The role of natural antisense transcripts in plant gene regulation

在高等生物中，负责蛋白质生产的遗传信息被转录成细胞核中的RNA分子，然后被加工成信使RNA并输出到细胞质中翻译成氨基酸序列。编码蛋白质的基因只占大多数生物体基因组的一小部分。因此，理论上，染色体上应该有足够的空间让单个基因彼此分离，以确保转录过程不会相互干扰。当我们分析模式植物拟南芥中蛋白质编码基因的分布时，我们惊奇地发现，相当数量的基因（总共956个）与其他基因部分重叠，以正反义排列。我们将这些基因座命名为聚合重叠基因对（cop）。这与历史上的观点形成了对比，即重叠基因的趋同排列是有害的，并且在进化过程中被优先移除。当我们检查来自大约1,400个个体实验的表达数据时，我们发现956个cop的正义转录本和重叠反义转录本通常存在于相同的组织类型中。这意味着这两个转录本可以排列形成双链RNA （dsRNA）分子。我们知道dsRNA分子可以被一种降解机制识别，这种降解机制被称为RNA干扰（RNAi），它会导致两个转录本的破坏。然而，我们最近发表的对正义和反义转录本分布的分析表明，没有迹象表明cop转录本有特定的降解。我们从动物系统中了解到，反义转录物除了为RNAi系统形成dsrna之外，还可以发挥其他作用。然而，在植物中，没有这样的替代途径被描述。因此，我们想知道，植物有哪些替代途径，哪些反义转录本是必需的。我们对956个cop的组织和表达的分析帮助我们做出了一个有根据的猜测，假设反义转录本的两种替代功能。(i)在cop中，我们发现具有一个以上转录本的基因比例远高于预期，这要么是因为原始RNA拷贝上的不同区域被移除（备选剪接变体），要么是因为转录本长度不同（备选聚腺苷酸化变体）。这意味着需要反义转录物来修饰义转录物的加工。（ii）我们还经常发现，虽然正义基因和反义基因都被转录，但它们的行为往往是拮抗的：在许多组织类型中，一种转录类型高，另一种转录类型低。对于特定应激处理会增加sense转录物水平的cop，反义转录物水平会降低。然而，很快，反义转录水平增加，而义转录水平降低。类似的拮抗模式可以在白天周期性改变转录水平的COPs基因（生物钟基因）中看到。这些观察结果表明了一种“转录干扰”模型，其中一个基因的转录影响另一个基因的转录，这可能用于在诱导后重置转录水平。我们想测试反义转录本是否以及如何影响义转录本的产生或组成。为此，我们选择了个体cop作为我们提出的两种机制的潜在模式基因。我们使用拟南芥，因为对于我们的每个模型cop，都有一个突变系，插入了反义基因。这些细胞系允许我们测定如果反义基因失活，是否以及如何改变义转录或转录过程。我们认为我们的工作将有助于扩大目前对基因表达的认识，这些认识主要集中在转录和转录稳定性上，而没有考虑反义转录物的调控作用。

In higher organisms, genetic information that is responsible for the production of a protein, is transcribed into an RNA molecule in the nucleus, which is then processed into a messenger RNA and exported into the cytoplasm for translation into an amino acid sequence. Protein-encoding genes only constitute a minority of the genomes of most organisms. In theory, there should therefore be sufficient space for individual genes to be separated from each other on the chromosomes, to ensure that transcription processes do not interfere with each other. When we analysed the distribution of protein-encoding genes in the model plant Arabidopsis thaliana, we found, to our surprise, that a considerable number of genes (a total of 956) partially overlap with other genes, in a sense-antisense arrangement. We named these loci Convergently Overlapping Gene pairs (COPs). This contrasts historical arguments that the convergent arrangement of overlapping genes is detrimental and has been preferentially removed during evolution. When we examined expression data from about 1,400 individual experiments, we found that both the sense and the overlapping antisense transcripts of the 956 COPs are often present in the same tissue types. This implies that the two transcripts could align to form double-stranded RNA (dsRNA) molecules. We know that dsRNA molecules can be recognised by a degradation mechanism, termed RNA interference (RNAi), which causes the destruction of both transcripts. Our analysis of the distribution of sense and antisense transcript, which we have recently published, shows, however, no indication for a specific degradation of COPs transcript. We know from animal systems that antisense transcripts can fulfil alternative roles than forming dsRNAs for the RNAi system. In plants, however, no such alternative pathways have been described. We therefore would like to find out, which alternative pathways plants have, for which antisense transcripts are required. Our analysis of the organisation and expression of the 956 COPs helped us to make an educated guess to postulate two alternative functions for antisense transcripts. (i) Among COPs, we find a much higher than expected proportion of genes that have more than one transcript, either because different regions are removed from the original RNA copy (alternative splice variants) or because the transcripts differ in length (alternative polyadenylation variants). This implies that antisense transcripts are required to modify sense transcript processing. (ii) We also frequently find that, while both the sense and antisense genes are transcribed, they often behave antagonistically: In many tissue types, one transcript type is high and the other low. For COPs where a particular stress treatment increases sense transcript levels, the antisense transcript level is reduced. Very soon, however, antisense transcript levels increase and sense transcript levels are reduced. A similar antagonistic pattern can be seen for COPs genes whose transcript levels change periodically during the day (circadian clock genes). These observations suggests a 'trancription interference' model, where transcription of one gene affects transcription of the other, which may be used to reset transcript levels after induction. We want to test if and how antisense transcripts can influence the production or composition of sense transcripts. For this we have selected individual COPs as potential model genes for each of the two mechanisms that we propose. We use Arabidopsis, because for each of our model COPs, there is a mutant line, with an insertion in the antisense gene. These lines allow us to assay if and how sense transcription or transcript processing is altered if the antisense gene in inactive. We think that our work will help to widen the current perception of gene expression, which focuses on transcription and transcript stability, but does not consider the regulatory effects from antisense transcripts.