免费医学论文发表-转录组和小RNA组分析揭示了重组贝戈莫病毒如何逃避RDRγ介导的病毒基因沉默，并在混合感染中胜过其亲本病毒

抽象

番茄黄叶卷曲病毒（TYLCV，海棠病毒属，双子座病毒科）可引起栽培番茄的严重病害。双子座病毒通过滚动循环和重组依赖性机制复制环状单链基因组DNA，在混合感染中经常产生重组体。环状双链复制中间体也可作为Pol II双向转录的模板。IS76 是 TYLCV 的重组衍生物，在从相关贝戈莫病毒获得的双向启动子/复制起点区域具有短序列，在混合感染中优于 TYLCV，并打破了番茄 Ty-1 品种的抗病性。Ty-1 编码一种γ分支 RNA 依赖性 RNA 聚合酶 （RDRγ），该聚合酶与 Dicer 样 （DCL） 介导的指导基因沉默的小干扰 （si） RNA 的生物发生有关。在这里，我们分析了在早期和晚期感染阶段感染 TYLCV、IS1 或其组合的 Ty-76 抗性和对照植物的转录组和小 RNA组。我们发现 RDRγ 提高了两种病毒全基因组中 21、22 和 24 nt siRNA 的产生率，并调节了 DCL 活性，有利于 22 和 24 nt siRNA。与亲代 TYLCV 相比，IS76 向感染阶段的转变更快，有利于沉默抑制蛋白和外壳蛋白基因的右转录，从而逃避 RDRγ 活性并促进其在单次和混合感染中的 DNA 积累。在共感染的 Ty-1 植物中，IS76 有效地竞争宿主复制和转录机制，从而损害 TYLCV 的复制和转录，并迫使其消除，从而进一步增加 siRNA 的产生。RDRγ 在 Ty-1 植物中组成型过表达，这与海棠病毒耐药性相关，而产生 siRNA 的 DCL（DCL2b/d、DCL3、DCL4）和与 siRNA 扩增相关的基因 （α-clade RDR1） 和功能 （Argonaute2） 在 TYLCV 和 IS76 感染的易感植物中上调至相似水平。总的来说，IS76 重组促进了复制并促进了沉默抑制蛋白和外壳蛋白的表达，这使得重组病毒能够逃避 RDRγ 增强的病毒 siRNA 产生指导转录和转录后沉默的负面影响。

作者摘要

在植物中，由产生 siRNA 的 Dicers 和结合 siRNA 的 Argonautes 介导的内源性和抗病毒 RNAi 以在转录和/或转录后方式沉默植物和病毒基因，可以被产生二级 siRNA 前体的 α 分支的 RNA 依赖性 RNA 聚合酶 （RDR） 扩增。为了建立成功的感染，病毒逃避或抑制抗病毒RNAi。在这里，我们进行了小的RNA组和转录组分析，以揭示重组ssDNA贝戈莫病毒如何逃避抑制性siRNA，克服由γ分支RDR介导的Ty-1番茄品种的耐药性，并在混合感染中胜过亲本病毒。我们发现，携带双向启动子和复制起始元件的基因间区域内的重组事件促进了病毒DNA复制，并促进了RNAi抑制蛋白和外壳蛋白基因的右向转录。这使得重组病毒能够逃避 RDRγ 增强产生的 22 和 24 nt siRNA 的负面影响，从而有效抑制亲本病毒，从而在混合感染中消除其。

数字

Fig 10Fig 11Fig 12图1图2图3Fig 4Fig 5Fig 6Fig 7Fig 8Fig 9Fig 10Fig 11Fig 12图1图2图3

引文： Jammes M， Golyaev V， Fuentes A， Laboureau N， Urbino C， Plissonneau C， et al. （2024） 转录组和小RNAome分析揭示了重组贝戈莫病毒如何逃避RDRγ介导的病毒基因沉默，并在混合感染中胜过其亲本病毒。PLoS 病理学 20（1）： e1011941 中。 https://doi.org/10.1371/journal.ppat.1011941

编辑 器： David M. Bisaro，美国俄亥俄州立大学

收到： 20年2023月3日;接受： 2024年12月2024日;发表： <>月 <>， <>

版权所有： ? 2024 Jammes et al.这是一篇根据知识共享署名许可条款分发的开放获取文章，该许可允许在任何媒体上不受限制地使用、分发和复制，前提是注明原作者和来源。

数据可用性： 本研究产生和分析的 sRNA-seq 和 mRNA-seq 数据已分别作为 BioProjects PRJNA1016149 （SRR26047678-SRR26047725） 和 PRJNA1018203 （SRR26081500-SRR26081547） 存放在 NCBI SRA（短读段存档）数据库中。

资金： 该研究得到了ANR的支持，该项目由PRIMA项目“预防和控制感染地中海蔬菜的新型和侵入性双子座病毒”，该项目由MP与M.M.P.，C.U.和CP（https://anr.fr/Project-ANR-18-PRIM-0003）合作协调。M.J.的博士工资由ANR项目资助，并由CIRAD（https://www.cirad.fr/）延长了4个月。A.F.得到了CIRAD南方行动旅行补助金的支持。资助者在研究设计、数据收集和分析、发表决定或手稿准备方面没有任何作用。

利益争夺： 作者声明没有竞争利益。

介绍

番茄黄叶卷曲病 （TYLCD） 是由几种单链 （ss）DNA 病毒引起的，这些病毒属于双子座病毒科（双子病毒）的贝戈莫病毒属。TYLCD是全球番茄种植的主要威胁之一，因为它有严重的叶片症状、植物发育迟缓和花流产，导致番茄产量下降。在地中海盆地，该病主要由番茄黄叶卷曲病毒（TYLCV）、番茄黄叶卷曲撒丁岛病毒（TYLCSV）及其重组体引起。

双子座病毒通过滚环和重组依赖性机制在植物细胞核中复制，并将滚环复制的环状ssDNA产物包封成双子体（孪生二十面体）病毒粒子[1,2]。两种复制机制的环状双链 （ds）DNA 中间体也可作为 Pol II 介导的病毒基因转录的模板。作为典型的单方贝戈莫病毒，TYLCV 和 TYLCSV 具有 Pol II 从病毒粒子和 ~2.8 Kbp 的环状 dsDNA 互补链双向转录的 1 个基因。向右（病毒粒子链）基因V2和V1编码外壳蛋白（CP/V3）[4,2]和强沉默抑制因子（V5），也与运动有关[11\u1]。左向（互补链）基因（C4-to-C1）编码复制起始蛋白（rep/C12）[2]、转录激活因子和沉默抑制因子（TrAP/C13）[14,3]、复制增强因子（REn/C15）[4]和沉默抑制因子（C5）也与运动和致病性有关[6,16,18–1]].基因间区域包含复制起点，并且如单粒和二分贝戈莫病毒所示，双向启动子分别驱动 Pol II 转录，分别翻译 Rep/C4 和 C1 的 C4-C2 mRNA 以及翻译 V1 和 CP 的 V2-V2 mRNA;另一个单向启动子驱动C3-C2 mRNA的Pol II转录，TrAP/C3和REn/C3蛋白从中翻译，与C1-C4 mRNA共端19'-共端[21\u21]。一些证据表明，在细胞感染的早期阶段，向左转录占主导地位，有利于病毒DNA的滚环复制，而向右转录在后期被激活，有利于病毒DNA的包封。如二分贝戈莫病毒所示，右转转录被抑制自身mRNA左转的病毒Rep的协同作用激活[22,21]，而病毒TrAP反式激活CP mRNA的右转转录[23,<>]，从而导致包封病毒环状ssDNA的病毒CP过表达。

双子座病毒通过以韧皮部为食的昆虫媒介（如粉虱、蚜虫和叶蝉）以持续循环的方式传播。TYLCV和其他海棠病毒仅由烟粉虱传播。由于其体积小、生物潜力大、寄主范围大和容易产生杀虫剂抗性，B.烟粉虱是一种非常难以控制的害虫。这就是为什么植物抗性育种是迄今为止预防和控制海棠病毒病的最有效策略。在TYLCD的案例中，来自野生茄属物种的1个抗性基因（Ty-6至Ty-1）已被引入栽培西红柿（Solanum lycopersicum）的基因组中，其中Ty-24抗性植物是栽培最多的[26\u<>]。

番茄Ty-1基因编码来自γ分支（RDRγ）的RNA依赖性RNA聚合酶[27]。该基因与模式植物拟南芥（RDR3、RDR4 和 RDR5）的 RDRγ 基因相似，尚未证明其功能。A 的 α 分支 RDR。拟南虫（RDR1、RDR2 和 RDR6）参与 RNA 干扰 （RNAi），这是一种进化上保守的机制，可调节基因表达并防御大多数真核生物中的转座子、转基因和病毒等侵入性核酸。RNAi 由小干扰 （si） RNA 指导，这些 RNA 由双链 （ds）RNA 前体的 Dicer 或 Dicer 样 （DCL） 家族蛋白产生，并与 Argonaute （AGO） 家族蛋白结合，形成 RNA 诱导的沉默复合物。在植物中，siRNA的dsRNA前体由正义和反义转录或反向重复序列的转录以及RDR1、RDR2或RDR6在其特定的单链（ss）RNA模板上合成互补链的活性产生[28]。DCL将各自的dsRNA底物加工成21 nt（DCL4）、22 nt（DCL2）和24 nt（DCL3）siRNA，然后根据AGO的大小和5'末端核苷酸身份对它们进行分类[29]。在 A.感染二分贝戈莫病毒（DCL4、DCL2和DCL3）的拟南芥分别产生21、22和24 nt病毒siRNA，覆盖整个病毒基因组的两条链[30,31]。大多数贝戈病毒siRNA的产生与RDR1、RDR2或RDR6或植物特异性DNA依赖性RNA聚合酶Pol IV和Pol V的活性无关[30,31]，这表明siRNA前体是由Pol II介导的环状病毒dsDNA的双向通读转录产生的[31,32].尽管如此，RDR21和DCL6依赖性通路仍会产生少量的4 nt贝戈莫病毒siRNA，这些次级siRNA参与RNAi的细胞间扩散[30,31]。在 A.拟南芥、RDRγ基因RDR3、RDR4和RDR5彼此相邻，它们在siRNA生物发生、基因沉默或抗病毒防御中的功能（如果有的话）仍然未知[33]。在水稻（Oryza sativa）中，γ分支RDR3参与转座子和其他富含重复序列的基因组区域的调节，产生21和24 nt siRNA，生化证据表明其在ssRNA和ssDNA模板上的聚合酶活性[34]与番茄α分支RDR1的活性相似[35]。

Solanum lycopersicum plants infected with TYLCV accumulate 21, 22 and 24 nt siRNAs derived from both strands of the entire virus genome [36–38], indicating that antiviral RNAi is mediated by at least three tomato DCLs. The tomato Ty-1 gene-encoded RDRγ mediates resistance against TYLCV by enhancing production of virus-derived 22 and 24 nt siRNAs on expense of 21 nt siRNAs [38] and increasing cytosine methylation of viral DNA [39], suggesting its involvement in 24 nt siRNA-directed transcriptional silencing of viral genes and possibly posttranscriptional silencing of viral mRNAs directed by 22 nt siRNAs [38]. This hypothesis is consistent with the findings that the resistance against TYLCV is compromised in Ty-1 plants co-infected with cucumber mosaic virus or a begomoviral betasatellite, which are known to encode suppressors of posttranscriptional and transcriptional gene silencing [40].

In Morocco, an invasive recombinant between the IL strain of TYLCV (TYLCV-IL) and TYLCSV was detected in 2010 in the Ty-1 resistant plants exhibiting typical symptoms of TYLCD [41]. In this unusual recombinant, called TYLCV-IS76, a short sequence of the intergenic region of TYLCV-IL between position 1 (the origin of replication and recombination break-point) and position 84 was replaced with the homologous although slightly shorter sequence of TYLCSV (1–76). Extended surveys conducted from 2012 revealed that TYLCV-IS76 had almost totally replaced its parental viruses in the Souss region of Morocco from where it probably originated. Interestingly, the invasion of TYLCV-IS76 coincided with the deployment of Ty-1 resistant tomato cultivars in this country [41]. Under laboratory conditions, TYLCV-IS76 is positively selected in the Ty-1 plants where it accumulates at higher levels than its parental viruses and, more intriguing, has a strong deleterious effect on TYLCV-IL, leading to disappearance of this parental virus at late stages of coinfection [42,43]. The molecular mechanisms underlying partial evasion of RDRγ-mediated resistance by TYLCV-IS76 and its strong deleterious impact on TYLCV-IL in mixed infection of Ty-1 plants are unknown. In this study, we began to uncover these mechanisms by comparative transcriptome and sRNAome profiling of susceptible vs Ty-1 resistant tomato plants infected with TYLCV-IL, TYLCV-IS76 or combination thereof at early and late stages of infection.

Results and discussion

We inoculated 14-days old tomato seedlings of a Ty-1 resistant (R) cultivar (Pristyla) and a nearly isogenic susceptible (S) one with the infectious clones of TYLCV-IL isolate RE4 (AM409201; hereafter IL), TYLCV-IS76 isolate G8 (LN812978; hereafter IS76) or their combination (IL+IS76). Viral DNA loads were measured with quantitative (q)PCR, while loads, production rates and profiles of viral mRNAs and virus-derived siRNAs were analysed by Illumina sequencing of total RNA from systemically infected leaf tissues collected at 10 and 30 days post-inoculation (dpi). Two biological replicates were analysed for each condition.

Consistent with the previous studies [39,41,42], the Ty-1 resistance gene encoding RDRγ had a negative impact on viral DNA accumulation. Indeed, following single virus infection at both 10 and 30 dpi, the loads of viral DNA in R plants carrying the functional RDRγ were much lower than those in nearly isogenic S plants lacking the functional RDRγ (Fig 1A). Notably, whereas the ratio of viral loads between S and R plants was ~20 for IL, it was only ~5.5 for IS76 at both time points, indicating that the recombinant IS76 was able to evade the defence mediated by RDRγ better than its parent IL. At 10 dpi, the DNA loads were higher for IS76 than IL ~2 times in S plants and ~8 times in R plants. By 30 dpi, IS76 and IL accumulated their DNA at similar levels in S plants, whereas in R plants the DNA loads were ~4 times higher for IS76 (Fig 1A), owing to evasion of RDRγ-mediated resistance.

thumbnail Download:

PPTPowerPoint slide

PNGlarger image

TIFForiginal image

Fig 1. Total viral DNA, mRNA and small (s)RNA accumulation in susceptible (S) and Ty-1 resistant (R) tomato plants infected with TYLCV-IL, its recombinant derivative TYLCV-IS76, or their combination (IL+IS76) at 10 and 30 days post inoculation (dpi).

