http://orcid.org/0000-0002-6426-4692
Figures
Abstract
The whitefly Bemisia tabaci MEAM1 is a notorious vector capable of transmitting many plant viruses, resulting in serious crop loss and food shortage around the world. To investigate potential sRNA-mediated regulatory mechanisms in whiteflies that are affected by virus acquisition and transmission, we conducted small RNA (sRNA) deep sequencing and performed genome-wide profiling of piwi-interacting RNAs (piRNAs) in whiteflies that were fed on tomato yellow leaf curl virus (TYLCV)-infected or non-infected tomato plants for 24, 48, and 72 h. In the present study, piRNA reads ranging from 564,395 to 1,715,652 per library were identified and shown to distribute unevenly in clusters (57 to 96 per library) on the whitefly (B. tabaci MEAM1) genome. Among them, 53 piRNA clusters were common for all treatments. Comparative analysis between libraries generated from viruliferous and non-viruliferous whiteflies identified five TYLCV-induced and 24 TYLCV-suppressed piRNA clusters. Approximately 62% of piRNAs were derived from non-coding sequences including intergenic regions, introns, and untranslated regions (UTRs). The remaining 38% were derived from coding sequences (CDS) or repeat elements. Interestingly, six protein coding genes were targeted by the TYLCV-induced piRNAs. We identified a large number of piRNAs that were distributed in clusters across the whitefly genome, with 60% being derived from non-coding regions. Comparative analysis revealed that feeding on a virus-infected host caused induction and suppression of only a small number of piRNA clusters in whiteflies. Although piRNAs primarily regulate the activity of transposable elements, our results suggest that they may have additional functions in regulating protein coding genes and in insect-virus interactions.
Citation: Shamimuzzaman M, Hasegawa DK, Chen W, Simmons AM, Fei Z, Ling K-S (2019) Genome-wide profiling of piRNAs in the whitefly Bemisia tabaci reveals cluster distribution and association with begomovirus transmission. PLoS ONE 14(3): e0213149. https://doi.org/10.1371/journal.pone.0213149
Editor: Youjun Zhang, Chinese Academy of Agricultural Sciences, CHINA
Received: October 31, 2018; Accepted: February 17, 2019; Published: March 12, 2019
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: The datasets of raw and processed sequencing data are available in the GEO (Gene Expression Omnibus) database of NCBI under accession number GSE111343 and in supplementary information files.
Funding: USAID Feed-the-Future program (58–0210-3–012) to KSL and ZF. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Small RNAs (sRNAs) are 19–31 nucleotide (nt) non-coding regulatory elements commonly found in plants, animals, including insects [1–4]. They play important roles in regulating gene expression, transposable elements (TE), or parasite immunity [2,5–7]. Major classes of sRNAs include small interfering RNAs (siRNAs) and microRNAs (miRNAs), which are primarily involved in silencing gene expression [1,2]. Another class of sRNAs known as PIWI-interacting RNAs (piRNAs) has been identified in vertebrate and invertebrate animals [4,5,8,9]. These piRNAs ranging from 26 to 31 nt are considered to regulate the activity of transposable elements as well as to silence the expression of protein coding genes [9–11].
The biogenesis of piRNAs has been studied in fruit fly (Drosophila melanogaster), honey bee (Apis melifera), mosquito (Anopheles gambiae), and other insects [4,11–13]. Three core proteins known as P-element induced wimpy testis (Piwi), Aubergine (Aub) and Argonaute-3 (Ago3) are required for piRNA biogenesis [7,8,12]. piRNAs are generated from distinct chromosomal regions referred to as piRNA clusters [7,8] and can vary in size from 1 kb to 250 kb and may contain as few as 10 to many thousands of piRNAs [13–15]. Two mechanisms of piRNA biogenesis has been described in Drosophila and other insects [7,13,14]. In first mechanism, long single-stranded piRNA precursors in the sense or antisense orientation are transcribed from piRNA clusters, and then processed into primary piRNAs by the cytoplasmic endonuclease, Zucchini (Zuc) [8,16,17]. The primary piRNAs, in association with Piwi proteins, will then form a RNA-induced silencing complex (RISC), which leads to piRNA-directed cleavage of mRNAs [8,14]. In the second mechanism, secondary piRNAs are generated through an amplification loop cycle, referring to as the ping-pong cycle [13,14]. Here, the target transcripts are cleaved by Aub-piRISCs, resulting in 5′- cleaved products being released and degraded in the process, whereas 3′- cleaved products are loaded onto Ago3 and subsequently processed into secondary piRNAs [13,18]. The hallmarks of the ping-pong mediated amplification of piRNAs are evidenced with a strong uridine (U) bias at the 5’ end for Aub-associated piRNAs and an adenosine (A) at position 10 for Ago3-associted piRNAs [7,8,13]. A 10 base-pair overlap is often seen between the complementary primary and secondary piRNAs.
