Identification and Characterization of Micrornas in the White-Backed

Insect Science (2016) 23, 452–468, DOI 10.1111/1744-7917.12343

ORIGINAL ARTICLE Identiﬁcation and characterization of microRNAs in the white-backed planthopper, Sogatella furcifera

Zhao-Xia Chang1,†, Nan Tang1,†, Lin Wang1, Li-Qing Zhang1, Ibukun A. Akinyemi1 and Qing-Fa Wu1,2 1School of Life Sciences, University of Science and Technology of China and 2CAS Key Laboratory of Innate Immunity and Chronic Disease, School of Life Sciences and Medical Center, University of Science and Technology of China, Hefei, China

Abstract MicroRNAs (miRNAs) are a novel class of small, non-coding endogenous RNAs that play critical regulatory roles in many metabolic activities in eukaryotes. Reports of the identification of miRNAs in Sogatella furcifera (white-backed planthopper), the insect that acts as the only confirmed vector of the southern rice black-streaked dwarf virus (SRBSDV), are limited. In this study, a total of 382 miRNAs were identified in S. furcifera, including 106 conserved and 276 novel miRNAs, using high-throughput sequencing based on two small RNA libraries from viruliferous and non-viruliferous S. furcifera, and these miRNAs belonged to 52 conserved miRNA families and 58 S. furcifera-specific families, respectively. Comparison with miRNAs from 26 insect species and five other species in miRBase showed that more than half of the conserved miRNA families are highly conserved in Hexapoda, while other miRNAs are only conserved in non- dipterans. Furthermore, 4 117 target genes predicted for the 382 identified miRNAs could be categorized into 45 functional groups annotated by Gene Ontology. Compared with non-viruliferous cells, eight up-regulated miRNAs and four down-regulated miRNAs were identified in cells inoculated with SRBSDV,among which miR-14 and miR-n98a may be involved in the immune response to SRBSDV infection. Analyses of the identified miRNAs will provide insights into the roles of these miRNAs in the regulation and expression of genes involved in the metabolism, development and viral infection of S. furcifera. Key words evolution; functional shift; genomic cluster; microRNA; Sogatella furcifera; SRBSDV

Introduction Nilaparvata lugens (brown planthopper). Its feeding on rice causes damage, such as yellowing, wilting and stunt- Sogatella furcifera (white-backed planthopper) is an im- ing, reduced grain production as well as brown ear symp- portant migrating pest of rice in East Asia. It is one of toms and rusty or black cracked kernels. The direct feed- the three major rice planthoppers of the Delphacidae and ing of S. furcifera on rice phloem sap via sucking not only is considered the most notorious crop pest in Asia (Xue causes damage to the crop but also transmits plant viruses et al., 2010; Zhang et al., 2010; Wang et al., 2012), along such as the southern rice black-streaked dwarf virus (SRB- with Laodephax striatellus (small brown planthopper) and SDV) (Tu et al., 2013) in a persistent and propagative manner. Gene expression is regulated by a class of small RNAs Correspondence: Qing-Fa Wu, School of Life Sciences, Uni- (sRNAs) including microRNAs (miRNAs), small interfer- versity of Science and Technology of China, Hefei 230027, ing RNAs (siRNAs) and piwi-interacting RNAs (piRNAs) China. Email: [email protected] (Zamore & Haley, 2005; Carthew & Sontheimer, 2009). †These authors contributed equally to this work. miRNAs have been indicated as vital participants in

452 C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. MicroRNAs in Sogatella furcifera 453

Fig. 1 Characteristics of small RNAs (sRNAs). (A and B) Length distribution of total sRNAs (A) and unique sRNAs (B) in Sogatella furcifera cultured cells. (C and D) The number of southern rice black-streaked dwarf virus (SRBSDV)-derived small RNAs (reads per million) in non-viruliferous cells (C) and viruliferous cells (D). (E and F) 3 position of one small interfereing RNA (siRNA) relative to the 5 position of the other siRNA in 21 nt virus-derived small interfering RNA (vsiRNA) pairs and 22 nt vsiRNA pairs.

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 454 Z. X. Chang et al.

post-transcriptional gene regulation in organisms because ddH2O three times, then ground in liquid nitrogen. The they were first reported in humans, fruit flies and pulverized leaves were dissolved in a His-Mg solution nematodes (Bartel, 2004). The discovery of new miRNAs (100 mmol/L L-histidine, 10 mmol/L MgCl2,pH6.2), in different groups of organisms has received increased followed by centrifugation at 4600 × g for 15 mins. The attention because of their influence on the stability and supernatant was filtered with 0.45 μm and 0.22 μm filter translation efficiency of mRNA through the regulation of discs and stored at 4 °C temporarily until further use. many important pathways (Fabian et al., 2010). miRNAs regulate gene expression through specific base-paring with their targets in eukaryotes, often at conserved Cell culture and virion inoculation sequences in the 3 untranslated region (UTR) of target mRNAs (Bartel, 2004, 2009). Insect miRNAs are involved S. furcifera was reared at 26 °C at 70% relative humidity in many physiological processes, such as apoptosis, cell on rice seedlings grown in a man-made climate chamber division and differentiation (Brennecke et al., 2003; Stark under a photoperiod of 16 h lightness and 8 h darkness. et al., 2003; Leaman et al., 2005), as well as immunity Embryos of S. furcifera were collected upon reaching the (Zhou et al., 2011). Regulation of gene expression by red-eye spot stage on the seventh day after oviposition and miRNAs appears to be carried out in a seemingly complex sterilized with 70% ethanol. The sterilized embryos were pattern, as a single miRNA can regulate hundreds of washed three times with Tyrode’s solution (140 mmol/L target genes, and a single gene can, conversely, be targeted NaCl, 3 mmol/L KCl, 2 mmol/L CaCl2, 1 mmol/L MgCl2, simultaneously by multiple miRNAs (Liu et al., 2012). 0.4 mmol/L NaH2PO4, 12 mmol/L NaHCO3, 6 mmol/L Insects are the largest group of animals and are im- D-glucose) and crushed into small fragments with a ster- portant in biological and agricultural research due to ile mortar and pestle, followed by digestion with 0.25% their capacity as agricultural pests; many reports on the trypsin (Invitrogen, Carlsbad, CA, USA) in Tyrode’s solu- miRNAs of insects have been published (Sandmann & tion and centrifugation for 5 mins at 200 × g. The pellets Cohen, 2007; Zhang et al., 2009; Skalsky et al., 2010; were re-suspended in Luria-Bertani medium (LBM) and Zhou et al., 2014). However, there are no such reports incubated at 25 °C (Kimura, 1986; Kimura & Omura, for S. furcifera, mainly because of the lack of a com- 1988). The cells were transferred to a new flask after plete genome sequence for S. furcifera. The use of next- reaching 90% confluence, and the primary cell line was generation sequencing (NGS) has gained popularity for passaged eight times. The cells were then inoculated with the identification of large numbers of miRNAs. In this SRBSDV virions and harvested at 7 days post-infection. study, two small RNA libraries were constructed from cultured cells of S. furcifera subjected to SRBSDV or mock inoculation (viruliferous and non-viruliferous, re- Small RNA library construction spectively) to gain insight into the regulatory role of miRNAs in host-pathogen interactions. A total of 382 Total RNA was extracted from S. furcifera cells using miRNAs were identified, and 4117 target genes were pre- TRIzol reagent (Invitrogen) following the manufacturer’s dicted in our study. The conserved and novel miRNAs instructions. Small RNA libraries were prepared with the identified are similar in terms of the functions of their TruSeq Small RNA kit (Illumina, San Diego, CA, USA) putative target genes. Furthermore, the protein encoded according to the TruSeq Small RNA Sample Preparation by the miR-14 target may be involved in the hedgehog Guide. Briefly, 2 μg of total RNA was ligated to a 3 signaling pathway in the response to virus infection. The adapter, and a 5 adapter was ligated to the products, fol- results of this study provide new information and a bet- lowed by reverse transcription into complementary DNA ter understanding of the regulation of genes involved in (cDNA) using SuperScript II Reverse Transcriptase. The metabolism and virus infection in S. furcifera. cDNAs were amplified via polymerase chain reaction (PCR) and subjected to gel electrophoresis to recover the products ranging in size from 140 bp to 160 bp. Qualified Materials and methods libraries were sequenced on an Illumina HiSeq 2500.

