The broad impact of environmental exposures on repetitive element expression in cellular biology

该项目的目的是确定重复元素(RE)在环境暴露的生物学结果中所起的作用。虽然众所周知，RES的表达会随着环境暴露的变化而变化，但对RES对细胞和生物体生物学的影响的机械性见解是一个尚未深入探索的研究领域。我们特别感兴趣的是研究RES通过与相邻基因形成融合转录本而改变相邻基因表达的程度。我们选择使用RNAseq来研究这个问题。第一个任务是开发一套生物信息学方法，允许对RE表达进行全基因组分析。在过去的一年里，这一直是该集团的主要关注点。我们发起了一个来自纽约西奈山的Eric Nestler博士的合作，因为他的团队之前已经证明，可卡因处理的小鼠的大脑中RES的表达发生了变化。此外，内斯特勒博士的团队之前已经生成了全基因组转录组RNA-SEQ数据，我们可以使用这些数据来测试我们的新生物信息学方法。 王博士使用软件tophat-Fusion研究了生物信息学管道的几个版本，该软件最初的开发目的是识别癌症中染色体易位导致的融合转录本。王博士对这一工具进行了调整和修改，以最大限度地满足我们对重复序列分析的需要。经过几个月的工作和采用非常严格的标准，管道获得的结果使我们能够识别数据集中表达的175K个独特的RE基因座，其中约1500个在接触可卡因后差异表达。此外，该管道包括允许识别融合RNAseq读取的工具，我们将其定义为在490个基因中包含连接到蛋白质编码外显子的RE的部分读取。我们使用实时荧光定量聚合酶链式反应证实，随机选择的13个基因(在490个基因中)检测到的融合RNAseq读数确实来自转录本，其中RE序列与来自蛋白质编码基因的外显子融合(这里称为RE-融合转录本)。我们用定量RT-PCR证实了其中一个基因，一种名为ARHGEF10的Rho鸟嘌呤交换因子，在可卡因暴露后差异表达。最值得注意的是，当表达重新融合转录本或相应的野生型转录本的工程基因在小鼠的大脑中分别异位过度表达时，只有重新融合转录本显著改变了可卡因的奖励行为。这一结果提供了证据，证明在大脑中表达重新融合转录本会在细胞内引起生物反应。一份报告这些发现的手稿最近已经提交出版。这些实验指出，RES在细胞对环境暴露的生物反应中具有作用，并提出了有多少其他环境因素可以将它们的影响传递给细胞的机制。 虽然这条tophat Fusion管道有效地识别了RE-Fusion转录本，但在年初就清楚地发现，有许多类型的RE-Fusion转录本没有用这种方法检测到。为了有效地消除假阳性，我们在管道中加入了一个步骤，要求通过正常的剪接事件将RE融合到非重复序列。然而，这失去了我们检测许多融合事件的能力，最明显的是那些发生在RE内并持续到侧翼非重复序列的转录事件。因此，为了检测细胞内所有可能类别的再融合事件，我们在过去的一年里花费了相当多的时间来开发生物信息学工作流程来实现这一结果。我们进行了各种尝试，试图简单地更改tophat-Fusion上的参数，但我们发现，使用该软件执行分析所需的计算能力超出了NIH的计算基础设施的能力。然后使用Bowtie设计了一个新的管道，在那里我们能够估计大约50%的哺乳动物基因能够表达RE融合转录本。然而，我们仍然发现，有些再融合事件是我们无法检测到的。因此，我们计划在明年与智利加多利察大学和塔尔卡大学(智利)的Gonzalo Riadi博士合作。Riadi博士是一名计算机科学家，在全基因组重复成分分析方面拥有专门的专业知识；他愿意与我们合作，开发必要的工具来检测细胞内所有的再融合转录本。我们将继续使用我们现有的生物信息学工作流程来评估再融合转录本在经过环境试剂处理的样本中的差异表达，并将纳入任何一旦可用而开发的新工具。

The purpose of this project is to determine the role that repetitive elements (REs) play in the biological outcome of environmental exposures. While it is known that the expression of REs changes in response to environmental exposures, mechanistic insights into the impact of REs on the biology of cells and organisms is an area of research that has not been explored in depth. We are specifically interested in studying the extent to which REs alter the expression of adjacent genes through the formation of fusion transcripts with those genes. We chose to use RNAseq to study this problem. The first task was to develop a set of bioinformatics methods that would allow for the genome-wide analysis of RE expression. This has been a major focus of the group over the past year. We initiated a collaboration Dr. Eric Nestler from Mount Sinai in New York since his group had previously demonstrated that the expression of REs is altered in the brains of mice treated with cocaine. Also, Dr. Nestler's group had previously generated genome-wide transcriptome RNA-seq data that we could use to test our new bioinformatics methods. Dr. Wang worked on several versions of a bioinformatics pipeline using the software TopHat-Fusion, which was initially developed to identify fusion transcripts resulting from chromosomal translocations in cancer. Dr. Wang has adapted and modified this tool in order to best suit our needs for the analysis of repetitive sequences. After several months of work and the adoption of very stringent criteria, the results obtained with the pipeline allowed us to identify 175K unique RE loci that were expressed in the data set, with about 1,500 of these being differentially expressed