The functions of piRNAs are highly conserved among a number of animal and insect species [5,8,9,12,13], which are primarily involved in silencing the activity of transposable elements (TEs) in germline cells but are also active in other cell types [4,7,8,19,20]. TEs represent an important part of repetitive elements and can drive intra-genomic variation and genome diversification. The movement of TEs from one part of a genome to another occurs via an RNA or DNA intermediate, and thus, are classified as retrotransposons (class I) and DNA transposons (class II), respectively [21,22]. The Class I TEs are composed of long terminal repeat (LTR) and non-LTR (NLTR) retrotransposons. Families of LTR retrotransposons, including gypsy, copia and pao, are abundant in diverse insect species [7,11,23]. Similarly, NLTR, including long interspersed nuclear elements (LINEs) and short interspersed nuclear elements (SINEs), are also well studied transposable elements in insects [8,13]. On the other hand, the Class II DNA transposons, including maverick, mariner, hAT, harbinger, and P elements, are also abundant in insects [7,13,24]. Thus, piRNAs can regulate the activity of both retrotransposons and DNA transposons. Furthermore, some piRNAs are non-repetitive and non-transposon-related, which can regulate expression of protein-coding genes [8,13]. In addition to the functions of piRNAs in controlling TE activities and silencing the expression of protein-coding genes, the piRNA pathway has also been shown to be involved in antiviral response in several insects (D. melanogaster, A. gambiae and A. aegypti [6,7,25,26]. Altogether, piRNAs play diverse roles in the pathways associated with insect development, reproduction and immune responses.
The whitefly Bemisia tabaci Middle-East Asia Minor 1 (MEAM1) is one of the most widely distributed whitefly cryptic species, which is capable of transmitting hundreds of plant viruses to cause devastating losses on important agricultural crops, including beans, cassava, cotton, melon, sweetpotato and tomato [27,28]. Whitefly-transmitted viruses are classified in the genera of Begomovirus, Crinivirus, Ipomovirus, Torradovirus and Carlavirus. One of the most important begomoviruses is tomato yellow leaf curl virus (TYLCV), which has also been the focus for numerous other studies to understand the whitefly-begomovirus interactions [29–31].
In the present study, our preliminary analysis of sRNA sequence datasets revealed an enrichment of sRNAs (ranging from 26–31 nt) in addition to the typical miRNAs, which led us to explore the piRNA profiling. We performed genome-wide identification of piRNA clusters in whitefly (B. tabaci, MEAM1) and analyzed their expression levels in whiteflies feeding on TYLCV-infected or uninfected tomato plants for 24, 48, and 72 h. With ~45% of the whitefly genome sequence comprised of repeat elements, including miniature inverted repeat transposable elements (MITEs), LINEs, SINEs, LTRs and DNA transposons [28], piRNAs, involved in regulation of transposable elements, may play an important role in maintaining whitefly genome integrity, reproduction and antiviral immune responses. The piRNA pathway has been linked to other epidemiologically important phenotypes in mosquitoes [20,26,32]. The role of piRNAs in antiviral immune responses in both Aedes aegypti and A. albopictus has also been reported by multiple research groups [20,25,26]. Knowledge of the mechanism on how the piRNA pathway regulates reproduction and antiviral immune responses in whitefly could be useful for both fundamental and applied research. Our enhanced understanding of the piRNA pathway in whitefly could facilitate the identification of novel targets for vector control using RNAi and genome editing technologies.
Materials and methods
Whiteflies and feeding assays
The same whitefly materials under the identical treatments used in our previous transcriptome analysis [31] were used in the present study. Briefly, a whitefly B. tabaci MEAM1 colony was maintained at the USDA-ARS, U.S. Vegetable Laboratory on broccoli (Brassica oleracea L. var. botrytis) in a greenhouse (26°C ± 5°C). The MEAM1 colony was confirmed using PCR primers to amplify the mitochondrial cytochrome oxidase 1 gene [19]. Approximately 1,500 adult whiteflies were transferred to TYLCV-infected or uninfected tomato (Solanum lycopersicum cv. Moneymaker) plants and were allowed to feed for 24, 48, or 72 h, as described by Hasegawa et al. [31]. At the end of each time point, 200–500 living whiteflies were collected and immediately stored at -80°C until processing. Three biological replicates were performed for each treatment. Those viruliferous whiteflies were demonstrated in our previous study to contain TYLCV in acquisition periods of 24, 48 and 72 h, whereas those whiteflies fed on non-infected tomato plants were negative for TYLCV [31].
RNA isolation and library preparation
Total RNA was isolated with a TRIzol Reagent (Thermo Fisher Scientific, USA) according to the manufacturer’s protocol, and purified with the Direct-zol RNA MiniPrep kit (Zymo, USA). sRNAs were then polyacrylamide gel electrophoresis (PAGE)-purified and used to construct sRNA libraries as described [33,34]. A total of 18 sRNA libraries were constructed and sequenced on an Illumina HiSeq 2500 system.
Processing of sRNA raw reads
Raw sRNA reads were processed to remove adapters and low quality sequences using an in-house perl script included in the VirusDetect package [35]. For piRNA analysis, reads shorter than 25 nt and longer than 40 nt were also removed. The remaining size-selected reads were aligned to the Rfam database (Version 12.1) (http://rfam.sanger.ac.uk/) to filter out any reads that were mapped to rRNAs, tRNAs, and snRNAs. Raw and processed sequencing data are available in the GEO (Gene Expression Omnibus) database of the National Center for Biotechnology Information (NCBI) under accession number GSE111343.
