Nucleosome positioning as a determinant of chromatin structure

真核生物基因组中只有很小一部分编码蛋白质或直接参与调节蛋白质的合成。基因组其余部分的功能在很大程度上仍然未知。现在已经确定，基因组在细胞核内包装或折叠成染色质结构的方式对调节或控制遗传活性做出了重大贡献。在细胞核内，染色质折叠的默认状态是30 nm的高阶染色质纤维，DNA包装成这种结构的详细方式是潜在DNA序列功能潜力的主要决定因素。我们之前的研究强烈表明，调节DNA如何折叠成高阶纤维的一个因素是DNA序列本身的长程组织。这是因为DNA序列影响了染色质的结构单元--核小体的相对位置。当这些核小体沿着DNA有规律地分布沿着时，所产生的核小体链可以折叠成非常紧凑、无活性的结构类型。间隔较不规则的核小体链的折叠效率较低，并且可能包含较小破坏的区域，这些区域可以提供对潜在序列的访问，因此具有功能重要性。我的研究方向是测量DNA序列中核小体倾向于结合的位点的位置。因此，我们测量了DNA序列本身的内在属性。我们对基因组中包含非常好的特征基因的区域进行了这种映射，以便我们可以将我们关于核小体定位信号的发现以及它们可能决定的染色质结构类型与相关背景联系起来。对于某些基因，我们还确定了核小体在细胞中相同DNA上的位置，并且我们发现体外和体内核小体定位图是相关的，证实了DNA序列参与确定染色质结构。有趣的是，大多数影响核小体定位的序列并不位于编码蛋白质的区域，而是位于中断和侧翼基因的序列中。体外和体内核小体定位之间的对应关系使我们考虑是否可以利用体外映射数据集来使用计算方法详细模拟染色质组织。我们已经发现，模拟的染色质结构是敏感的核小体密度，也许最重要的是，在一个适当的核小体密度的模拟染色质的“结构”，甚至更密切相关的体内定位状态比原来的体外数据。后一种观察结果强调了体外定位数据的价值，并有力地证明了这种计算方法的合理性。对于这个建议，我们打算通过研究三个完整的基因区域，人和鸡β-珠蛋白结构域和人H19/Igf 2印迹结构域来扩展我们的体外核小体作图。在这样做的过程中，我们希望证明，我们在基因水平上检测到的特征和关系也是明显的，也许甚至更清楚地确定，在一个染色质组织的功能相关的水平。我们还将增加我们的核小体定位序列的数据库的数量级。我们建议的一个主要重点是继续和扩大我们的计算研究，以便我们能够确定我们在体外测量的信息如何被翻译成细胞中的染色质结构。此外，如此大的定位序列数据库的可用性呈现了实际鉴定DNA序列的什么特征导致核小体定位的前景。如果我们知道这一点，我们就可以开发出能够根据染色质结构“读取”DNA序列的算法，这一目标无疑将扩大我们对基因组及其工作方式的理解。

Only a very small fraction of a eukaryotic genome codes for proteins or is directly involved in regulating their synthesis. The function of the remainder of the genome remains largely unknown. It is now well established that the way in which the genome is packaged or folded into chromatin structures within the nucleus of the cell makes a major contribution towards regulating or controlling genetic activity. Within the nucleus the default state of chromatin folding is the 30 nm, higher-order chromatin fibre and the detailed way in which the DNA is packaged into this structure is a major determinant of the functional potential of the underlying DNA sequence. Our previous research strongly suggests that one factor which regulates how DNA is folded into the higher-order fibre is the long-range organisation of the DNA sequence itself. This happens because the DNA sequence influences the relative locations of the building blocks of chromatin, the nucleosomes. When these are regularly spaced along the DNA the resulting chain of nucleosomes can be folded into a very compact, inactive type of structure. Chains of nucleosomes which are less regularly spaced are less efficiently folded and are likely to incorporate regions of minor disruption which could provide access to the underlying sequence and thus be of functional importance. My research is directed towards measuring the locations of sites in the DNA sequence where nucleosomes prefer to bind. Thus, we measure an intrinsic property of the DNA sequence itself. We carry out this mapping on regions of the genome that contain very well characterised genes so that we can relate our findings about nucleosome positioning signals, and the kinds of chromatin structures they might dictate, in a relevant context. For some genes, we have also determined the locations of nucleosomes on the same DNA in the cell and we find that the in vitro and in vivo nucleosome positioning maps are related, confirming the involvement of DNA sequence in determining chromatin structure. Interestingly, most of the sequences that influence nucleosome positioning do not lie in the regions that code for proteins but are found in the sequences which interrupt and flank the genes. The correspondence between in vitro and in vivo nucleosome positioning has led us to consider whether the in vitro mapping datasets can be exploited to simulate chromatin organisation in detail using computational approaches. We have found that simulated chromatin structures are sensitive to nucleosome density and, perhaps most significantly, that at an appropriate nucleosome density the simulated chromatin 'structures' are even more closely related to the in vivo positioning state than is the original in vitro data. This latter observation underlines the value of the in vitro positioning data and strongly justifies this type of computational approach. For this proposal we intend to extend our in vitro nucleosome mapping by studying three entire gene regions, the human and chicken beta-globin domains and the human H19/Igf2 imprinted domain. In doing this we hope to demonstrate that the features and relationships we detect at gene level are also apparent, and perhaps even more clearly identified, at a more functionally relevant level of a chromatin organisation. We will also increase our database of nucleosome positioning sequences by an order of magnitude. A major focus of our proposal is to continue and expand our computational studies so that we can identify how the information we measure in vitro is translated into chromatin structure in the cell. Furthermore, the availability of such a large database of positioning sequences presents the prospect of actually identifying what features of the DNA sequence cause nucleosome positioning. If we knew this, we could develop algorithms that would 'read' DNA sequence in terms of chromatin structure, a goal which would undoubtedly expand our understanding of the genome and how it works.