(A) Viral DNA loads measured by quantitative PCR (qPCR). The qPCR data were normalized using the tomato 25S rRNA gene. (B) Loads of total viral mRNAs measured by Illumina RNA-seq in reads per million (RPM) of total (plant + viral) mRNA reads. (C) Loads of total viral small (s)RNAs measured by Illumina sRNA-seq in reads per million (RPM) of total (plant + viral) sRNA reads in the size range from 15 to 34 nts. In all panels, bar graphs plot the loads for two biological replicates per each condition, with the standard error shown with a capped vertical line and the mean value indicated above. Bars for IL and R are colour-coded in purple and yellow, respectively. Ratios of the mean values for each virus (IL, IS76) and their combination (IL+IS76) in S vs R plants (S/R) are given below each graph. In the case of viral mRNAs and sRNAs, their production rates (the total mRNA or total sRNA load in RPM divided by the respective virus DNA load)—“mRNA/DNA” and “sRNA/DNA”—are also indicated below the graphs.

https://doi.org/10.1371/journal.ppat.1011941.g001

In mixed infection (IL+IS76) of S and R plants at 10 dpi, the DNA loads of IL were respectively 5.1 and 10 times lower than in singly infected S and R plants, whereas the DNA loads of IS76 were respectively 3.3 and 2.5 times lower than in singly infected S and R plants (Fig 1A). By 30 dpi, the DNA loads of IL were respectively 3.2 and 286 times lower than in singly infected S and R plants. In sharp contrast, the DNA loads of IS76 by 30 dpi were similar between single and mixed infection of both S and R plants (Fig 1A). Thus, IL had only a transient negative impact on IS76 in both S and R plants, whereas IS76 had a strong negative impact on IL at both early and late stages of co-infection in S plants and nearly eliminated IL from co-infected R plants, confirming its remarkable competitiveness in mixed infections [42,43].

RDRγ modulates production rates of viral mRNAs and strongly enhances production rates of viral sRNAs from both viruses

Using the Illumina sequencing data (S1 and S2 Datasets), we first measured the collective loads of viral mRNAs and viral sRNAs in reads per million (RPM) of total (plant+viral) mRNA and total (plant+viral) sRNA reads, respectively. Whereas viral mRNA loads overall correlated relatively well with viral DNA loads (Fig 1B vs 1A), viral sRNA loads did not correlate with viral DNA loads (Fig 1C vs 1A).

尽管病毒 DNA 和 mRNA 载量之间存在明显的相关性，但病毒 mRNA 的生产率（即病毒 mRNA 载量除以病毒 DNA 载量）的计算确实揭示了 S 和 R 植物之间以及时间点之间的差异（图 1B，图表下方）。

在10 dpi的单次感染中，S植物中IL和IS76的病毒mRNA产生率（mPR分别为7.1和6.8）低于R植物（mPR=12和9.5）。因此，这两种病毒似乎通过增加病毒mRNA的产生，部分补偿了RDRγ对病毒DNA复制的强烈负面影响。当 30 dpi 时，S 和 R 植物中两种病毒的 mRNA 产生率都下降了，并且仅观察到 IL 的代偿作用（分别为 mPR = 3.3 和 5.4），而 IS76 没有观察到代偿作用（mPR = 4.8 和 4.7）（图 1B）。

In mixed infections at 10 dpi, the rates of mRNA production from IL were similar between S and R plants (mPR = 7.2 and 7.5, respectively). Hence, in the presence of IS76, IL was not able to cope with the negative impact of RDRγ by increasing its mRNA production rate. In contrast, the presence of IL did not affect the rates of mRNA production from IS76 which became even higher in both S and R plants (mPR = 13 and 11, respectively), compared to the respective plants singly infected with IS76 (mPR = 6.8 and 9.5, respectively). By 30 dpi, the rates of mRNA production from both viruses dropped down in both S and R plants. Compared to single infections at this time-point, IL only slightly affected IS76 in S plants (mPR = 4.4 vs 4.8) and R plants (mPR = 5.3 vs 4.7). In contrast, IS76 had a substantial negative impact on IL in both S (mPR = 1.0 vs 3.3) and R (mPR = 3.4 vs 5.4) plants (Fig 1B).

The overall production rates of viral sRNAs (i.e., viral sRNA loads in RPM divided by viral DNA loads) in R plants were drastically higher than in S plants, irrespective of the conditions (Fig 1C, below the graphs). Notably, the positive effect of RDRγ on the sRNA production rate (sPR) was more pronounced for IL. Indeed, in single infections at both 10 and 30 dpi the ratio of sPRs in R vs S plants was much higher for IL (1050/65 and 1097/79, respectively), compared to IS76 (472/47 and 636/77). In mixed infections, IL did not have any substantial effect on the rates of sRNA production from IS76. In contrast, IS76 modulated those from IL, with the most pronounced effect observed in R plants at 30 dpi (sPR = 6971 vs 1097 in single infection).

Taken together, RDRγ modulates the overall production rates of viral mRNAs and strongly enhances the overall production rates of viral sRNAs from both viruses in both single and mixed infections at both time-points.

IS76 undergoes faster transition to overexpression of the rightward genes

Mapping of Illumina mRNA-seq 100 nt paired-end reads on the reference genomes of IL and IS76 revealed the three Pol II transcription units previously reported for begomoviruses, one rightward (virion strand) unit for the V2-V1 mRNA and two leftward (complementary strand) units for 3’-coterminal C1-C4 and C2-C3 mRNAs (Figs 2 and 3 and S3 Dataset). Consistent with a previous mRNA-seq study of TYLCV [37], reads were not homogeneously distributed along the length of each mRNA, due to sequence-specific biases in Illumina library preparation and sequencing protocols leading to either underrepresentation or overrepresentation of certain sequences. Indeed, the map patterns are similar for each viral mRNA between S and R plants and between 10 and 30 dpi. The relative abundance of all reads (in RPM) representing each viral mRNA differed substantially, showing that V2-V1 mRNA is the most abundant for both viruses in all conditions, followed by the second most abundant C2-C3 mRNA and the least abundant C1-C4 mRNA (Figs 2, 3 and S1). Similar mRNA profiles were observed for TYLCV-IL (isolate Almeria) in susceptible tomato (cv. Moneymaker) at 7, 14 and 21 dpi [37].

thumbnail Download:

PPTPowerPoint slide

PNGlarger image

TIFForiginal image

Fig 2.

Single-nucleotide resolution maps of viral mRNA reads in susceptible (S) and Ty-1 resistant (R) tomato plants infected with TYLCV-IL or its recombinant derivative TYLCV-IS76 at 10 (A) and 30 (B) days post inoculation (dpi). For each condition, Illumina mRNA-seq 100 nt paired-end reads were mapped onto the reference sequences of IL and IS76 genomes with zero mismatches (see S3 Dataset for more details of mapping). Histograms plot the numbers of viral reads at each nucleotide position of the IL and IS76 genomes (2781 and 2773 bp in length, respectively): blue bars above the axis represent virion strand (rightward) reads starting at each respective position, while red bars below the axis represent complementary strand (leftward) reads ending at each respective position. The viral genome organization is shown schematically above the histograms, with ORFs of the viral rightward (V1, V2) and leftward (C1-to-C4) genes shown with blue and red arrows, respectively, and capped and polyadenylated viral mRNAs (V2-V1, C1-C4 and C2-C3) shown as solid blue and red lines.

https://doi.org/10.1371/journal.ppat.1011941.g002

thumbnail Download:

PPTPowerPoint slide

PNGlarger image

TIFForiginal image

Fig 3.

Single-nucleotide resolution maps of viral mRNA reads in susceptible (S) and Ty-1 resistant (R) tomato plants co-infected with TYLCV-IL and its recombinant derivative TYLCV-IS76 at 10 (A) and 30 (B) days post inoculation (dpi). For each condition, Illumina 100 nt paired-end reads were mapped onto the reference sequences of IL and IS76 genomes with zero mismatches (see S3 Dataset for more details of mapping). Histograms plot the numbers of viral reads at each nucleotide position of the IL and IS76 genomes (2781 and 2773 bp in length, respectively): blue bars above the axis represent virion strand (rightward) reads starting at each respective position, while red bars below the axis represent complementary strand (leftward) reads ending at each respective position. The viral genome organization is shown schematically above the histograms, with ORFs of the viral rightward (V1, V2) and leftward (C1-to-C4) genes shown with blue and red arrows, respectively, and capped and polyadenylated viral mRNAs (V2-V1, C1-C4 and C2-C3) shown as solid blue and red lines.

https://doi.org/10.1371/journal.ppat.1011941.g003

The most striking difference between IS76 and IL is that in all conditions IS76 accumulates its V2-V1 mRNA at relatively higher levels, compared to IL (Figs 2, 3 and S1). This suggests that the recombination region of IS76 modulates the bidirectional promoter in favour of rightward transcription. To estimate the activities of the bidirectional promoter driving leftward transcription of C1-C4 mRNA and rightward transcription of V2-V1 mRNA as well as the monodirectional promoter driving transcription of C2-C3 mRNA, we calculated the production rate of each viral mRNA (i.e., the mRNA load in RPM divided by the mRNA transcription unit length in nucleotides and by the viral DNA load). The results revealed that the ratio of production rates of V2-V1 mRNA vs C1-C4 mRNA was higher for IS76 than IL in all conditions (Figs 4 and 5). The higher ratio of rightward-to-leftward transcription rates may reflect a faster replication cycle of IS76 in which encapsidation of viral DNA by the viral CP (translated from V2-V1 mRNA) begins earlier than for IL. This hypothesis is consistent with the accumulation dynamics of viral DNA: IS76 accumulates (and hence replicates) its DNA faster than IL in both S and R plants (Fig 1A).

thumbnail Download:

PPTPowerPoint slide

PNGlarger image

TIFForiginal image

图 4.

在接种后 1 （A） 和 76 （B） 天 （dpi） 时，单纯感染 TYLCV-IL 及其重组衍生物 TYLCV-IS10 的易感 （S） 和 Ty-30 抗性 （R） 番茄植株中病毒 mRNA 和病毒基因组互补 （C） 支架的病毒 mRNA 和 sRNA 的产生率 （dpi）。代表IL和IS2的每个mRNA（V1-V1、C4-C2、C3-C76）的Illumina mRNA-seq读段和代表病毒粒子和每个病毒基因组的互补链的Illumina sRNA-seq读段（大小范围为20至25 nts）以总（植物+病毒）reads的百万分之一（RPM）计算。将所得计数除以qPCR测量的相应病毒DNA的载量，如果是病毒mRNA，则除以核苷酸中每个转录单元的长度。在每个面板中，条形图绘制了病毒 DNA（分别为 IL 和 IS76 的黄色和紫色条）的负载量以及向右 （V2-V1） 和向左（C1-C4、C2-C3）mRNA（分别为蓝色和红色条）以及来自病毒粒子和互补林分的 sRNA（分别为蓝色和红色条）的产生率。在所有情况下，每个条件的载样量为两个生物学重复，标准误差用带帽的垂直线和上面指示的平均值显示。

https://doi.org/10.1371/journal.ppat.1011941.g004

thumbnail 下载：

PPT格式PowerPoint 幻灯片

巴布亚新几内亚放大图像

TIFF格式原始图像

图 5.

在接种后 1 （A） 和 76 （B） 天 （dpi） 时，在共感染 TYLCV-IL 及其重组衍生物 TYLCV-IS10 的易感 （S） 和 Ty-30 抗性 （R） 番茄植株中，病毒 mRNA 和病毒基因组互补 （C） 林分衍生的病毒 mRNA 和 sRNA 的产生率 （dpi）。代表IL和IS2的每个mRNA（V1-V1、C4-C2、C3-C76）的Illumina mRNA-seq读段和代表病毒粒子和每个病毒基因组的互补链的Illumina sRNA-seq读段（大小范围为20至25 nts）以总（植物+病毒）reads的百万分之一（RPM）计算。将所得计数除以qPCR测量的相应病毒DNA的载量，如果是病毒mRNA，则除以核苷酸中每个转录单元的长度。在每个面板中，条形图绘制了病毒 DNA（分别为 IL 和 IS76 的黄色和紫色条）的负载量以及向右 （V2-V1） 和向左（C1-C4、C2-C3）mRNA（分别为蓝色和红色条）以及来自病毒粒子和互补林分的 sRNA（分别为蓝色和红色条）的产生率。在所有情况下，每个条件的载样量为两个生物学重复，标准误差用带帽的垂直线和上面指示的平均值显示。

https://doi.org/10.1371/journal.ppat.1011941.g005

有趣的是，在10 dpi的单次感染植物中，R植物中IL的V2-V1 mRNA与C1-C4 mRNA的产生率比S植物低2.2倍（mPR比= 5.5对12），而IS76的该比率几乎不受影响（mPR比= 20.2对22.5）。因此，RDRγ 通过阻碍从左转录到右转录的转变，对 IL 的复制周期产生负面影响。

从左向右转录的转变由基因间区域顺式元件控制，最明显的是被病毒TrAP反转激活的右向启动子中的元件[44,45]。此外，左向启动子中的短迭代序列重复序列（iterons）调节这种转变，因为病毒Rep结合这些iterons，从而抑制其自身mRNA的左向转录，如二分begomovirus所证明的那样[22]。最后，如另一种二分贝戈莫病毒[21]所示，Rep介导的左向转录抑制进一步增强了TrAP/C2反转激活的V1-V2 mRNA的右向转录，从而促进了包封所需的病毒CP的过表达。

Notably, the production rates of C1-C4 mRNA from both viruses in all conditions correlated with those of C2-C3 mRNA (Fig 4). Indeed, the C2-C3/C1-C4 ratios are all within a relatively narrow range (3.5 to 5.7), suggesting that C1-C4 mRNA transcription driven by the bidirectional promoter and C2-C3 mRNA transcription driven by the downstream monodirectional promoter are co-regulated. The mechanism of this co-regulation remains to be investigated.

Collectively, the results reveal a mechanism explaining how IS76 copes better than IL with the negative impact of RDRγ on virus replication. While RDRγ slows down the transition of IL to the cell infection stage favouring the rightward transcription over the leftward transcription, IS76 undergoes this transition almost as fast as in the absence of functional RDRγ.

