2018-12-17

Next-generation sequencing reveals two populations of damage-induced small RNAs at endogenous DNA double-strand breaks

题目：

新一代测序技术揭示了在内源DNA双链断裂中两种损伤诱导的RNA

发表刊物及时间：

Nucleic Acids Research, Volume 46, Issue 22, 14 December 2018, Pages 11869–11882, https://doi.org/10.1093/nar/gky1107

作者：

[Franziska Bonath](javascript:;) [Judit Domingo-Prim](javascript:;) [Marcel Tarbier](javascript:;)[Marc R Friedländer](javascript:;) [Neus Visa](javascript:;)

通讯作者及单位：

Marc R Friedländer

Science for Life Laboratory, Department of Molecular Biosciences, The Wenner-Gren Institute, Stockholm University, SE-106 91 Stockholm, Sweden

Neus Visa

Department of Molecular Biosciences, The Wenner-Gren Institute, Stockholm University, SE-106 91 Stockholm, Sweden

摘要：

Recent studies suggest that transcription takes place at DNA double-strand breaks (DSBs), that transcripts at DSBs are processed by Drosha and Dicer into damage-induced small RNAs (diRNAs), and that diRNAs are required for DNA repair. However, diRNAs have been mostly detected in reporter constructs or repetitive sequences, and their existence at endogenous loci has been questioned by recent reports. Using the ==homing endonuclease（自制的核酸内切酶）== I-PpoI, we have investigated diRNA production in genetically unperturbed human and mouse cells. I-PpoI is an ideal tool to clarify the requirements for diRNA production because it induces DSBs in different types of loci: the repetitive 28S locus, unique genes and intergenic loci. We show by extensive sequencing that the rDNA locus produces substantial levels of diRNAs, whereas unique genic and intergenic loci do not. Further characterization of diRNAs emerging from the 28S locus reveals the existence of two diRNA subtypes. Surprisingly, Drosha and its partner DGCR8 are dispensable for diRNA production and only one diRNAs subtype depends on Dicer processing. Furthermore, we provide evidence that diRNAs are incorporated into Argonaute. Our findings provide direct evidence for diRNA production at endogenous loci in mammalian cells and give insights into RNA processing at DSBs.

最近的研究表明，DNA双链断裂（DSBs）中会发生转录，DSB的转录本由Drosha和Dicer加工成损伤诱导的小RNA（diRNA），用于修复DNA。但是，最近diRNA大多多数在报告构建体或重复序列当中被检测到，并且最近的报道针对它们在内源基因座的存在性也提出了质疑。我们在遗传未受干扰的人和小鼠的细胞中利用自制的内切酶I-Ppol研究了diRNA的产生。I-Ppol 是一个阐明diRNA产生的理想工具，因为它在不同类型的基因座中诱导DSB：重复的28S基因座、独立的基因和基因间区域。我们用大量的测序展示了rDNA基因座会产生高水平的diRNA，然而独立的基因和基因间区则不会。通过进一步描绘28S基因座产生的diRNA揭示了两种diRNA亚型的存在。令 人惊奇的是，Drosha和它的伴侣DGCR8在diRNA的产生中是非必需的，只有一种diRNA亚型是依赖 于Dicer加工。而且，我们有证据表明diRNA被整合到Argonaute当中。我们的研究结果为哺乳动物细胞可以产生内源基因座的双链RNA提供了直接证据，并为双链断裂过程中的RNA 加工提供了见解。

==homing endonuclease（自制的核酸内切酶）==：自己表达纯化出来的核酸内切酶，而不是从商业化公司购买获得

图表选析

image.png

Figure 1. DiRNAs originate from endogenous DSBs in human cells. (A) Schematic representation of the 28S rDNA (upper part) and definition of areas around DSBs (lower part). The arrow indicates the location of the I-PpoI target site. The lower panel defines the position and polarity of the transcripts analysed. Upstream and downstream are relative to the I-PpoI cleavage site. Sense and antisense are according to the canonical transcription start of the rDNA locus (arrow). Sense transcripts have the same polarity as the 28S rRNA. (B) Small RNA reads from HeLa cells transfected with a plasmid that expressed I-PpoI analysed at different times after transfection, as indicated. The reads were mapped to the 28S rDNA locus and normalized to total collapsed 28S rRNA read counts. The grey vertical line in each plot indicates the position of the I-PpoI cleavage site in the 28S locus. The top (blue) and bottom (red) parts of the plot show the coverage in the sense and antisense strands, respectively. The reads in the sense strand (blue) are assumed to be predominantly degradation products of rRNA. (C) Quantification of diRNA reads in the 28S rDNA locus expressed as reads per million (RPM) of total collapsed reads. The plot includes only upstream, antisense reads. (D) Length distributions of antisense small RNAs derived from the 28S rDNA locus, upstream region, at different times after transfection with the I-PpoI expression plasmid. The black bars highlight reads of 21 and 22 nt. The percentage of the 21–22 nt population is displayed. Enrichment of the 21–22 nt fraction is significant at 16 h (P = 1.85 × 10−5) and 24 h (P = 1.77 × 10−11) in an exact binomial statistical test.

