I. Colorful History Nowadays, epigenetics is in full swing, and DNA methylation and demethylation are actually a few waves of a tributary in the long river of epigenetics. DNA methylation is an early discovered and recognized epigenetic modification, which is the addition of a methyl group (-CH3) to the 5th carbon atom of the pyrimidine ring of the base cytidine (C). This modification does not occur on all Cs of the DNA sequence, but mainly on Cs that appear together with CG (it was later discovered: in embryonic stem cells, many methylation occurs on other Cs as well). Don’t look at this small modification, but it adds extra information to the DNA, making the genetic information of the limited genome show a rich diversity of expression. The question arises, what exactly is this extra information? First of all, DNA methylation is simply understood as a “lock”, and all the parts marked by DNA methylation are mostly genes that need to be “dusted off” and “imprisoned”, such as The genome’s “mischief makers”, transposons, are controlled by the “lock” of methylation. These “mischief makers” will jump around in the genome and mess up the genome, causing many problems, such as tumor development. DNA methylation is most consistent with the definition of epigenetic inheritance; DNA methylation is very stable and heritable at the somatic level, which means that the “daughter” cells formed by somatic cell division not only inherit the genomic DNA sequence of the “mother” cells, but also make very faithful copies of the “mother” cells. In other words, the “daughter” cells formed by somatic cell division not only inherit the genomic DNA sequence of the “mother” cell, but also copy the methylation pattern of the “mother” cell genomic DNA very faithfully. However, during germ cell formation and the differentiation of the fertilized egg into the embryo, the methylation of genomic DNA undergoes a massive “remodeling” with a very short time window. For example, the fertilized egg apparently undergoes a “cleaning” of the genomic DNA methylation of the sperm before dividing, and then begins a rapid rebuilding soon after the fertilized egg divides. In addition, methylation remodeling occurs during the formation of germ cells (sperm and eggs). This raises an important fundamental question: How is DNA methylation “cleaned” in such a short period of time? The key to answering this question is to find a DNA demethylase that can remove the methylation groups from DNA. However, this problem is easy to propose, but not easy to solve. From the beginning, it took many people more than a decade of work, but it has been pending! The reason is that the process of DNA demethylation is limited to the post-fertilization period of the egg before division and the germ cell generation process, with a very short time window and limited cell sources, and it is impossible to obtain sufficient number of cells or tissues for molecular biology operations to identify the gene, let alone biochemical operations to purify the enzyme. This DNA demethylation process has been a great mystery for many years. The question is important, but “the dog bites the hedgehog, and there is no way to get to the bottom of it”. There was an interlude. A Canadian biologist of Jewish origin, Professor Szyf, made a big joke about DNA methylation, but he went off the beaten track and ended up as an alternative scientist, a marginal figure in the field. His lab discovered a DNA demethylase, which was published in Nature in 1999 and caused a sensation in the circle, but it could not be duplicated by other labs, and Szyf gave detailed conditions in JBC, but later no one reported that it could be duplicated. The study of DNA demethylation was silent for several years from then on. However, there are still some visionary scientists who still think about this problem and do not stop exploring. 2005, Shi Yang’s lab discovered histone demethylase and catalytic mechanism, which caused a sensation and encouraged people to search and identify DNA demethylase with confidence. Shang Yongfeng’s lab at Peking University reported gene-specific “active DNA demethylation” in ovarian cancer cells, and Zhu Health’s lab in the US discovered and identified DNA demethylases in plants. However, mammalian DNA methylesterases are still “unidentified” and “unknown”. In 2008, two papers published in Nature by German laboratories and French laboratories reported localized active DNA demethylation in carrier cells. At this point, Bestor at Columbia University, a subleader in DNA methylation, came out and commented on the matter in Cell with a sarcastic tone (The colorful history of active DNA demethylation). Earlier, developmental biologist Niehrs of Heidelberg reported in Nature