Comparative Characterization of MicroRNAs from the Liver Flukes Fasciola gigantica and F. hepatica

MicroRNAs (miRNAs) are key regulators of gene expression at the post-transcription level. The present study specifically explored and compared the miRNA expression profiles of F. gigantica and F. hepatica using an integrated sequencing and bioinformatics platform and quantitative real-time PCR. Nineteen and 16 miRNA candidates were identified from F. gigantica and F. hepatica, respectively. The two parasites shared 11 miRNAs, with 8 also showing similarity to miRNAs of Schistosoma japonicum. Another 8 miRNAs were identified as F. gigantica-specific and 5 as F. hepatica-specific, most of which were novel. Predicted target analysis with 11465 mRNA and EST sequences of F. hepatica and F. gigantica revealed that all of the miRNAs had more than one target, ranging from 2 to 398 with an average of 51 targets. Some functions of the predicted targets were only found in F. gigantica, such as “transcription regulator”, while some others were only found in F. hepatica, such as “reproduction” and “response to stimulus”, indicating the different metabolism and gene regulation patterns of the two parasites. The present study represents the first global comparative characterization of miRNA expression profiles of F. gigantica and F. hepatica, which has provided novel valuable resources for a better understanding of the two zoonotic trematodes.


Introduction
The parasitic flatworms Fasciola hepatica and F. gigantica are common liver flukes which cause fasciolosis, a vector-borne disease with wide latitudinal, longitudinal and altitudinal dispersal in the world [1][2][3]. Fasciolosis can affect human beings and all ruminant animals of agricultural importance [4,5]. The geographically wide distribution of these parasites makes fasciolosis one of the most important helminth diseases with a global distribution [6][7][8][9]. In addition to livestock health significance, fasciolosis is also an important food-borne zoonosis in humans. Conservative estimates suggest that 2.4-17 million people are infected with Fasciola spp. worldwide with a further 180 million at risk of infection. Most of these infected or ''at-risk'' populations are located in the Americas, parts of Europe, South Africa, the Middle East and Asia [10][11][12]. Having considerable human and animal health significance, flukes of the genus Fasciola are among the largest and best known digenean trematodes, with F. gigantica and F. hepatica as the two key representatives capable of causing significant public health problems in humans [13]. The transcriptome of F. hepatica was recently investigated and reported, providing an improved insight into biological processes and microRNA (miRNA) targets in F. hepatica [14]. miRNAs are 18-25 nucleotide non-coding RNAs that regulate gene expression at the post-transcriptional level. They are now considered as a key mechanism of gene regulation in organ development, cellular differentiation, proliferation, function and even regulation of the immune system [15][16][17]. miRNAs are essential for the complex life cycles of pathogenic parasites, allowing them to respond to environmental and developmental signals [18,19]. The investigation of miRNA function in such pathogens can provide foundations for their control [20]. Due to the significance of miRNAs in parasite development and adaptation, and different environmental niches, there are likely to be differences in miRNA sequences and expression patterns between F. hepatica and F. gigantica, which will be of benefit in designing new tools for control. However, despite their socioeconomic importance and the zoonotic significance of the two species, miRNA research on F. hepatica and F. gigantica has not previously been reported. Therefore, we compared the miRNA expression profiles of F. hepatica and F. gigantica here, by nextgeneration sequencing, bioinformatic analysis and stem-loop realtime reverse transcription polymerase chain reaction (qRT-PCR).

Ethics Statement
The buffalo and beef cattle from which F. gigantica and F. hepatica adults were collected respectively, were taken from two local abattoirs (Nanning Slaughterhouse, Nanning City, Guangxi Zhuang Autonomous Region, China; Wensu Slaughterhouse, Wensu County, Xinjiang Uygur Autonomous Region, China, respectively). These animals were being processed as part of the normal work of the two abattoirs.

