Profile of Exosomal and Intracellular microRNA in Gamma-Herpesvirus-Infected Lymphoma Cell Lines

Exosomes are small vesicles released from cells, into which microRNAs (miRNA) are specifically sorted and accumulated. Two gamma-herpesviruses, Kaposi sarcoma-associated herpesvirus (KSHV) and Epstein—Barr virus (EBV), encode miRNAs in their genomes and express virus-encoded miRNAs in cells and exosomes. However, there is little information about the detailed distribution of virus-encoded miRNAs in cells and exosomes. In this study, we thus identified virus- and host-encoded miRNAs in exosomes released from KSHV- or EBV-infected lymphoma cell lines and compared them with intracellular miRNAs using a next-generation sequencer. Sequencing analysis demonstrated that 48% of the annotated miRNAs in the exosomes from KSHV-infected cells originated from KSHV. Human mir-10b-5p and mir-143-3p were much more highly concentrated in exosomes than in cells. Exosomes contained more nonexact mature miRNAs that did not exactly match those in miRBase than cells. Among the KSHV-encoded miRNAs, miRK12-3-5p was the most abundant exact mature miRNA in both cells and exosomes that exactly matched those in miRBase. Recently identified EXOmotifs, nucleotide motifs that control the loading of miRNAs into exosomes were frequently found within the sequences of KSHV-encoded miRNAs, and the presence of the EXOmotif CCCT or CCCG was associated with the localization of miRNA in exosomes in KSHV-infected cells. These observations suggest that specific virus-encoded miRNAs are sorted by EXOmotifs and accumulate in exosomes in virus-infected cells.


Introduction
Exosomes are small membrane vesicles of 50-130 nm in diameter that are released by many cultured cells [1]. They contain not only cellular proteins but also mRNA and small RNA [2]. Small RNA including microRNA (miRNA) is particularly abundant in exosomes. Small RNA in exosomes has been thought to have various cellular functions, suggesting that exosomes are a tool for delivering miRNA to distant cells.
The genomes of large DNA viruses such as herpesviruses include sequences that encode miRNAs [3]. The genomes of two human gamma herpesviruses, Kaposi sarcoma-associated herpesvirus (KSHV, or human herpesvirus 8, HHV-8) and Epstein-Barr virus (EBV), encode at least 12 and 25 pri-miRNAs, respectively [4,5]. Virus-encoded miRNAs have been demonstrated to have various functions and to be associated with oncogenesis and viral replication [3,[6][7][8]. It has also been demonstrated that viral small RNA including miRNA is present along with host-encoded miRNA in exosomes released from virus-infected cells [9][10][11]. However, there is no information about the distributions of miRNA in cells and exosomes in virusinfected cells. Moreover, the detailed functions of the virus-encoded miRNA found in exosomes are unknown [11].
Recently, it was demonstrated that some miRNAs are selectively sorted into exosomes, and sequence motifs (EXOmotifs) that control their localization into exosomes were identified within miRNAs [12]. The human hnRNPA2B1 protein plays an important role in miRNA sorting into exosomes by binding specifically to the EXOmotifs in miRNAs. In this study, we revealed the distribution of virus-and host-encoded miRNAs between the exosomes and cellular fraction in KSHV-or EBV-infected lymphoma cell lines using a next-generation sequencer. The miRNA profile revealed that virus-encoded miRNAs were expressed at high levels in exosomes. In addition, we identified EXOmotifs that effectively promoted the loading of miRNAs into exosomes within virus-and host-encoded miRNAs in the virus-infected cells.