Identification of piRNA clusters in whitefly
The high-quality cleaned sRNA reads were aligned to the whitefly reference genome using Bowtie [36] allowing up to two mismatches. Replicates were combined and genome-mapped reads were used to identify piRNA clusters using the proTRAC (Version 2.3) pipeline [37] with default parameters. proTRAC predicts potential piRNAs and piRNA clusters based on typical characteristics such as the number of loci with uridine (U) at position 1 or adenosine (A) at position 10, the number of loci within the typical piRNA length (26–31 nt), strand bias, and sliding window of 5 kb. Whitefly gene annotation and repeat sequence annotation obtained from the whitefly genome database [28] were utilized while running the proTRAC pipeline. The proTRAC output provides the normalized quantitative expression of piRNA clusters in reads per million mapped reads (RPM). This pipeline also generates a fasta file for individual clusters that contains all piRNA sequences. Differentially expressed piRNA clusters at each time points were identified using edgeR [38], with a cutoff of adjusted p-value < 0.05. In addition, we calculated the overall percentage for sense versus antisense strand derived piRNAs, and single versus multiple loci derived piRNAs.
piRNA cluster associated genes and transposable elements
Genes and repeat elements mapped to the piRNA clusters were retrieved from proTRAC output. We calculated the percentage of CDS-, repeat elements- and non-coding sequence-derived piRNAs utilizing the piRNA fasta files generated by proTRAC. Repeat annotation from the whitefly genome database [28] was utilized to divide piRNA cluster mapped repeat elements into different classes and families. We identified commonly expressed piRNA clusters across all time points using bedtools intersect [39]. Similarly, piRNA clusters that were induced or suppressed in whiteflies fed on uninfected versus TYLCV-infected plants were identified.
Results
Sequencing and initial processing of sRNA reads
To shed light on the whitefly piRNA profiles in response to feeding on TYLCV-infected tomato, deep sequencing of sRNAs was performed using RNAs isolated from whiteflies fed on either TYLCV-infected or uninfected tomato for 24, 48, and 72 h. We performed three biological replicates for each time point. Sequencing of 18 sRNA libraries produced a total of approximately 175 million reads with a range of approximately 5 to 13 million reads per library (S1 Table) and Pearson correlation coefficient (S2 Table) indicates high reproducibility of sequencing libraries. After trimming and removing low quality reads, rRNAs, tRNAs, and snRNAs, we found that reads with 28–31 nt in length were the most abundant in the datasets (Fig 1A), indicating that the libraries were enriched for reads corresponding to piRNAs. Reads that had lengths within the range of expected piRNAs of 26–40 nt were selected for further analysis.
A. Distribution of sRNA read lengths. Size distribution of sequence reads with peaking at 28–31 nt which is usual size range for piRNAs. Horizontal axis provides read length and vertical axis provides total number of reads at specific length. 24 h-H, 48 h-H, and 72 h-H represents whiteflies fed on uninfected tomato for 24, 48 and 72 hours, respectively. 24 h-V, 48 h-V, and 72 h-V represents whiteflies fed on TYLCV-infected tomato for 24, 48 and 72 hours, respectively. Fig 1B. Distribution of piRNAs in different genomic regions. Percentage of all identified piRNAs mapping to different genomic features. Red color denotes piRNAs derived from coding sequences (CDS), green color denotes piRNAs derived from repeat elements, and blue denotes piRNAs derived from non-coding sequences including intergenic regions, introns and UTRs. C). Distribution of piRNAs based on strand. Green color denotes percentage of piRNAs derived from sense strand whereas brown color denotes percentage of piRNAs derived from antisense strand. D). Distribution of piRNAs based on mapping to number of genomic loci. Blue color denotes piRNAs mapped to multiple genomic loci whereas brown color denotes piRNAs mapped to single unique genomic loci. 24 h-H, 48 h-H, and 72 h-H represents whiteflies fed on uninfected tomato for 24, 48 and 72 hours, respectively. 24 h-V, 48 h-V, and 72 h-V represents whiteflies fed on TYLCV-infected tomato for 24, 48 and 72 hours, respectively.
Expression of core components of piRNA pathway genes in Bemisia tabaci
Core components of piRNA pathway genes, Ago-3 (Bta04637), Aub (Bta08949), and piwi (Bta00007 and Bta00198) are conserved in the B. tabaci genome [28]. We retrieved the expression values of the above genes from a recently published transcriptome dataset [31] that was generated using the same pool of total RNA from which sRNAs were isolated. We found that all core component genes of the piRNA pathway were expressed at all three time points (Table 1). The expression of Ago-3 ranged in16 to 18 RPKM, and Aub in 83 to 93 RPKM. Both piwi genes were also expressed, although at lower levels (Table 1), indicating a functional piRNA pathway is present in the whitefly B. tabaci. P-values (Table 1) obtained from edgeR [38] analysis indicate that expression level of these piRNA pathway genes did not change significantly across all three time points.
Identification of piRNA clusters
Whitefly genome-mapped reads were used to identify piRNA clusters using proTRAC pipeline [37]. We detected 57 to 96 clusters of piRNAs at each of the three different time points and treatments (Table 2 and S2 Table). The total number of piRNAs at each time point ranged from 564,395 to 1,715,652 (Table 2). The results obtained from the proTRAC pipeline also provides a detailed graphical view of each cluster. A representative view of an identified graphical cluster was shown in S1 Fig. On average, each cluster contained 13,998 piRNAs. We found piRNA clusters are derived from either positive or negative strand (S2 Table), suggesting they are mono directionally transcribed. The percentage of sense strand-derived piRNAs gradually decreased (from 60% to 49%) over the three tested time periods with acquisition of TYLCV, whereas the antisense strand-derived piRNAs increased (from 40% to 51%) over the same period time (Fig 1C). Additionally, we identified piRNAs matching to single genomic loci as well as multiple genomic loci. The percentage of piRNAs matching to a single genomic loci ranged from 41% to 47% whereas the percentage of piRNAs matching to multiple genomic loci ranged from 53% to 59% (Fig 1D). When mapped to the whitefly B. tabaci MEAM1 genome, piRNAs were shown to be derived from either coding sequences (CDS), transposable elements, or non-coding sequences. We found that ~ 62% of whitefly piRNAs were derived from non-coding sequences, which included intergenic regions, introns, and UTRs (Fig 1B). The potential roles for non-coding sequence derived piRNAs are still unknown. Approximately 38% of piRNAs were derived from CDS and repeat elements (Fig 1B).