It is worth noting that the V2 ORF-encoded protein is a strong suppressor of antiviral RNAi and gene silencing [5–11]. Favouring expression of this protein via enhanced transcription of V2-V1 mRNA at earlier stages of cell infection would allow the recombinant IS76 to better suppress antiviral RNAi and, in particular, to counteract the repressive viral sRNAs whose production rates are strongly enhanced by RDRγ for both viruses at both 10 and 30 dpi (Fig 1C).

Our assertions that IS76 replicates faster than IL and that RDRγ slows down viral replication, which are based on the differences in viral DNA accumulation dynamics and leftward-to-rightward transcription rate ratios, were also supported by our Southern blot hybridization analysis of viral DNA forms with strand-specific probes. Indeed, in S plants at 30 dpi, when the dsDNA intermediates of replication were above the detection threshold for both viruses and accumulated at comparable levels, circular ssDNA products of rolling-circle replication accumulated at higher levels for IS76 than IL (S2 Fig). Furthermore, the ssDNA-to-dsDNA ratio of IS76 was higher in S plants, compared to R plants at this time point (S2D Fig).

IS76 is more transcriptionally active and reduces the transcription rates of all IL mRNAs in the course of coinfection

In coinfected S plants at 10 dpi, the production rates of all mRNAs from IL were similar to those observed in S plants singly infected with IL, while the presence of IL resulted in a ~2 times increase in the production rate of each mRNA of IS76 (Figs 5 vs 4). On the other hand, in coinfected R plants at 10 dpi the individual production rates of IS76 mRNAs were only slightly higher than those observed in R plants singly infected with IS76, and the presence of IS76 resulted in a ~2 times decrease in the individual production rate of each IL mRNA (Figs 5 vs 4). By 30 dpi the negative effect of IS76 on mRNA production from IL becomes evident in both S and R plants. In coinfected S plants, the production rates of all IL mRNAs were proportionally decreased, each about 3 times, compared to singly infected S plants (Figs 5 vs 4). In contrast, the individual mRNA production rates of IS76 were comparable in singly infected and coinfected S plants at 30 dpi. In coinfected R plants, where only residual amounts of IL DNA were detected by 30 dpi, the individual production rates of IS76 mRNAs were comparable to those observed in singly infected R plants and DNA accumulation of IS76 almost reached the levels observed in singly infected R plants. On the other hand, the presence of IS76 reduced the individual production rates of all IL mRNAs in R plants, although to a lesser extent than in S plants.

Notably, mixed infection did not have any drastic effect on the rightward-to-leftward mRNA production rate ratio of IL or IS76 (Figs 5 vs 4).

RDRγ boosts the production rates of viral siRNAs from both strands of the entire virus genome, with the most pronounced effects at the promoter and terminator regions of both viruses

在IL和IS76的参考基因组上绘制Illumina sRNA读数显示，病毒sRNA来源于S和R植物中整个病毒基因组的两条链，并且在两个时间点（图6、7、8以及S3和S4数据集），并且由三个主要大小类别（图9和S4数据集;见下文）。在 10 dpi 和 30 dpi 的单次感染和共感染 S 植物中，正义和反义极性的病毒 sRNA 的热点集中在 Pol II 转录单元内，并且在具有双向启动子的基因间区域 （IR）、含有 poly（A） 位点的 Pol II 终止子区域以及较小程度上 C2-C3 启动子区域（在 C2 和 C4 ORF 之间）。相比之下，在 76 dpi 和 10 dpi 下单独感染或合并感染 IL 和 IS30 的 R 植物中，病毒 sRNA 热点更均匀地分布在整个病毒基因组中，包括 IR 和终止子区域（图 6 和 7）。这些结果与先前的研究基本一致，这些研究分析了易感番茄植株中不同时间点的TYLCV-IL sRNAs[36\u37]，并比较了易感（创钱者）和Ty-1抗性（Tygress）植株中TYLCV-IL（分离株Almeria）的sRNA谱[38]。

thumbnail 下载：

PPT格式PowerPoint 幻灯片

巴布亚新几内亚放大图像

TIFF格式原始图像

图 6.

接种后 20 （A） 和 25 （B） 天 （dpi） 感染 TYLCV-IL 或其重组衍生物 TYLCV-IS1 的易感 （S） 和 Ty-76 抗性 （R） 番茄植株中 10-30 nt 病毒小 （s）RNA 的单核苷酸分辨率图。对于每种情况，将大小范围为20至25 nts的Illumina sRNA-seq读数定位到IL和IS76基因组的参考序列上，且零错配（有关每种大小类别的病毒sRNA的映射和映射的更多详细信息，请参见S4数据集）。直方图绘制了 IL 和 IS76 基因组每个核苷酸位置的病毒读数（长度分别为 2781 和 2773 bp）：轴上方的蓝色条表示从每个相应位置开始的病毒粒子链（向右）读数，而轴下方的红条表示在每个相应位置结束的互补链（向左）读数。病毒基因组组织示意性地显示在直方图上方，病毒右向（V1、V2）和向左（C1-to-C4）基因的ORF分别用蓝色和红色箭头显示，加帽和多聚腺苷酸化的病毒mRNA（V2-V1、C1-C4、C2-C3）显示为蓝色和红色实线。

https://doi.org/10.1371/journal.ppat.1011941.g006

thumbnail 下载：

PPT格式PowerPoint 幻灯片

巴布亚新几内亚放大图像

TIFF格式原始图像

图 7.

接种后 20 （A） 和 25 （B） 天 （dpi） 共感染 TYLCV-IL 及其重组衍生物 TYLCV-IS1 的易感 （S） 和 Ty-76 抗性 （R） 番茄植株中 10-30 nt 病毒小 （s）RNA 的单核苷酸分辨率图谱。对于每种情况，将大小范围为20至25 nts的Illumina sRNA-seq读长定位到IL和IS76基因组的参考序列上，且零错配（有关混合感染中读取映射和计数的详细信息以及每种大小类别的病毒sRNA的图谱，请参见S4数据集）。直方图绘制了 IL 和 IS76 基因组每个核苷酸位置的病毒读数（长度分别为 2781 和 2773 bp）：轴上方的蓝色条表示从每个相应位置开始的病毒粒子链（向右）读数，而轴下方的红条表示在每个相应位置结束的互补链（向左）读数。病毒基因组组织示意性地显示在直方图上方，病毒右向（V1、V2）和向左（C1-to-C4）基因的ORF分别用蓝色和红色箭头显示，加帽和多聚腺苷酸化的病毒mRNA（V2-V1、C1-C4、C2-C3）显示为蓝色和红色实线。

https://doi.org/10.1371/journal.ppat.1011941.g007

thumbnail Download:

PPTPowerPoint slide

PNGlarger image

TIFForiginal image

Fig 8.

Production rates of sRNAs from different regions of virion (V) and complementary (C) stands of the viral genome in susceptible (S) and Ty-1 resistant (R) tomato plants infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or a combination thereof (IL+S76) at 10 (A) and 30 (B) days post inoculation (dpi). Illimina sRNA-seq reads in the size range from 20 to 25 nts representing virion and complementary strands of each transcription unit (V2-V1, C1-C4, C2-C3) and two parts of the intergenic region with the rightward (IR1) and the rightward (IR2) promoters of IL and IS76 genomes were counted in reads per million (RPM) of total (plant + viral) reads (see Material and Methods for further details of read counting in mixed infection). The resulting counts were divided by the length of each region in nucleotides and the load of respective viral DNA measured by qPCR and then multiplied by 10000. Bar graphs plot sRNA loads for the rightward (V2-V1) and leftward (C1-C4, C2-C3) mRNA transcription units and two parts of the intergenic region (IR1 and IR2). The loads of sRNAs derived from the virion and complementary stands of each region of the viral genome are represented with blue and red bars, respectively. In all cases, the loads are for two biological replicates per each condition, with the standard error shown with a capped vertical line and the mean value indicated above.

https://doi.org/10.1371/journal.ppat.1011941.g008

从整个病毒基因组的两条链产生病毒 sRNA，包括“非转录”IR（在左侧 C1-C4 和右侧 V2-V1 单元的 Pol II 转录起始位点之间）表明，病毒 siRNA 的 dsRNA 前体是由 Pol II 介导的通读转录产生的，远远超出了之前针对二分贝戈莫病毒提出的向左和向右方向的 poly（A） 信号 [30，31]. 为了支持这一假设，我们对IL和IS76感染植物的Illumina测序分析显示，低丰度长RNA读数覆盖了向右和向左转录单元的反义链以及IR的两条链，这可能代表了推定通读转录本的残余（S1图和S3数据集).覆盖整个环状病毒基因组的向左和向右通读转录本可能通过与极性相反和/或彼此配对的 mRNA 形成病毒 sRNA 的 dsRNA 前体。只有后一种事件可能产生源自 Pol II 转录单元外 IR 的病毒 sRNA 的 dsRNA 前体。与 S 植物相比，R 植物中两种病毒的 IR 中正义和反义极性的病毒 sRNA 富集表明 RDRγ 可能间接促进双向通读转录。此外，RDRγ 可能会将通读转录本转换为 dsRNA。在 S 植物中，IR 衍生的 sRNA 的 dsRNA 前体可能由α分支 RDR（RDR1、RDR2 和/或 RDR6）使用 Pol II 通读转录或 Pol V 或 Pol IV 转录产生的 ssRNA 底物产生。尽管如此，α分支RDR的潜在活性无法解释病毒sRNA热点集中在mRNA转录单元中的原因。事实上，Illumina对来自A的病毒sRNA进行了测序。感染二分海棠病毒的野生型和RDR1/RDR2/RDR6三突变株在病毒sRNA积累、大小、极性或热点谱方面没有显著差异[31]。值得注意的是，后者的图谱类似于我们在 S 植物中观察到的 IL 和 IS76 的各自图谱，因为 IR 和 Pol II 终止子区域在 sRNA 热点中耗尽。

接下来，我们计算了来自 Pol II 转录单元的每条链和包含右向 （IR1） 和向左 （IR2） 启动子元件的 IR 的两部分的 sRNA 的产生率。为此，将来自相应区域每条链的 sRNA 读段计数（以 RPM 为单位）除以核苷酸中相应区域的长度，然后除以病毒 DNA 载量。对所得 sRNA 产生率的比较显示，与 S 植物相比，R 植物在两种病毒在两个时间点的单一感染和混合感染中都显着增加（图 8），表明病毒 sRNA 产生的 RDRγ 依赖性增强会影响病毒基因组的所有区域和链。引人注目的是，这种增强在 IR 的两个部分和两条链中比在 Pol II 转录单元中更为明显（图 8），这与 sRNA 单核苷酸分辨率图中观察到的差异一致（图 6 和 7）。

来自转录单元或 IR 的病毒 sRNA 的产生率与病毒 mRNA 的产生率无关

在 10 dpi 的单次感染 S 植物中，在 C2-C3 和 C1-C4 单位中观察到最高的 sRNA 产生率，其次是 V2-V1 单位，没有任何实质性的正向或反向链偏差（图 8A）。然而，虽然 V2-V1 单位在 IL- 和 IS76 感染的 S 植物中以相似的速率产生 sRNA，但 IL 的 C1-C4 和 C2-C3 单位产生 sRNA 的速率高于 IS76。这与IL感染植物中V2-V1 mRNA的产生率相似，C1-C4和C2-C3 mRNA的产生率更高（图4）。因此，从左向转录单元以较高速率产生的两种极性的病毒 sRNA 不会降低（而是增加）C1-C4 或 C2-C3 mRNA 的产生速率，而 C76-C1 或 C4-C2 mRNA 可能会被相反极性的 sRNA 靶向切割和降解。此外，对于 IL 和 IS3，C76-C8 和 C2-C1 单元以相似的速率产生两种极性的病毒 sRNA，它们以截然不同的速率产生各自的 mRNA。此外，在单独感染IL或IS1的S植物中，IR两个部分的sRNA的产生率相当（图4A）。由于向右 V4-V<> mRNA 转录的速率远高于向左的 C<>-C<> mRNA 转录速率（图 <>），因此源自基于 IR 的双向启动子的病毒 sRNA 似乎不调节 Pol II 介导的双向转录。

在单独感染的 S 植物中，当 30 dpi 时，两种病毒的病毒 sRNA 产生速率都有所增加。对于IL基因组的每个区域，这种增加几乎是成比例的（~1.5倍），而对于IS76来说，这种增加是不成比例的，其中IR和V2-V1单元都以~2-3倍的速率产生sRNA，而其向左的单位以~1.2-1.5倍的速率产生sRNA（图8B与8A）。与 10 dpi 类似，在 30 dpi 时，来自 IL 和 IS76 不同区域的病毒 sRNA 的相对产生率（图 8B）与各自病毒 mRNA 的相对产生率无关（图 4B）。尽管如此，病毒 sRNA 生成率在 10 到 30 dpi 之间的总体增加确实与病毒 mRNA 生成率的整体下降相吻合。根据我们的假设，在 S 植物细胞感染后期，病毒 sRNA 的产生增加可能是由于产生 sRNA 前体的通读转录增加。

在 10 dpi 的单体感染 R 植株中，来源于 IS76 基因组所有区域的病毒 sRNA 的产生率几乎成比例地（~2 至 2.5 倍）低于 IL 基因组的产生率，表明 IS76 可以更好地逃避 RDRγ 活性促进 sRNA 的产生。与 S 植物相比，两种病毒都从病毒基因组的互补（反向）链中产生相对更丰富的 sRNA，尤其是在 V2-V1 单元和 IR 的两个部分内（图 8A）。Voorburg等[38]先前在TYCLV感染的Ty-1抗性植物和易感植物中观察到sRNA热点谱和链偏倚的类似改变。值得注意的是，尽管靶向 V2-V1 启动子和 V76-V4 mRNA 的病毒 sRNA 的产生率相差 2-2.5 倍，但 IL 和 IS2 之间 V1-V2 mRNA 的产生率相当（图 1A）。此外，IL 左向启动子和左向转录单元的病毒 sRNA 产生率高得多，与左向 mRNA 的产生率高得多，反之亦然。

By 30 dpi in R plants singly infected with IL, the rates of sRNA production slightly increased in both strands of the IR, while they were slightly decreased in both strands of all the three transcription units (Fig 8). At the same time, production rates of all IL mRNAs were strongly and almost proportionally (2–3 times) decreased (Fig 4B). In the case of IS76, the sRNA production rates were increased almost proportionally in both strands of each region of the virus genome but still remained lower than those in the respective regions and strands of the IL genome. This increase did coincide with decreased production rates of all viral mRNAs (Fig 4B). The sRNA strand biases observed at 10 dpi for both viruses in R plants (see above) were also observed at 30 dpi.