Exosome isolation
Exosomes were isolated following a previously described ultracentrifugation protocol [15]. After incubation for 72 h at 37°C with 5% CO 2 , cells were centrifuged at 300×g for 5 min. The supernatant was then passed through a 0.22-μm filter. This filtered supernatant was transferred to a fresh tube (50 mL) and centrifuged at 2,000×g for 30 min. The supernatant obtained from this procedure was then transferred to ultracentrifuge tubes and spun in a SW32Ti swinging bucket rotor (Beckman Coulter, Brea, CA, USA) at 12,000×g for 30 min at 4°C. The supernatant was again transferred to new ultracentrifuge tubes and spun for 70 min at 110,000×g. The supernatant was then discarded and the pellet was suspended in 1 ml of sterile phosphate-buffered saline (PBS). Samples were then transferred to 1.5-ml microtubes and supplemented with 200 μL of ExoQuick-TC (System Biosciences, Mountain View, CA, USA). After incubation at 4°C overnight, the mixture was centrifuged at 1,500×g for 30 min. The supernatant was then discarded and the pellet was centrifuged at 1,500×g for 5 min. Finally, the resulting pellet was suspended in 100 μL of sterile PBS.

Electron microscopy
A 6-μL aliquot of exosomes was absorbed onto glow-discharged 300-mesh heavy-duty carboncoated Cu grids (Veco grids; Nisshin EM, Tokyo, Japan) for 2 min and the excess was blotted Research Program on Emerging and Re-emerging Infectious Diseases (grant numbers 15fk0108011h0503 and 16fk0108119j0001) from the Agency for Medical Research and Development, and Grants-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science (grant number 15K08509). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing Interests:
The authors have declared that no competing interests exist. on filter paper (Whatman; GE Healthcare, Piscataway, NJ, USA). The grids were then washed twice with MilliQ water and negatively stained with 2% uranyl acetate. Data were collected using an H7700 transmission electron microscope (Hitachi, Tokyo, Japan) operating at 80 kV and 10,000× magnification.

RNA extraction
Total RNA was extracted from cells or exosomes using the High Pure miRNA Isolation Kit (Roche Diagnostics, Boehringer Mannheim, Germany), in accordance with the manufacturer's protocol. The extracted RNA was analyzed using small RNA chips on a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Small RNA profiles contained small RNA of between 4 and 40 nucleotides in length, which is consistent with miRNA.

Next-generation sequencing (NGS)
The TruSeq Small RNA-Seq Sample Prep Kit (Illumina, San Diego, CA, USA) was used in accordance with the manufacturer's protocol to prepare a small RNA library for sequencing. The quality and yield after sample preparation were measured using Bioanalyzer with a High Sensitivity DNA kit (Agilent Technologies) and corresponded to the expected 150 bp. DNA sequencing was carried out using Miseq (Illumina, San Diego, CA, USA) with MiSeq reagent kit v3, in accordance with the manufacturer's protocol. Sequence reads were analyzed with CLC Genomics Workbench (version 9.0; Qiagen). After adaptor trimming, reads of less than 15 or more than 26 nucleotides in length were removed, and all reads of 15-25 nucleotides in length were analyzed against miRBase release 21 retrieved from the miRNA database (http:// www.mirbase.org/). Homo_sapiens.GRCh37.57.ncrna was used as a comprehensive noncoding RNA database (http://www.ncrna.org/). All annotated reads matching pre-miRNA were counted as miRNA reads.

Real-time RT-PCR for miRNA
Copy numbers of human miRNAs were measured by ABI TaqMan 1 MicroRNA Assays in accordance with the manufacturer's protocol (Applied Biosystems, Foster City, CA, USA). In addition, the quantification of 17 KSHV-encoded miRNAs and a human cellular miRNA, miR21, was performed using the miScript Reverse Transcription Kit with miScript Primer Assays and miScript SYBR Green PCR Kit from Qiagen.

Statistical analysis
Data were analyzed by Student's t-test, Mann-Whitney U-test, or Chi-square test using SPSS software (IBM, Armonk, NY, USA).

Accession number
Sequence data of the small RNAs analyzed by NGS in this study were deposited in the DNA Data Bank of Japan (DDBJ; accession number: DRA004793; Bioproject PRJDB4918).