CDS and repeat elements derived piRNAs
Apart from the non-coding genomic-region derived piRNAs, there was a significant percentage of piRNAs that were derived from gene coding sequences and repeat elements which were of major interests for this study. The percentage of CDS-derived piRNAs in whiteflies fed on uninfected tomato were 14.6%, 21.0%, and 19.6% of all piRNAs for 24 h, 48 h and 72 h, respectively, whereas the percentages were 15.0%, 22.6%, and 26.3% in whiteflies fed on TYLCV-infected tomato (S2 Fig). These piRNAs were derived from genes that were predicted to be involved in diverse growth, developmental and metabolic pathways (S3 Table).
piRNAs are well known for their involvement in silencing the activity of transposable elements. Overall 19.4% of all identified piRNAs were derived from different classes of repeat elements (Fig 1B and S4 Table). For whiteflies fed on uninfected and TYLCV-infected tomato for 24 h, approximately 16% of piRNAs were derived from repeat elements, whereas at 48 h, it increased to ~ 22% (S2 Fig). Similarly, we found approximately 21% and 23% of piRNAs derived from repeat elements at 72 h for whiteflies fed on uninfected and TYLCV-infected tomato, respectively (S2 Fig).
Further analysis of repeat elements associated with piRNA clusters provided a comprehensive view of silencing of both class I (retrotransposons) and class II (DNA transposons) transposable elements. Two major families of retrotransposons are LTR and non-LTR retrotransposons. We found approximately 10%, 9% and 8% of LTR repeat elements were targeted by piRNA clusters at 24 h, 48 h, and 72 h, respectively, in whiteflies fed on uninfected tomato (Fig 2). Similarly, in whiteflies fed on TYLCV-infected tomato, we found 10%, 10%, and 13% of LTR repeat elements were targeted by piRNA clusters at 24 h, 48 h, and 72 h, respectively (Fig 2). We also identified piRNA clusters associated with non-LTR repeat elements that included LINEs and SINEs for each time point. Overall approximately 6–8% of non-LTR repeat elements were targeted by piRNA clusters across all time points. Our analysis identified a small percentage of DNA transposons targeted by piRNA clusters. They ranged between 4% and 7% of the total repeat elements targeted by piRNA clusters at all time points (Fig 2). Additionally, we found a major class of repeat elements targeted by piRNA clusters, called MITE elements, which ranged from 21% to 23% of the total repeat elements targeted by piRNA clusters across three different time points (Fig 2).
Different color denotes different families of transposable elements. It shows the percentage of specific repeat family present at specific time points after whiteflies fed on uninfected/TYLCV-infected tomato. 24 h-H, 48 h-H, and 72 h-H represents whiteflies fed on uninfected tomato for 24, 48 and 72 hours, respectively. 24 h-V, 48 h-V, and 72 h-V represents whiteflies fed on TYLCV-infected tomato for 24, 48 and 72 hours, respectively.
Comparison of identified piRNA clusters across three time points
To identify common and treatment-specific piRNA clusters, we performed comparative analysis across all time points using genomic coordinates of the identified piRNA clusters. We found 53 clusters that were commonly expressed across all time points in whiteflies fed on TYLCV-infected or uninfected tomato, and had normalized expression values that ranged from a few hundred to a few thousand RPM (S5 Table). We performed statistical test using edgeR to determine differentially expressed piRNA clusters at three time points separately. P-values (S5 Table and Table 3) obtained from edgeR analysis indicate that one of these common piRNA clusters showed significant differential expression. The top 10 highly expressed piRNA clusters were expressed at 2,606 to 10,708 RPM (Table 3) and the edgeR obtained p-values are greater than 0.05.
Induced piRNA clusters in whitefly acquired TYLCV
To evaluate whether the expression of certain piRNA clusters may be affected upon TYLCV acquisition, we identified five piRNA clusters that were exclusively induced and 24 piRNA clusters that were suppressed in whiteflies fed on TYLCV-infected tomato (Fig 3 and Table 4). Low p-values obtained from edgeR analysis indicate these expression differences are significant. Among those five piRNA clusters that were induced in whiteflies when fed on TYLCV-infected tomato (Table 4), their expression ranged from 592 to 2,387 RPM and targeted six protein coding genes as well as a number of transposable elements (Tables 5 and 6). Three of the protein coding genes (Bta07031, Bta05801, Bta02620) have unknown functions, while the other three genes (Bta07186, Bta07187, Bta07188) were annotated as encoding a COMM domain-containing protein, a sentrin-specific protease, and a mitochondrial carrier protein. In addition, these induced piRNA clusters targeted both class I retrotransposons and class II DNA transposons. The piRNA targets included multiple members of different families of transposable elements (Table 6).
Most piRNA clusters are common, and only handful of them are induced in whiteflies fed on TYLCV-infected tomato. 24 h-H, 48 h-H, and 72 h-H represents whiteflies fed on uninfected tomato for 24, 48 and 72 hours, respectively. 24 h-V, 48 h-V, and 72 h-V represents whiteflies fed on TYLCV-infected tomato for 24, 48 and 72 hours, respectively.