that a DNA damage-inducible protein, Gadd45a, promotes DNA demethylation, but immediately drew a tit-for-tat rebuttal. The Pfeifer lab at the Beckman Institute in California published an article in PLoS Genet denying Niehrs’ results with the tit-for-tat title: GADD45A does not promote DNA demethylation! However, several subsequent reports from different labs confirmed Niehrs’ results. Nevertheless, there are still some problems: 1) Gadd4a knockout does not hinder embryonic development, nor does it affect the demethylation process before fertilization, 2) CADD4a is not really a demethylation enzyme, but only initiates a DNA repair mechanism called Nuclear acid excision repair (NER), which is mainly used to repair DNA damage caused by UV light. NER, which is mainly used to repair DNA damage caused by UV light), cuts up a section of methylated DNA and resynthesizes the new strand. It is understandable that this mechanism plays a role in local or individual gene-specific demethylation, but I am afraid that it is unacceptable to be involved in genome-wide reprogramming of methylation as a universal mechanism. To use an analogy, if you want to change the wall color, would you rather take down the bricks on the wall and replace them with new ones, or just remove the paint paint from the bricks? Especially for a large area of the whole building to change? After all, it’s a matter of efficiency on the one hand and economy on the other. Finally, in 2009-2010, DNA demethylase took center stage and became a shining new star. The discovery and identification of DNA demethylases represent two different research models (two research models, two and a half mechanisms, seven molecules). Research model 1: Hypothesis driven This is the classical scientific research model that requires first formulating a hypothesis and then designing an experiment to test it. Hypothesis is often wrong and therefore a high-risk strategy for us in general. Success in this strategy requires not only intuition, imagination and courage, but also experience and a solid foundation, and of course, logical reasoning skills are essential. Appreciating high-level scientific research can sometimes be as intellectually enjoyable as reading a Sherlock Holmes mystery. In the case of DNA demethylation research, we can appreciate two hypothesis driven exemplars. First up, Anjana Rao, formerly an immunologist in the Department of Pathology at Harvard University, now at the La Jolla Institute for Metabolic Research in California, was elected to the American Academy of Sciences in 2008. She was not originally from the epigenetics circle, but everyone is everyone, and when they strike, they strike big, just like Shi Yang, who was not specialized in histone methylation modification, but was inspired to break into the epigenetics territory and still made an unexpected and important discovery. Rao’s idea that DNA demethylation might be by a mechanism of oxygenation (hydroxylation or addition of monooxygenation) may have been inspired by the study of a model organism, the African trypanosome. There is an enzyme in trypanosomes, JBP1, which adds an oxygen (hydroxylation) to the methyl group of thymine to form a hydroxyl group, and then other enzymes attach a sugar group to this hydroxyl group, thus forming a unique structure (named J-structure) that plays a role in the regulation of gene expression, although the original methyl group is not removed. Thymine and methylcytosine have very similar structures (both are pyrimidine rings) with the same position of the methyl group, and Rao hypothesized that similar enzymes and mechanisms may exist in higher organisms. Based on this hypothesis, they performed multiple rounds of homology searches (PSI-blast) through protein sequences and finally found three genes, Tet1, 2, and 3, in both human and murine protein databases. Moreover, they found that these genes are prevalent in the postnatal phylum, suggesting the importance and conserved nature of this functional mechanism. Through cytological experiments and biochemical analyses, they obtained two important findings: 1. they confirmed that all three proteins have oxygenated modifications of methylated cytosine, but with different intensities of activity, which should be the expected result of their plan; 2. the methyl cytosine was catalyzed by the oxygenation reaction to form a new modified form, hydroxyl-methyl-cytidine (hm-C). cytidine (hm-C), a modification that imparts new epigenetic properties and information to DNA; this second discovery was unexpectedly unplanned and may have even more profound implications. Immediately afterwards, they rapidly pushed their work further, linking this modification to hematopoietic cell differentiation (Science 