Virion preparation Sequencing of mRNA ends The SRBSDV virion was extracted from infected rice leaves as described by Kimura (1986). Briefly, the rice Total RNA from S. furcifera adults was isolated using leaves were washed sequentially with 70% ethanol and TRIzol reagent and treated with Baseline-ZEROTM

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 MicroRNAs in Sogatella furcifera 455

Fig. 2 Characteristics of microRNAs (miRNAs). (A) The numbers represent the miRNAs predicted by different methods. (B) The numbers represent the miRNAs that passed filtering as described in the experimental procedures. (C) Analysis of miRNA nucleotide bias at each position in Sogatella furcifera. (D) Analysis of the miRNA length distribution. (E) Percentage of expression for the top 25 miRNAs.

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 456 Z. X. Chang et al.

DNase (Epicentre, Madison, WI, USA). Poly(A) RNA Bioinformatics analyses of miRNAs was enriched from approximately 20 μg of total RNA using the FastTrackTM MAG Micro mRNA Low-quality reads shorter than 18 nt and longer than Isolation Kit (Invitrogen) following the manufac- 30 nt, as well as contaminants formed through adapter- turer’s instructions. The poly(A) RNA was employed adapter ligation, were removed using the cutadapt pro- as a template for synthesizing first-strand cDNA gram, version 1.8 (−m 18 –M 30). Clean reads were using a 5 biotin-modified tagging primer (5 biotin- mapped onto the virus genome with a one nucleotide mis- TCCTCTCTATGGGCAGTCGGTGATTTTTTTTTTTT match using Bowtie (Langmead et al., 2009). The clean TTTTTTT-3). The cDNA was purified with 1.8 × reads were mapped to Rfam (version 11) using Bowtie and Agencourt RNAclean XP beads (Beckman Coulter, then divided into six categories: rRNA, tRNA, miRNA, Brea, CA, USA), and double-stranded cDNA was snRNA, other annotated sequences and unmapped se- synthesized using RNaseH and DNA polymerase quences. The sRNAs (miRNA and unmapped sRNAs) (Fermentas, Waltham, MA, USA), then purified using were aligned to the known miRNAs and miRNA precur- 1.8 × AMPure XP beads (Beckman Coulter). After sors in miRBase 21 to obtain the conserved miRNAs. digestion of the double-stranded cDNA with NlaIII The reads were then mapped to the complete genome (NEB, Ipswitch, MA, USA), biotin-labeled fragments of S. furcifera sequenced in our laboratory (unpublished) were purified using streptavidin-coated magnetic with Bowtie. Prediction of S. furcifera miRNA was per- beads (Invitrogen) and ligated to an adapter with a formed using miRDeep2 (Friedlander et al., 2012), mireap CATG overhang at the 3 end (Adapter-F: CCATCT- (http://sourceforge.net/projects/mireap/) and a genome- CATCCCTGCGTGTCTCCGACTCAGCATG; Adapter wide homology-based computational approach. R:CTGAGTCGGAGACACGCAGGGATGAGATGG). The genome-wide homology-based computational ap- The ligated product was amplified for 20 cycles using proach was carried out using previous methods (Singh phusion DNA polymerase (NEB), and the integrity was & Nagaraju, 2008) with some modifications. To identify observed in a 2% agarose gel. Products ranging from potential miRNA homologues, mature miRNA sequences 200 bp to 350 bp were recovered with the MinElute Gel present in miRBase were searched against the S. furcifera extraction kit (Qiagen, Hilden, Germany) and measured genome using Basic Local Alignment Search Tool – n on an Agilent 2100 Bioanalyzer (Agilent, Santa Clara, (BLASTn). The secondary structure of the hits and neigh- CA, USA). The library was sequenced using an Ion boring sequences was analyzed with RNAfold (Hofacker Torrent Personal Genome Machine (Life Technologies, et al., 2004), and pre-miRNA sequences were extracted Pleasanton, CA, USA). by considering a sliding window of 100 nt with an incre- ment of 10 nt from ± 80 nt of the miRNA hit position. The pre-miRNA sequences were also subjected to secondary structure analysis using RNAfold (free energy ࣘ Real-time reverse-transcription PCR -18 kcal/mol and bulge size ࣘ 7). To identify conserved and specific miRNAs in S. furcifera, miRNAs were traced Total RNA (50 ng) was reverse-transcribed and by searching for homologues using the BLAST program. quantified using the Bulge-LoopTM miRNA primer set Mature miRNAs and pre-miRNAs were used as queries (RiboBio, Guangzhou, China) for miRNAs versus U6 against miRBase (word size = 4; match reward = 5; mis- small nuclear RNA as a control. Total RNA (500 ng) was match penalty = -4) (Marco et al., 2010). reverse-transcribed and quantified using gene-specific The S. furcifera-specific miRNAs were grouped into primers for target genes versus 18S rRNA as a control. families with CD-hit (Fu et al., 2012) using a threshold The primer sequences used for these assays are shown of 90% sequence identity (Formey et al., 2014). miRNA in Table S5. Reverse transcription was performed expression was counted based on perfectly matched reads using the RevertAid First Strand cDNA Synthesis Kit falling within the miRNA position on the precursor, plus (Fermentas) according to the manufacturer’s instructions. 2 nt upstream and 5 nt downstream. To compare functional Reverse transcription quantitative PCR (RT-qPCR) arms with Drosophila melanogaster,weusedathird-party was conducted on a LightCycler 96 (Roche) with dataset of small RNA reads from D. melanogaster that FastStart Essential DNA Green Master mix. All reactions followed the same counting methods (Ruby et al., 2007). were performed in triplicate, and the expression level The UTRs of the sequenced reads were assembled with of RNAs was calculated using the 2−Ct method Trinity (Grabherr et al., 2011). Target genes of the miR- (Livak & Schmittgen, 2001). NAs were predicted using MiRanda (Enright et al., 2003),

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 MicroRNAs in Sogatella furcifera 457

