A Bayesian Framework to Identify Methylcytosines from High-Throughput Bisulfite Sequencing Data

High-throughput bisulfite sequencing technologies have provided a comprehensive and well-fitted way to investigate DNA methylation at single-base resolution. However, there are substantial bioinformatic challenges to distinguish precisely methylcytosines from unconverted cytosines based on bisulfite sequencing data. The challenges arise, at least in part, from cell heterozygosis caused by multicellular sequencing and the still limited number of statistical methods that are available for methylcytosine calling based on bisulfite sequencing data. Here, we present an algorithm, termed Bycom, a new Bayesian model that can perform methylcytosine calling with high accuracy. Bycom considers cell heterozygosis along with sequencing errors and bisulfite conversion efficiency to improve calling accuracy. Bycom performance was compared with the performance of Lister, the method most widely used to identify methylcytosines from bisulfite sequencing data. The results showed that the performance of Bycom was better than that of Lister for data with high methylation levels. Bycom also showed higher sensitivity and specificity for low methylation level samples (<1%) than Lister. A validation experiment based on reduced representation bisulfite sequencing data suggested that Bycom had a false positive rate of about 4% while maintaining an accuracy of close to 94%. This study demonstrated that Bycom had a low false calling rate at any methylation level and accurate methylcytosine calling at high methylation levels. Bycom will contribute significantly to studies aimed at recalibrating the methylation level of genomic regions based on the presence of methylcytosines.     

DNA methylation: a profile of methods and applications

Ever since methylcytosine was found in genomic DNA, this epigenetic alteration has become a center of scientific attraction, especially because of its relation to gene silencing in disease. There is currently a wide range of methods designed to yield quantitative and qualitative information on genomic DNA methylation. The earliest approaches were concentrated on the study of overall levels of methylcytosine, but more recent efforts havefocused on the study ofthe methylation status of specific DNA sequences. Particularly, optimization of the methods based on bisulfite modification of DNA permits the analysis of limited CpGs in restriction enzyme sites (e.g., combined bisulfite restriction analyses and methylation-sensitive single nucleotide primer extension) and the overall characterization based on differential methylation states (e.g., methylation-specific PCR, MethyLight, and methylation-sensitive single-stranded conformational polymorphism) and allows very specific patterns of methylation to be revealed (bisulfite DNA sequencing). In addition, novel methods designed to search for new methylcytosine hot spots have yielded further data without requiring prior knowledge of the DNA sequence. We hope this review will be a valuable tool in selecting the best techniques to address particular questions concerning the cytosine methylation status of genomic DNA.     

Histone deacetylase inhibitors decrease reelin promoter methylation in vitro

We investigated the effects of agents that induce reelin mRNA expression in vitro on the methylation status of the human reelin promoter in neural progenitor cells (NT2). NT2 cells were treated with the histone deacetylase inhibitors, trichostatin A (TSA) and valproic acid (VPA), and the methylation inhibitor aza-2'-deoxycytidine (AZA) for various times. All three drugs reduced the methylation profile of the reelin promoter relative to untreated cells. The acetylation status of histones H3 and H4 increased following treatment with VPA and TSA at times as short as 15 min following treatment; a result consistent with the reported mode of action of these drugs. Chromatin immunoprecipitation experiments showed that these changes were accompanied by changes occurring at the level of the reelin promoter as well. Interestingly, AZA decreased reelin promoter methylation without concomittantly increasing histone acetylation. In fact, after prolonged treatments with AZA, the acetylation status of histones H3 and H4 decreased relative to untreated cells. We also observed a trend towards reduced methylated H3 after 18 h treatment with TSA and VPA. Our data indicate that while TSA and VPA act to increase histone acetylation and reduce promoter methylation, AZA acts only to decrease the amount of reelin promoter methylation.     

DNA repair replication,DNA breaks and sister-chromatid exchange in human cells treated with adriamycin in vitro

The effects of adriamycin (AM) on DNA repair replication, the frequency of sister-chromatid exchange (SCE), the rate of cell proliferation and the frequency of DNA strand breaks were studied in human cells in vitro. No repair replication was observed in lymphocytes exposed to AM in concentrations up to 10^?3 moles/1. DNA repair replication induced by UV and alkylating agents was not affected by a concentration of AM that completely inhibited cell proliferation (10^?6 moles/1).Fibroblasts exposed to AM at 10^?4 moles/1 in the presence of hydroxyurea showed an increase of strand breaks and cross-links in DNA. When AM was added to UV-irradiated fibroblasts, there was an increase of DNA strand breaks in addition to the breaks caused by UV alone. Similar effects were observed in lymphocytes.A dose-dependent increase of SCE was observed in lymphocytes exposed to low concentrations of AM (<10^?7 moles/1). At higher concentrations the increase of SCE levelled off, and cell proliferation became severely inhibited. There was no evidence of removal of SCE-inducing damage in cells exposed to AM during G₀ or G₁. The level of SCE induced in the third cell cycle after treatment with AM was not different from that induced during the first two cell cycles.These results suggest that the various genotoxic and cytotoxic effects of AM are caused by different types of cellular damage. Moreover, AM-induced DNA damage persists for several cell cycles in human cells in vitro and seems to be resistant to repair activity.     

