Identification and Analysis of the Porcine MicroRNA in Porcine Cytomegalovirus-Infected Macrophages Using Deep Sequencing

Porcine cytomegalovirus (PCMV; genus Cytomegalovirus, subfamily Betaherpesvirinae, family Herpesviridae) is an immunosuppressive virus that mainly inhibits the immune function of T lymphocytes and macrophages, which has caused substantial damage in the farming industry. In this study, we obtained the miRNA expression profiles of PCMV-infected porcine macrophages via high-throughput sequencing. The comprehensive analysis of miRNA profiles showed that 239 miRNA database-annotated and 355 novel pig-encoded miRNAs were detected. Of these, 130 miRNAs showed significant differential expression between the PCMV-infected and uninfected porcine macrophages. The 10 differentially expressed pig-encoded miRNAs were further determined by stem-loop reverse-transcription polymerase chain reaction, and the results were consistent with the high-throughput sequencing. Gene Ontology analysis of the target genes of miRNAs in PCMV-infected porcine macrophages showed that the differentially expressed miRNAs are mainly involved in immune and metabolic processes. This is the first report of the miRNA transcriptome in porcine macrophages and an analysis of the miRNA regulatory mechanisms during PCMV infection. Further research into the regulatory mechanisms of miRNAs during immunosuppressive viral infections should contribute to the treatment and prevention of immunosuppressive viruses.


Introduction
Porcine cytomegalovirus (PCMV) is a member of the genus Cytomegalovirus, subfamily Betaherpesvirinae, family Herpesviridae. It is an icosahedral virus with a double-stranded linear DNA genome. PCMV is now widely distributed around the world. China, Britain, the United conducted a comprehensive analysis of the miRNA expression profile of porcine macrophages infected with PCMV. We also predicted host targets of the pig-encoded miRNAs that were significantly differentially expressed (DE).

Materials and Methods
Virus and cells for high-throughput sequencing PCMV Sichuan strain and porcine macrophages (provided by the Animal Biotechnology Center of Sichuan Agricultural University) were used in this study. Cell cultures were performed in Corning 25-cm 2 culture bottles (Corning, Shanghai, China) containing modified RPMI-1640 nutrient solution (Thermo Fisher Scientific, Waltham, UK) supplemented with 10% fetal bovine serum and 50 mg/ml penicillin/streptomycin antibiotic solution (Gibco, Beijing, China), and were maintained under standard conditions of 5% CO 2 and 37°C. Porcine macrophages were infected with PCMV at 10 PFU per cell, and then PCMV-infected and uninfected porcine macrophages were collected at 72 h postinfection (hpi). Our study used pooled samples and experiments were performed in triplicate. All infected and control samples were stored until use in a single RNA isolation assay.

RNA isolation
The total RNA from the PCMV-infected and uninfected porcine macrophages described above was extracted using Trizol reagent (Invitrogen, Shanghai, China), in accordance with the manufacturer's instructions. The purity and concentration (ratio OD 260 /OD 280 ) of total RNA samples were determined with a NanoDrop ND-1000 spectrophotometer (Nano Drop Inc., Wilmington, DE, USA).

Small RNA library construction and sequencing
The total RNA from each sample was sequentially ligated to 3 0 and 5 0 small RNA adapters. cDNA was then synthesized and amplified using Illumina's proprietary RT primers and amplification primers. Subsequently, 120-140-bp PCR-amplified fragments were extracted and purified from the PAGE gel. Next, the completed libraries were quantified using an Agilent 2100 Bioanalyzer. The samples were diluted to a final concentration of 8 pM, and cluster generation was performed on an Illumina cBot using a TruSeq Rapid SR cluster kit (#GD-402-4001, Illumina), following the manufacturer's instructions. Finally, high-throughput sequencing was performed on an Illumina HiSeq 2000 using TruSeq Rapid SBS Kits (#FC-402-4002, Illumina), in accordance with the manufacturer's instructions [24,25].