IS76 衍生的 siRNA 可能会加强混合感染中 IS76 对 IL 的负面影响

与单一感染相比，在 S 或 R 植物合并感染的任何时间点存在 IL 的情况下，来自 IS76 基因组任何区域或链的 sRNA 的产生率没有显着差异（图 8）。同时，在IL干扰IS76复制的早期时间点，在存在IL的情况下，所有IS76 mRNA的产生率都较高（图5）。相反，IS76 对 S 与 R 植物和 10 与 30 dpi 中 IL 基因组的 sRNA 产生率具有强烈且对比鲜明的影响。事实上，在30 dpi的合并感染的S植物中，IL基因组的所有区域和链的sRNA产生速率都大大降低，除了具有右启动子的IR部分，其速率在单一和混合感染中非常相似。这些改变与所有IL mRNA的产生率的大幅下降相吻合。与此形成鲜明对比的是，在合并感染的R植株中，当共感染的R植株产生30 dpi时，IL基因组所有区域和链的sRNA产生率远高于单独感染的R植株（图8），这也与所有IL mRNA的产生率下降相吻合。在 IS76 存在下，IL DNA 积累的随之急剧减少表明 IL 无法应对 RDRγ 介导的病毒基因组所有区域和链的 sRNA 产生的增加。除了 IL 基因组衍生的 sRNA 外，来自 IS76 基因组所有区域的高丰度 sRNA（图 7 和图 8）可以进一步抑制 IL 基因表达，因为它们中的大多数与转录单元内的 IL 基因组和重组区域外的 IR 具有 100% 的同一性（参见 S5 数据集).与 S 和 R 植物的单一感染相比，这种跨靶标抑制与混合感染中所有 IL mRNA 的产生率较低一致。IL 衍生的 sRNA 对 IS76 基因表达的相互靶向也是可能的，尤其是在 S 植物感染的早期阶段，当这些 sRNA 以相对较高的水平积累时。然而，IS76 似乎比 IL 更好地逃避抑制性 sRNA，使所有 mRNA 的产生率甚至高于单一感染。

RDRγ 可提高病毒 siRNA 的所有三个功能大小类别的产生率，并调节 Dicer 活性，以支持 22 和 24 nt siRNA

与之前对TYLCV-IL [36–38]和其他begomovirus（例如[31]）的研究一致，在所有条件下都观察到了源自IL和IS21基因组两条链的病毒siRNA（22、24和76 nt）的三个主要（和功能）大小类别（图9和S4数据集）。它们的相对丰度在S和R植物之间有很大差异。在 S 植物中，21 nt 类在 10 dpi 和 30 dpi 时均占两种病毒的主导地位，其次是丰度第二高的 22 nt 类和丰度更低的 24 nt 类。在所有条件下，R植物中，两种极性的22和24 nt siRNA的相对比例均显著高于S植物（图9）。Voorburg等报道了感染TYLCV-IL的易感植物和Ty-1抗植物在病毒sRNA大小谱上的类似差异[38]。

thumbnail 下载：

PPT格式PowerPoint 幻灯片

巴布亚新几内亚放大图像

TIFF格式原始图像

图 9. 接种后 1 天和 76 天 （dpi） 感染 TYLCV-IL、其重组衍生物 TYLCV-IS76 或其组合 （IL+S10） 的易感 （S） 和 Ty-30 抗性 （R） 番茄植株中来自完整病毒基因组的病毒 sRNA 的大小图谱 （dpi）。

计算映射到完整病毒基因组的病毒粒子（向右）和互补（向左）链的20至25 nts大小范围内的Illumina sRNA-seq读长，计算总6-20 nt病毒读长（设置为25%）中100个个体大小类别的百分比（%），并绘制为条形图，蓝色和红色条形代表向右和向左的链， 分别。在所有面板中，百分比是每种条件下两次生物学重复的百分比，标准误差用带帽的垂直线和上面指示的平均值显示。

https://doi.org/10.1371/journal.ppat.1011941.g009

在 A.拟南-二分贝戈莫病毒系统，21、22和24 nt病毒siRNA的生物发生分别由DCL4、DCL2和DCL3介导，它们都有助于病毒基因的转录后和转录沉默[30,31]。因此，除了通过靶向 IL 和 IS76 的所有三种番茄 DCL 促进病毒 siRNA 的整体产生外，RDRγ 还更显着地增强了 DCL2 和 DCL3 的活性。如上所述，RDRγ 可能通过促进整个病毒基因组的双向通读转录或直接通过将病毒 ssRNA 转录本转化为 dsRNA 来间接发挥作用。在S植物中缺乏功能性RDRγ的情况下，显性21 nt类的热点在所有条件下都局限于转录单元，这表明DCL4可能优先处理由mRNA组成的dsRNA和相反极性的通读转录本。相比之下，丰度较低的 22 nt 类的热点以及更明显的低丰度 24 nt 类的热点也扩散到 IR 和终结者区域，尤其是在 30 dpi 处（S4 数据集）。因此，DCL3 和 DCL2 可能优先处理由正义和反义极性的通读转录本组成的 dsRNA。事实上，IR 的两个部分（只有通读转录本可能形成 dsRNA）产生的 22 和 24 nt siRNA 比例（图 10 和 11）相对高于 Pol II 单元（S4 图 ）。在 R 植物中，源自 IR 的 21 nt siRNA 的比例大幅降低，有利于 22 nt siRNA，在较小程度上支持 24 nt siRNA，因此 22 nt 类占主导地位（图 10 和 11）。尽管 22 nt 和 21 nt siRNA 的热点在整个 IL 和 IS76 基因组的两条链上都变得更加均匀地分布（S4 数据集）。在Pol II单元中也观察到sRNA大小分布的类似变化，尽管21 nt siRNA的比例降低幅度较小（S4图）。这些发现表明，除了促进双向转录（可能为所有三种 DCL 生成 dsRNA 底物）外，RDRγ 还可能将 ssRNA 模板转化为 DCL2 优先处理的 dsRNA 底物，而 DCL3 不太优先处理的 dsRNA 底物。

thumbnail 下载：

PPT格式PowerPoint 幻灯片

巴布亚新几内亚放大图像

TIFF格式原始图像

图 10. 在接种后 1 天和 1 天 （dpi） 时，感染 TYLCV-IL、其重组衍生物 TYLCV-IS1 或其组合 （IL+S76） 的易感 （S） 和 Ty-76 抗性 （R） 番茄植株中含有右启动子的基因间区域 10 （IR30） 衍生的病毒 sRNA 的大小图谱 （dpi）。

Illumina sRNA-seq读长在20至25 nts的大小范围内，映射到病毒基因组，计数IR1的病毒粒子（向右）和互补（向左）链，计算总6-20 nt病毒读长（设置为25%）中100个个体大小类别的百分比（%），并绘制为条形图，蓝色和红色条形代表向右和向左的链， 分别。在所有面板中，百分比是针对每种条件的两个生物学重复，标准误差用带帽的垂直线显示，平均值如上所示。

https://doi.org/10.1371/journal.ppat.1011941.g010

thumbnail 下载：

PPT格式PowerPoint 幻灯片

巴布亚新几内亚放大图像

TIFF格式原始图像

Fig 11. Size profiles of viral sRNAs derived from the leftward promoter-containing intergenic region 2 (IR2) in susceptible (S) and Ty-1 resistant (R) tomato plants infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or a combination thereof (IL+S76) at 10 and 30 days post inoculation (dpi).

Illumina sRNA-seq reads in the size range from 20 to 25 nts mapped to the viral genome the virion (rightward) and complementary (leftward) strands of the IR1 were counted and percentages (%) of 6 individual size-classes in the total 20–25 nt viral reads (set to 100%) were calculated and plotted as bar graphs, with blue and red bars representing rightward and leftward strands, respectively. In all panels, the percentages are for two biological replicates per each condition with the standard error shown with a capped vertical line and the mean value indicated above.

https://doi.org/10.1371/journal.ppat.1011941.g011

Notably, the proportion of 24 nt siRNAs derived from all genome regions is higher in IL-derived sRNAs compared to IS76-derived sRNAs in most conditions, except for coinfected R plans at 30 dpi where the proportions of 24 nt siRNAs produced from residual IL and highly abundant IS76 are comparable (Figs 9, 10, 11 and S4). Thus, IS76 evades DCL3 activity better than IL and, at the same time, attracts other two DCLs generating 21 and 22 nt siRNAs better than IL (except nearly eliminated IL).

Collectively, IS76 is transcribed by Pol II more efficiently than IL owing to both recombination region elements and more efficient evasion of DCL3-mediated transcriptional silencing generating 24 nt siRNAs. More efficient transcription of IS76 facilitates replication of its DNA and accelerates transition from replication to encapsidation, but at the same time attracts DCL2 and DCL4 that mediate post-transcriptional silencing. These properties of IS76 explain its selective advantage and competitiveness in mixed infections with IL, both in the absence and presence of functional RDRγ.

Interestingly, the higher abundance of complementary strand-derived sRNAs observed within the IR and the V2-V1 unit in R plants is largely due to increased proportion of the 24 nt siRNAs that exhibit the complementary strand bias in both R and S plants in most conditions (S4 Dataset). Strand biases in viral sRNA profiles along the viral genome likely result from differential sequence-specific stability of sRNAs produced by DCLs in a form of duplexes from longer dsRNA precursors and then sorted by AGO proteins. AGOs form stable complexes with guide strands of the sRNA duplexes, and discard their passenger strands, leading to degradation of the latter. As the guide strand is selected by AGOs based on the size, 5′-nucleotide identity and other sequence features [29], differences in size and nucleotide composition of viral siRNA duplexes processed by DCLs from different types of dsRNA precursors might result in local hotspots and strand biases.

Comparison of 5’-terminal nucleotide identities of viral sRNAs did not reveal any substantial difference between IL and IS76 (S2D Dataset). In both viruses, 21 and 22 nt siRNAs predominantly possess 5’U (60–70%) followed by 5’A (~20%), suggesting their preferential association with AGO1 (5’U) and less preferential with AGO2 (5’A), whereas 24 nt siRNAs possess predominantly 5’A and 5’U (~40–50% each), suggesting their preferential association with AGO4 clade proteins (5’A) and an as-yet unknown AGO (5’U). The sRNA 5’-nucleotide profiles of IL and IS76 were similar between S and R plants at both time-points, indicating that RDRγ does not influence the sorting of viral sRNAs by AGOs.

Ty-1 gene is constitutively overexpressed in R plants

To complement the viral siRNA profiling results, we analysed our mRNA-seq data for expression levels of Ty-1 and other tomato genes implicated in siRNA biogenesis and function.

Consistent with previous findings for Ty-1 gene of the Ty-1/ty-1 hybrid Tygress [27], this resistance gene was expressed at higher levels in our R plants (Ty-1/ty-1 hybrid Pristyla), compared to S plants, in both mock-inoculated and virus-infected plants at both 10 and 30 dpi (Fig 12). No effect of viral infection on Ty-1 expression levels in R plants was observed. As those levels were comparable between 10 and 30 dpi, Ty-1 overexpression in R plants appears to be constitutive and sufficient to confer virus resistance. In contrast to R plants, expression of this gene was elevated between 10 and 30 dpi in mock-inoculated S plants and was further upregulated by both IL and IS76 as well as mixed infection, compared to mock control at 30 dpi. Notably, in both single and mixed infections of S plants, this gene was upregulated to similar levels and these levels did not reach the levels of Ty-1 overexpression in R plants (Fig 12). At 10 dpi, expression of this gene in S plants was not altered by viral infection. As discussed previously by Verlaan et al. [27], it is not clear if the resistance conferred by the Ty-1 allele is due to a higher transcriptional level as comparted to that of ty-1 alleles, or the difference in amino acid sequence of the Ty-1 allele-encoded protein. Nonetheless, ty-1 upregulation by TYLCV-IL observed in cv. Moneymaker (ty-1/ty-1) by Verlaan et al. [27] and confirmed here for our susceptible ty-1/ty-1 cultivar, nearly isogenic to the resistant Ty-1/ty-1 hybrid Pristyla, would imply that RDRγ variants encoded by ty-1 alleles could potentially contribute to antiviral defence at later stages of viral infection when their expression is upregulated. However, this contribution is not sufficient to confer resistance to TYLCV, either due to weaker functionality or insufficient accumulation of the protein even after upregulation of ty-1 allele expression.

thumbnail Download:

PPTPowerPoint slide

PNGlarger image

TIFForiginal image

Fig 12. Silencing-related tomato genes differentially expressed in susceptible (S) and Ty-1 resistant (R) tomato plants mock-inoculated vs infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or a combination thereof (IL+S76) at 10 or 30 days post inoculation (dpi).

Charts plot the counts of Illumina mRNA-seq reads representing mRNAs of the RNA-dependent RNA polymerase (RDR) family genes RDRγ (Ty-1) and RDR1, the Dicer like (DCL) family genes DCL2b, DCL2d, DCL3 and DCL4, the Argonaute (AGO) family gene AGO2a and the Domain Rearranged Methyltransferase (DRM) family gene DRM1L in reads per million (RPM) of total mRNA-seq reads. The counts are for two biological replicates per each condition, with the standard error shown with a capped vertical line and the unfilled boxes positioned at the mean value levels and connected with solid lines (blue for S plants and red for R plants). The gene accession numbers (according to the annotated tomato reference genome ITAG4.1 available on Sol Genomics Network www.solgenomics.net) are given below the gene names.

https://doi.org/10.1371/journal.ppat.1011941.g012

IL and IS76 upregulate to similar levels the tomato genes implicated in siRNA biogenesis and function

RDR1 gene encoding an α-clade RDR, known to be induced by RNA virus and viroid infections or salicylic acid treatment of tomato plants (reviewed in [33]), was found to be upregulated to similar levels in IL-, IS76- and [IL+IS76]-infected S plants but not in R plants (Fig 12). Tomato RDR1 can synthesize complementary RNA on ssRNA and ssDNA substrates in vitro [35] and its homolog in Arabidopsis thaliana is implicated in biogenesis of RNA virus-derived and endogenous siRNAs [46–48]. Thus, RDR1 upregulation may contribute to the antiviral defense in S plants, but the disease caused by tomato begomoviruses indicates that such defense is evaded and particularly by IL and IS76. However, the fact that RDR1 gene expression was not altered in Ty-1 resistant plants (Fig 12), along with the above-mentioned findings in begomovirus-infected Arabidopsis wild-type vs RDR1/RDR2/RDR6 triple mutant plants, where no substantial difference in loads or size, polarity and hotspot profiles of viral siRNAs were observed [31], would suggest a minor contribution of RDR1 (and perhaps other α-clade RDRs) to defense against TYLCV, compared to RDRγ whose anti-TYLCV activity is only partially evaded by IS76.