PITA (Kertesz et al., 2007) and RNAhybrid (Rehmsmeier Alignment of the reads from the viruliferous sample et al., 2004). The default parameters for each program that mapped to the SRBSDV genome sequence showed were employed to predict the gene targets, except for the that 150 701 reads were derived from SRBSDV,which ac- following values: (i) miRanda: score threshold (145), en- counted for 0.7% of the total reads. The 21 and 22 nt sR- ergy threshold (-10) kcal/mol (Zhang & Verbeek, 2010); NAs were dominant, accounting for 39% and 42% of the (ii) PITA: seed region (length: 7–8, no mismatches, no clean reads, respectively (Fig. 1D). In contrast, 395 reads G : U pairs); and (iii) RNAhybrid: G ࣘ -20 kcal/mol. showed similarities to the SRBSDV genome, representing The DESeq package (Anders & Huber, 2010) was 0.002% of the total reads for the non-viruliferous sample, used to detect differentially expressed miRNAs between and the peak of these sRNAs occurred at a length of 18 samples. The variance was estimated with the estimate- nt (Fig. 1C), which suggests that these small RNAs are Dispersions program (method = "blind" and sharing- non-specific products or contaminants. The (+) and (−) Mode = "fit-only"), and differentially expressed miRNAs virus-derived small RNAs exhibited an equal ratio in the were identified using the nbinomTest program with de- viral genome sequence (Figs. S1A, S1B), and analysis of fault parameters. Differentially expressed miRNAs with a 21–22 nt siRNAs revealed a 2-nt overhang at the 3 end of P-value < 0.05 were deemed to be significant for down- the potential guide strand (Figs. 1E, and 1F), suggesting stream analysis. that these virus-derived small RNAs were produced from dsRNAs by the Dicer enzyme (Cenik et al., 2011; Wu et al., 2015). This finding suggested that the RNA silenc- Functional annotation and enrichment analysis ing pathway plays important roles in combating SRBSDV infection in cultured cells. The assignment of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) annotations for the gene sets was performed using InterProScan Identification of miRNAs in S. furcifera (Jones et al., 2014). Gene enrichment in annotation terms (GO or KEGG) was measured using the Fisher exact test. To annotate all known sRNAs, the two libraries were The false discovery rate (FDR) was calculated based on queried against the Rfam database (version 11). Approx- the method of Benjamini and Hochberg (1995). WEGO imately 4% of the sRNAs were derived from tRNAs and (Web Gene Ontology Annotation Plot) was employed for less than 1% of the sRNAs were derived from rRNAs, plotting the GO annotation results (Ye et al., 2006). snRNAs and other noncoding sRNAs (Fig. S2A, S2B). All of the reads classified to the un-annotated and miRNA categories were pooled together, and a total of 17 879 742 Results reads were used to predict mature miRNAs in subsequent analysis. Deep sequencing and virus-derived small RNA analysis We applied a comprehensive strategy to identify miRNAs in S. furcifera by combining sequence-based The cultured S. furcifera cells inoculated with either miRNA prediction methods (miRDeep2 and mireap) and SRBSDV or mock were harvested at 7 days post-infection. homology-based methods. Finally, 1285 miRNAs were The cDNA libraries constructed from viruliferous and predicted using miRDeep2 and mireap, while 124 miR- non-viruliferous S. furcifera cells were subjected to se- NAs were predicted via the homology-based approach, quencing using the Illumina sequencing platform, which with 66 of the predicted miRNAs overlapping with each generated 20 837 648 and 22 354 767 raw reads, re- other (Fig. 2A). To identify miRNAs with high confi- spectively. Adaptor sequences and low-quality reads were dence, the following criteria were used to filter the com- trimmed to obtain 15 056 464 and 17 154 261 clean reads bined 1337 unique miRNA candidates: (i) each miRNA with lengths of 18–30 nt (Table S1). The majority of the was supported by at least one perfect miRNA star read; small RNAs were between 20 and 24 nt (Figs. 1A and 1B), or (ii) a homolog of the miRNA was conserved in at least with similarities in length distribution being observed one other species. Ultimately, 382 miRNAs were identi- in the two libraries. The 22 nt sRNA peak represents fied based on showing at least one read on both miRNA the abundant dicer-derived products, which accounted for strands or the existence of a homologous miRNA in an- 17.05% and 16.67% of the total sequence reads in the vir- other species (Fig. 2B), including 106 conserved miRNAs uliferous and non-viruliferous libraries, respectively. The with homologs in other species, belonging to 52 conserved 26–28 nt peak represents endogenous piRNAs originating miRNA families and 276 S. furcifera-specific miRNAs from the cultured embryo cells (Fig. S2C, S2D). (Table S2).

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 458 Z. X. Chang et al.

The majority of the miRNAs were 22 nt (37%) or mologs than S. furcifera (Fig. 3B). The results of the var- 23 nt (27%) in length (Fig. 2D), which is similar to the data ious comparisons showed that some of the S. furcifera from other insect species in miRBase 21 (Fig. S3). Exactly miRNAs are specific to non-dipterans (Figs. 3A, and 3B). 53.8% of the miRNAs exhibited uridine (U) at the first In addition to the conserved miRNAs with homologs in nucleotide position, while 9.1% of the miRNAs exhibited other species, there were 276 S. furcifera-specific miR- guanine (G) at the first position (Fig. 2C), similar to pre- NAs which were grouped into 58 families by CD-hit (Fu vious reports (Seitz et al., 2011; Wang, 2013). The four et al., 2012) using a threshold of 90% sequence iden- most abundant miRNAs (sfu-bantam [3 331 203 reads], tity (Formey et al., 2014). Thus, the S. furcifera-specific sfu-miR-13b [776 030], sfu-miR-2765a [696 298 reads] miRNAs included not only singleton sequences but also and sfu-miR-9a [608 440 reads]) (Fig. 2E) accounted for expanded families that display at least two mature miR- 61% of the total reads for all miRNAs. The highest ex- NAs (Table S3). All of this information was used to repression among the S. furcifera-specific miRNAs was construct the phylogenetic history of the miRNA families observed for sfu-miR-n72, with 293 594 reads, which found in S. furcifera (Fig. 3D). The expanded miRNA fam- accounted for 3% of the total miRNA reads. Approxi- ilies were mainly associated with signal transduction. For mately 67% (55/82) of the abundantly expressed miRNAs example, targets of the mir-n107 family were enriched in (> 1000 reads) were conserved, and the majority (94/109) small guanosine triphosphatase-mediated signal transduc- of the miRNAs showing low expression (< 20 reads) were tion, and the mir-n127 family, showing the most members, S. furcifera specific. likely participates in the regulation of proteolysis.