Effects of caffeine-pretreatment on SCE and chromosome aberrations induced by monofunctional- and bifunctional-mitomycin C in bloom syndrome lymphocytes

Bloom syndrome (BS) lymphocytes, which are characterized by a high incidence (75.4 per cell) of SCE, were treated with caffeine (CAF) during the first cell cycle and with monofunctional-(M-MC) and bifunctional-(MC)mitomycin C during the second cycle. The effect on the SCE level was synergistic. The CAF-pretreated cells in combination with M-MC and MC post-treatments, had significantly higher (SCE values 152.5 and 167.9 SCE per cell, resp.) than those treated with M-MC or MC alone during the second cycle (101.1 and 116.4 SCE per cell, resp.). M-MC and MC in the presence of BrdU (without CAF) for 2 cell cycles increased SCE to 157.6 and 169.4 per cell (about twice the control level). M-MC + CAF and MC + CAF treatments for 2 cell cycles did not produce a synergistic effect on the SCE frequency in BS cells; the SCE level was not significantly greater than that with M-MC or MC alone. Normal cells treated with MC and CAF for 2 cycles had a maximum SCE frequency of 156 per cell. This suggests that cells with SCE frequencies above this level may not be able to survive, i.e., this is the “saturation” level of SCE. However, CAF alone had almost no effect on SCE in either BS or normal cells and did not produce multiple chromosome aberrations. The lack of CAF effect on BS cells suggests that the lesions in DNA strands of BS cells which lead to SCE are double-strand lesions. In normal cells CAF is known to significantly slow down DNA-chain growth; the reduced rate of DNA-chain growth in BS is an inherent defect of the cells. Therefore, though CAF enhanced SCE and chromosome aberrations (shattered chromosomes) in combination with alkylating agents, CAF alone did not significantly increase the SCE rate in either BS cells or in normal cells. Thus, processes which may induce SCE are not only related to retarded rate of DNA-chain growth, but also to breaks in the template strand permitting double-strand exchanges to occur.     

Correspondence between effects of 5-azacytidine on SCE formation, cell cycling and DNA methylation in Chinese hamster cells

The effects of 5-azacytidine (5-Aza-C), alone and in combination with mitomycin C, were measured on sister-chromatid exchange (SCE) formation and DNA methylation in different genomic regions of Chinese hamster ovary cells and in Chinese hamster cells containing amplified, dihydrofolate reductase sequences and resistant to methotrexate. 5-Aza-C, when present for the penultimate preharvest cell cycle, induced SCEs in a manner consistent with a directly measured reduction in deoxycytosine methylation in cellular DNA. At higher 5-Aza-C concentrations, cell cycling was inhibited and both SCE induction and DNA demethylation tended to level off. Under appropriate conditions, 5-Aza-C also potentiated the induction of SCEs by mitomycin C. 5-Aza-C-induced DNA demethylation could also be detected in the vicinity of different DNA sequences with the use of comparative HpaII/MspI digestion, DNA blotting, and molecular probes. The efficiency of an individual demethylation event in inducing SCE induction appeared to be very low, compared with alkylating agents such as 8-methoxypsoralen, suggesting that SCE induction by 5-Aza-C might be an indirect effect from long range changes induced in cellular DNA or chromatin conformation.     

Determination of DNA sequences containing methylcytosine in Bacillus subtilis Marburg.