Among four types of tomato DCLs (DCL1 to DCL4) involved in sRNA biogenesis, two of the four genes encoding DCL2 variants (DCL2b and DCL2d) and single genes encoding DCL3 and DCL4 were found to be upregulated to similar levels in S plants infected with IL, IS76 or both viruses, compared to mock control at 30 dpi (Fig 12). DCL2b, being upregulated in virus-infected S plants, showed unaltered expression in R plants. This tomato gene is known to mediate both biogenesis of 22 nt endogenous sRNAs and defence against RNA virus infection [49]. DCL2d, being upregulated most pronouncedly in S plants, was also upregulated in R plants at 30 dpi, albeit much less pronouncedly. Thus, both variants of DCL2 might contribute to the biogenesis of 22 nt viral siRNAs and anti-TYLCV defense in S plants. The fact that production rates of 22 nt siRNAs from both IL and IS76 were boosted in R plants, compared to S plants (Fig 9), suggests that lower expression levels of the DCL2 variants in R plants are sufficient for increased production of 22 nt siRNAs when RDRγ is boosting production of dsRNA substrates for DCLs.

DCL4 and DCL3, both showing the expression profiles similar to that of DCL2b (Fig 12), are known to mediate the biogenesis of respectively 21 nt [50] and 24 nt [51] endogenous sRNAs in tomato plants. These DCL genes likely mediate the biogenesis of TYLCV-derived 21 and 24 nt siRNAs, respectively, as shown for their homologues in A. thaliana infected with a bipartite begomovirus [30,31]. As argued above for DCL2, expression levels of DCL4 and DCL3 appear to be sufficient for robust siRNA biogenesis in R plants. It can also be suggested that upregulation of these and other silencing-related genes depends on viral loads, which are much higher in S plants than in R plants at 30 dpi. Consistent with this hypothesis, no upregulation of silencing-related genes was observed in virus-infected S plants at 10 dpi.

Among AGO family genes, only AGO2a was upregulated upon late virus infection (30 dpi) in S plants and to a lesser extent in R plants (Fig 12). IL, IS76 and their combination upregulated AGO2a expression to similar levels. AGO2a is known to be co-upregulated together with AGO1a, DCL2b and DCL2d upon RNA virus infection in tomato [52] and to confer defense against RNA viruses in N. benthamiana [52,53]. Consistent with upregulation of AGO2a gene expression by IL and IS76 at 30 dpi the proportion of AGO2-associated 5’A-sRNAs of the 21 and 22 nt classes derived from both viruses were increased between 10 dpi (15–21%) and 30 dpi (18–26%) on the expense of AGO1-associated 5’U-sRNAs of these size classes (65–73% vs 56–69%) (S2D Dataset).

Finally, DRM1L –one of the tomato paralogs of Domain-Rearranged Methyltransferase 2 (DRM2) that mediates 24 nt siRNA-directed DNA methylation in A. thaliana [28]–was upregulated upon virus infection in S plants and to a lesser extent in R plants (Fig 12). Like in the case of AGO2a and other silencing related genes, IL, IS76 or their combination up-regulated DRM1L expression to comparable levels. A role of DRM1L in cytosine methylation of plant and viral DNA in tomato remains to be investigated.

Concluding remarks

In this study, we began to elucidate the molecular mechanisms underlying the Ty-1 resistance-breaking phenotype and selective advantage of the recombinant virus IS76 as well as the strong negative impact of IS76 on its major parent TYLCV-IL in the Ty-1 plants expressing RDRγ, an RNA-dependent RNA polymerase of the γ-clade, whose function in antiviral RNAi is poorly understood. Compared to previous studies of TYLCV-IL in susceptible (S) tomato plants using transcriptomics and sRNAomics [37] and in Ty-1 resistant (R) vs S plants using only sRNAomics [38], we used both transcriptomics and sRNAomics for a comprehensive comparative study of TYLCL-IL, its recombinant derivative IS76 and combination thereof in both S and R tomato plants at early and late stages of infection. We found that, independent of virus identity, constitutive overexpression of RDRγ in R plants boosts the production rates of all three functional classes of viral siRNAs (21, 22 and 24 nt) from both strands of the entire virus genome and modulates DCL activities in favour of the 22 and 24 nt classes. Based on our in-depth analysis of sRNA and mRNA sequencing data for S and R plants, this is likely achieved by indirect and direct activities of RDRγ. In our current model, RDRγ might indirectly interfere with processing and polyadenylation of viral mRNAs, which would enhance readthrough transcription of circular dsDNA far beyond the poly(A) signals in both leftward and rightward directions. Both leftward and rightward readthrough transcription might span the IR and proceed even further, thus producing genome-length and longer transcripts. The leftward and rightward readthrough transcripts might pair to each other forming the dsRNAs preferentially processed by DCL3, or be converted by RDRγ to the dsRNAs preferentially processed by DCL2 and less preferentially by DCL3. Both in the presence and absence of functional RDRγ, readthrough transcripts might also pair to mature mRNAs of opposite polarities and the resulting dsRNAs would preferentially be processed by DCL4 and less preferentially by DCL2 and DCL3. This model for RDRγ-dependent and RDRγ-independent biogenesis of begomoviral siRNAs remains to be further validated using biochemical approaches.

Based on our comparative analysis of the production rates of viral siRNAs with those of viral mRNAs, IS76 appears to evade RDRγ activities and repressive siRNAs much better than IL. This is likely achieved by faster replication and accelerated transition to cell infection stages favouring the rightward transcription of viral silencing suppressor (V2) and CP genes. In our current model, V2 overexpression at earlier stages of cell infection might suppress transcriptional silencing of viral dsDNA and posttranscriptional silencing of viral mRNAs, while CP overexpression might facilitate encapsidation of viral ssDNA, followed by movement and reinfection of new cells. In mixed infection, more efficient replication and accelerated transition to overexpression of the rightward genes might provide the competitive advantage for IS76 observed in both S and R plants, while better evasion of repressive siRNAs might allow IS76 to keep high production rates of its mRNAs, even when RDRγ boosts the production rates of siRNAs from both viruses. In contrast, IL is less competitive in both S and R plants and might not evade the repressive activity of additional highly-abundant viral siRNAs derived from the transcription units and the IR sequences outside of the recombination region of IS76 which share 100% identity with the respective sequences of the IL genome. This outcompetes IL from co-infected R plants while IS76 reaches the accumulation levels of its DNA and siRNAs comparable to those in singly infected R plants.

It remains to be investigated how the alterations in the recombination region of IS76, which include 19 SNPs and 3 indels of 2, 3 and 9 nucleotides, might facilitate its replication and accelerate the transition from leftward to rightward transcription. These alterations surround the CAAT box of the core promoter driving rightward transcription and might also affect other cis-elements required for basal activity of this promoter and its transactivation by the viral protein TrAP/C2 [45]. Interestingly, a TATA-associated composite element (TACE) conserved in many genera of Geminiviridae, which often contains a TrAP-responsive conserved late element (CLE) or its variants with GC-rich sequences [45], is not affected by IS76 recombination, whereas an additional CLE located at the upstream position of IL is mutated in IS76 (S5 Dataset). Our results indicate that the mutation of the upstream CLE motif does not affect and even enhances the rightward promoter activity of IS76, suggesting that the TACE itself functions as a TrAP-responsive element in both TYLCV-IL and TYLCV-IS76. As CLE was proposed to bind an as-yet unidentified host transcriptional repressor protein, while TrAP interaction with this protein would de-repress the promoter activity [45], the removal of one of the two CLEs present in IL through IS76 recombination might facilitate TrAP-mediated de-repression of the rightward promoter.

The IR-based cis-elements regulating both replication efficiency and rightward-to-leftward transcription ratio might also be affected by cytosine methylation potentially directed by 24 nt viral siRNAs. In fact, a total number of cytosines on both strands of the recombination region is higher in IL (S5 Dataset). It remains to be investigated if those cytosines present in the recombination region of IL (but absent in IS76) are indeed methylated in a substantial fraction of viral circular dsDNA, thereby interfering with its transcription or replication, and whether or not RDRγ promotes cytosine methylation of viral dsDNA by boosting production of 24 nt siRNAs. Previously, cytosine methylation of TYLCV-IL DNA in susceptible tomato plants lacking the functional RDRγ was studied using bisulfite sequencing and the results revealed substantial methylation at CG, CHG and CHH sites within the entire IR as well as the V2 ORF and two parts of the C1 ORF flanking the C4 ORF, although no correlation was found between the methylation hotspots and the sRNA hotspots profiles [37]. It should be noted that the bisulfite sequencing approach used by Piedra-Aguilera et al. [37] could not distinguish between circular and linear forms of viral dsDNA. Circular dsDNA is used not only for Pol II transcription of viral mRNAs and rolling circle replication producing multiple copies of circular ssDNA, but also for recombination-dependent replication generating linear dsDNA of heterogeneous length [1]. These linear dsDNA molecules including concatemers with more-than-one copies of the viral genome might be transcribed by Pol II in both directions to produce dsRNA precursors of viral siRNAs and might also be targeted by 24 nt viral siRNAs for cytosine methylation as proposed earlier [32]. Thus, linear viral dsDNA would serve as a decoy diverting the RNAi machinery from actively transcribed circular dsDNA generating viral mRNAs. The proportion of heterogeneous linear dsDNA in total viral DNA we measured by qPCR might vary for both viruses depending on the time-point of infection or co-infection and the presence of functional RDRγ, which may contribute to the observed discrepancy between the total viral DNA loads and the viral siRNA loads (and hence their production rates estimated here).

Material and methods

Plant material

Solanum lycopersicum cultivar “Pristyla” carrying the Ty-1 resistance gene in a heterozygous state (Ty-1/ty-1) (Gautier Semences, France) and a nearly isogenic susceptible cultivar (ty-1/ty-1) [42] were used. Seeds were sown in a nursery pot and young seedlings were transplanted in individual pots and placed in a S3 containment growth chamber with 14 h light at 26°C and 10 h dark at 24°C. They were watered with a solution containing 15:10:30 NPK fertilizer and oligoelements.

Viral infectious clones

Two agroinfectious clones for TYLCV-IS76 [MA:SouG8:10] (GenBank accession number LN812978) and TYLCV-IL [RE:STG4:04] (GenBank accession number AM409201) were previously constructed using the binary vector pCAMBIA2300 and mobilized to the Agrobacterium tumefaciens strain C58 MP90 [41,42].

Agroinoculation and sampling

Agroinfiltration or co-agroinfiltration of 14-day old seedlings with agrobacteria preparations were performed as described in Belabess et al. [42]. Two groups of tomato plants were agroinfected, one for sampling at 10 days post inoculation (dpi) and another one for sampling at 30 dpi.

The following plants were inoculated for sampling at 10 dpi. For susceptible plants, 14 seedlings were agroinoculated with TYLCV-IL, 14 with TYLCV-IS76 and 16 were co-agroinoculated with TYLCV-IL and TYLCV-IS76. For Ty-1 resistant plants, 13 seedlings were agroinoculated with TYLCV-IL, 15 with TYLCV-IS76 and 18 were co-agroinoculated with TYLCV-IL and TYLCV-IS76.

The following plants were inoculated for sampling at 30 dpi. For susceptible plants, 8 seedlings were agroinoculated with TYLCV-IL, 8 with TYLCV-IS76 and 12 were co-agroinoculated with TYLCV-IL and TYLCV-IS76. For Ty-1 resistant plants, 7 seedlings were agroinoculated with TYLCV-IL, 8 with TYLCV-IS76 and 11 were co-agroinoculated with TYLCV-IL and TYLCV-IS76.

As negative controls, 3 seedlings of each cultivar and for each sampling date were mock-inoculated with a preparation of agrobacteria containing an empty pCAMBIA2300 plasmid.

At 10 and 30 dpi, youngest leaves were cut from the apex of each plant and immediately frozen in dry ice before storage at -80°C. The collected leaves were pooled in two biological replicates for each condition, based on quantitative (q)PCR analysis of viral DNA loads (see below).

For the 10 and 30 dpi sampling groups, the infection status of each plant was preliminary assessed at 18 and 30 dpi, respectively, by qPCR analysis of a pool of five 4-mm diameter leaf discs collected from the youngest leaf for which five leaflets were visible (one disc per leaflet). Total DNA from the leaf disc samples was extracted using the Dellaporta protocol [54] modified as follows. Leaf tissue was ground in 400 μL extraction buffer (100 mM Tris-HCl pH 8.0, 50 mM EDTA, 500 mM NaCl, 1% SDS, 0.5 mM Na2SO3, and 0.1 mg/ml RNase A), incubated at 65°C for 10 min and centrifuged (16,000 g 10 min). One volume of cold isopropanol was added to 300 μL of the supernatant and nucleic acids were precipitated by centrifugation (16,000 g, 20 min); the pellet was washed with 70% ethanol and then resuspended in 250 μL sterile bidistilled water and stored at -20°C.

Quantification of viral DNA loads

The load of viral DNA in each sample was measured by real-time qPCR. Each qPCR reaction was performed in a volume of 10 μL containing 2 μL of total DNA diluted 1:20, the LightCycler 480 SYBR Green I qPCR master mix (Roche, Germany), and primers. The primers for quantification of TYLCV-IL and TYLCV-IS76 [42] were added at concentrations 800 nM and 300 nM, respectively. The primers for the house-keeping tomato 25S rRNA gene [42], used as internal control for normalization of virus quantification with respect to plant DNA, were added at a concentration of 300 nM. Two technical repeats were performed for each DNA sample. The qPCR reactions were run in 384-well plates using the LightCycler 480 (Roche, Germany) with the following cycling conditions: 95°C for 10 min followed by 40 cycles each consisting of a denaturation step at 95°C for 10 sec, a hybridization step at 63°C for 40 sec for TYLCV-IL or for 20 sec for TYLCV-IS76, and an elongation step at 72°C for 15 sec. The qPCR results were analysed with the LinReg computer program [55], which calculates the initial concentration N0 for each sample, expressed in fluorescence units. This N0 value was normalized by the plant DNA concentration (N0 25S) and the amplicon size and then multiplied by 100.