Evolutionary analysis of the identified S. furcifera Genomic clusters of miRNAs in S. furcifera miRNAs miRNA clusters are often defined as a group of miRNAs Evolutionary analysis is important to reveal the func- that are within 10 kb of each other on the same genomic tional dynamics of miRNAs in insects, especially for strand (Marco et al., 2013), from which mature miRNAs newly emerging miRNAs. Further analyses were carried may be produced from a polycistronic transcript. We ob- out to study the conservation of the miRNA families in served that almost 23% of the S. furcifera miRNAs were 26 species of Hexapoda and five other species (Fig. 3C). within 1 kb of another miRNA, and 50% of the miRNAs The results showed that mir-10, mir-9, mir-46 were highly contained clusters within a distance of 5 kb (Fig. 4A). conserved among the insect species, implying that these The mean number of miRNAs per cluster was approxi- conserved miRNAs exhibit highly important and similar mately two for 1 kb clusters and almost four for larger functions. More than half of the miRNA families found groups (10–30 kb). Hence, it is possible that more than in S. furcifera were conserved in insect species with mir- half of all miRNAs found in S. furcifera are expressed from 2 and mir-2765 families observed to be expanded in S. polycistronic transcripts, which vary in length by up to furcifera (Fig. 3C). mir-2 is widespread in invertebrates, 10 kb. and recent research suggests that the mir-2 family reg- The proportion of clustered miRNAs in S. furcifera as ulates insect metamorphosis by controlling the juvenile well as Tribolium castaneum and D. melanogaster was hormone signaling pathway (Lozano et al., 2015). miR- calculated to determine whether clustering is evolution- 2765 was recently shown to be down-regulated in female arily conserved in various insects. Although the birth and mosquitoes compared with males (Jain et al., 2015). In S. death rates of insect miRNAs are high, elements that are furcifera, both mir-2 and mir-2765 were highly expressed, conserved between two species tend to maintain their indicating that miRNA duplication could be conducive to clustering features (Fig. 4B). At a clustering distance of the maintenance of these functional miRNAs. We also 100 bp to 10 kb, more of the conserved miRNA pairs observed that mir-2765, mir-1175, mir-2796 and mir-71 that were clustered in S. furcifera maintained their link- are not detected in D. melanogaster while mir-2796 is age in T. castaneum than in D. melanogaster. Within 10 not present in the dipteran species (Fig. S4), suggest- kb, these proportions are similar for T. castaneum and D. ing that it is conserved only in non-dipterans. To further melanogaster (Fig. 4B), with 18 miRNAs showing con- determine the conservation of the S. furcifera miRNAs, served linkage between S. furcifera and T. castaneum. the percentages of detectable homologs in other insects These findings suggested that S. furcifera clusters are bet- were compared with fruit fly miRNAs. S. furcifera and D. ter conserved in T. castaneum than in D. melanogaster. melanogaster share the highest percentage of homologs miRNAs residing in genomic clusters that generate with Aedes aegypti, but D. melanogaster shares more ho- polycistronic primary-miRNA transcripts represents a

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 MicroRNAs in Sogatella furcifera 459

Fig. 3 Conservation of Sogatella furcifera microRNAs (miRNAs). (A) Venn diagram showing the number of homologs of S. furcifera miRNAs in different species. Dipterans (aae, cqu); other insects (bmo, tca, ame). (B) Percentage of miRNAs in Drosophila (black) and S. furcifera (white boxes) with detectable homologs in other species. (C) S. furcifera miRNA families. Conservation of miRNA families in 26 insect species and five other species. (D) Gain or loss of miRNA families. In accordance with the rules of miRBase, the abbreviation represents the species name. aae: Aedes aegypti;aga:Anopheles gambiae;ame:Apis mellifera;api:Acyrthosiphon pisum; bmo: Bombyx mori; cqu: Culex quinquefasciatus;dan:Drosophila ananassae;der:Drosophila erecta;dgr:Drosophila grimshawi; dme: Drosophila melanogaster; dmo: Drosophila mojavensis; dpe: Drosophila persimilis; dps: Drosophila pseudoobscura;dse:Drosophila sechellia; dsi: Drosophila simulans; dvi: Drosophila virilis;dwi:Drosophila willistoni; dya: Drosophila yakuba; hme: Heliconius melpomene; lmi: Locusta migratoria;mse:Manduca sexta; ngi: Nasonia giraulti; nlo: Nasonia longicornis;nvi:Nasonia vitripennis; pxy: Plutella xylostella; tca: Tribolium castaneum;dre:Danio rerio; ptr: Pan troglodytes;bta:Bos taurus;has:Homo sapiens;cfa:Canis familiaris.

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 460 Z. X. Chang et al.

Fig. 4 Genomic clustering of insect microRNAs (miRNAs). (A) The number of miRNAs in clusters (solid line) and the number of clusters (dashed line) for different genomic distances in Sogatella furcifera. (B) The number of S. furcifera miRNAs in clusters showing conservation of the clustering observed in Drosophila (solid line) or Tribolium castaneum (dashed line). (C) Conservation of mir-71/mir-2/mir-13 cluster in multiple insect species. strategy for bringing cohorts of target genes under co- but absent in A. aegypti, contains the mir-1174/1175 clus- ordinate control by miRNAs (Mohammed et al., 2014). ter. However, miR-750 and miR-1174 exhibit the same The mir-2/13/71 cluster is a good example of high con- seed sequences (Fig. 5B). It is likely that this cluster servation of organization within insects. In Drosophila, exhibits functional properties, and the key miRNA that this cluster is fragmented and lacks mir-71 (Fig. 4C). regulates its targets is mir-1175. The evolutionary conservation and clustering propensity It is notable that some clusters consist of S. furcifera- of the mir-2/13/71 family across protostomes could indi- specific families. For example, two S. furcifera-specific cate common functional roles across the member species miRNA families, mir-n78 and mir-n98, are clustered of Protostomia (de Souza Gomes et al., 2013). The mir- in two locations (Fig. 5B). Within cluster_52, there 750/1175 cluster, which is present in four insect species (S. are members of mir-n78 and mir-n98 families, whereas furcifera, Bombyx mori, T. castaneum and Apis mellifera) within cluster_39, two additional miRNAs (mir-n77a and

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 MicroRNAs in Sogatella furcifera 461

Fig. 5 Genomic locations of Sogatella furcifera microRNA (miRNA) clusters. (A) Genomic locations of mir-750/1175 clusters in different insect genomes. The numbers on the bottom indicate the nucleotide distance between miRNAs for each miRNA. (B) Two of the miRNA clusters containing S. furcifera-specific miRNA families. mir-n77b) are observed at the end of the cluster. The find- tially accumulated arm of the hairpin precursor, whereas ing that many S. furcifera-specific families are close to the opposite arm is referred to as a miRNA star. Moreover, each other within genomic locations suggested that some it has been reported that the dominant arm can switch in miRNAs achieved specific expansion via segment dupli- different tissues and different organisms (Chiang et al., cation. 2010; Griffiths-Jones et al., 2011). Similar to Drosophila miRNAs, most S. furcifera miRNAs display dominant 5 or 3 arm usage (Fig. 6A). For some miRNAs, reads could Functional shifts in insect miRNAs be detected on both the 5 arm and 3 arm, which suggests that these miRNAs may have switched arm usage miRNAs can be processed from the 5 or 3 arm of the (the arm from which the dominant mature miRNA is pro- RNA duplex. A mature miRNA is defined as the preferen- duced) during evolution.

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 462 Z. X. Chang et al.

Fig. 6 Arm usage bias in insect microRNAs (miRNAs). (A) Proportion of reads detected on the 5 arm of miRNAs with respect to the total number of reads from the miRNA in Sogatella furcifera. (B) Comparison of “relative arm usage” between S. furcifera and Drosophila. miRNAs within the 3/5 and 5/3 quadrants show a switch in arm usage. The dashed line indicates the theoretical expectation for conserved arm usage between the two species. Dotted lines delimit the boundaries of the dashed lines to a less than 10-fold difference in arm usage.