Suppressed piRNA clusters when whitefly acquired TYLCV
Among those 24 piRNA clusters that were suppressed in whiteflies when fed on TYLCV (Table 4), expression values of the piRNA clusters ranged from 615 to 24,372 RPM, and targeted a total of 17 protein coding genes with diverse biological functions (S6 Table). Eight of them have unknown functions and the remaining genes are not predicted to be involved in any single pathway. In addition to protein coding genes, these piRNA clusters also target both class I and II transposable elements (S7 Table). Class I retrotransposons targeted by these piRNA clusters include LTR and Non-LTR elements. Moreover, a number of class II DNA TE families were also targeted by these piRNA clusters.
Discussion
piRNAs are widely known to be involved in silencing the activity of transposable elements. Numerous studies on the piRNA pathway have broadened our understanding of the role of piRNA-mediated regulation of TEs in genetic, biological and developmental processes [40–42]. Additionally, piRNAs also regulate the activity of protein coding genes reported for a number of insect species. More recently different groups have reported that piRNAs target the non-coding sequences as well [8,13]. We found that the highest percentage of piRNAs in whitefly is derived from non-coding sequences (Fig 1B), which makes it challenging to infer their potential functions. However, transposable elements and protein coding genes are also targeted by piRNAs.
It was quite interesting to understand that piRNA clusters could be derived either from a positive or from a negative stranded DNA in the whitefly genome. Generation of different orientation of piRNAs was likely due to the two different mechanisms of piRNA biogenesis [7,13,14]. First, piRNA precursors are transcribed from piRNA clusters, and then processed into primary piRNAs [8,16,17], which lead to piRNA-directed cleavage of mRNAs [8,14]. On the other hand, secondary piRNAs are generated through the ping-pong cycle [13,14], which lead to an amplification of piRNA regeneration. The percentage of sense strand-derived piRNAs gradually decreased over time, from 60% at 24 h to 49% at 72 h post virus acquisition in the viruliferous whiteflies, whereas the antisense strand-derived piRNAs increased from 40% to 51% during the same period of time. Although such trends are interesting, we are still unsure about their implications. Recent studies have demonstrated the adaptive antiviral immunity in the mosquito A. aegyti [43] as well as many of other arthropods [44] suggesting likely involvement of piRNAs in facilitating TYLCV transmission by the vector whitefly B. tabaci.
In this study, we found that piRNAs were derived from both class I retrotransposons and class II DNA transposons. For class I, three classes of LTR retrotransposons, gypsy, copia and pao-Bel are the most prevalent sources of piRNAs (S4 Table). Previous studies in mosquito and other insects showed the same dominant sources for LTR retrotransposon-derived piRNAs as observed in whitefly [7,11,23]. We also found piRNAs derived from non-LTR retrotransposons, LINEs and SINEs in whitefly. In addition to class I retrotransposons, our study identified piRNAs derived from different families of DNA transposons (S4 Table), including MITE, hAT, TcMar, Maverick, Academ and MULE-MuDR. The highest percentage of piRNAs are derived from MITE DNA transposons when compared among individual TE families (Fig 2). Overall, our results demonstrated that TE-piRNA production in whitefly shared the trends of piRNA biogenesis in other insect species. The mechanism of piRNA-mediated silencing is complex and likely depends on both transcriptional and posttranscriptional mechanisms. Perhaps the mechanism of silencing might differ among different families of transposable elements.
An increasing number of studies reported that piRNAs also target protein coding genes and silence their expression [8,13]. In addition to CDS, they can also target introns and UTRs. Even though gene-derived piRNAs were predominantly mapped to the sense strand in several mosquito species [7,13], we found piRNAs derived from genes in whiteflies are produced from both sense and antisense strands. We observed that the piRNA targeted genes (S3 Table) were not associated with any particular pathway, but instead, seemed to be involved in diverse biological processes.
Our study identified a large number of piRNA clusters. We divided the piRNA clusters into three different categories such as common piRNA clusters, and induced and suppressed piRNA clusters in the presence of TYLCV in whiteflies. We identified 53 common piRNA clusters that are expressed in whiteflies that fed on tomatoes with or without TYLCV infection (S5 Table). These piRNA clusters showed consistent expression across all treatments and did not show any differential expression and targeted numerous classes of transposable elements and protein coding genes of different biological functions. It is likely that these common piRNA clusters are not involved in the acquisition, circulation or transmission of TYLCV in whitefly B. tabaci.