miRNA-seq data analysis
After sequencing, the total raw sequencing reads were filtered using the CHASTITY quality control filter. The clean reads that passed quality filtering were then further filtered by searching for reads whose 3 0 -ends aligned to at least 6 nt of the 3 0 -adapter. The adapter sequences were trimmed and the adapter-trimmed reads (15 nt) were aligned to the known S. scrofa pre-miRNA sequences in miRBase 19 (http://www.mirbase.org/) using Novoalign software (v2.07.11) with at most one mismatch. The unmapped reads and longer reads (length >15 nt) were discarded as unusable. Any sequences that matched the mature miRNA region, ±4 nt within a known pre-miRNA hairpin, were grouped and the 5-prime (5p) or 3-prime (3p) arms of the precursor hairpin were also indicated. The RNA sequence reads of rRNA, snRNA, tRNA, and snoRNA were excluded [26].
To correct for the difference in tag counts between samples, the tag counts were scaled to tpm (the copy number of transcripts per million) based on the total number of tags aligned to known S. scrofa pre-miRNAs in miRBase 19.
To predict novel pig-encoded and virus-encoded miRNAs, the miRDeep2 web server (http:// www.mdc-berlin.de/en/research/research_teams/systems_biology_of_gene_regulatory_elements/ projects/miRDeep/) was used. The fold-back secondary structures of the pig-encoded miRNAs were predicted using Mfold software (http://mfold.rna.albany.edu/). We arranged all the sequence data from the 3 0 adapter-trimmed files, and the adapter-trimmed sequences whose length was <17 bp or had mismatches >1 bp were excluded. The significantly DE miRNAs for the infected sample vs. the uninfected sample were determined by fold change filtering. (A fold change 2.0 indicates that the specific miRNA in a PCMV-infected sample is upregulated, while a fold change 0.5 indicates that the specific miRNA in a PCMV-infected sample is downregulated.)

Target prediction and functional enrichment of DE miRNAs
The database microT-CDS (http://diana.imis.athena-innovation.gr/DianaTools/index.php? r=microT_CDS/index), the positional relationships between miRNAs, and genomic annotation sets were used for the prediction of targets of the 130 DE miRNAs [27]. The target genes of the DE miRNAs were evaluated using Gene Ontology (GO). The GO project's (http://www. geneontology.org) online server was used to provide a controlled vocabulary of terms for describing gene product characteristics and gene product annotation data from GO consortium members. We used Fisher's exact test to determine whether there was more overlap between the DE list and the GO annotation list than would be expected by chance. The p-value denotes the significance of the enrichment of GO terms in the DE genes. The GO terms with p < 0.05 were considered significant. The WEGO software (http://wego.genomics.org.cn/cgibin/wego/index.pl) was used to produce histograms of the GO annotations [28].

Stem-loop RT-PCR
The total RNA from each sample was extracted using Trizol reagent (Invitrogen), in accordance with the manufacturer's instructions. After RNA had been reverse-transcribed using an RT-PCR kit (GeneCopoeia, Guangzhou, China), the RT products were amplified using specific primers (Table 1), and the amplification conditions were as follows: 94°C for 3 min, followed by 40 cycles of 95°C for 20 s, 60°C for 40 s, and 72°C for 20 s, with a final elongation at 72°C for 5 min. RT-PCR reactions were carried out using the SYBR Green PCR Core Reagents Kit (Applied Biosystems, Foster City, CA, USA) on an ABI Prism 7900 Sequence Detection System (Applied Biosystems), in accordance with the manufacturer's instructions.