Choice of leaf samples for pooling

The plant leaf samples collected at 10 and 30 dpi were pooled according to the viral load in each plant estimated by qPCR analysis of the leaf discs collected at 18 dpi and 30 dpi, respectively, and processed as described above. For each condition, two pools of the leaf samples with similar viral loads were assembled. For single infection, the plants with the most similar and representative (close to mean) viral loads were selected and homogenously divided in two batches; to do this, the samples were ranked from the sample exhibiting the lowest virus concentration to the one with the highest concentration and selected alternatively to form the two pools. In mixed infection with TYLCV-IL and TYLCV-IS76, the criteria used for homogeneity was the ratio of the viral loads between TYLCV-IS76 and TYLCV-IL. The plants with the most similar ratio were selected and divided homogenously after ranking and alternative selection as described above. Due to the contrasted weight of available leaves at 10 and 30 dpi, the leaf samples collected at 10 dpi were from 6 plants, while the leaf samples collected at 30 dpi were from 3 plants. Each leaf pool was ground in liquid nitrogen and the resulting powder was divided for DNA and RNA extraction and stored at -80°C until use.

DNA extraction for qPCR and Southern blot hybridization

Total DNA from the pooled leaf samples was extracted using a CTAB method of Doyle and Doyle [56]. A 0.5 ml aliquot of the CTAB buffer (100 mM Tris pH 8.0, 1.4 M NaCl, 50 mM EDTA pH 8.0, 2% CTAB, and 0.2% mercaptoethanol added before use) preheated at 60°C was added to ~0.1 g leaf tissue ground in liquid nitrogen. The mixture was incubated at 60°C for 1 hr and then centrifuged for 10 min at 10,000 rpm at room temperature (RT). The supernatant was mixed with equal volume of chloroform:isoamylalcohol (24:1). The mix was shaken for 3 min and then centrifuged for 10 min at 5,000 rpm at RT. The supernatant was transferred to a new tube and 0.66 volume of cold isopropanol was added. The tubes were stored at 4°C overnight and then centrifuged at 10,000 rpm for 10 min at RT. The supernatant was discarded and 0.5 ml of washing buffer (76% ethanol, 10 mM ammonium acetate) was added. The tubes were incubated for 20 min at RT and then centrifuged at 10,000 rpm for 5 min at RT. The supernatant was discarded and the pellet was air dried. Then, 100 μl of H2O and 1 μl of RNase A (10 mg/ml) were added and the mixture was incubated for 1 hr at 37°C. Two volumes of H2O were added, and the DNA was precipitated with 0.3 volumes of 3M sodium acetate and 2.5 volumes of cold absolute ethanol, followed by incubation for 15 min at -80°C and centrifugation at 10,000 rpm for 10 min at RT. The supernatant was discarded and the pellet was air dried at RT. The pellet was resuspended in 50 μl of H2O and the tubes stored at -20°C.

RNA extraction and validation for Illumina sequencing

Total RNA extraction from the pooled leaf samples was performed using a CTAB-LiCl method as described by Golyaev et al. [57]. The integrity of high and low molecular weight RNA was evaluated by electrophoresis on respectively a 1.2% agarose-formaldehyde gel, followed by EtBr staining, and a 15% polyacrylamide-urea gel, followed by blot hybridization with a plant miR160-specific probe, as described previously [58].

Illumina sequencing and bioinformatic analysis of viral mRNAs and viral sRNAs

Illumina sequencing was performed at Fasteris (www.fasteris.com) using the same total RNA extracts for library preparations with the Illumina stranded mRNA and the Illumina TruSeq small RNA protocols.

The mRNA libraries were multiplexed and sequenced in two flowcells of NovaSeq 6000, one flowcell with the samples from the plants collected at 10 dpi and the other one with the samples from the plants collected at 30 dpi, yielding 25’921’135 to 45’871’986 and 20’213’119 to 40’363’645 100 nt paired-end reads, respectively, and Q30 = 89.36 to 91.87 and Q30 = 89.85 to 92.00, respectively.

The sRNA libraries were multiplexed and sequenced in two flowcells of NovaSeq 6000. The first flowcell with the samples from the plants collected at 10 dpi were sequenced with 50 nt paired-end reads yielding 28’814’844 to 54’420’088 reads with Q30 = 96.46 to 96.83 for the forward read used for our follow-up analysis. The second flowcell with the samples from the plants collected at 30 dpi were sequenced with 75 nt single-end reads, yielding 37’719’583 to 50’968’605 reads with Q30 = 96.86 to 97.43.

In all cases, the libraries were de-multiplexed, followed by adapter trimming with Trimmomatic. The resulting reads were mapped using Burrow-Wheeler Aligner (BWA) 0.7.12 [59] onto the reference sequences of TYLCV-IL (AM409201) and TYLCV-IS76 (LN812978) with and without mismatches. Mapped viral reads were sorted by polarity (forward, reverse) and, in the case of sRNAs, also by size (from 15 to 34 nts) and 5’-terminal nucleotide identity (5’A, 5’U, 5’G, 5’C), and then counted (S1 Dataset for mRNA counts and S2 Dataset for sRNA counts). Single-nucleotide resolution maps of viral mRNA and sRNA reads (S3 and S4 Datasets, respectively) were generated using MISIS-2 [60].

To quantify viral mRNA and viral sRNA loads for each virus (or its selected region or strand), we used the reads aligned without mismatches. The viral read counts in each library were normalized in reads per millions (RPM) of total (viral + plant) reads.

In mixed infection, the number of reads derived from each virus (or its selected region or strand) was calculated using reads aligned without mismatches at the SNP positions present along the genome of TYLCV-IS76 and TYLCV-IL. Noteworthy, we purposely used the wild type TYLCV-IS76 infectious clone but not the laboratory generated one, TYLCV-IS76’, both of which having the same competitiveness properties [42]. Unlike TYLCV-IS76’, the wild type recombinant can be distinguished from TYLCV-IL not only by 19 SNPs and three indels of 2, 3 and 9 nts in the recombination region (between the replication origin at position 1 and the recombination breakpoint at position 84 of IL or position 76 of IS76) but also by other SNPs scattered along the viral genome (17 SNPs in the V2-V1 transcription unit, 3 SNPs in the C2-C3 transcription unit, 7 SNPs in the C1-C4 transcription unit, 2 SNPs in the intergenic region upstream of the replication origin and 1 SNP in the intergenic region downstream the recombination breakpoint (S5 Dataset). Thus, the number of reads derived from each virus (or its selected region or strand) was counted at each SNP using MISIS-2 [60] and a percentage of reads derived from each virus (or its selected region) was calculated. The average percentage at all SNPs of the viral genome (or its selected region) was applied on all parts of the genome (or its selected region) that contain no SNPs to estimate the number of reads derived from the entire genome of each virus (or its selected region) or each strand of the viral genome (or its selected region).

For viral mRNA, a production rate for each viral mRNA was calculated by dividing the mRNA counts in RPM by the DNA loads (measured by qPCR) and by the length of the mRNA from cap to poly(A) site.

For viral sRNAs, the production rate of sRNAs in the size range from 20 to 25 nts derived from the viral genome (or its selected region) and each strand of the viral genome (or its selected region) was calculated dividing the sRNA counts in RPM by the viral DNA load and, in the case of selected regions of the virus genome, by the length of each region.

Mean values with standard deviations of all those loads and production rates were calculated for the two biological replicates per each condition.

For profiling the expression of silencing-related genes, the mRNA-seq reads were mapped with BWA on the tomato reference genome ITAG4.1 available on Sol Genomics Network (www.solgenomics.net). Read were counted with the HTSeq count tool [61]. The data were then analyzed using DicoExpress [62]. Only the genes presenting a CPM (count per million) greater than or equal to 5 for at least half of the conditions were kept for further analysis. The read counts of the selected genes were normalized using the TMM method of the EdgeR package [63]. The differential analysis was performed by applying a negative binomial generalized linear model (GLM) with the EdgeR package. A gene was considered to be differentially expressed if the FDR (false discovery rate) was less than or equal to 0.05.

Southern blot hybridization analysis with strand-specific probes

Samples of total DNA (0.5 μg) were resolved in a 0.8% agarose gel. The gel was stained with ethidium bromide for 15 min and photographed under UV light. Following denaturation and neutralization steps, DNA was transferred by capillary blotting to Hybond N+ membrane (GE Healthcare/Amersham) as described in the Hybond N+ manual. The transferred DNA was fixed to the membrane by using an UV-crosslinker (Stratagene). Blot hybridization was performed as described previously for small RNA analysis [64]. Briefly, the blot membrane was sequentially hybridized at 35°C overnight in UltraHyb-oligo buffer (Ambion) with short DNA oligonucleotides end-labelled with 32P gamma ATP by T4 polynucleotide kinase and purified through MicroSpin G-25 columns (GE Healthcare), following the manufacturers’ recommendations. The first probe (5’-ATCATTTCCACGCCCGTCTCGAAGGTTCGCCGA) hybridized to the complementary strand of both TYLCV-IL and TYLCV-IS76, while the second probe (5’-AAGTTCAGCCTTCGGCGAACCTTCGAGACGGGC) hybridized to the virion strand of both TYLCV-IL and TYLCV-IS76. The membrane was washed 3 times with 2X SCC, 0.5% SDS for 30 min at 35°C, and then exposed for 3 to 14 days to a phosphor screen, followed by scanning in a PhosphorImager (GE Healthcare). For the second hybridization the membrane was stripped with 0.5X SSC, 0.5% SDS for 30 min at 80°C and then with 0.1X SSC, 0.5% SDS for 30 min at 80°C. The four blot membranes shown in S2 Fig were hybridized and exposed simultaneously.

Supporting information

Supplementary Methods, Results and Discussion for S2 Fig.

Showing 1/10: ppat.1011941.s001.docx

Skip to figshare navigation

S1 Text. For viral DNA methylation analysis, total plant DNA was digested with cytosine methylation-dependent enzyme McrBC (NewEngland Biolab) in a total volume of 25 μL containing 2.5 μL reaction buffer, 0.25 μL albumine, 0.25 μL GTP, 0.5 μg total DNA and 15 U McrBC. The reaction was carried at 37oC for 1 hr, followed by enzyme inactivation at 65oC for 25 min. As a positive control, 0.2 μg plasmid containing a single methylated cytosine (supplied in the NewEngland Biolab McrBC kit) was mixed with 0.4 μg total plant DNA and digested in parallel as describe above. For each sample, a second aliquot of total DNA (0.5μg) was treated in parallel under the above conditions but without McrBC. Both McrBC-digested and undigested (buffer-incubated) total DNA samples were loaded side- by-side on the 0.8 % agarose gel for Southern blot hybridization analysis.Southern blot hybridization analysis with strand-specific probes revealed that circular dsDNA of IL and IS76 is resistant to McrBC digestion in S plants at 30 dpi where this form of viral DNA is above the detection threshold for both viruses. The results obtained for R plants where circular dsDNA of IS76 (but not IL) is detectable are not conclusive, although it appears to be less resistant to McrBC. However, we cannot exclude unspecific activity of McrBC digesting non-methylated dsDNA under our conditions, because McrBC was unexpectedly able to digest viral ssDNA that is produced by rolling circle replication and is not supposed to be a substrate for cytosine methylation directed by siRNAs.

1 / 10

Download

figshare

S1 Text. Supplementary Methods, Results and Discussion for S2 Fig.

For viral DNA methylation analysis, total plant DNA was digested with cytosine methylation-dependent enzyme McrBC (NewEngland Biolab) in a total volume of 25 μL containing 2.5 μL reaction buffer, 0.25 μL albumine, 0.25 μL GTP, 0.5 μg total DNA and 15 U McrBC. The reaction was carried at 37°C for 1 hr, followed by enzyme inactivation at 65°C for 25 min. As a positive control, 0.2 μg plasmid containing a single methylated cytosine (supplied in the NewEngland Biolab McrBC kit) was mixed with 0.4 μg total plant DNA and digested in parallel as describe above. For each sample, a second aliquot of total DNA (0.5 μg) was treated in parallel under the above conditions but without McrBC. Both McrBC-digested and undigested (buffer-incubated) total DNA samples were loaded side-by-side on the 0.8% agarose gel for Southern blot hybridization analysis. Southern blot hybridization analysis with strand-specific probes revealed that circular dsDNA of IL and IS76 is resistant to McrBC digestion in S plants at 30 dpi where this form of viral DNA is above the detection threshold for both viruses. The results obtained for R plants where circular dsDNA of IS76 (but not IL) is detectable are not conclusive, although it appears to be less resistant to McrBC. However, we cannot exclude unspecific activity of McrBC digesting non-methylated dsDNA under our conditions, because McrBC was unexpectedly able to digest viral ssDNA that is produced by rolling circle replication and is not supposed to be a substrate for cytosine methylation directed by siRNAs.

https://doi.org/10.1371/journal.ppat.1011941.s001

(DOCX)

S1 Fig.

Counts of viral mRNA reads in susceptible (S) and Ty-1 resistant (R) tomato plants infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or a combination thereof (IL+S76) at 10 (A) and 30 (B) days post inoculation (dpi). Illumina mRNA-seq reads representing the virion (rightward) and complementary (leftward) strands of the Pol II transcription units (V2-V1, C1-C4, C2-C3) and two parts of the intergenic region (IR1 and IR2) were counted in reads per million (RPM) of total (plant + viral) mRNA reads and the resulting counts plotted as bar graphs. Blue and red bars represent the rightward and leftward reads, respectively. In all cases, the counts are for two biological replicates per each condition, with the standard error shown with a capped vertical line and the mean value indicated above.

https://doi.org/10.1371/journal.ppat.1011941.s002

(PDF)

S2 Fig. Southern blot hybridization analysis of McrBC-treated and control non-treated DNA from susceptible (S) and Ty-1 resistant (R) plants mock-inoculated or infected with IL, IS76 or IL+IS76 at 10 and 30 days post-inoculation (dpi).