In addition to common clusters, it was interesting to observe that five piRNA clusters were induced when whiteflies fed on TYLCV-infected tomato. However, the expression of these specific piRNA clusters was not consistent across three time points, suggesting a potential influence from the virus circulation or trafficking through different orgasm in the viruliferous insect vector as measured post acquisition for 24, 48 and 72 h. These specific piRNA clusters targeted six protein coding genes (Table 5). We found that half of the genes have unknown functions, while the other three were annotated as encoding a COMM domain-containing protein, a sentrin-specific protease, and a mitochondrial carrier protein. Based on our previous published transcriptome data [31], these genes are expressed at very low levels, as they are potentially targeted by piRNAs. However, it is still not clear how these proteins are involved in virus transmission or anti-viral immunity in whitefly. The COMM domain family is a multifunctional protein, which could regulate transcription factor NF-kappa-B and copper metabolism. It has been reported that COMM domain-containing proteins modulate the activity of cullin-RING E3 ubiquitin ligase (CRL) complexes [45], and promote the ubiquitination and proteasomal degradation of cellular proteins. Perhaps silencing the expression of this gene in whitefly by piRNAs facilitates deubiquitination of certain cellular proteins involved in virus transmission. Other induced piRNA clusters target sentrin-specific protease 1 (SENP1), which is involved in the small ubiquitin-like modifier (SUMO) pathway. This protein is associated with deconjugation of SUMO from targeted proteins [46,47]. As SENP1 expression is suppressed by the action of piRNA, it could facilitate the conjugation of SUMO to the target proteins, suggesting that piRNA mediated suppression of SENP1 might be playing a role in anti-viral defense, thus to facilitate circulation or trafficking of the virus particles in a viruliferous insect vector. Another gene that was suppressed by the induced piRNA clusters upon acquisition of TYLCV was the mitochondrial carrier protein gene. Mitochondrial carrier proteins transfer molecules across the membranes in the mitochondria. Suppression of the gene encoding mitochondrial carrier protein by a piRNA cluster may facilitate virus trafficking across the membrane to enhance virus transmission. We also looked into the transposable elements targeted by these piRNA clusters. We found that both classes of TEs were targeted by piRNAs, including Class I: LTR/Gypsy, Copia and LINEs/Penelope, R1, L1, and Jockey and class II DNA transposons: MITE, hAT, and TcMar (Table 6). Future in-depth studies are needed to understand the role of piRNAs during virus transmission and immunity.
We also identified 24 piRNA clusters that were induced in whiteflies fed on uninfected tomato (Table 4). There were 15 genes targeted by these piRNA clusters (S6 Table), with one-third of them having unknown functions. The remaining genes were not associated with any particular pathway. Our analysis also found that both class I and II transposable elements were targeted by these piRNA clusters (S7 Table). Class I retrotransposons that were targeted include LTR/Gypsy, Copia, Pao, LINE/DRE, Jockey, Penelope, LOA, CR1, RTE, Dong-4, and SINE, whereas class II DNA transposons include MITE, DNA/Academ, Kolobok-T2, Mariner, MULE-MuDR, Maverick, and hAT. There were multiple members of each of those families of transposable elements. Further studies are needed to pinpoint how the suppression of these piRNAs by TYCLV is associated with anti-viral defense mechanisms. Moreover, a detailed follow up functional study is also required to elucidate the mechanisms on how piRNAs are associated with virus transmission and anti-viral defense in whitefly.
This study provides genome-wide profiling of piRNAs in whiteflies. We found that piRNA clusters were distributed across the whitefly genome, with more than 60% of identified piRNA clusters being derived from non-coding regions with unknown function. Comparative analysis revealed that feeding on a virus-infected host caused induction and suppression of only a small number of piRNA clusters in whiteflies. Majority of the identified piRNA clusters were commonly expressed across all time points in whiteflies fed on TYLCV-infected or uninfected tomato. Apart from piRNAs targeting non-coding regions, other piRNAs target protein coding genes and transposable elements. Although piRNAs primarily regulate the activity of TEs, our results suggest that they may have additional functions in regulating protein coding genes and in insect-virus interactions. This report on piRNA profiling of whitefly will serve as a valuable resource to the whitefly research community. Additionally, this study could also open new avenues to target piRNA-associated genes in whitefly using RNAi or genome editing technologies for vector management.
Supporting information
S1 Fig. Distribution of piRNAs in different genomic regions.
Representative piRNA clusters identified by the proTRAC program. It provides detailed information about the piRNA clusters including the genomic coordinates, read coverage, genes and repeat elements falling within the cluster.
https://doi.org/10.1371/journal.pone.0213149.s001
(TIFF)
S2 Fig. Distribution of piRNAs in different genomic regions.
Percentage of piRNAs mapping to different genomic features. Red color denotes piRNAs derived from coding sequences (CDS), green color denotes piRNAs derived from repeat elements, and blue denotes piRNAs derived from non-coding sequences including intergenic regions, introns and UTRs. 24 h-H, 48 h-H, and 72 h-H represents whiteflies fed on uninfected tomato for 24, 48 and 72 hours respectively. 24 h-V, 48 h-V, and 72 h-V represents whiteflies fed on TYLCV-infected tomato for 24, 48 and 72 hours, respectively.
https://doi.org/10.1371/journal.pone.0213149.s002
(TIFF)
S1 Table. Summary of reads in 18 small RNA libraries.
https://doi.org/10.1371/journal.pone.0213149.s003
(DOCX)
S2 Table. Pearson's correlation coefficients for three biological replicates and Clusters of piRNAs at each of the three different time points and treatments.
https://doi.org/10.1371/journal.pone.0213149.s004
(XLSX)
S3 Table. Predicted piRNAs target genes involved in diverse growth, developmental and metabolic pathways.
https://doi.org/10.1371/journal.pone.0213149.s005
(XLSX)
S4 Table. piRNAs targeting different classes of repeat elements.
https://doi.org/10.1371/journal.pone.0213149.s006
(XLSX)
S5 Table. piRNA clusters commonly expressed across all time points in whiteflies fed on TYLCV-infected or uninfected tomato.
https://doi.org/10.1371/journal.pone.0213149.s007
(XLSX)
S6 Table. Predicted genes targeted by suppressed piRNA clusters.
https://doi.org/10.1371/journal.pone.0213149.s008
(DOCX)
S7 Table. Transposable elements targeted by suppressed piRNA clusters.
https://doi.org/10.1371/journal.pone.0213149.s009
(DOCX)
References
- 1. Bartel DP. MicroRNAs: Genomics, biogenesis, mechanism, and function. Cell. 2004. pp. 281–297.