图1. DiRNA起源于人细胞中的内源性DSB。 （A）28S rDNA（上部）的示意图和并定义了DSB周围的区域（下部）。箭头表示I-PpoI靶向的位置。 下图定义了分析的转录本的位置和极性。上游和下游相对于I-PpoI切割位点。正向和方向是根据rDNA基因座的经典转录起始（箭头）。正向转录本具有与28S rRNA相同的极性。 （B）如所示，在转染后不同时间分析 表达I-PpoI的质粒转染的HeLa细胞的小RNA读数。将读数定位于28S rDNA基因座并标准化为完全折叠的28S rRNA读数计数。每个图中的灰色垂直线表示I-PpoI切割位点在28S基因座中的位置。图中的顶部（蓝色） 和底部（红色）部分分别显示有义链和反义链的覆盖范围。有义链（蓝 色）中的读数被认为主要是rRNA的降解产物。 （C）28S rDNA基因座中diRNA 读数的定量表示为每百万读数（RPM）的总折叠读数。该图仅包括上游反义读数。 （D）在用I-PpoI表达质粒转染后不同时间源自28S rDNA基因 座上游区域的反义小RNA的长度分布。黑条突出显示21和22 nt的读数。 显示21-22 nt种类的百分比。在精确的二项式统计检验中，在21小时（P = 1.85×10-5）和24小时（P = 1.77×10-11），21-22nt组分的富集是显着的。

image.png

Figure 2. DiRNAs originate from endogenous DSBs in murine cells. (A) Small RNA reads from mESC cells transfected with a plasmid that expressed I-PpoI analysed 36 h after transfection. The reads were mapped to the 28S rDNA locus and normalized to spike-ins. The grey vertical lines indicate the position of the I-PpoI cleavage site in the 28S locus. The top (blue) and bottom (red) part of the plot show the coverage in the sense and antisense strands, respectively. (B) Length distributions of antisense small RNAs derived from the 28S rDNA locus, upstream region, at 36 h after transfection with the I-PpoI expression plasmid. The black bars highlight reads of 21 and 22 nt. The percentage of the 21–22 nt population is displayed. (C) Quantification of diRNA reads in the 28S rDNA locus expressed as collapsed reads per million ==spike-ins==. The plot includes only upstream, antisense reads.

图2. 在鼠类细胞中， diRNAs 源于内源性 DSBs. (A)表达 I-PpoI 的质粒转染小鼠胚胎干细胞细胞 36h 后小 RNAs 读数的分析。读数被映射到 28S rDNA 位点， 并标准化为尖峰。 灰色垂直线表示 I-PpoI 裂解位点 在 28S 位点的位置。 上半部分（蓝色） 和底部（红色） 部分分别显示了正义和反义链的覆盖范围 (B)在 I-PpoI 表达质粒转染后 36h， 来自上游区域 28S rDNA 位点的反义小 RNA 的长度分布。黑色柱突出显示 21 和 22nt 的读数。显示了 21-22nt 分布的百分比。 (C)量化 28S rDNA 位点的 diRNA 读数， 表达为每百万个spike-in折叠读数。 只包括上游的反义读数。

==spike-ins==：没有特别翻译，往往是一段已知序列，在实验后根据这段序列的数量对其它序列进行定量校准，相当于外标的control

image.png

Figure 3. DiRNAs arise from repetitive, but not from non-repetitive genomic loci in mammalian cells. (A) Schematic definition of repetitive ribosomal (rDNA), genic and intergenic regions used in (B and C). I-PpoI sites in genic regions can be located in exons or introns. I-PpoI sites defined as intergenic do not overlap with any annotated gene. (B and C) Total read counts of collapsed reads mapped to the 28S rDNA locus or to genic and intergenic loci in HeLa cells (B) or mESCs (C). The genome coordinates of the loci included in the analysis are listed in Supplementary Tables S2 and S3. For an overview of underlying samples see Supplementary Table S1. Read counts per locus are listed in Supplementary Table S4. (D) The efficiency of I-PpoI cleavage at selected loci was quantified by qPCR using primers spanning the I-PpoI motif on genomic DNA and expressed as percentage of DSBs. The analysis was carried out 24 h (for Ryr2, Dab1 and SLCO5A1) or 36 h (for AAMDC, chr7 and 28S rDNA) after transfection. The ARPP locus is not cleaved by I-PpoI and was used for normalization.