Profile of miRNAs by NGS
To reveal the miRNA profiles in exosomes and cells, small RNAs were extracted from exosomes and cells separately. Exosomes were derived from the supernatant of TY-1, BCBL-1, LCL, and Bjab cell cultures using Exo-Quick; transmission electron microscopy showed the presence of exosomes in the final pellet of Exo-Quick (Fig 1). No viral particle was observed in the exosome samples with electron microscopy. PCR analysis for KSHV and EBV DNA showed no or faint band in the exosome samples, suggesting no or low level of contamination of virus particles in the isolated exosomes (Fig 1). The small RNAs from the exosomes and cells were sequenced using a next-generation sequencer (Table 1). Small RNAs isolated from cells and exosomes were subjected to NGS, yielding averages of 1,712,881 and 1,536,143 reads for cellular and exosomal RNAs, respectively. The sequences that could be aligned to the reference genome represented 29.76-60.28% of exosomal miRNA reads and 54.81-88.42% of cellular miRNA reads (Table 1 and S1 Table). In KSHV-infected cell lines, TY-1 and BCBL-1, 48% of the annotated miRNAs in the exosomes originated from KSHV, while the proportion was 43-44% in cells. In EBV-transformed LCL, 7% of annotated miRNAs in the exosomes originated from EBV, while the proportion was 15.7% in cells.

Distribution of miRNAs in cells and exosomes
All mature annotated miRNAs were ranked according to their read counts in cells and exosomes of TY-1 and BCBL-1 (Tables 2 and 3). The most abundant mature miRNAs in exosomes were miRK12-4-3p in TY-1 and miRK12-8-3p in BCBL-1. In both of these cell lines, KSHV miRNAs were the most abundant mature miRNA. Eight and seven KSHV-encoded miRNAs ranked in the top 20 in exosomes of TY-1 and BCBL1, respectively. These KSHV-encoded miRNAs also had high rankings among the cellular miRNAs. Identified miRNAs were plotted based on the percentages of total read counts (Fig 2). miR-K12-4-3p, miR-K12-8-3p, miR-K12-2-5p, and miR-K12-3-5p were frequently expressed in both exosomes and cells of TY-1 and BCBL-1   (Fig 2A). miR-92a ranked second for exosomes but twelfth and thirteenth for cells in TY-1 and BCBL-1, respectively. Higher expression of miR-92a was also observed in LCL and Bjab cells (S2 and S3 Tables). miRNAs with different levels of expression between exosomes and cells are listed in S4 and S5 Tables. A volcano plot shows that 22 miRNAs were identified as exhibiting significantly different expression between cells and exosomes (P<0.05, Fig 3). The levels of miR143-3p, miR-486-5p, and miR-10b-5p were more than 200 times higher in exosomes than in cells (Fig 3). The high expression levels of these host-encoded miRNAs in exosomes were confirmed by real-time PCR (ABI TaqMan 1 MicroRNA Assays) in KSHV-infected cells (Fig  4). In addition, higher expression of miR16 and miR21 in cells than in exosomes was demonstrated by real-time PCR, except for in Bjab cells. In EBV-infected LCL, high expression of miR10b-5p, miR143, and miR145 was confirmed in exosomes by real-time PCR (Fig 4).