Parasites
Adults of F. gigantica and F. hepatica were collected from the gallbladders of slaughtered animals with naturally acquired infections. These worms were cleaned and stored as described previously [19]. Briefly, the worms were transferred to sterile physiological saline (37uC) in a sterile beaker, washed extensively with saline on a rotary shaker, transferred to Dulbecco's modification of Eagle's medium (DMEM) and incubated at 37uC (10% CO 2 ) for 3 h to allow the flukes to regurgitate all the gut contents from their digestive tracts. Subsequently, the flukes were transferred to RNase-free screw-top cryotubes containing RNAlater (Sigma), kept at 4uC overnight and then stored at 280uC. The specific identity of each worm was verified as F. hepatica or F. gigantica by morphological features and by isolating genomic DNA [21] and conducting PCR-coupled, bidirectional sequencing (Shenggong Co. Ltd. Shanghai, China) of the internal transcribed spacers (ITS-1 and ITS-2) of nuclear ribosomal DNA [22].

Small RNA Preparation and High-throughput Sequencing
Total RNA from three flukes of each species was prepared using Trizol reagent (Invitrogen) according to the manufacturer's protocol with some modifications. For the flukes that were preserved in RNAlater solution, 50% isopropanol were used at the precipitation step to gain a clear mix solution. Ten mg total RNA were separated by electrophoresis on a Novex 15% TBE-Urea gel (Invitrogen Co. Ltd) and RNA in the region of 18-30 nt was purified. A reverse transcription PCR (RT-PCR) kit (Invitrogen Co. Ltd) was used for reverse transcription of the small RNAs to form complementary DNA (cDNA). Finally, this RT-PCR product was purified using a 6% TBE PAGE gel (Invitrogen Co. Ltd) and sequenced with a Solexa Genome Analyzer (Illumina HiSeq 2000) at Huada Genomics Institute Co. Ltd, China.

High-throughput Sequencing and Computational Analysis
Clean reads were analyzed as described previously [19]. Briefly, adaptors, reads smaller than 18 nt and low quality sequences were trimmed from the raw data, and the reads were then searched against GenBank Rfam database (version 10.1) to remove noncoding RNAs including rRNA, tRNA, snRNA, snoRNA and other ncRNAs. RepeatMasker (http://www.repeatmasker.org) was used to eliminate repetitive sequences in the clean reads. The genome of Schistosoma japonicum (http://lifecenter.sgst.cn/ schistosoma/cn) was used as the reference genome for short read alignment using the program SOAP [23]. miRNA candidates were identified with precursors which had standard fold-back structure. A stem-loop hairpin was considered typical when the mature miRNAs presented in one arm instead of the loop of hairpin precursors. The free energy hybridization threshold of a stem-loop hairpin was set lower than 218 kcal/mol. Conserved miRNAs homologues were identified by matching the miRNA candidates with miRNAs deposited in the miRBase database (version 16.0). miRNAs with no match were identified as ''novel'' miRNAs.
A total of 3055 and 8410 mRNA and EST sequences of F. hepatica and F. gigantica respectively were downloaded from NCBI, and were combined into a single dataset. Potential targets of miRNAs of the two parasites were predicted with RNAhybrid software [24]. To reduce false-positive results, two extra parameters were imposed on the analysis: 1) the ggG was set as lower than 225 kcal/mol; 2) the P-value was set as #0.01. The Gene Ontology database (GO, http://www.geneontology.org/) was used for functional analysis of predicted targets. In addition, published studies focusing on miRNA function were also searched to inform molecular function in the present work.

Analysis of Novel miRNA Expression
The expression levels of novel miRNAs in adult F. hepatica and F. gigantica were analyzed with the modified stem-loop real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) using SYBR Green as described previously [25]. All of the primers were synthesized by Shenggong Co, Ltd., China. The PCR was performed using an ABI PRISMH 7300 Sequence Detection System and SYBR Green PCR Master Mix (TOYOBO). A 20 ml reaction mixture included 5 ml cDNA for each species (at 1:20 dilution), 5 mM forward and reverse primers, and 10 ml 26 SYBR Green PCR Master Mix. The cel-miR-lin4 was added into the mixture as the endogenous control in all real time PCR reactions. The primer pairs were as follows: forward 59-ACACTCCAGCTGGGTCCCTGAGACCTCAAGTG-39 and reverse 59-CTCAACTGGTGTCGTGGAGTCGGCAATT-CAGTTGAGTCACACTT-39. The cycling conditions were as follows: 95uC 5 min, followed by 30 cycles of 95uC for 15 s, 65uC for 15 s, and 72uC for 32 s. The quantification of each miRNA relative to the endogenous control was calculated using the equation: N = 2 -DCt , DCt = Ct miRNA -Ct lin4 [26]. All reactions were carried out in triplicate.