The relative arm usage for S. furcifera and D. et al., 2014), to obtain longer 3 UTRs of genes, mRNAs melanogaster was plotted as shown in Figure 6B. De- with poly(A) tails were enriched for library preparation viations of this measure from zero indicate that there is and sequenced with an ion-torrent sequencer. A total of abiastoward5 (positive values) or 3 (negative values) 2 416 740 reads longer than 50 nt were obtained. These arm usage for a given miRNA. The data clearly show that sequences were assembled into 3883 contigs, which were nine miRNAs have undergone a switch in arm preferences assumed to come from the authentic end of the mRNAs. during insect evolution. For example, the 3 arm of mir- These contigs were combined with 21 254 known tran- 10 produces the dominant product in D. melanogaster, scripts to obtain the 3 UTRs. The average length of the whereas the 5 arm dominates S. furcifera.Thesame 3 UTRs was 300 nt, which was longer than the orig- switch was observed for mir-965, mir-971, mir-981 and inal mean length (289 nt) of these genes. There were mir-281. In contrast, mir-87, mir-283, mir-33 and mir-219 362 genes for which longer 3 UTR sequences were ob- exhibit a dominant 5 product in D. melanogaster and a tained, and the average length of the extension of the 3 dominant 3 arm in S. furcifera (Fig. 6B). Mature miRNAs UTR was 179 nt. produced from the 5 and 3 arms of the same precursor To obtain reliable gene targets of the miRNAs, three dif- hairpin exhibit distinct sequences and regulate different ferent computational tools (MiRanda, PITA and RNAhy- gene sets; therefore, significant shifts in arm usage are brid) were applied together, and only the targets predicted predicted to alter the target genes and function of a given supported by at least two programs were retained for miRNA. downstream analysis. Exactly 4117 genes were predicted We examined arm usage among the clustered miRNAs to be the targets of 382 miRNAs. The average number at different clustering distances. The results of this anal- of miRNA targets recorded was 116, and the largest was ysis showed that clustered miRNAs tend to exhibit the identified for miR-n131 at 417. A total of 44 580 miRNA same dominant arm (Table S4). A chi-square test of the target pairs were obtained from the gene 3 UTR, which result showed that the association between clustered miR- included 12 246 pairs for 106 conserved miRNAs and NAs and use of the same arm is stronger in clusters with 32 334 pairs for 276 novel miRNAs, with an average of a greater distance (Table S4). As described above, par- 115 and 117 main target miRNAs, respectively. The novel alogs tend to produce functional miRNAs from the same miRNAs appeared to exhibit more targets than the con- arm. This result suggests that common motifs in the pri- served miRNAs in S. furcifera, which suggests that con- mary transcript may concurrently affect the arm choice of served miRNAs display stronger specificity than novel multiple miRNAs in a cluster. miRNAs. The annotation of unique target genes of the conserved mIRNA target prediction and novel miRNAs compared with the Gene Ontology (GO) database provides useful information for under- As most miRNA target sites are predicted to be located standing gene functions and specific processes that have in the 3 UTR region (Bartel, 2004; Bartel, 2009; Peterson occurred throughout the evolution of S. furcifera.The

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 MicroRNAs in Sogatella furcifera 463

GO terms identified in our analysis for all of the pu- Discussion tative target genes subjected to GO functional classification suggest that the functions of the genes targeted miRNAs are a family of small, endogenously initiated by conserved and novel miRNAs are similar, mainly be- non-coding RNAs that extensively regulate gene ex- ing associated with the binding (GO: 0005488), cell part pression through either mRNA cleavage or translational (GO: 0044464), cellular process (GO: 0009987) and repression, thus playing important roles in the develop- metabolic process (GO: 0008152) categories (Fig. S5). ment and physiology of organisms (He & Hannon, 2004). However, the immune system process (GO: 0006955) cat- No miRNAs have been previously reported in S. furcifera. egory was only present among the novel miRNA targets, In this study, miRNAs were identified from cultured S. which suggests that some novel miRNAs participate in S. furcifera cells using both high-throughput sequencing furcifera immunity. methods (mireap and miRDeep2) and homology-based prediction method. mireap identified 1148 miRNA candidates, but only 310 (27 %) miRNAs were supported by an Differential expression of miRNAs in viruliferous miRNA star or conservation in other species. miRDeep2 and non-viruliferous cells identified 602 miRNA candidates, and 330 (55 %) passed the filtering procedure; 260 out of these 330 miRNAs Identification of the differential expression of miR- overlapped with the 310 miRNAs predicted by mireap. In NAs between viruliferous and non-viruliferous S. fur- this regard, the performance of miRDeep2 was better than cifera cells with DESeq showed that eight miRNAs that of mireap. Only 50 and 70 miRNAs were exclusively were significantly up-regulated, while four miRNAs were predicted by either mireap or miRDeep2, respectively, down-regulated in the viruliferous cells (P-value < 0.05) which suggests that these different computational tools (Fig. 7A). miR-14 and miR-2796 are conserved miRNAs are complementary to each other. To identify miRNAs and exhibited the most abundant differentially expressed that were not expressed in the cultured embryo cells, miRNAs, showing 1.28-fold and 1.33-fold increases in homology-based method was also used to scan the viruliferous cells, respectively. The other 10 miRNAs genome sequence, and 74 miRNAs were identified. Inter- were S. furcifera-specific miRNAs. miR-n98a is a novel S. estingly, 72 of the 74 miRNAs overlapped with the mireap furcifera-specific miRNA that was down-regulated in vir- and miRDeep2 results, but there were still two conserved uliferous cells to 0.6-fold level in non-viruliferous cells. miRNAs (mir-iab-8 and let-7) that were identified only The changes in the expression of miR-14 and miR-n98a by the homology-based method. These results indicated were validated through qPCR using the same total RNA that it is necessary to employ complementary methods samples employed for the construction of the sequencing for miRNA discovery in new species. In this study, a total libraries (Fig. S6A). of 382 miRNAs were identified, including 106 conserved According to our predictions, 832 target genes are reg- and 276 novel miRNAs. Because many miRNAs were ulated by the eight up-regulated miRNAs, and 407 tar- not expressed in the cultured embryo cells, it is certain get genes are regulated by the four down-regulated miR- that more miRNAs, and especially S. furcifera-specific NAs. For three of the putative target genes of miR-14 miRNAs, will be identified when additional small RNAs and miR-n98a, RT-qPCR measurements were conducted from different cell types and developmental stages of S. in both viruliferous and non-viruliferous cells, in which furcifera become available. these target genes exhibited distinct binding free ener- It is now well established that miRNAs function by tar- gies in association with miRNAs (Table S5). Compared geting complementary sequences in mRNA transcripts, with the expression level in non-viruliferous cells, the usually in the 3 UTR, to inhibit protein translation or expression levels of the miR-14 target genes were de- induce target degradation. Many computational methods creased 12%–20%, while the expression levels of miR- and algorithms have been developed to predict targets, n98a target genes were increased 27%–97% in virulifer- although none developed so far has been capable of cap- ous cells (Fig. S6B). These target genes of up-regulated turing all true targets (Witkos et al., 2011). These ap- miRNAs and down-regulated miRNAs were clustered proaches differ in their algorithmic style, and various based on GO category (Fig. 7B). GO enrichment analy- features or factors are weighed differently, causing dif- sis showed that “nucleocytoplasmic transport”, with four ferences in their predictions (Peterson et al., 2014). Al- genes, was significantly enriched for up-regulated miR- though the intersection of all programs may achieve the NAs, whereas the “nucleotide biosynthetic process” cat- highest specificity, it also results in the lowest sensitiv- egory, with three genes, was enriched for down-regulated ity (Sethupathy et al., 2006). miRanda (Enright et al., miRNAs. 2003), which considers matching along the entire miRNA

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 464 Z. X. Chang et al.