2009) and then also finding that its abnormalities (Tet2 mutations, deletions) were strongly associated with hematologic tumorigenesis (Nature 2010). A star figure in epigenetics, Yi Zhang, also quickly stepped in at this time and discovered that Tet1-mediated hydroxymethyl cytosine modification plays an important function in maintaining the self-renewal (renewal) function of stem cells (Nature 2010). In 2011, DNA hydroxymethylation modification work became a regular feature of Nature (Nature 18 Jul 2010; Nature 30 Mar 2011; Nature 03 Apr 2011; Nature, 13 April 2011, Cell 14 April 2011). Old Dog, New Tricks This is also a typical Hypothesis driven finding. Their theoretical model is that uracil, formed by the deamination of methylcytosine, can only occur in RNA, and if it occurs in DNA, it is either considered by the cell to be synthesized with the wrong ingredients, or it is considered to be chemically damaged by the base, in short, the cell will trigger a base excision repair (base excision repair). reaction (base excision repair, BER), which has similarities to NER. At this point, a molecule, AID (base cytosine deaminase), which was discovered in the early 1990s to play an important role in the ability of the immune system to develop, came into the view of those exploring DNA demethylases. AID was originally cloned in the laboratory of Honjo T, a leading Japanese immunologist, and the cloning and functional elucidation of AID moved significantly forward for B-cell development and antibody diversity. Such an old molecule was found to have new and important functions. the system used by M. Azim Surani and Wolf Reik this time was primitive germ cells. In AID knockout cells, both labs found that sperm demethylation was significantly affected, but still some of it occurred, suggesting the existence of other mechanisms. Finally, they both confirmed that the demethylation involved in AID is accomplished by initiating base excision repair through deamination of methylated cytosines. It is like pulling out certain parts of the whole building in many parts and replacing them with new ones, not economical enough, but it does happen and we can only trust the facts rather than experience and common sense. Research Model 2: Systematic Filtering What if there are no clues, no clues, no ideas (hypotheses)? Cast a net and fish for them systematically. Stephen J. Elledge made his fortune by building a powerful RNAi screening system, a very large and powerful “net”. He built this super fishing system like an oversized fishing boat, equipped with an oversized net (to catch all genes) and automated operation (barcode system + gene chip), and became the king of the sea in functional genome research. In 2009, Yi Zhang’s lab in North Carolina designed an ingenious experiment to screen with fertilized egg cells and successfully opened a small gap. First they devised a method to detect the methylation status of genomic DNA in living cells. There is a class of proteins in cells that specifically recognize and bind methylated DNA (methyl DNA binding protein, MBD), and if MBD is linked (fused) to a green fluorescent protein, it is equivalent to adding a fluorescent tag to MBD, allowing the location of MBD to be directly observed and tracked under a microscope. Generally, MBD-GFP is bound to methylated DNA on the genome and appears as a green spot (foci) under the fluorescence microscope; when genomic methylation is removed, MBD-GFP has nowhere to bind, so it is dispersed in the nucleus, and under the microscope, those original green fluorescent spots disappear and become a diffuse green color. In this way, the change of genomic methylation can be judged by observing the change of fluorescence signal. With this judgment indicator it is possible to screen. The selected genes are “removed” one by one by RNA interference technology, and then the fluorescence signal is observed. If the fluorescence signal in the fertilized egg cells does not disappear or disappears very slowly after a certain gene is interfered with, it suggests that this gene may be involved in the genomic demethylation process. The screening was very laborious, they screened 4000 genes, and finally they got a positive candidate, which turned out to be a transcriptional elongation factor ELP3. the exact mechanism is not clear. Recently, Song Hongjun, a Chinese scientist at Johns Hopkins University School of Medicine, published a research paper in the latest issue of Cell that unifies the Tet1-catalyzed hydroxymethylation reaction with the AID-catalyzed deamination reaction. The once iron plate of DNA demethylation has finally been pried open; indeed, there is a lot of gold and silver underneath, don’t you want to squeeze in and grab some?