Total DNA extracted from tomato plants was digested with McrBC or incubated in digestion buffer without McrBC and then separated on 1% agarose gel (4 separate gels for S and R plants at 10 and 30 dpi, respectively). As control, plasmid DNA with one methylated cytosine site was spiked into total DNA from the R plant infected with IS76 at 30 dpi and loaded on one of the 4 gels. Following electrophoresis, the gels were stained with ethidium bromide and then DNA was transferred to nylon membranes by blotting and denatured. The membranes were successively hybridized with 32P-labelled DNA oligonucleotide probes specific for the complementary and virion strands of viral DNA and, following each hybridization, exposed together to a phosphor screen for 1 hour to 2 weeks and scanned on a PosphorImager. Note that after the first hybridization, the membranes were stripped to remove the first probe and then hybridized with the second probe. Pictures of EtBr-stained gels of the samples from 10 dpi and 30 dpi are shown in panels (A) and (C), respectively, while the respective membrane scans are shown in panels (B) and (D). Positions of plant genomic DNA (gDNA), undigested and digested plasmid DNA, viral circular double-stranded DNA (dsDNA) and viral circular single-stranded DNA (ss) are indicated.

https://doi.org/10.1371/journal.ppat.1011941.s003

(PDF)

S3 Fig.

Counts of viral sRNAs in susceptible (S) and Ty-1 resistant (R) tomato plants infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or a combination thereof (IL+S76) at 10 (A) and 30 (B) days post inoculation (dpi). Illumina sRNA-seq reads representing the virion (rightward) and complementary (leftward) strands of the Pol II transcription units (V2-V1, C1-C4, C2-C3) and two parts of the intergenic region (IR1 and IR2) were counted in reads per million (RPM) of total (plant + viral) sRNA reads and the resulting counts plotted as bar graphs. Blue and red bars represent the rightward and leftward reads, respectively. In all cases, the counts are for two biological replicates per each condition, with the standard error shown with a capped vertical line and the mean value indicated above.

https://doi.org/10.1371/journal.ppat.1011941.s004

(PDF)

S4 Fig.

Size profiles of viral sRNAs derived from the transcription units V2-V1 (A), C1-C4 (B) and C2-C3 (C) in susceptible (S) and Ty-1 resistant (R) tomato plants infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or a combination thereof (IL+S76) at 10 and 30 days post inoculation (dpi). Illumina sRNA-seq reads in the size range from 20 to 25 nts mapped to the viral genome the virion (rightward) and complementary (leftward) strands of each transcription unit were counted and percentages (%) of 6 individual size-classes in the total 20–25 nt viral reads (set to 100%) were calculated and plotted as bar graphs, with blue and red bars representing rightward and leftward strands, respectively. In all panels, the percentages are for two biological replicates per each condition, with the standard error shown with a capped vertical line and the mean value indicated above.

https://doi.org/10.1371/journal.ppat.1011941.s005

(PDF)

S1 Dataset. Counts of Illumina mRNA-seq reads from susceptible (S) and Ty-1 resistant (R) tomato plants mock-inoculated or infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or their combination (IL+IS76) at 10 and 30 days post inoculation (dpi).

The Illumina 100 nt paired-end reads from each library (two biological replicates per condition: pool 1 and pool 2) were mapped without (A) or with (B) mismatches to the reference sequences of the Solanum lycopersicum genome (nuclear, chloroplast and mitochondrion) and the viral genomes (IL and IS76), sorted by polarity (forward, reverse, total) and counted. The counts of plant and viral reads mapped without mismatches were then normalized per million of total reads (RPM) in each library (C).

https://doi.org/10.1371/journal.ppat.1011941.s006

(XLSX)

S2 Dataset. Counts of Illumina small RNA-seq reads from susceptible (S) and Ty-1 resistant (R) tomato plants mock-inoculated or infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or their combination (IL+IS76) at 10 and 30 days post inoculation (dpi).

The Illumina 15–34 nt reads from each library (two biological replicates per condition: pool 1 and pool 2) were mapped without (A) or with (B) mismatches to the reference sequences of the Solanum lycopersicum genome (nuclear, chloroplast and mitochondrion) and the viral genomes (IL and IS76), sorted by size (15 nt through 34 nt) and polarity (forward, reverse, total) and then counted. The counts of plant and viral reads mapped without mismatches were then normalized per million of total reads (RPM) in each library (C) and were also sorted by 5’-terminal nucleotide identity (5’A, 5’C, 5’G, 5’U) and then counted in percentage of total (D).

https://doi.org/10.1371/journal.ppat.1011941.s007

(XLSX)

S3 Dataset. Single nucleotide resolution maps of Illumina mRNA-seq reads representing viral transcripts from susceptible (S) and Ty-1 resistant (R) tomato plants infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or their combination (IL+IS76) at 10 and 30 days post inoculation (dpi).

For each condition, Illumina 100 nt paired-end reads of the two biological replicates (1 and 2) were mapped onto the reference sequences of IL and IS76 using BWA and the resulting BAM files were analysed by MISIS-2 (Seguin et al. 2016 [60]) to generate tables of reads mapped to each reference sequence with zero mismatches. The reference sequences were extended at the 3-end by 99 nts from the 5’-terminal sequence to allow for mapping RNAs derived from the circular viral genome including the first and last nucleotide of the linear reference. In each table, the first column gives nucleotide positions of the corresponding viral genome sequence. In the next columns, the positions of 5′-terminal nucleotide of sense RNAs and 3′-terminal nucleotide of antisense RNAs along the reference sequence are indicated, and the read counts are given for each RNA mapped with zero mismatches to the forward (rightward) strand (columns fwd1 and fwd2) and the reverse (leftward) strand (columns rev1 and rev2), along with the total counts of reads mapped at the respective positions of the forward (rightward) and reverse (leftward) strands in the two replicates divided by 2 (i.e., average counts). In each table file on the right side, histograms of the average counts of rightward and leftward reads are inserted with the rightward reads colored in blue and the leftward reads colored in red. In the case of mixed infections (IL+IS76), the number of reads derived from each virus was counted at each SNP using MISIS-2 (Seguin et al. 2006 [60]) and a percentage of reads derived from each virus (or its selected region) was calculated. The average percentage at all SNPs of the viral genome (or its selected region) was applied on all parts of the genome (or its selected region) that contain no SNPs to estimate the number of reads derived from the entire genome of each virus (or its selected region) or each strand of the viral genome (or its selected region).

https://doi.org/10.1371/journal.ppat.1011941.s008

(XLSX)

S4 Dataset. Single-nucleotide resolution maps of viral 20–25 nt small (s)RNAs in susceptible (S) and Ty-1 resistant (R) tomato plants infected with TYLCV-IL, its recombinant derivative TYLCV-IS76 or their combination (IL+IS76) at 10 and 30 days post inoculation (dpi).

For each condition, Illumina 20–25 nt reads of the two biological replicates were combined and mapped onto the reference sequences of IL and IS76 using BWA and the resulting BAM files were analysed by MISIS-2 (Seguin et al. 2016 [60]) to generate tables of reads mapped to each reference sequence with zero mismatches and sorted by size and polarity. The reference sequences were extended at the 3’-end by 33 nts from the 5’-terminal sequence to allow for mapping sRNAs derived from the circular viral genome (at the junction of the first and last nucleotide of the linear reference). The counts of reads mapped to the extended sequence were then added to the 5’-sequence. In each table, the first column gives nucleotide positions of the corresponding viral genome sequence. In the next columns, the positions of 5′-terminal nucleotide of sense sRNAs and 3′-terminal nucleotide of antisense siRNAs along the reference sequence are indicated, and the read counts are given for each sRNA of 20-, 21-, 22-, 23-, 24- and 25-nt classes mapped with zero mismatches to the forward (rigthward) strand (columns 20 rightward, 21 rightward, 22 rightward, 23 rightward, 24 rightward, 25 rightward) and the reverse (leftward) strand (columns 20 leftward, 21 leftward, 22 leftward, 23 leftward, 24 leftward, 25 leftward), along with the total counts of 20–25 nt sRNAs mapped on the forward (rightward) and reverse (leftward) strands. In each table file on the right side, histograms of three major size-classes of siRNAs (21, 22, and 24 nt rightward and leftward reads) are inserted with the rightward reads colored in blue and the leftward reads colored in red. In the case of mixed infections (IL+IS76), the number of reads derived from each virus was counted at each SNP using MISIS-2 (Seguin et al. 2006 [60]) and a percentage of reads derived from each virus (or its selected region) was calculated. The average percentage at all SNPs of the viral genome (or its selected region) was applied on all parts of the genome (or its selected region) that contain no SNPs to estimate the number of reads derived from the entire genome of each virus (or its selected region) or each strand of the viral genome (or its selected region).

https://doi.org/10.1371/journal.ppat.1011941.s009

(XLSX)

S5 Dataset. Reference sequences of the viral genome IL and IS76 and their pairwise alignment.

The start and stop codons of viral ORFs are coloured in red and underlined, the CAAT and TATA-boxes of the promoters coloured in brick red, the TATA-associated composite element (TACE) and conserved late elements (CLE) highlighted in green and cyan, respectively, the iterons highlighted in grey and SNPs and indels highlighted in yellow.

https://doi.org/10.1371/journal.ppat.1011941.s010

(PDF)

Acknowledgments

We are thankful to Martine Granier and Sophie Le Blaye for technical assistance.

References

1.Geminiviruses Jeske H. In: de Villiers EM, Hausen HZ. Editors. TT Viruses. Current Topics in Microbiology and Immunology, vol 331. Springer, Berlin, Heidelberg; 2009. pp. 185–226. https://doi.org/10.1007/978-3-540-70972-5_11.

2.Hanley-Bowdoin L, Bejarano ER, Robertson D, Mansoor S. Geminiviruses: masters at redirecting and reprogramming plant processes. Nat Rev Microbiol. 2013;11:777–88. pmid:24100361

View ArticlePubMed/NCBIGoogle Scholar

3.Palanichelvam K, Kunik T, Citovsky V, Gafni Y. The capsid protein of tomato yellow leaf curl virus binds cooperatively to single-stranded DNA. J Gen Virol. 1998;79:2829–2833. pmid:9820160

View ArticlePubMed/NCBIGoogle Scholar

4.Hallan V, Gafni Y. Tomato yellow leaf curl virus (TYLCV) capsid protein (CP) subunit interactions: implications for viral assembly. Arch Virol. 2001;146:1765–73. pmid:11699961

View ArticlePubMed/NCBIGoogle Scholar

5.Rojas MR, Jiang H, Salati R, Xoconostle-Cázares B, Sudarshana MR, Lucas WJ, et al. Functional analysis of proteins involved in movement of the monopartite begomovirus, Tomato yellow leaf curl virus. Virology. 2001;291:110–25. pmid:11878881

View ArticlePubMed/NCBIGoogle Scholar

6.Luna AP, Morilla G, Voinnet O, Bejarano ER. Functional analysis of gene-silencing suppressors from tomato yellow leaf curl disease viruses. Mol Plant Microbe Interact. 2012;25:1294–306. pmid:22712505

View ArticlePubMed/NCBIGoogle Scholar

7.Zrachya A, Glick E, Levy Y, Arazi T, Citovsky V, Gafni Y. Suppressor of RNA silencing encoded by Tomato yellow leaf curl virus-Israel. Virology. 2007;358:159–65. pmid:16979684

View ArticlePubMed/NCBIGoogle Scholar

8.Glick E, Zrachya A, Levy Y, Mett A, Gidoni D, Belausov E, et al. Interaction with host SGS3 is required for suppression of RNA silencing by tomato yellow leaf curl virus V2 protein. Proc Natl Acad Sci U S A. 2008;105:157–61. pmid:18165314

View ArticlePubMed/NCBIGoogle Scholar

9.Wang B, Li F, Huang C, Yang X, Qian Y, Xie Y, et al. V2 of tomato yellow leaf curl virus can suppress methylation-mediated transcriptional gene silencing in plants. J Gen Virol. 2014;95: 225–230. pmid:24187017

View ArticlePubMed/NCBIGoogle Scholar

10.Wang B, Yang X, Wang Y, Xie Y, Zhou X. Tomato Yellow Leaf Curl Virus V2 Interacts with Host Histone Deacetylase 6 To Suppress Methylation-Mediated Transcriptional Gene Silencing in Plants. J Virol. 2018;92:e00036–18. pmid:29950418

View ArticlePubMed/NCBIGoogle Scholar

11.Wang L, Ding Y, He L, Zhang G, Zhu JK, Lozano-Duran R. A virus-encoded protein suppresses methylation of the viral genome through its interaction with AGO4 in the Cajal body. Elife. 2020;9:e55542. pmid:33064077

View ArticlePubMed/NCBIGoogle Scholar

12.Desbiez C, David C, Mettouchi A, Laufs J, Gronenborn B. Rep protein of tomato yellow leaf curl geminivirus has an ATPase activity required for viral DNA replication. Proc Natl Acad Sci U S A. 1995;92:5640–4. pmid:7777563

View ArticlePubMed/NCBIGoogle Scholar

13.Noris E, Jupin I, Accotto GP, Gronenborn B. DNA-binding activity of the C2 protein of tomato yellow leaf curl geminivirus. Virology. 1996;217:607–12. pmid:8610454

View ArticlePubMed/NCBIGoogle Scholar

14.Dong X, van Wezel R, Stanley J, Hong Y. Functional characterization of the nuclear localization signal for a suppressor of posttranscriptional gene silencing. J Virol. 2003;77:7026–33. pmid:12768021

View ArticlePubMed/NCBIGoogle Scholar

15.Settlage SB, See RG, Hanley-Bowdoin L. Geminivirus C3 protein: replication enhancement and protein interactions. J Virol. 2005;79(15):9885–95. pmid:16014949

View ArticlePubMed/NCBIGoogle Scholar

16.Jupin I, De Kouchkovsky F, Jouanneau F, Gronenborn B. Movement of tomato yellow leaf curl geminivirus (TYLCV): involvement of the protein encoded by ORF C4. Virology. 1994;204:82–90. pmid:8091687

View ArticlePubMed/NCBIGoogle Scholar

17.Xie Y, Zhao L, Jiao X, Jiang T, Gong H, Wang B, Briddon RW, Zhou X. A recombinant begomovirus resulting from exchange of the C4 gene. J Gen Virol. 2013;94(Pt 8):1896–1907. pmid:23720217

View ArticlePubMed/NCBIGoogle Scholar

18.Padmanabhan C, Zheng Y, Shamimuzzaman M, Wilson JR, Gilliard A, Fei Z, et al. The tomato yellow leaf curl virus C4 protein alters the expression of plant developmental genes correlating to leaf upward cupping phenotype in tomato. PLoS One. 2022;17:e0257936. pmid:35551312

View ArticlePubMed/NCBIGoogle Scholar

19.Mullineaux PM, Rigden JE, Dry IB, Krake LR, Rezaian MA. Mapping of the polycistronic RNAs of tomato leaf curl geminivirus. Virology, 1993; 193:414–23. pmid:8438578

View ArticlePubMed/NCBIGoogle Scholar

20.Frischmuth S, Frischmuth T, Jeske H. Transcript mapping of Abutilon mosaic virus, a geminivirus. Virology. 1991;185:596–604. pmid:1962440

View ArticlePubMed/NCBIGoogle Scholar

21.Shivaprasad PV, Akbergenov R, Trinks D, Rajeswaran R, Veluthambi K, Hohn T, et al. Promoters, transcripts, and regulatory proteins of Mungbean yellow mosaic geminivirus. J Virol. 2005;79:8149–63. pmid:15956560

View ArticlePubMed/NCBIGoogle Scholar

22.Eagle PA, Orozco BM, Hanley-Bowdoin L. 1994. A DNA sequence required for geminivirus replication also mediates transcriptional regulation. Plant Cell 6: 1157–70. pmid:7919985

View ArticlePubMed/NCBIGoogle Scholar

23.Sunter G, Bisaro DM. Transactivation of geminivirus AR1 and BR1 gene expression by the viral AL2 gene product occurs at the level of transcription. Plant Cell. 1992;4:1321–31. pmid:1446172

View ArticlePubMed/NCBIGoogle Scholar

24.Zakay Y, Navot N, Zeidan M, Kedar N, Rabinowitch H, Czosnek H, et al. Screening Lycopersicon accessions for resistance to tomato yellow leaf curl virus: Presence of viral DNA and symptom development. Plant Disease. 1991; 75:279–281. https://www.apsnet.org/publications/plantdisease/backissues/Documents/1991Abstracts/PD_75_279.htm.