Fig. 7 (A) Comparison of microRNA (miRNA) expression levels in viruliferous and non-viruliferous Sogatella furcifera cells. (B) Specific-enriched Gene Ontology (GO) terms for target genes of differentially expressed miRNAs. sequence, unlike most miRNA target predictors, is the dual oxidase (DUOX) activation, which is an important most sensitive of these tools (Witkos et al., 2011). RNAhy- component of the innate host defense against bacteria brid (Rehmsmeier et al., 2004) considers the free energy and viruses (Lee et al., 2015). Some significant genes in between a miRNA and a mRNA with a user-defined seed the Hh signaling pathway are differentially expressed in region, while PITA (Kertesz et al., 2007) uses target-site dengue virus-refractory Aedes aegypti compared with a accessibility as the major feature for miRNA target pre- susceptible strain, suggesting activation of the Hh signal- diction. Thus, we considered sequence alignment, the free ing pathway in A. aegypti in response to dengue infection energy of binding and target site accessibility. To maxi- (Chauhan et al., 2012). The patched (Ptc) protein is a pos- mize the predictive power of these tools, we chose the itive effector of Hh and was predicted as the target gene of predicted targets supported by at least two of three meth- sfu-miR-14 by miRanda and PITA. Analogous to miR-14, ods (miRanda, RNAhybrid and PITA). This integrated which has been identified as a modulator of Hh signaling approach will increase the likelihood that the indicated in Drosophila (Kim et al., 2014), sfu-miR-14 may be in- interaction is real and functional. volved in the Hh signaling pathway in the response to virus The southern rice black-streaked dwarf virus (SRB- infection. miR-n98a is a S. furcifera-specific miRNA that SDV) represents a severe threat to rice production in was down-regulated in viruliferous cells. The predicted East Asian countries. This virus is vectored by migrat- target genes of miR-n98a include SFU-20.387, a putative ing S. furcifera in a circulative, propagative and persistent rab5-interacting protein (rab5ip). rab5ip may function in manner. miRNAs have been proven to play a key role in endocytic vesicles as a receptor for Rab5-GDP and partic- host-pathogen interactions (Mead & Tu, 2008; Padman- ipate in the activation of Rab5 (Hoffenberg et al., 2000) abhan et al., 2009; Chawla & Sokol, 2011). For example, which plays a role in hepatitis C virus (HCV) genome down-regulation of aae-miR-2940 reduces metallopro- replication (Stone et al., 2007), suggesting that miR-n98a tease levels to restrict West Nile virus (WNV) replication might regulate the viral replication of SRBSDV in the cul- in mosquitoes (Slonchak et al., 2014), and up-regulation tured cells. Therefore, this information provides a foun- of miR-34 may be associated with the anti-pathogen and dation for deciphering the relationships between miRNAs immune responses to dengue virus in Aedes albopictus and their targets in S. furcifera in the response to SRBSDV. (Liu et al., 2015). Differential expression of miRNAs was calculated using DESeq because this program is able to address high-throughput data without replicates (Anders Conclusion & Huber, 2010). In this study, 12 miRNAs were identified as differentially expressed miRNAs, suggesting that these This study identified the first set of miRNA in So- miRNAs might play roles in the response to SRBSDV gatella furcifera. The results of high-throughput se- infection. miR-14 is a conserved miRNA that was up- quencing of small RNA libraries from viruliferous and regulated in cells inoculated with SRBSDV.The hedgehog non-viruliferous cells were analyzed using computa- (Hh) signaling pathway plays an important role in host re- tional tools, resulting in the identification of 106 con- sistance to enteric infection by modulating uracil-induced served and 276 novel miRNAs in S. furcifera. The

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 MicroRNAs in Sogatella furcifera 465 conservation, genomic organization and functional shifts Cenik, E.S., Fukunaga, R., Lu, G., Dutcher, R., Wang, Y., Tanaka of the 382 identified miRNAs among insects were exten- Hall, T.M. and Zamore, P.D. (2011) Phosphate and R2D2 sively characterized. Bioinformatics analysis suggested restrict the substrate specificity of Dicer-2, an ATP-driven that the conserved miRNAs exhibit stronger specificity ribonuclease. Molecular Cell, 42, 172–184. than the novel miRNAs, but the functions of the putative Chauhan, C., Behura, S.K., Debruyn, B., Lovin, D.D., Harker, target genes were similar. The up and down-regulation of B.W., Gomez-Machorro, C., Mori, A., Romero-Severson, J. some genes as compared with non-viruliferous cells may and Severson, D.W. (2012) Comparative expression profiles be actively involved in the immune response against SRB- of midgut genes in dengue virus refractory and susceptible SDV infection. The information obtained in this study will Aedes aegypti across critical period for virus infection. PLoS provide insights into the roles of miRNAs in the regulation ONE, 7, e47350. and expression of genes involved in the metabolism, de- Chawla, G. and Sokol, N.S. (2011) Micrornas in Drosophila velopment, immune response and virus infection of S. fur- development. International Review of Cell and Molecular cifera. The small RNA data are available at NCBI (http:// Biology, 286, 1–65. www.ncbi.nlm.nih.gov/bioproject/PRJNA309771). Chiang, H.R., Schoenfeld, L.W., Ruby, J.G., Auyeung, V.C., Spies, N., Baek, D., Johnston, W.K., Russ, C., Luo, S.J., Babi- arz, J.E., Blelloch, R., Schroth, G.P.,Nusbaum, C. and Bartel, Acknowledgments D.P.(2010) Mammalian microRNAs: experimental evaluation of novel and previously annotated genes. Genes & Develop- This work was supported by the Strategic Priority Re- ment, 24, 992–1009. search Program of the Chinese Academy of Sciences De Souza Gomes, M., Donoghue, M.T., Muniyappa, M., (Grant XDB11040400), the Ministry of Science and Tech- Pereira, R.V., Guerra-Sa, R. and Spillane, C. (2013) Com- nology of China (Grant 2014CB138405), the National putational identification and evolutionary relationships of the Natural Science Foundation of China (Grant 31571305) microRNA gene cluster miR-71/2 in protostomes. Journal of and the National Science and Technology Major Project Molecular Evolution, 76, 353–358. (2013ZX10004605-001-002). The funders played no role Enright, A.J., John, B., Gaul, U., Tuschl, T., Sander, C. and in the study design, data collection and analysis, decision Marks, D.S. (2003) MicroRNA targets in Drosophila. Genome to publish, or preparation of the manuscript. Biology,5,R1. Fabian, M.R., Sonenberg, N. and Filipowicz, W. (2010) Disclosure Regulation of mRNA translation and stability by mi- croRNAs. Annual Review of Biochemistry, 79, 351– The authors have no conflicts of interest, or involvement, 379. financial or otherwise, that might potentially bias the Formey, D., Sallet, E., Lelandais-Briere, C., Ben, C., Bustos- work. Sanmamed, P., Niebel, A., Frugier, F., Combier, J.P., Debelle, F., Hartmann, C., Poulain, J., Gavory, F., Wincker, P., Roux, References C., Gentzbittel, L., Gouzy, J. and Crespi, M. (2014) The small RNA diversity from Medicago truncatula roots under Anders, S. and Huber, W.(2010) Differential expression analysis biotic interactions evidences the environmental plasticity of for sequence count data. Genome Biology, 11, R106. the miRNAome. Genome Biology, 15, 457. Bartel, D.P. (2004) MicroRNAs: genomics, biogenesis, mecha- Friedlander, M.R., Mackowiak, S.D., Li, N., Chen, W. and Ra- nism, and function. Cell, 116, 281–297. jewsky, N. (2012) miRDeep2 accurately identifies known and Bartel, D.P. (2009) MicroRNAs: target recognition and regula- hundreds of novel microRNA genes in seven animal clades. tory functions. Cell, 136, 215–233. Nucleic Acids Research, 40, 37–52. Benjamini, Y. and Hochberg, Y. (1995) Controlling the false Fu, L.M., Niu, B.F., Zhu, Z.W., Wu, S.T. and Li, W.Z. (2012) discovery rate – a practical and powerful approach to mul- CD-HIT: accelerated for clustering the next-generation se- tiple testing. Journal of the Royal Statistical Society Series quencing data. Bioinformatics, 28, 3150–3152. B-Methodological, 57, 289–300. Grabherr, M.G., Haas, B.J., Yassour, M., Levin, J.Z., Thompson, Brennecke, J., Hipfner, D.R., Stark, A., Russell, R.B. and Cohen, D.A., Amit, I., Adiconis, X., Fan, L., Raychowdhury, R., Zeng, S.M. (2003) Bantam encodes a developmentally regulated Q., Chen, Z., Mauceli, E., Hacohen, N., Gnirke, A., Rhind, N., microRNA that controls cell proliferation and regulates the Di Palma, F., Birren, B.W., Nusbaum, C., Lindblad-Toh, K., proapoptotic gene hid in Drosophila. Cell, 113, 25–36. Friedman, N. and Regev, A. (2011) Full-length transcriptome Carthew, R.W. and Sontheimer, E.J. (2009) Origins and mecha- assembly from RNA-Seq data without a reference genome. nisms of miRNAs and siRNAs. Cell, 136, 642–655. Nature Biotechnology, 29, 644–652.