after cocaine exposure. Moreover, the pipeline included tools that allowed the identification of fusion RNAseq reads, which we defined as those containing part of a RE connected to a protein-coding exon, in 490 genes. We confirmed using real-time PCR that the fusion RNAseq reads detected for 13 randomly-selected genes (out of the 490) were indeed derived from transcripts where a RE sequence was fused with an exon from a protein coding gene (herein called a RE-fusion transcript). We confirmed that one of these genes, a rho guanine exchange factor called Arhgef10, was differentially expressed after cocaine exposure using quantitative RT-PCR. Most notably, when engineered genes expressing either the RE-fusion transcript or the corresponding wild-type transcript were individually ectopically overexpressed in the brain of mice, only the RE-fusion transcript significantly altered cocaine reward behavior. This result provided evidence that expression of a RE-fusion transcript within the brain causes a biological response within the cell. A manuscript reporting these findings has been recently submitted for publication. These experiments point out that REs have a role in the biological response of the cells to environmental exposures and suggests a mechanism of how many other environmental agents could impart their effects on the cell. While this TopHat Fusion pipeline worked effectively to identify RE-fusion transcripts, it became clear early in the year that there were many types of RE-fusion transcripts that were not being detected with this approach. To effectively eliminate false positives, we incorporated a step in the pipeline that required that the RE be fused to a non-repetitive sequence through a normal splicing event. However, this eliminated our ability to detect many fusion events, most notably those that arose from transcription that initiated within a RE and continued into the flanking non-repetitive sequence. Therefore, in an attempt to detect all possible categories of RE-fusion events within the cell, we spent considerable time over the past year developing bioinformatics workflows to achieve this result. We made various attempts on simply trying to change parameters on TopHat-Fusion but we found that the computational power required to perform the analysis using this software was beyond the capacity of the computing infrastructure at the NIH. A new pipeline was then devised using Bowtie, where we were able to estimate that about 50% of the mammalian genes are capable of expressing RE-fusion transcripts. Nevertheless, we are still finding that there are certain RE-fusion events that we are not able to detect. Therefore, over the next year we plan to collaborate with Dr Gonzalo Riadi from the Pontificia Universidad Catolica de Chile and the Universidad de Talca (Chile). Dr. Riadi is a computer scientist with specific expertise in the genome-wide analysis of repetitive elements; he is willing to work together with us to develop the tools necessary to detect all RE-fusion transcripts within a cell. We will continue to use our existing bioinformatics workflow for evaluating the differential expression of RE-fusion transcripts in samples that have been treated with environmental agents and will incorporate any new tools that are developed once they are available.