Results
Profile Characteristics of Short RNAs from F. hepatica and F. gigantica Deep sequencing yielded 16.41 and 10.36 million high-quality reads longer than 18 nt for F. hepatica and F. gigantica, respectively. Exons and introns accounted for a very small percentage of the clean reads, which indicated high integrity of the RNA in the samples. Among the reads, 14.4 (54.01%) million were common between the two species, and 8.7 and 3.5 million reads were F. hepaticaand F. gigantica-specific, which represented 2.9 and 0.48 million unique reads.
Non-coding RNAs (ncRNA), including rRNA, tRNA, snRNA, snoRNA, repeats, exons and introns, accounted for 4.3 (26.06%) and 2.5 (25.09%) million of the high-quality reads from F. hepatica and F. gigantica, respectively. The percentages of different kinds of ncRNA were similar for the two species. Repeat-associated small RNAs focused on two types of repeat: LINE/RTE:0 and LINE/ RTE:1 in both species.
A total of 8 and 5 miRNAs were identified as F. giganticaand F. hepatica-specific, respectively. For F. gigantica, one of these miRNAs matched with miR-4006b (cin-miR-4006b) of Ciona intestinalis, while for F. hepatica, one of these miRNAs matched with miR-1957 (mmu-miR-1957) of Mus musculus. The remaining miRNA candidates have not had match with known miRNAs, and were therefore novel miRNAs.
A distinguishing characteristic of the miRNAs was that most precusors had only one mature miRNA originating from the 5p or the 3p arm, respectively, except miR-2e which had mature miRNAs on both arms of its precursor. However, although the miR-2e had a star sequence, its star sequence had only one copy (Figure 1).

Target and Function Prediction
A total of 11465 mRNA and EST sequences from combined databases representing F. hepatica and F. gigantica, were used for target prediction of all miRNA candidates. It was found that all of the miRNAs matched with more than one target, ranging from 2 to 398, with an average of 51 targets (Table S1). For F. gigantica, some miRNAs had significantly high target numbers. For example, 3 novel miRNAs had target numbers of 114, 236, and 398. The accession numbers of targets are shown in Table S1. To identify miRNA differences between the two parasites, the best matched targets of the species-specific miRNAs of each parasite were selected for functional prediction and comparative analysis. All the targets in F. hepatica and F. gigantica were similar in catalytic and binding functions, but ''transcription regulator'' function was only found in the targets of F. gigantica. There were some significant differences in the biological processes of the identified targets: Metabolic process were similar between the two parasites, but some other processes, including immune system processing, locomotion, reproduction and response to stimulus, were only found in the targets of F. hepatica (Figure 2).

miRNA Quantification
The 6 novel miRNAs shared by the two parasites were selected for qRT-PCR analysis (Figure 3). The relative expression levels (REL) of three miRNAs were higher in F. hepatica than in F. gigantica: REL of miR-124 was 2.5260.65 in F. hepatica, which was 2.29 times higher than that in F. gigantica (1.1060.36); REL of miR-novel-04 was 2.6960.36 in F. hepatica, which was 2.69 times higher than that in F. gigantica (1.0060.05); REL of miR-novel-02 was 1.83 times higher in F. hepatica (1.4660.25) than in F. gigantica (0.0860.21).
REL of the other 3 miRNAs was found to be lower in F. hepatica than that in F. gigantica. The REL of miR-10 was 2.33 fold higher in F. gigantica (2.2160.16) than that in F. hepatica (0.9560.05); and the let-7 was 1.14 fold higher F. gigantica (1.1460.30) than that in F. hepatica (1.0060.16). For miR-novel-03, a 1.43 times higher expression level was found in F. gigantica (1.4360.07) than in F. hepatica (1.0060.06).