Results
Overview of the high-throughput sequencing data PCMV is an important immunosuppressive virus; however, an understanding of its immune escape mechanisms is still lacking. To investigate the miRNA transcriptome in PCMV-infected cells, as well as to clarify the regulatory mechanisms of miRNAs during PCMV infection, we used a high-throughput sequencing technology on libraries of small RNAs from PCMVinfected and uninfected porcine macrophages. In this study, 3,987,148 clean reads and 3,575,737 adapter-trimmed reads (reads that have passed quality filtering, adapter filtering, and length filtering with length 15 nt) were generated by uninfected porcine macrophages, and 7,567,884 clean reads and 7,295,330 adapter-trimmed reads were generated by infected cells (NCBI GEO Accession number: GSE76156). Of these, 5,543,830 and 2,215,640 reads from PCMV-infected and uninfected porcine macrophages, respectively, aligned to known S. scrofa pre-miRNAs in miRBase 19 and the Ensembl database.
The size distributions of the adapter-trimmed reads were similar in PCMV-infected and uninfected porcine macrophage libraries, and most of the reads were 22 nt in length (Fig 1). In addition to miRNAs, a series of noncoding small RNAs (sRNAs) were also detected in PCMVinfected and uninfected samples.

Analysis of the pig-encoded miRNA expression profile
All filtered data were processed using Illumina's Genome Analyzer Pipeline. When these data were aligned with the S. scrofa reference genome, 594 miRNAs were detected. Of these, 239 mature miRNAs were already annotated in miRBase 19, and 355 novel pig-encoded miRNAs were discovered in this study (S1 Table).

Differential expression of porcine macrophage miRNAs
Of the 594 detected miRNAs, 404 (68%) were coexpressed in PCMV-infected and uninfected porcine macrophages; however, 123 (20.7%) and 67 (11.2%) were expressed in the PCMVinfected and uninfected porcine macrophages, respectively (S2 Table, Table 2). The data on the differential expression of porcine macrophage miRNAs are displayed in a scatter plot (Fig 2). The deep sequencing data also showed that 130 unique pig-encoded miRNAs were significantly DE between the uninfected and PCMV-infected samples. Of these 130 miRNAs, 73 (56.2%) were significantly upregulated in PCMV-infected samples, while 57 (43.8%) were downregulated in them (Fig 3).

Target prediction and gene functional annotation
To explore the regulatory roles of the DE miRNAs from this study, web prediction programs were used to predict their putative target genes (S3 Table). The miRNA regulatory network map generated by the web prediction programs showed that multiple miRNAs share the same target genes (S1 Fig). There were also many miRNAs targeting immune-related genes. For example, ssc-miR-19a targets TNFα, IL-12α, and IL-20; ssc-miR-19b targets TNFα and IL-20; ssc-miR-7 targets IL-1β, IL-2, and TGF-β1; ssc-let-7a targets IFN-delta-1; ssc-miR-30a-5p targets IL-12α and IL-12β; and ssc-miR-24-3p targets IL-1α. The results of the GO annotation reflected the regulatory roles that the pig-encoded miR-NAs play in host physiological processes, including the regulation of cellular components, molecular functions, and biological processes. The target genes of the 130 DE miRNAs are mainly related to biological regulation of cellular, metabolic, and immune system processes (Fig 4, S4 Table). Of the target genes of DE miRNAs, 14.4% are related to immune system processes, which suggests that the changes in the miRNA expression profile during viral infection may affect immune function.

Confirmation of the relative abundance of pig-encoded miRNAs
To confirm the relative abundance of pig-encoded miRNAs in PCMV-infected and uninfected porcine macrophages, we designed stem-loop real-time PCR primers to detect 10 DE miRNAs. We then quantified the expression level of these miRNAs, and the results obtained via RT-PCR were highly consistent with the high-throughput sequencing data (Table 3, S5 Table).