图3. DiRNA来自哺乳动物细胞中的重复区域，而不是非重复的基因组位点。 （A）（B和C）中使用的重复核糖体（rDNA），基因和基因间区域的示意图。基因区域中的I-PpoI位点可位于外显子或内含子中。定义为基因间的I-PpoI位点不与任何注释基因重叠。 （B和C）在HeLa细胞（B）或mESC（C）中映射到28S rDNA基因座或基因和基因间基因座的折叠读数的总读数。包括在分析中的基因座的基因组坐标列于补充表S2和S3中。有关基础样本的概述，请参见补充表S1。每个基因座的读数计数列于补充表S4中。 （D）通过qPCR使用跨越基因组DNA上的I-PpoI基序的引物定量I-PpoI切割在所选基因座处的效率，并表示为DSB的百分比。在转染后24小时（对于Ryr2，Dab1和SLCO5A1）或36小时（对于AAMDC，chr7和28S rDNA）进行分析。 ARPP基因座未被I-PpoI切割并用于标准化。

image.png

Figure 6. Endogenous diRNAs are incorporated into Argonaute. (A) Coverage of collapsed small RNA reads mapping to the 28S rDNA normalized to ==reads per million== (RPM,每百万读数) in Input and Argonaute RIP samples from either mock transfected (uncut) or I-PpoI expressing cells (cut). Coverage from three (antibody A) or two (antibody B) biological replicates were averaged. The grey vertical lines indicate the I-PpoI cleavage site. (B) Argonaute incorporation rates for miRNAs (black), snoRNAs (blue), tRNAs (green) or reads mapping to the antisense (diRNAs, red) or sense strand of the 28S rDNA (yellow). The Argonaute incorporation rate is estimated by fold-change (RIP/Input) normalized to the median miRNA fold-change. The plot shows averages from three (antibody A) or two (antibody B) biological replicates. Grey bars indicate the standard deviations. (C) Length distributions of diRNAs produced at the 28S rDNA (upstream, antisense) in Input or Argonaute RIP samples, as indicated in the figure. The percentage of the 21–22 nt population in each sample is displayed.

图 6.内源性 diRNA 结合 Argonaute。（A）小 RNA 读数匹配到 28S rDNA 的的覆盖率， 其标准化为每百万读数作为 Input。 和来自 模拟转染（未切割）或 I-PpoI 表达细胞（切割） 的 Argonaute RIP 样品。将来自三个（抗体 A）或两个（抗体 B）生物学重复的覆 盖率取平均值。灰色垂直线表示 I-PpoI 切割位点。（B） miRNA （黑色）， snoRNA（蓝色）， tRNA（绿色） 或映射到 28S rDNA （黄色）的反义（diRNA，红色）或有义链的读数与 Argonaute 结 合率。通过标准化为中值 miRNA 倍数变化的倍数变化（RIP /输 入）估计 Argonaute 结合率。该图显示来自三个（抗体 A）或两 个（抗体 B）生物学重复的平均值。灰色条表示标准偏差。 （C） 如图所示，在 Input 或 Argonaute RIP 样品中 28S rDNA（上游， 反义）产生的 diRNA 的长度分布。 显示每个样本中 21-22nt 种群 的百分比。

==注意：此图的B和C标反了==

image.png

Figure 7. Model for the biogenesis of Dicer-dependent and Dicer-independent diRNAs at endogenous DSBs in the mammalian genome. See main text for details.

图7. 在哺乳动物基因组中内源 DSB的Dicer依赖和Dicer非依赖的diRNA的生物发生模型。详细见原文

We have revealed the existence of two different diRNA populations that result from the processing of dilncRNAs at repetitive, transcribed loci and we propose a model for their biogenesis. According to this model, the free DNA ends at the DSB recruit RNAPII, which results in the synthesis of dilncRNAs (Figure 7, box 1). We have been able to detect dilncRNAs extending from DSBs in both the 28S rDNA locus and the Ryr2 gene. The newly made dilncRNAs are able to anneal with already existing transcripts made at the same locus, and double-stranded RNAs are produced. These double-stranded RNAs can be processed by Dicer into diRNAs with the characteristic length of 21–22 nt (Figure 7, box 2).

我们已经揭示了两个不同的diRNA群体的存在，这些diRNA群体是由重复的转录基因座上的dilncRNA 的加工产生的，我们提出了它们的生物发生的模型。根据该模型，游离DNA终止于DSB招募RNAPII， 其导致dilncRNA的合成（图7，第一格）。我们已经能够检测到28S rDNA基因座和Ryr2基因中从 DSB延伸的dilncRNA。新生成的dilncRNA能够与已存在的同一个基因座产生的转录物结合，并产生 双链RNA。这些双链RNA可由Dicer加工成具有21-22nt特征长度的diRNA（图7，第二格）。 如其他人所示（15,51），DNA-RNA杂合体在DSBs形成，这意味着dilncRNAs也与模板DNA结合。此 外，我们的模型提出，一部分dilncRNAs可能被不同的RNA内切酶和RNA外切酶降解，这导致在Dicer 缺失细胞中观察到的具有大范围不同长度的diRNA-Is群体（图7，第三格）。因此，野生型细胞的 diRNA长度谱是diRNA-Ds和diRNA-Is的谱的组合。

翻译小组：

叶名琛、渠梦葳、王俊豪、邓俊玮、黄敬潼、黄子亮、常彦琪、李碧琪、陈凯星、 郑凌伶