- 2. Ghildiyal M, Zamore PD. Small silencing RNAs: An expanding universe. Nature Reviews Genetics. 2009. pp. 94–108. pmid:19148191
- 3. Kim VN, Han J, Siomi MC. Biogenesis of small RNAs in animals. Nature Reviews Molecular Cell Biology. 2009. pp. 126–139. pmid:19165215
- 4. Aravin AA, Lagos-Quintana M, Yalcin A, Zavolan M, Marks D, Snyder B, et al. The small RNA profile during Drosophila melanogaster development. Developmental Cell. 2003. pp. 337–350. pmid:12919683
- 5. Malone CD, Hannon GJ. Small RNAs as guardians of the genome. Cell. 2009. pp. 656–668. pmid:19239887
- 6. van Rij RP, Berezikov E. Small RNAs and the control of transposons and viruses in Drosophila. Trends in Microbiology. 2009. pp. 163–171. pmid:19299135
- 7. Biryukova I, Ye T. Endogenous siRNAs and piRNAs derived from transposable elements and genes in the malaria vector mosquito Anopheles gambiae. BMC Genomics. 2015;16. pmid:25879960
- 8. Liu P, Dong Y, Gu J, Puthiyakunnon S, Wu Y, Chen XG. Developmental piRNA profiles of the invasive vector mosquito Aedes albopictus. Parasites and Vectors. 2016;9: 1–15.
- 9. Kim VN. Small RNAs just got bigger: Piwi-interacting RNAs (piRNAs) in mammalian testes. Genes and Development. 2006. pp. 1993–1997. pmid:16882976
- 10. Zheng K, Xiol J, Reuter M, Eckardt S, Leu NA, McLaughlin KJ, et al. Mouse MOV10L1 associates with Piwi proteins and is an essential component of the Piwi-interacting RNA (piRNA) pathway. Proc Natl Acad Sci. 2010;107: 11841–11846. pmid:20534472
- 11. Wang W, Ashby R, Ying H, Maleszka R, Forêt S. Contrasting sex-and caste-dependent piRNA profiles in the transposon depleted haplodiploid honeybee Apis mellifera. Genome Biol Evol. 2017;9: 1341–1356. pmid:28472327
- 12. Thomson T, Lin H. The biogenesis and function of PIWI proteins and piRNAs: Progress and prospect. Annu Rev Cell Dev Biol. 2009;25: 355–376. pmid:19575643
- 13. George P, Jensen S, Pogorelcnik R, Lee J, Xing Y, Brasset E, et al. Increased production of piRNAs from euchromatic clusters and genes in Anopheles gambiae compared with Drosophila melanogaster. Epigenetics and Chromatin. 2015;8. pmid:26617674
- 14. Brennecke J, Aravin AA, Stark A, Dus M, Kellis M, Sachidanandam R, et al. Discrete small RNA-generating loci as master regulators of transposon activity in Drosophila. Cell. 2007;128: 1089–1103. pmid:17346786
- 15. Arensburger P, Hice RH, Wright JA, Craig NL, Atkinson PW. The mosquito Aedes aegypti has a large genome size and high transposable element load but contains a low proportion of transposon-specific piRNAs. BMC Genomics. 2011;12. pmid:22171608
- 16. Voigt F, Reuter M, Kasaruho A, Schulz EC, Pillai RS, Barabas O. Crystal structure of the primary piRNA biogenesis factor Zucchini reveals similarity to the bacterial PLD endonuclease Nuc. RNA. 2012;18: 2128–2134. pmid:23086923
- 17. Nishimasu H, Ishizu H, Saito K, Fukuhara S, Kamatani MK, Bonnefond L, et al. Structure and function of Zucchini endoribonuclease in piRNA biogenesis. Nature. 2012;491: 284–287. pmid:23064230
- 18. Nagao A, Mituyama T, Huang H, Chen D, Siomi MC, Siomi H. Biogenesis pathways of piRNAs loaded onto AGO3 in the Drosophila testis. RNA. 2010;16: 2503–2515. pmid:20980675
- 19. Houwing S, Kamminga LM, Berezikov E, Cronembold D, Girard A, van den Elst H, et al. A role for Piwi and piRNAs in germ cell maintenance and transposon silencing in zebrafish. Cell. 2007;129: 69–82. pmid:17418787
- 20. Morazzani EM, Wiley MR, Murreddu MG, Adelman ZN, Myles KM. Production of virus-derived ping-pong-dependent piRNA-like small RNAs in the mosquito soma. PLoS Pathog. 2012;8. pmid:22241995
- 21. Miller WJ, McDonald JF, Pinsker W. Molecular domestication of mobile elements. Genetica. 1997;100: 261–70. pmid:9440279
- 22. Kapitonov V V., Jurka J. A universal classification of eukaryotic transposable elements implemented in Repbase. Nature Reviews Genetics. 2008. pp. 411–412. pmid:18421312
- 23. Bargues N, Lerat E. Evolutionary history of LTR-retrotransposons among 20 Drosophila species. Mob DNA. 2017;8. pmid:28465726
- 24. McCullers TJ, Steiniger M. Transposable elements in Drosophila. Mob Genet Elements. 2017;7: 1–18. pmid:28580197
- 25. Hess AM, Prasad AN, Ptitsyn A, Ebel GD, Olson KE, Barbacioru C, et al. Small RNA profiling of Dengue virus-mosquito interactions implicates the PIWI RNA pathway in anti-viral defense. BMC Microbiol. 2011;11. pmid:21356105