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 466 Z. X. Chang et al.

Griffiths-Jones, S., Hui, J.H.L., Marco, A. and Ronshaugen, quencing for microRNAs profiling in Apis mellifera: compar- M. (2011) MicroRNA evolution by arm switching. EMBO ison between nurses and foragers. Insect Molecular Biology, Reports, 12, 172–177. 21, 297–303. He, L. and Hannon, G.J. (2004) MicroRNAs: small RNAs with Liu, Y.X., Zhou, Y.H., Wu, J.Y., Zheng, P.M., Li, Y.J., Zheng, a big role in gene regulation. Nature Reviews Genetics,5, X.Y., Puthiyakunnon, S., Tu, Z.J. and Chen, X.G. (2015) The 522–531. expression profile of Aedes albopictus miRNAs is altered by Hofacker, I.L., Priwitzer, B. and Stadler, P.F. (2004) Prediction dengue virus serotype-2 infection. Cell Bioscience, 5, 16. of locally stable RNA secondary structures for genome-wide Livak, K.J. and Schmittgen, T.D. (2001) Analysis of relative surveys. Bioinformatics, 20, 186–190. gene expression data using real-time quantitative PCR and Hoffenberg, S., Liu, X., Nikolova, L., Hall, H.S., Dai, W., the 2-࢞࢞CT method. Methods, 25, 402–408. Baughn, R.E., Dickey, B.F., Barbieri, M.A., Aballay, A., Stahl, Lozano, J., Montanez, R. and Belles, X. (2015) MiR-2 fam- P.D. and Knoll, B.J. (2000) A novel membrane-anchored Rab5 ily regulates insect metamorphosis by controlling the juve- interacting protein required for homotypic endosome fusion. nile hormone signaling pathway. Proceedings of the National Journal of Biological Chemistry, 275, 24661–24669. Academy of Sciences of the United States of America, 112, Jain, S., Rana, V., Tridibes, A., Sunil, S. and Bhatnagar, R.K. 3740–3745. (2015) Dynamic expression of miRNAs across immature and Marco, A., Hui, J.H., Ronshaugen, M. and Griffiths-Jones, adult stages of the malaria mosquito Anopheles stephensi. S. (2010) Functional shifts in insect microRNA evolution. Parasit Vectors, 8, 179. Genome Biology and Evolution, 2, 686–696. Jones, P., Binns, D., Chang, H.Y., Fraser, M., Li, W., Marco, A., Ninova, M., Ronshaugen, M. and Griffiths-Jones, Mcanulla, C., Mcwilliam, H., Maslen, J., Mitchell, A., Nuka, S. (2013) Clusters of microRNAs emerge by new hairpins in G., Pesseat, S., Quinn, A.F.,Sangrador-Vegas, A., Scheremet- existing transcripts. Nucleic Acids Research, 41, 7745–7752. jew, M., Yong, S.Y., Lopez, R. and Hunter, S. (2014) In- Mead, E.A. and Tu, Z.J. (2008) Cloning, characterization, and terProScan 5: genome-scale protein function classification. expression of microRNAs from the Asian malaria mosquito, Bioinformatics (Oxford, England), 30, 1236–1240. Anopheles stephensi. BMC Genomics, 9, 244. Kertesz, M., Iovino, N., Unnerstall, U., Gaul, U. and Segal, Mohammed, J., Siepel, A. and Lai, E.C. (2014) Diverse modes E. (2007) The role of site accessibility in microRNA target of evolutionary emergence and flux of conserved microRNA recognition. Nature Genetics, 39, 1278–1284. clusters. RNA, 20, 1850–1863. Kim, K., Vinayagam, A. and Perrimon, N. (2014) A rapid Padmanabhan, C., Zhang, X.M. and Jin, H.L. (2009) Host small genome-wide microRNA screen identifies miR-14 as a mod- RNAs are big contributors to plant innate immunity. Current ulator of Hedgehog signaling. Cell Reports, 7, 2066–2077. Opinion in Plant Biology, 12, 465–472. Kimura, I. (1986) A study of rice dwarf virus in vector cell Peterson, S.M., Thompson, J.A., Ufkin, M.L., Sathyanarayana, monolayers by fluorescent-antibody focus counting. Journal P., Liaw, L. and Congdon, C.B. (2014) Common features of of General Virology, 67, 2119–2124. microRNA target prediction tools. Frontiers in Genetics,5, Kimura, I. and Omura, T. (1988) Leafhopper cell cultures as a 23. means for phytoreovirus research. Advances in Disease Vector Rehmsmeier, M., Steffen, P., Hochsmann, M. and Giegerich, Research, 5, 111–135. R. (2004) Fast and effective prediction of microRNA/target Langmead, B., Trapnell, C., Pop, M. and Salzberg, S.L. (2009) duplexes. RNA, 10, 1507–1517. Ultrafast and memory-efficient alignment of short DNA se- Ruby, J.G., Stark, A., Johnston, W.K., Kellis, M., Bartel, D.P.and quences to the human genome. Genome Biology, 10, R25. Lai, E.C. (2007) Evolution, biogenesis, expression, and tar- Leaman, D., Chen, P.Y., Fak, J., Yalcin, A., Pearce, M., Un- get predictions of a substantially expanded set of Drosophila nerstall, U., Marks, D.S., Sander, C., Tuschl, T. and Gaul, microRNAs. Genome Research, 17, 1850–1864. U. (2005) Antisense-mediated depletion reveals essential and Sandmann, T. and Cohen, S.M. (2007) Identification of novel specific functions of microRNAs in Drosophila development. Drosophila melanogaster microRNAs. PLoS ONE, 2, e1265. Cell, 121, 1097–1108. Seitz, H., Tushir, J.S. and Zamore, P.D. (2011) A 5-uridine am- Lee, K.A., Kim, B., Bhin, J., Kim Do, H., You, H., Kim, plifies miRNA/miRNA* asymmetry in Drosophila by pro- E.K., Kim, S.H., Ryu, J.H., Hwang, D. and Lee, W.J. (2015) moting RNA-induced silencing complex formation. Silence, Bacterial uracil modulates Drosophila DUOX-dependent gut 2, 4. immunity via Hedgehog-induced signaling endosomes. Cell Sethupathy, P., Megraw, M. and Hatzigeorgiou, A.G. (2006) A Host Microbe, 17, 191–204. guide through present computational approaches for the iden- Liu, F., Peng, W., Li, Z., Li, W., Li, L., Pan, J., Zhang, S., Miao, tification of mammalian microRNA targets. Nature Methods, Y.,Chen, S. and Su, S. (2012) Next-generation small RNA se- 3, 881–886.