Discussion
Similar to HCMV, PCMV also belongs to the herpes virus family and has the biological characteristic of latency activation [5]. Because of the high homology between human and pig genomes, further research on PCMV should also contribute to the prevention and treatment of  HCMV [1]. Additionally, PCMV is the major threat to pig-human xenotransplantation, and the immunosuppressive therapy necessary after surgery may aggravate an infection of PCMV; therefore, further research into the molecular mechanisms of PCMV infection should contribute to the development of xenotransplantation technology [29].
To better understand the roles of miRNAs in immune suppression mechanisms and pathological processes, a new-generation high-throughput sequencing platform (Illumina HiSeq 2000) was used in this study to determine the miRNA transcriptome in PCMV-infected cells. Unfortunately, PCMV-encoded miRNAs were not detected in this study, so whether PCMVencoded miRNAs exist in infected hosts needs further confirmation.
In this study, PCMV-infected porcine macrophages were used as a model for researching the regulatory function of miRNAs during the infection of an immunosuppressive virus, and a series of mammalian-encoded miRNAs were detected. Under the stress of PCMV infection, 73 miRNAs were upregulated and 57 miRNAs were downregulated in porcine macrophages compared with their expression levels in the uninfected group, of which four specific upregulated miRNAs (ssc-miR-424-3p, ssc-miR-novel-chr15_16769, ssc-miR-novel-chr16_17559, ssc-miR-novel-JH118654-1_42487) were expressed only in PCMV-infected porcine macrophages. This shows that virally infected cells might affect the expression of immune-related miRNA target genes through regulating the expression level of specific miRNAs, thereby regulating the immune process. The miRNA expression level is known to be significantly reduced in cancer tissue relative to that in healthy tissue [30], but in the present study, the miRNA expression level was significantly increased after PCMV infection (adapter-trimmed reads of PCMV-infected and uninfected macrophages numbered 3,575,737 and 7,295,330, respectively). This suggests that the host miRNA expression profile was affected by viral infection. Interestingly, viral infection in porcine macrophages triggered the significant differential expression of some specific miRNAs, which has also been confirmed in other studies [31,32]. This indicates that these specific miR-NAs may play a similar regulatory function during viral infection.
In this study, a series of miRNA isomers (isomiR) were present in the miRNA expression profile. Although the phenomenon of isomiR variability is mainly due to the variability of the Dicer or Drosha cleavage positions on pre-miRNA hairpins, it may be related to the genomic variation under viral infection or tumorigenesis. Alternatively, different algorithms or different sequencing techniques may also lead to length and sequence heterogeneity of miRNAs [33,34].
miRNA is known to be an extremely important part of the regulatory mechanisms of gene expression [35]. The expression profile and functional studies of miRNA have also revealed that miRNAs are involved in the regulation of almost all known cellular processes, and changes   [36,37].
Although the role of miRNAs in cell development and tumorigenesis has been described in earlier studies, the importance of these small RNA molecules in the development and function of the immune system has only just begun to attract attention [38][39][40]. Recent studies have shown that miRNA expression is an early cellular response against exogenous or endogenous signal pressures. Additionally, to regulate the processes of specific immunity, miRNAs are also involved in the regulatory mechanisms of the innate immune response, including the proliferation of neutrophils and mononuclear cells [41]. The regulatory roles of miRNAs in immune processes include assisting in the development and differentiation of B and T lymphocytes and  Comparison of the levels of miRNA expression detected by qRT-PCR and high-throughput sequencing. "+" and "−" indicate upregulated and downregulated miRNAs, respectively. The fold change cutoffs of the upregulated miRNAs and the downregulated miRNAs were 2 and 0.5, respectively. qRT-PCR Ct threshold is 0.015.
doi:10.1371/journal.pone.0150971.t003 natural killer cells, and inducing the proliferation of mononuclear cells and neutrophils [42][43][44]. Therefore, the identification and characterization of the regulatory miRNAs involved in the immune response and exploration of their pathological roles in host-pathogen interactions are essential for the treatment and control of viral diseases. Currently, widespread immunosuppressive viruses are causing huge losses to the farming industry. The discovery of a regulatory function for miRNAs in immunity provides us with a novel approach to prevent and treat viral diseases [45]. Pig is a major animal model for life science research; therefore, the discovery of a regulatory role for miRNAs during infection with an immunosuppressive virus in pigs should contribute to our understanding of how to overcome human infection with immunosuppressive viruses.

Conclusion
In this study, we obtained the miRNA expression profile of PCMV-infected porcine macrophages by high-throughput sequencing. A total of 239 miRNA database-annotated and 355 novel pig-encoded miRNAs were detected. Of these, 130 miRNAs showed significant differential expression between the PCMV-infected porcine macrophages and the control samples.