石原 悟

Abstract CYP19A1 encodes aromatase, which converts testosterone to estrogen, and is induced during placental maturation. To elucidate the molecular mechanism underlying this function, histone methylation was analyzed using the placental cytotrophoblast cell line, JEG3. Treatment of JEG3 cells with 3-deazaneplanocin A, an inhibitor of several methyltransferases, resulted in increased CYP19A1 expression, accompanied by removal of the repressive mark H3K27me3 from the CYP19A1 promoter. However, this increase was not observed in cells treated with GSK126, another specific inhibitor for H3K27me3 methylation. Expression of TFAP2C, which encodes AP-2γ, a transcription factor that regulates CYP19A1, was also elevated on 3-deazaneplanocin A treatment. Interestingly, TFAP2C messenger RNA (mRNA) was readily degraded in JEG3 cells but protected from degradation in the presence of 3-deazaneplanocin A. TFAP2C mRNA contained N6-methyladenosines, which were reduced on drug treatment. These observations indicate that the TFAP2C mRNA undergoes adenosine methylation and rapid degradation, whereas 3-deazaneplanocin A suppresses methylation, resulting in an increase in AP-2γ levels. We conclude that the increase in AP-2γ expression via stabilization of the TFAP2C mRNA is likely to underlie the increased CYP19A1 expression.

Human CYP19 spans a region of chromosome 15 of approximately 130 kb and encodes aromatase, an enzyme required for estrogen synthesis. In the human granulosa cell-line KGN, there are seven open chromatin regions within the CYP19 locus. In this study, we demonstrate that two of these regions ~40 kb upstream and ~15 kb downstream of the CYP19 promoter are cohesin-loading sites, physically interacting with the promoter to negatively and positively regulate transcription, respectively. These observations suggest that CYP19 expression is controlled by a balance between the upstream silencer and downstream enhancer. When cohesin is depleted, CYP19 expression is elevated since the silencer is 2.5-fold further from the promoter than the enhancer and most likely depends on cohesin-mediated tethering to influence expression.

<title>Abstract</title> The ‘open’ and ‘compact’ regions of chromatin are considered to be regions of active and silent transcription, respectively. However, individual genes produce transcripts at different levels, suggesting that transcription output does not depend on the simple open-compact conversion of chromatin, but on structural variations in chromatin itself, which so far have remained elusive. In this study, weakly crosslinked chromatin was subjected to sedimentation velocity centrifugation, which fractionated the chromatin according to its degree of compaction. Open chromatin remained in upper fractions, while compact chromatin sedimented to lower fractions depending on the level of nucleosome assembly. Although nucleosomes were evenly detected in all fractions, histone H1 was more highly enriched in the lower fractions. H1 was found to self-associate and crosslinked to histone H3, suggesting that H1 bound to H3 interacts with another H1 in an adjacent nucleosome to form compact chromatin. Genome-wide analyses revealed that nearly the entire genome consists of compact chromatin without differences in compaction between repeat and non-repeat sequences; however, active transcription start sites (TSSs) were rarely found in compact chromatin. Considering the inverse correlation between chromatin compaction and RNA polymerase binding at TSSs, it appears that local states of chromatin compaction determine transcription levels.

Ligation-mediated polymerase chain reaction (LM-PCR) is a common technique for amplification of a pool of DNA fragments. Here, a double-stranded oligonucleotide consisting of two primer sequences in back-to-back orientation was designed as an adapter for LM-PCR. When DNA fragments were ligated with this adapter, the fragments were sandwiched between two adapters in random orientations. In the ensuing PCR, ligation products linked at each end to an opposite side of the adapter, i.e. to a distinct primer sequence, were preferentially amplified compared with products linked at each end to an identical primer sequence. The use of this adapter in LM-PCR reduced the impairment of PCR by substrate DNA with a high GC content, compared with the use of traditional LM-PCR adapters. This result suggested that our method has the potential to contribute to reduction of the amplification bias that is caused by an intrinsic property of the sequence context in substrate DNA. A DNA preparation obtained from a chromatin immunoprecipitation assay using pulldown of a specific form of histone H3 was successfully amplified using the modified LM-PCR, and the amplified products could be used as probes in a fluorescence in situ hybridization analysis. (C) 2017 Elsevier Inc. All rights reserved.