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 MicroRNAs in Sogatella furcifera 467

Singh, J. and Nagaraju, J. (2008) In silico prediction and sively parallel pyrosequencing-based transcriptome analyses characterization of microRNAs from red flour beetle (Tri- of small brown planthopper (Laodelphax striatellus), a vector bolium castaneum). Insect Molecular Biology, 17, 427– insect transmitting rice stripe virus (RSV). BMC Genomics, 436. 11, 303. Skalsky, R.L., Vanlandingham, D.L., Scholle, F., Higgs, S. and Zhang, Y. and Verbeek, F.J. (2010) Comparison and integration Cullen, B.R. (2010) Identification of microRNAs expressed of target prediction algorithms for microRNA studies. Journal in two mosquito vectors, Aedes albopictus and Culex quin- of Integrative Bioinformatics, 7, 1–13. quefasciatus. BMC Genomics, 11, 119. Zhang, Y., Zhou, X., Ge, X., Jiang, J.H., Li, M.W., Jia, S.H., Slonchak, A., Hussain, M., Torres, S., Asgari, S. and Khromykh, Yang, X.N., Kan, Y.C., Miao, X.X., Zhao, G.P., Li, F. and A.A. (2014) Expression of mosquito microRNA Aae-miR- Huang, Y.P.(2009) Insect-specific microRNA involved in the 2940-5p is downregulated in response to West Nile Virus development of the silkworm Bombyx mori. PLoS ONE,4, infection to restrict viral replication. Journal of Virology, 88, e4677. 8457–8467. Zhou, G., Wang, T., Lou, Y., Cheng, J., Zhang, H. and Xu, Stark, A., Brennecke, J., Russell, R.B. and Cohen, S.M. (2003) J.H. (2014) Identification and characterization of microRNAs Identification of Drosophila microRNA targets. PLoS Biol- in small brown planthopper (Laodephax striatellus)bynext- ogy, 1, e60. generation sequencing. PLoS ONE, 9, e103041. Stone, M., Jia, S., Heo, W.D., Meyer, T. and Konan, K.V.(2007) Zhou, R., O’hara, S.P. and Chen, X.M. (2011) MicroRNA regu- Participation of rab5, an early endosome protein, in hepatitis lation of innate immune responses in epithelial cells. Cellular C virus RNA replication machinery. Journal of Virology, 81, and Molecular Immunology, 8, 371–379. 4551–4563. Tu, Z., Ling, B., Xu, D.L., Zhang, M.X. and Zhou, G.H. (2013) Effects of southern rice black-streaked dwarf virus on the Accepted March 28, 2016 development and fecundity of its vector, Sogatella furcifera. Virology Journal, 10, 145. Supporting Information Wang, B. (2013) Base composition characteristics of mammalian miRNAs. Journal of Nucleic Acids, 2013, 951570. Additional Supporting Information may be found in the Wang, Y., Fan, H.W., Huang, H.J., Xue, J., Wu, W.J., Bao, online version of this article at the publisher’s web-site: Y.Y., Xu, H.J., Zhu, Z.R., Cheng, J.A. and Zhang, C.X. (2012) Chitin synthase 1 gene and its two alternative splic- Fig. S1 Characteristics of virus-derived small RNAs. ing variants from two sap-sucking insects, Nilaparvata (A) Percentages of (+) and (−) virus-derived small RNAs lugens and Laodelphax striatellus (Hemiptera: Delphaci- from SRBSDV. (B) Abundance of virus-derived small dae). Insect Biochemistry and Molecular Biology, 42, 637– RNAs from the 10 segments of SRBSDV genome. 646. Fig. S2 (A and B) Frequency of sRNA types observed Witkos, T.M., Koscianska, E. and Krzyzosiak, W.J. (2011) Prac- in viruliferous and non-viruliferous S. furcifera libraries. tical aspects of microRNA target prediction. Current Molec- (C and D) Frequency of 26–28 nt sRNA types observed ular Medicine, 11, 93–109. in viruliferous and non-viruliferous S. furcifera libraries. Wu, Q.F., Ding, S.W., Zhang, Y.J. and Zhu, S.F, (2015) Identifi- Fig. S3 The length distribution miRNA and miRNA* cation of viruses and viroids by next-generation sequencing in multiple insect species. The abbreviation represents and homology-dependent and homology-independent algo- the species name. aae: Aedes aegypti;ame:Apis mellif- rithms. Annu Review of Phytopathology, 53, 425–444. era; api: Acyrthosiphon pisum;bmo:Bombyx mori;dme: Xue, J., Bao, Y.Y., Li, B.L., Cheng, Y.B., Peng, Z.Y., Liu, H., Drosophila melanogaster; pxy: Plutella xylostella; tca: Xu, H.J., Zhu, Z.R., Lou, Y.G., Cheng, J.A. and Zhang, C.X. Tribolium castaneum. (2010) Transcriptome analysis of the brown planthopper Ni- Fig. S4 Alignment of miR-2796 precursors of multi- laparvata lugens. PLoS ONE, 5, e14233. ple insect species using Mega 5.02 software. Phyloge- Ye, J., Fang, L., Zheng, H.K., Zhang, Y., Chen, J., Zhang, Z.J., netic tree was constructed based on the alignment using Wang, J., Li, S.T., Li, R.Q., Bolund, L. and Wang, J. (2006) the neighbor-joining method. The abbreviation represents WEGO: a web tool for plotting GO annotations. Nucleic Acids the species name. mse: Manduca sexta;hme:Heliconius Research, 34, W293–W297. melpomene;bmo:Bombyx mori; tca: Tribolium casta- Zamore, P.D. and Haley, B. (2005) Ribo-gnome: the big world neum;sfu:Sogatella furcifera;nvi:Nasonia vitripennis; of small RNAs. Science, 309, 1519–1524. ame: Apis mellifera; api: Acyrthosiphon pisum. Zhang, F.J., Guo, H.Y., Zheng, H.J., Zhou, T.J, Zhou, Y.J., Wang, Fig. S5 GO classification of the S. furcifera miRNA S.Y., Fang, R.X., Qian, W. and Chen, X.Y. (2010) Mas- target genes.

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468 468 Z. X. Chang et al.

Fig. S6 (A) RT-qPCR analyses of miR-14 and miR-98a Table S2 The list of miRNAs found in S. furcifera (miR- expression levels in viruliferous and non-viruliferous S. n represents S. furcifera-specific miRNAs) furcifera cells. (B) The expression levels of three puta- Table S3 The list of miRNA clusters identified in S. tive target genes of miR-14 and miR-n98a determined by furcifera. RT-qPCR analyses in viruliferous and non-viruliferous S. Table S4 Arm usage among the clustered miRNAs at furcifera cells. different clustering distances. Table S1 Summary of the two small RNAs libraries Table S5 Three of the putative target genes of miR-14 from S. furcifera cultured cells and miR-n98a used for RT-qPCR analyses.

C 2016 The Authors Insect Science published by John Wiley & Sons Australia, Ltd on behalf of Institute of Zoology, Chinese Academy of Sciences, 23, 452–468