contribute to oncogenesis or tumor progression. The oncogenic potential of these rearrangements has been ascribed to the resulting chromosome arm imbalances affecting the dosage of tumor suppressor genes or protooncogenes. Because DNA hypomethylation has been linked to rearrangements in the pericentromeric regions of chromosome 1 and 16 in two types of non-cancer cell populations, we examined methylation of normally highly methylated satellite DNA sequences in these regions in Wilms tumors. Hypomethylation was found to be frequent in juxtacentromeric (satellite 2) sequences and, especially, in centromeric (satellite alpha) sequences of chromosome 1. Hypomethylation of satellite 2 DNA of chromosome 16 showed a high degree of concordance with that of satellite 2 DNA of chromosome 1. We discuss the relationship of this satellite DNA hypomethylation in Wilms tumors to chromosome aberrations, as determined by assays for loss of heterozygosity.

DNA demethylation and pericentromeric rearrangements of chromosome 1.

Rearrangements in the vicinity of the centromere of chromosome 1 are over-represented in many types of human cancer and are a characteristic feature of a rare genetic disease called ICF (immunodeficiency, centromeric heterochromatin instability, and facial anomalies). Evidence is presented that implicates DNA hypomethylation in the formation of these pericentromeric chromosomal anomalies. The DNA methylation inhibitors 5-azadeoxycytidine and 5-azacytidine, but not other tested genotoxins, induced the preferential formation of pericentromeric rearrangements of chromosome 1 at a very high frequency in a pro-B-cell line (FLEB14) and at a lower frequency in a mature B-cell line (AHH-1). These abnormal chromosomes appear identical to the diagnostic chromosomal aberrations in the ICF syndrome. A major component of the pericentromeric DNA in chromosome 1, satellite 2, was shown to be hypomethylated in an ICF B-cell line, although DNA from this cell line did not display detectable overall hypomethylation. It is hypothesized that demethylation in certain DNA regions, including in pericentromeric satellite DNA, helps lead to pericentromeric chromosomal rearrangements in lymphocytes from ICF patients and in normal lymphoblastoid cells incubated in vitro with DNA demethylating agents.

Reporter gene expression upon stable transfection when only a TATA box or a TATA box plus Sp1 sites are present 5' to the gene.

Episomal plasmids for stable transfection of mammalian cell cultures were constructed that have a G418-resistance (neo) gene immediately downstream of a highly truncated promoter. These plasmids had a function hygromycin-resistance gene (hyg) as a selectable marker. Surprisingly, in LTK- cells, but not HeLa cells, stably transfected with these BK virus-based plasmids having no promoter elements adjacent to the neo gene, readthrough transcription, probably from about 1 kb upstream, gave almost as efficient expression of the neo gene as of the hyg gene with a full-length promoter immediately upstream. When the transfecting plasmids contained Epstein-Barr virus (EBV) DNA sequences for episomal maintenance and had multiple Sp1 sites and a TATA box as the only promoter elements 5' to the neo gene, only about 3-9% of HeLa transfectants were G418 resistant (G418R). In transfections with analogous plasmids lacking these promoter elements 5' to the neo gene, no G418R colonies were seen. The establishment of the G418R phenotype probably required integration of plasmid DNA into favorable chromosomal sites and was aided by the presence of the TATA box plus Sp1 sites as a subminimal promoter. The absence of detectable G418-resistance in most of the HeLa transfectant clones obtained with EBV-type plasmids, even at a high plasmid copy number and even when a TATA box and six Sp1 sites were present immediately upstream of the neo gene, indicates that these elements do not suffice for appreciable gene expression in vivo and that this is a suitable model system for studying DNA rearrangements that can potentiate expression of the neo gene.

Frequent detection of bcl-2/JH translocations in human blood and organ samples by a quantitative polymerase chain reaction assay.

Using an ultrasensitive assay involving the PCR, we have examined the frequency of a follicular lymphoma-associated translocation in peripheral blood from 132 individuals, most of whom were healthy blood donors. This translocation occurs between the bcl-2 proto-oncogene and the JH gene region and prolongs the life of lymphocytes. At a level of sensitivity of 1 translocation-bearing cell per 5 x 10(6) cells, almost one-half of healthy human adults had this translocation in the mononuclear fraction of peripheral blood. However, the range of frequency of these translocations spanned almost three orders of magnitude among translocation-positive individuals. Furthermore, there was a statistically significant increase with age in the percentage of individuals who were translocation positive. Such an age correlation was also seen for the percentage of blood donors with rather high translocation frequencies (> or = 20 per 5 x 10(6) peripheral blood mononuclear cells). However, the blood donor who had by far the highest concentration of this translocation was a healthy 35-year-old male containing approximately 900 apparently monoclonal, translocation-bearing cells per 5 x 10(6) peripheral blood mononuclear cells. Our findings suggest that some individuals who may be at risk for follicular lymphoma might be able to be identified by this PCR assay on peripheral blood. Also, these data may help explain the age dependence of the occurrence of this cancer.

Effect of internal direct and inverted Alu repeat sequences on PCR.

We have studied the effect of repeated DNA sequence, especially Alu repeats, on PCR. Alu repeats are sequences that are approximately 300 bp long and interspersed at a very high copy number throughout the human genome. We amplified part of the human low-density lipoprotein receptor gene containing two Alu repeat sequences in the same orientation, approximately 7.8 kb apart, with unique sequence primers outside these repeats. The major PCR product was a DNA fragment with an in vitro deletion between the Alu repeats. The formation of this product depended on the template concentration and the type of polymerase used. Such a product arose apparently as a result of a "jumping reaction" involving a primer whose extension was terminated prematurely within one Alu repeated followed by annealing of such an incompletely extended primer to the other, distant Alu repeat. No such jumping products were seen when a 0.8-kb region containing two nearby inverted Alu repeats within the human alpha-galactosidase A gene was subject to PCR with unique sequence primers annealing just outside these repeats.