Discussion
In the present study, we obtained 16.41 and 10.36 million high quality reads greater than 18 nt from the total RNA of F. hepatica and F. gigantica, respectively. Among these reads, 8.7 and 3.5 million reads were F. hepaticaand F. gigantica-specific, which represented 2.9 and 0.48 million unique reads, indicating 6.04 times higher numbers of unique reads from F. hepatica than from F. gigantica. However, the numbers of miRNA candidates were similar between the two species, with 19 and 16 miRNA candidates identified in F. gigantica and F. hepatica, respectively. This phenomenon indicated a higher level of redundancy in the expression of small RNAs in F. hepatica than in F. gigantica, and further indicated a potential different mechanisms of gene regulation between the two parasites.
Most of the miRNAs were shared by F. gigantica (57.89%, 11/ 19) and F. hepatica (68.75%, 11/16) and most of the shared miRNAs (72.73%, 8/11) were fully conserved with miRNA sequences from S. japonicum as opposed to miRNAs from other species deposited in the miRBase database. When miRNAs of another trematode species Schistosoma mansoni [20], were included for analyses, an additional miRNA (sma-miR-125a) was identified as being commonly expressed in F. gigantica, F. hepatica and S. mansoni. As S. japonicum, S. mansoni, F. gigantica and F. hepatica are all trematodes, they are likely to have some similar metabolic processes and therefore share similar miRNAs. In terms of species-specific miRNAs, only one miRNA from each species could be identified which possessed a known homologue in other species. This phenomenon might indicate species-specific characteristics of the miRNAs of the two parasites compared with other parasites. It has been reported that an ''intermediate genotype'' of Fasciola spp. might be a hybrid of the two species [5,27]. We speculate that the intermediate-Fasciola may contain ''speciesspecific'' miRNAs from F. gigantica and F. hepatica, because significant regulatory mechanisms of each species would be genetically inherited by the hybrids. Furthermore, there may also be some intermediate-Fasciola specific miRNAs emerging, because of the key regulatory, and rapidly evolving characteristics of miRNAs, as we have described from other parasites including Clonorchis sinensis and Orientobilharzia turkestanicum [19,28].
Target prediction analysis showed that some of the miRNAs had substantially higher target numbers than others in F. gigantica with 114, 236, and 398, whereas the three highest target number of miRNAs of F. hepatica were 64 (Fhe-miR-novel-10), 39 (Fhe-miR-novel-07) and 23 (Fhe-mmu-miR-1957). One reason for this phenomenon might be that there were more mRNA and EST sequences available for F. gigantica (8410) than for F. hepatica (3055). To attempt to overcome this limitation, the mRNA and EST sequences of F. gigantica and F. hepatica were combined before target prediction. Moreover, the highest target number of F. gigantica was 6.22 times higher than that of F. hepatica, while F. gigantica mRNA and EST sequences represented only 2.75 times higher numbers in the combined interrogated dataset. Overall, target analysis indicated that the miRNA Fgi-miR-novel-03 had more targets than any of the others, which might indicate higher importance of the miRNA for the parasite than other miRNAs. This phenomenon can be further verified by the genome copy number of Fgi-miR-novel-03: it occurs in six locations on the reference genome, while the others are only represented one or twice. Increased copy number within a genome is an effective approach to improve the expression of a regulator [29,30]. This result therefore indicated an obvious regulating difference of miRNAs of the two parasites. Furthermore, the metabolic differences between the two parasites were verified by the analysis of the functions of predicted targets: some targets and functions were found to be F. gigantica-specific, while some others were F. hepatica-specific.
In conclusion, the present study represents the first global comparative characterization of miRNA expression profiles of F. gigantica and F. hepatica, and has provided a novel, unique resource to facilitate fundamental and applied molecular investigations of the two key representatives of liver flukes, which in turn has implications for the better control of human and animal fasciolosis.