View ArticleGoogle Scholar

25.Zamir D, Ekstein-Michelson I, Zakay Y, Navot N, Zeidan M, Sarfatti M, et al. Mapping and introgression of a tomato yellow leaf curl virus tolerance gene, TY-1. Theoretical and Applied Genetics. 1994;88:141–146. pmid:24185918

View ArticlePubMed/NCBIGoogle Scholar

26.El-Sappah AH, Qi S, Soaud SA, Huang Q, Saleh AM, Abourehab MAS, et al. Natural resistance of tomato plants to Tomato yellow leaf curl virus. Front Plant Sci. 2022;13:1081549. pmid:36600922

View ArticlePubMed/NCBIGoogle Scholar

27.Verlaan MG, Hutton SF, Ibrahem RM, Kormelink R, Visser RG, Scott JW, et al. The Tomato Yellow Leaf Curl Virus resistance genes Ty-1 and Ty-3 are allelic and code for DFDGD-class RNA-dependent RNA polymerases. PLoS Genet. 2013;9:e1003399. pmid:23555305

View ArticlePubMed/NCBIGoogle Scholar

28.Borges F, Martienssen RA. The expanding world of small RNAs in plants. Nat Rev Mol Cell Biol. 2015;16:727–41. pmid:26530390

View ArticlePubMed/NCBIGoogle Scholar

29.Fang X, Qi Y. RNAi in Plants: An Argonaute-Centered View. Plant Cell. 2016;28:272–85. pmid:26869699

View ArticlePubMed/NCBIGoogle Scholar

30.Blevins T, Rajeswaran R, Shivaprasad PV, Beknazariants D, Si-Ammour A, Park HS, et al. Four plant Dicers mediate viral small RNA biogenesis and DNA virus induced silencing. Nucleic Acids Res. 2006;34:6233–46. pmid:17090584

View ArticlePubMed/NCBIGoogle Scholar

31.Aregger M, Borah BK, Seguin J, Rajeswaran R, Gubaeva EG, Zvereva AS, et al. Primary and secondary siRNAs in geminivirus-induced gene silencing. PLoS Pathog. 2012;8:e1002941. pmid:23028332

View ArticlePubMed/NCBIGoogle Scholar

32.Pooggin MM. How can plant DNA viruses evade siRNA-directed DNA methylation and silencing? Int J Mol Sci. 2013;14:15233–59. pmid:23887650

View ArticlePubMed/NCBIGoogle Scholar

33.Wassenegger M, Krczal G. Nomenclature and functions of RNA-directed RNA polymerases. Trends Plant Sci. 2006;11:142–51. pmid:16473542

View ArticlePubMed/NCBIGoogle Scholar

34.Jha V, Narjala A, Basu D, Sujith TN, Pachamuthu K, Chenna S, et al. Essential role of γ-clade RNA-dependent RNA polymerases in rice development and yield-related traits is linked to their atypical polymerase activities regulating specific genomic regions. New Phytol. 2021;232(4):1674–1691. pmid:34449900

View ArticlePubMed/NCBIGoogle Scholar

35.Schiebel W, Haas B, Marinkovi? S, Klanner A, S?nger HL. RNA-directed RNA polymerase from tomato leaves. II. Catalytic in vitro properties. J Biol Chem. 1993; 268: 11858–67. https://doi.org/10.1016/S0021-9258(19)50279-4. pmid:7685023

View ArticlePubMed/NCBIGoogle Scholar

36.Fuentes A, Carlos N, Ruiz Y, Callard D, Sánchez Y, Ochagavía ME, et al. Field Trial and Molecular Characterization of RNAi-Transgenic Tomato Plants That Exhibit Resistance to Tomato Yellow Leaf Curl Geminivirus. Mol Plant Microbe Interact. 2016;29:197–209. pmid:26713353

View ArticlePubMed/NCBIGoogle Scholar

37.Piedra-Aguilera á， Jiao C， Luna AP， Villanueva F， Dabad M， Esteve-Codina A， et al.番茄黄叶卷曲病毒（TYLCV）转录组、sRNA组和甲基组的集成单碱基分辨率图谱。科学代表 2019;9：2863。PMID：30814535

查看文章PubMed/NCBI公司Google 学术搜索

38.Voorburg CM， Bai Y， Kormelink R. 番茄黄叶卷曲病毒感染后易感和抗性 Ty-1 编码番茄植株的小 RNA 分析。前沿植物科学 2021;12：757165.PMID：34868151

查看文章PubMed/NCBI公司Google 学术搜索

39.Butterbach P、Verlaan MG、Dullemans A、Lohuis D、Visser RG、Bai Y 等。Ty-1 对番茄黄叶卷曲病毒的抗性涉及病毒基因组胞嘧啶甲基化增加，并受到黄瓜花叶病毒感染的影响。美国国家科学院院刊，2014 年;111:12942–7.PMID：25136118

查看文章PubMed/NCBI公司Google 学术搜索

40.Voorburg CM， Yan Z， Bergua-Vidal M， Wolters AA， Bai Y， Kormelink R. Ty-1，一种针对双子座病毒的通用抗性基因，因 β 卫星的共同复制而受到损害。Mol 植物病理。2020年21月;2(160):172–2019.Epub 格式 22 31756021 月 <>.PMID：<>

查看文章PubMed/NCBI公司Google 学术搜索

41.Belabess Z、Dallot S、El-Montaser S、Granier M、Majde M、Tahiri A 等。监测番茄黄叶卷曲病毒非经典重组体的出现动态及其亲本病毒在番茄中的置换。病毒学。2015;486:291–306.PMID：26519598

查看文章PubMed/NCBI公司Google 学术搜索

42.贝拉贝斯 Z、彼得施密特 M、格拉尼尔 M、塔希里 A、布伦扎尔 A、乌尔比诺 C。在摩洛哥南部取代其亲本病毒的非典型番茄黄叶卷曲病毒重组在实验条件下表现出高度选择性优势。J Gen Virol。2016;97:3433–3445.PMID：27902403

查看文章PubMed/NCBI公司Google 学术搜索

43.Jammes M， Urbino C， Diouf MB， Peterschmitt M. 改进侵袭性重组番茄黄叶卷曲病毒 -IS76 的出现情景。病毒学。2023;578:71–80.PMID：36473279

查看文章PubMed/NCBI公司Google 学术搜索

44.鉴定原生质体中番茄金色花叶病毒外壳蛋白启动子激活所需的最小序列。病毒学。2003;305:452–62.PMID：12573590

查看文章PubMed/NCBI公司Google 学术搜索

45.Cantú-Iris M、Pastor-Palacios G、Mauricio-Castillo JA、Ba?uelos-Hernández B、Avalos-Calleros JA、Juárez-Reyes A 等人。对一种新的begomovirus的分析揭示了几个Geminiviridae属的CP基因启动子中保守的复合元件：理解晚期基因复杂调控的线索。PLoS 一。2019;14：e0210485。PMID：30673741

查看文章PubMed/NCBI公司Google 学术搜索

46.Garcia-Ruiz H、Takeda A、Chapman EJ、Sullivan CM、Fahlgren N、Brempelis KJ 等。拟南芥 RNA 依赖性 RNA 聚合酶和 dicer 样蛋白在萝卜花叶病毒感染期间的抗病毒防御和小干扰 RNA 生物发生中的作用。植物细胞。2010;22:481–96.PMID：20190077

查看文章PubMed/NCBI公司Google 学术搜索

47.Wang XB， Wu Q， Ito T， Cillo F， Li WX， Chen X， et al. RNAi 介导的病毒免疫需要扩增拟南芥中病毒衍生的 siRNA。美国国家科学院院刊，2010 年;107:484–9.PMID：19966292

查看文章PubMed/NCBI公司Google 学术搜索

48.曹明， 杜萍， 王旭， 于玉清， 邱永华， 李伟， Gal-On A， 等.病毒感染通过拟南芥中一类不同的内源性siRNA引发宿主基因的广泛沉默。美国国家科学院院刊，2014 年;111:14613–8.PMID：25201959

查看文章PubMed/NCBI公司Google 学术搜索

49.王婷， 邓， 张旭， 王晖， 王汪， 刘旭， 等.番茄 DCL2b 是 22-nt 小 RNA 的生物合成、由此产生的二级 siRNA 以及宿主防御 ToMV 所必需的。园艺研究 2018;5：62。PMID：30181890

查看文章PubMed/NCBI公司Google 学术搜索

50.Yifhar T、Pekker I、Peled D、Friedlander G、Pistunov A、Sabban M 等。番茄反式作用短干扰 RNA 程序未能调节生长素反应 FACTOR3 和 ARF4 是枯叶综合征的基础。植物细胞。2012;24:3575–89.PMID：23001036

查看文章PubMed/NCBI公司Google 学术搜索

51.Kravchik M， Damodharan S， Stav R， Arazi T. 番茄 DCL3 沉默突变体的生成和表征。植物科学 2014;221–222:81–9.PMID：24656338

查看文章PubMed/NCBI公司Google 学术搜索

52.阿尔凯德 C、多奈尔 L、阿兰达 MA。转录组分析揭示了pepino花叶病毒株对AGO和DCL基因的差异调控。Mol 植物病理。2022;23:1592–1607.PMID：35852033

查看文章PubMed/NCBI公司Google 学术搜索

53.Ludman M， Burgyán J， Fátyol K. Crispr/Cas9 介导的 Argonaute 2 失活揭示了其对抗病毒反应的不同参与。科学代表 2017;7：1010。PMID：28432338

查看文章PubMed/NCBI公司Google 学术搜索

54.德拉波尔塔 SL、伍德 J、希克斯 JB。一种植物DNA小量制剂：第二版。植物分子生物学代表 1983;1:19–21.

查看文章Google 学术搜索

55.Ruijter JM、Ramakers C、Hoogaars WM、Karlen Y、Bakker O、van den Hoff MJ 等。扩增效率：在定量PCR数据分析中将基线和偏差联系起来。核酸研究 2009;37：E45。PMID：19237396

查看文章PubMed/NCBI公司Google 学术搜索

56.道尔 JJ，道尔 JL。从新鲜组织中分离植物DNA。焦点 1990;12:13–15.

查看文章Google 学术搜索

57.Golyaev V， Candresse T， Rabenstein F， Pooggin MM. 通过对干叶中的小RNA进行深度测序来重建植物病毒组和抗病毒RNAi表征。科学代表 2019;9：19268。PMID：31848375

查看文章PubMed/NCBI公司Google 学术搜索

58.Malpica-López N、Rajeswaran R、Beknazariants D、Seguin J、Golyaev V、Farinelli L 等。重新审视Tobamovirus复制酶复合蛋白在病毒复制和沉默抑制中的作用。Mol 植物微生物相互作用。2018;31:125–144.PMID：29140168

查看文章PubMed/NCBI公司Google 学术搜索

59.Li H， Durbin R. 使用 Burrows-Wheeler 变换进行快速准确的长读长比对。生物信息学。2010;26:589–95.PMID：20080505

查看文章PubMed/NCBI公司Google 学术搜索

60.MISIS-2：一种生物信息学工具，用于深入分析小 RNA 和表示病毒准物种中的共识主基因组。J Virol方法。2016;233:37–40.PMID：26994965

查看文章PubMed/NCBI公司Google 学术搜索

61.Anders S， Pyl PT， Huber W. HTSeq — 一个用于处理高通量测序数据的 Python 框架。生物信息学。2015;31:166–9.PMID：25260700

查看文章PubMed/NCBI公司Google 学术搜索

62.Lambert I， Paysant-Le Roux C， Colella S， Martin-Magniette ML. DiCoExpress：一种通过基于GLM模型内部对比的差异分析，处理从质量控制到共表达分析的多因素RNAseq实验的工具。植物方法。2020;16:68.PMID：32426025

查看文章PubMed/NCBI公司Google 学术搜索

63.罗宾逊医学博士、麦卡锡 DJ、史密斯 GK。edgeR：用于数字基因表达数据差异表达分析的 Bioconductor 软件包。生物信息学。2010;26:139–140.PMID：19910308

查看文章PubMed/NCBI公司Google 学术搜索

64.Akbergenov R、Si-Ammour A、Blevins T、Amin I、Kutter C、Vanderschuren H 等人。不同植物物种中双子病毒衍生的小RNA的分子表征。核酸研究 2006;34(2):462–71.PMID：16421273

查看文章PubMed/NCBI公司Google 学术搜索