- 26. Schnettler E, Donald CL, Human S, Watson M, Siu RWC, McFarlane M, et al. Knockdown of piRNA pathway proteins results in enhanced semliki forest virus production in mosquito cells. J Gen Virol. 2013;94: 1680–1689. pmid:23559478
- 27. Navas-Castillo J, Fiallo-Olivé E, Sánchez-Campos S. Emerging virus diseases transmitted by whiteflies. Annu Rev Phytopathol. 2011;49: 219–248. pmid:21568700
- 28. Chen W, Hasegawa DK, Kaur N, Kliot A, Pinheiro PV, Luan J, et al. The draft genome of whitefly Bemisia tabaci MEAM1, a global crop pest, provides novel insights into virus transmission, host adaptation, and insecticide resistance. BMC Biol. 2016;14. pmid:27974049
- 29. Luan JB, Wang XW, Colvin J, Liu SS. Plant-mediated whitefly-begomovirus interactions: Research progress and future prospects. Bulletin of Entomological Research. 2014. pp. 267–276. pmid:24548638
- 30. Ghanim M, Czosnek H. Interactions between the whitefly Bemisia tabaci and begomoviruses: Biological and genomic perspectives. Management of Insect Pests to Agriculture: Lessons learned from deciphering their genome, transcriptome and proteome. 2016. pp. 181–200.
- 31. Hasegawa DK, Chen W, Zheng Y, Kaur N, Wintermantel WM, Simmons AM, et al. Comparative transcriptome analysis reveals networks of genes activated in the whitefly, Bemisia tabaci when fed on tomato plants infected with tomato yellow leaf curl virus. Virology. 2018;513: 52–64. pmid:29035786
- 32. Vodovar N, Bronkhorst AW, van Cleef KWR, Miesen P, Blanc H, van Rij RP, et al. Arbovirus-derived piRNAs exhibit a ping-pong signature in mosquito cells. PLoS One. 2012;7. pmid:22292064
- 33. Chen YR, Zheng Y, Liu B, Zhong S, Giovannoni J, Fei Z. A cost-effective method for Illumina small RNA-Seq library preparation using T4 RNA ligase 1 adenylated adapters. Plant Methods. 2012; pmid:22995534
- 34. Li R, Gao S, Hernandez AG, Wechter WP, Fei Z, Ling KS. Deep sequencing of small RNAs in tomato for virus and viroid identification and strain differentiation. PLoS One. 2012; pmid:22623984
- 35. Bolger AM, Lohse M, Usadel B. Trimmomatic: A flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30: 2114–2120. pmid:24695404
- 36. Langmead B, Trapnell C, Pop M, Salzberg S. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 2009;10: R25. pmid:19261174
- 37. Rosenkranz D, Zischler H. proTRAC—a software for probabilistic piRNA cluster detection, visualization and analysis. BMC Bioinformatics. 2012;13. pmid:22233380
- 38. Robinson MD, McCarthy DJ, Smyth GK. edgeR: A Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2009; pmid:19910308
- 39. Quinlan AR, Hall IM. BEDTools: A flexible suite of utilities for comparing genomic features. Bioinformatics. 2010;26: 841–842. pmid:20110278
- 40. Lee YCG. The role of piRNA-mediated epigenetic silencing in the population dynamics of transposable elements in Drosophila melanogaster. PLoS Genet. 2015;11. pmid:26042931
- 41. Handler D, Meixner K, Pizka M, Lauss K, Schmied C, Gruber FS, et al. The genetic makeup of the Drosophila piRNA pathway. Mol Cell. 2013;50: 762–777. pmid:23665231
- 42. Simonelig M. Developmental functions of piRNAs and transposable elements: a Drosophila point-of-view. RNA Biol. 2011;8: 754–759. pmid:21712652
- 43.
Tassetto M, Kunitomi M, Whitfield ZJ, Dolan P, Vargas IS, Ribiero I, Chen T, Olson KE, and Andino R. Antiviral adaptive immunity and tolerance in the mosquito Aedes aegyti. bioRxiv preprint first posted online Oct. 9, 2018; http://dx.doi.org/10.1101/438911
- 44.
ter Horst AM, Nigg JC, and Falk BW. Endogenous viral elements are widespread in arthropod genomes and commonly give rise to piRNAs. bioRxiv preprint first posted online Aug. 20, 2018; http://dx.doi.org/10.1101/396382.
- 45. Mao X, Gluck N, Chen B, Starokadomskyy P, Li H, Maine GN, et al. COMMD1 (Copper Metabolism MURR1 Domain-containing protein 1) regulates Cullin RING ligases by preventing CAND1 (Cullin-associated Nedd8-dissociated protein 1) binding. J Biol Chem. 2011;286: 32355–32365. pmid:21778237
- 46. Gong L, Millas S, Maul GG, Yeh ET. Differential regulation of sentrinized proteins by a novel sentrin-specific protease. J Biol Chem. 2000;275: 3355–3359. pmid:10652325
- 47. Bailey D, O’Hare P. Characterization of the localization and proteolytic activity of the SUMO-specific protease, SENP1. J Biol Chem. 2004;279: 692–703. pmid:14563852