The methylcytosine-containing sequences in the DNA of Bacillus subtilis 168 Marburg (restriction-modification type BsuM) were determined by three different methods: (i) examination of in vivo-methylated DNA by restriction enzyme digestion and, whenever possible, analysis for methylcytosine at the 5' end; (ii) methylation in vitro of unmethylated DNA with B. subtilis DNA methyltransferase and determination of the methylated sites; and (iii) the methylatability of unmethylated DNA by B. subtilis methyltransferase after potential sites have been destroyed by digestion with restriction endonucleases. The results obtained by these methods, taken together, show that methylcytosine was present only within the sequence 5'-TCGA-3'. The presence of methylcytosine at the 5' end of the DNA fragments generated by restriction endonuclease AsuII digestion and the fact that in vivo-methylated DNA could not be digested by the enzyme XhoI showed that the recognition sequences of these two enzymes contained methylcytosine. As these two enzymes recognized a similar sequence containing a 5' pyrimidine (Py) and a 3' purine (Pu), 5'-PyTCGAPu-3', the possibility that methylcytosine is present in the complementary sequences 5'-TTCGAG-3' and 5'-CTCGAA-3' was postulated. This was verified by the methylation in vitro, with B. subtilis enzyme, of a 2.6-kilobase fragment of lambda DNA containing two such sites and devoid of AsuII or XhoI recognition sequences. By analyzing the methylatable sites, it was found that in one of the two PyTCGAPu sequences, cytosine was methylated in vitro in both DNA strands. It is concluded that the sequence 5'-PyTCGAPu-3' is methylated by the DNA methyltransferase (of cytosine) of B. subtilis Marburg.     

HL-60细胞内DNA甲基化作用与RNA聚合酶活力的关系

以 S-腺苷酰 - L-甲硫氨酸 ( SAM)为诱导物 ,在 1 0 μmol/L最佳浓度下可诱导 HL- 60细胞分化达 1 6%左右 .HPLC测定结果证明 ,诱导物处理后 HL- 60细胞 DNA甲基化水平升高 .通过 3 H-UTP同位素参入法 ,测定了不同处理时间和不同浓度 SAM对 HL- 60细胞 DNA模板体外转录活性的影响 ,发现体外活力下降 .比较了不同浓度α-鹅膏蕈碱存在下 RNA聚合酶活力的变化 ,结果表明 SAM处理后细胞中不同 RNA转录产物所占份额改变     

DNA methylation profiling in cancer: From single nucleotides towards the methylome

The genomic DNA methylation pattern (methylome) is a cell epigenetic program that controls the expression of genetic information. The methylation pattern substantially changes in early carcinogenesis. A detailed survey of the methylcytosine distribution in the genome in norm and pathology is of immense importance for a better understanding of the etiology of cancer and its early diagnosis. The techniques available make it possible to simultaneously examine many samples (high-throughput analysis) and to examine large genome loci or even the total methylome (large-scale analysis). The review considers the main trends in the development of new approaches to DNA methylation and describes the techniques most commonly used in the field, their application, and results. Emphasis is placed on the use of various DNA microarrays (oligonucleotide microarrays, BAC arrays, etc.) as a method of choice for epigenetic analysis of tumors. Alternative sequence-based techniques of methylation analysis are discussed. The use of large-scale analysis to identify new epigenetic markers and to develop an epigenetic classification of neoplasms is considered.     

The effect of chlorpromazine on SCE frequency in human chromosomes: An in vitro and in vivo study

Schizophrenic patients who were receiving, or who had received chlorpromazine showed SCE levels similar to those in a normal control population. Of 8 normal individuals whose lymphocytes were exposed in vitro to chlorpromazine (0.05–2.00 μg/ml) for two cell cycles, 4 showed a significant increase in SCE, 3 showed no increase and 1 a decrease compared with untreated lymphocytes. Lymphocytes from a further 8 donors treated with 2.0 μg/ml chlorpromazine prior to mitogen stimulation (G₀ lymphocytes) showed a similar SCE response. Only 3 of the 8 donors showed a significant increase in SCEs over the baseline level. When proliferating lymphocytes were exposed to chlorpromazine 38 h after culture initiation and prior to the addition of BrdUrd to the culture medium, metaphase chromosomes from only 3 of the 8 individuals studied showed increased levels of exchange. These results indicate that chlorpromazine can induce SCEs in vitro but that there is considerable variation in SCE response among individuals. Furthermore, our data emphasises the importance of using more than 1 or 2 donors when analysing SCE response in human chromosomes.     

Azelaic Acid Has Sensitizing Effect in the Chemotherapeutic Treatment of Several Melanoma Cell Lines

Chemotherapy for melanoma results in low response and must be reinforced with sensitizer compounds. We believed that azelaic acid (AZA) could modulate melanomas’resistance to antineoplastics. Therefore we tried to compare in vitro treatment with antineoplastics alone versus AZA treatment followed by antineoplastics. We carried out MTT assays to evaluate the cytotoxicity of melphalan, lomustine (CCNU), fotemustine, and 4-Hydroxyanisole (4-HA) on three melanoma lines (B16F10, SK-MEL-28, and SK-MEL-1), and the modulating effect of pretreatment with AZA (1 mM). AZA showed a dose-dependent antineoplastic activity on the three lines. Melphalan was the most active drug followed by CCNU, fotemustine, and 4-HA. The most sensitive line was B16F10 and the least sensitive was SK-MEL-1. Previous treatment with AZA of B16F10 reinforced the effect of melphalan (2.5 times), CCNU (10 times), and fotemustine (14 times); whereas for SK-MEL-28 and SK-MEL-1, only the cytotoxicity of CCNU and fotemustine increased. An antagonist effect was produced by 4-HA on all three lines. We concluded that AZA enhances in vitro cytotoxicity of CCNU and fotemustine.     

