Hydroxy-Sensitive Cut Counting (HSCC); simultaneous, genome-wide mapping of 5-methylcytosine and 5-hydroxymethylcytosine in mammals

人类基因组包含了创造人类所必需的所有指令（基因），从一个受精卵细胞开始。虽然所有细胞都有相同的基因组序列，但人体内有数百种明显不同的细胞类型，如脑细胞（神经元）、肾细胞和肝细胞。每种细胞类型必须“打开”不同的基因子集，以使每种组织类型正常运作。细胞标记基因开启或关闭的主要机制被称为“表观遗传学”。表观遗传学是指基因表达的可遗传变化，这些变化不是由潜在DNA序列的突变引起的。哺乳动物中研究最全面的表观遗传标记是DNA甲基化，其涉及将称为“甲基基团”的标签分子连接到胞嘧啶以产生5-甲基胞嘧啶（5 mC）。它通常发生在CpG二核苷酸背景下，尽管非CpG甲基化发生在胚胎干细胞中，但在哺乳动物中所有CpG的60-90%之间是甲基化的。DNA甲基化与基因沉默有关，对正常发育至关重要。基因组印记、X染色体失活、重复元件抑制、基因表达稳定性和致癌等关键过程依赖于或涉及DNA甲基化模式的动态变化。从这个角度来看，了解修饰的DNA在基因组中的位置非常重要。我们对哺乳动物DNA甲基化的大部分了解都是基于一套可以区分5 mC和未甲基化的胞嘧啶（C）的技术。这些方法包括使用甲基敏感性限制酶（可以切割未甲基化的DNA，但不能切割甲基化的DNA）和亚硫酸氢盐测序，这种方法可以同时精确定量几个相邻胞嘧啶的甲基化水平。这种分析被一种新型修饰DNA的鉴定所混淆，5-羟甲基胞嘧啶（5 hmC）在哺乳动物组织中以高水平存在。事实上，在小鼠大脑样本中，5 hmC是5 mC的40%。尽管在过去的30年里对DNA甲基化进行了深入的研究，但哺乳动物组织中5 hmC的存在一直被遗漏，特别是因为大多数用于鉴定甲基化的技术不能区分5 mC和5 hmC。我们提出了一种激进的新技术，称为羟基敏感切割计数（HSCC），同时分析5 mC，5 hmC和C在150 - 230万CCGG序列存在于整个小鼠和人类基因组分别。HSCC是基于这样的观察：如果内部C是5-羟甲基胞嘧啶，限制性内切酶MspI可以切割其靶位点CCGG，但如果是β-葡糖基-5-羟甲基胞嘧啶（ghmC）则不能。使用充分表征和市售的酶T4噬菌体β-葡糖基转移酶，我们将基因组中的所有5 hmC转化为ghmC，并在用MspI处理之前和之后消化样品。然后使用下一代测序法对每个MspI位点周围的序列进行测序。每个MspI位点的处理与未处理样品中的序列数是存在的5羟甲基胞嘧啶量的量度，允许全基因组5 hmC水平的半定量。所得到的5 hmC曲线将代表组织“标识符”;将是正常组织状态的读出的条形码。HSCC代表了对现有的“全基因组”技术的巨大改进，该技术使用抗体来收获含有5 hmC的DNA。这种“亲和”技术只能检测含有5 hmC的基因组区域，并且通常表现出强烈的序列背景偏倚。相反，HSCC将检测MspI位点，而不管5 hmC水平如何。这是至关重要的，因为如果我们要了解这个令人兴奋的新标记在哺乳动物生物学中的作用，那么知道5 hmC在哪里耗尽与知道它在哪里富集一样重要。

The human genome contains all the instructions (genes) necessary to create a human being, beginning from a single fertilised egg cell. Although the same genome sequence is present in all cells, there are hundreds of markedly different cell types, such as brain (neurons), kidney, and liver cells, in the human body. Each cell type must 'turn on' a different subset of genes for the correct functioning of each tissue type. A primary mechanism by which cells mark genes to be turned on or off is called, 'epigenetics'. Epigenetics refers to heritable changes in gene expression that are not caused by mutations in the underlying DNA sequence. The most comprehensively studied epigenetic mark in mammals is DNA methylation which involves the attachment of a tag-molecule called a 'methyl group' to cytosines to give 5-methylcytosine (5mC). It typically occurs in a CpG dinucleotide context, although non-CpG methylation occurs in embryonic stem cells, between 60-90% of all CpGs are methylated in mammals. DNA methylation is associated with gene silencing and is essential for normal development. Key processes including genomic imprinting, X-chromosome inactivation, suppression of repetitive elements, stability of gene expression and carcinogenesis either depend or involve dynamic changes in DNA methylation patterns. From this perspective it is important to know where modified DNA resides in the genome.Much of what we know about DNA methylation in mammals is based on a set of techniques which can distinguish between 5mC and an unmethylated cytosine (C). These include the use of methyl-sensitive restriction enzymes (which can cut unmethylated, but not methylated DNA) and bisulfite sequencing, a method that allows accurate quantification of methylation levels at several neighbouring cytosines simultaneously. This analysis has been confounded by the identification of a new type of modified DNA, 5-hydroxymethylcytosine (5hmC) that is present at high levels in mammalian tissues. Indeed, 5hmC is 40% as common as 5mC in mouse brain samples. Despite intense study of DNA methylation for the last 30 years, the presence of 5hmC in mammalian tissues had been missed especially as most the techniques used to identify methylation, do not differentiate between 5mC and 5hmC. We propose a radical, new technique termed Hydroxy Sensitive Cut Counting (HSCC) to simultaneously analyse for 5mC, 5hmC and C at the 1.5 -2.3 million CCGG sequences present throughout the mouse and human genomes respectively. HSCC is based on the observation that the restriction enzyme MspI, can cut its target site, CCGG, if the internal C is 5hydroxymethylcytosine, but not if it is beta-glucosyl-5-hydroxymethylcytosine (ghmC). Using a well characterised and commercially available enzyme T4 Phage beta-glucosyltransferase, we will convert all the 5hmC in the genome to ghmC, and digest the sample before and after treatment with MspI. The sequences surrounding each MspI site will then be sequenced using next generation sequencing. The number of sequences in the treated versus untreated samples for each MspI site is a measure of the amount of 5hydroxymethylcytosine present, allowing for semi-quantification of genome-wide 5hmC levels. The resulting 5hmC profiles will represent tissue 'identifier'; a barcode that will be a read-out of a normal tissue state. HSCC represents a dramatic improvement on the existing 'genome-wide' techniques that use antibodies to harvest DNA containing 5hmC. Such 'affinity' techniques can only assay regions of the genome which contain 5hmC and often exhibit strong sequence-context biases. In contrast, HSCC will assay MspI sites regardless of 5hmC level. This is vital, as knowing where 5hmC is depleted is as important as knowing where it is enriched if we are to understand the role of this exciting new mark in mammalian biology.