Phosphopeptides designed for 5-methylcytosine recognition

An artificial phosphopeptide has been developed through rational design of the interaction with 5-methylcytosine in duplex DNA. The peptide consists of two tandem zinc finger motifs, in one of which the glutamate was replaced with a phosphotyrosine, the phosphotyrosine in the peptide being effective for methylcytosine selectivity of DNA binding. The flexible modulation of the target methylated sequence by rearrangement of zinc finger peptides is possible, and the phosphopeptide provided us an important hint for expansion of the codes for the interactions of zinc fingers with DNA to methylated DNA sequences. The fluorescence-labeled phosphopeptide provided information on the methylation status of genomic DNA through fluorescence anisotropy after a 10 min incubation.     

Sister chromatid exchange and proliferation pattern after ultrasound exposure in vivo.

The sister chromatid exchange (SCE) frequency and cell-cycle specific metaphase patterns were investigated in PHA-stimulated lymphocytes in vitro after in vivo exposure of 15 patients to diagnostic ultrasound. No significant differences were found in SCE rates before and after ultrasound exposure. The evaluation of cell-cycle specific metaphase patterns, differentiating cells which passed one, two, or more cell cycles, gave no evidence for a cell-growth inhibitory effect. Since SCE is regarded as a highly sensitive method for the detection of mutagenic effects, our data confirm previous reports on the genetic safety of diagnostic ultrasound.     

Frequency of mutagen-induced sister chromatid exchanges in the course of 3 cell cycles

The induction of SCE by fotrine (0.125 and 0.250 microgram/ml) and thiophosphamide (5 micrograms/ml) during the first three cell cycles was studied in the Chinese hamster cells. No increase in the SCE number was observed after treatment with thiophosphamide and fotrine at the G2 stage (the first stage from the moment of fixation) as compared with the control variants. The maximal sensitivity of the cells to the SCE induction by the mutagens is marked at the G1 stage of the first cell cycle before the moment of fixation. The level of SCE remains approximately the same in the second cell cycle before the moment of fixation (20-32 h) and decreased down to the control level at the G1 stage of the third cell cycle (48-52 h).     

Prolonged Treatment with DNMT Inhibitors Induces Distinct Effects in Promoters and Gene-Bodies

Treatment with the demethylating drugs 5-azacytidine (AZA) and decitabine (DAC) is now recognised as an effective therapy for patients with Myelodysplastic Syndromes (MDS), a range of disorders arising in clones of hematopoietic progenitor cells. A variety of cell models have been used to study the effect of these drugs on the methylation of promoter regions of tumour suppressor genes, with recent efforts focusing on the ability of these drugs to inhibit DNA methylation at low doses. However, it is still not clear how nano-molar drug treatment exerts its effects on the methylome. In this study, we have characterised changes in DNA methylation caused by prolonged low-dose treatment in a leukemic cell model (SKM-1), and present a genome-wide analysis of the effects of AZA and DAC. At nano-molar dosages, a one-month continuous treatment halved the total number of hypermethylated probes in leukemic cells and our analysis identified 803 candidate regions with significant demethylation after treatment. Demethylated regions were enriched in promoter sequences whereas gene-body CGIs were more resistant to the demethylation process. CGI methylation in promoters was strongly correlated with gene expression but this correlation was lost after treatment. Our results indicate that CGI demethylation occurs preferentially at promoters, but that it is not generally sufficient to modify expression patterns, and emphasises the roles of other means of maintaining cell state.     

Modeling of induction and elimination of sister chromatid exchanges in time

Induction and elimination of sister chromatid exchanges (SCE) have been simulated during several cell cycles. Two models of SCE elimination are suggested. The first model postulates that the mutagen-induced lesions are not repaired, a lesion being only inherited by one daughter cell after DNA synthesis. According to the second model, lesions are completely repaired at the first S-phase. No SCE induction takes place during next cell cycles. SCE frequency ranges for both models are described by an equation, including the probability distribution function. The best correspondence in experimental and theoretical results was obtained using the model claiming repair of lesions during one cell cycle.