Mature miRNAs in exosomes and cells exactly or not exactly matching sequences registered in miRBase
The mature miRNAs included both exact mature miRNAs, which corresponded exactly to sequences registered in miRBase, and nonexact mature miRNAs, which had a deletion, addition, and/or mutation causing them to differ from the exact sequences. NGS revealed that exosomes contained nonexact mature miRNAs more frequently than cells ( Table 4). The proportion of exact mature miRNAs was 27-49% in exosomes, which was significantly lower than that in cells (34-56%, P<0.001, Chi-square test). KSHV miRNAs included nonexact mature miRNAs at a higher rate than EBV-and host-encoded miRNAs (Fig 5). Exact mature miRNAs were more common in cells than in exosomes for EBV-and host-encoded miRNAs, but not for KSHV-encoded miRNAs (Fig 5). High total read numbers of miRK12-4-3p and miRK12-8-3p were detected in TY1 and BCBL1, but they contained exact mature miRNAs at a low rate (Fig 6). NGS data indicated that the most abundant exact miRNA in KSHVencoded miRNA was miRK12-3-5p in TY1 and BCBL1 (Fig 6). High expression of exact mature miRK12-3-5p in the exosomes was confirmed by miScript PCR assay real-time PCR, which can efficiently detect exact mature miRNAs (Fig 7). A similar phenomenon was observed for EBV-encoded miRNAs. NGS revealed that these miRNAs included nonexact mature miRNAs at 38%-50% (Fig 5). The proportions of exact and nonexact mature miRNAs varied among the miRNAs (Fig 8). In addition, some EBV-encoded miRNAs differed in their expression levels between cells and exosomes (Fig 8). For example, miR-BART8-5p was found to be the most abundant EBV exact mature miRNA in cells, but was expressed at a low level in exosomes of LCL.

EXOmotifs and miRNA localization
A recent study identified several short sequence motifs (EXOmotifs) within miRNAs that guide their loading into exosomes [12]. NGS revealed that EXOmotifs were common within the sequences of the top 20 mature miRNAs found in exosomes of TY-1 and BCBL-1 (Tables 2 and 3). Next, to identify EXOmotifs within these miRNAs that efficiently promote sequestration into exosomes, we counted the reads of exosomal and intracellular mature miRNAs with specific EXOmotifs. Then, we calculated the fold change of miRNA expression between exosomes and cells for each EXOmotif by dividing the read number in exosomes by the read number in cells of miRNAs with the EXOmotif. The fold change between exosomes and cells varied among EXOmotifs (Fig 9). Two EXOmotifs, "CCCG" and "CCCT, " were associated with high fold change to exosomes of both KSHV-and host-encoded miRNAs. This supports the findings of a previous study in which these two motifs were characterized as EXOmotifs that strongly promote the transit of miRNAs into exosomes [12]. Several other EXOmotifs showed more than one fold change for host-encoded miRNAs, but not for KSHV-encoded miRNAs. Interestingly, two motifs (CGCC and TGCG) observed within KSHV-and EBV-encoded miR-NAs, but not within host-encoded miRNAs, showed more than four fold change to exosomes in KSHV-infected cells.

Discussion
In this study, we demonstrated the distribution of virus-and host-encoded miRNAs between the exosomes and the cellular fraction in KSHV-or EBV-infected lymphoma cell lines using a next-generation sequencer. Approximately half of the annotated miRNAs in the exosomes of KSHV-infected cells originated from KSHV, whereas 7% of the annotated miRNAs in the  exosomes of EBV-infected cells originated from EBV. Among the KSHV-encoded miRNAs, miR-K12-8-3p, miR-K12-4-3p, and miR-K12-2-5p were expressed highly in the exosomes of KSHV-infected cells. Moreover, miR-92a, mir-10b-5p, and mir-143-3p were identified as exosomal host-encoded miRNAs as their levels in exosomes were more than double those in cells. Exosomes contained mature miRNAs not exactly matching miRBase more frequently than cells. miR-K12-3-5p was identified as the most common KSHV-encoded exact mature miRNA in exosomes. The EXOmotifs "CCCG" and "CCCT" are likely to be associated with the localization of miRNAs in exosomes in KSHV-infected cells.
Previous studies on serum samples showed that exosomes contained virus-encoded miR-NAs (EBV, KSHV, JCV, and HCV) [15,[23][24][25]. However, the detailed profiles of virusencoded miRNAs in both exosomes and intracellularly have not been reported yet. This study showed the profile of exosomal and intracellular miRNAs in gamma-herpesvirus-infected lymphoma cell lines. As shown in Table 1, 48% of the annotated miRNAs in the exosomes originated from viruses, but the corresponding proportion was 43-44% in the cells of TY1 and BCBL1. This proportion is markedly higher than that for EBV miRNAs in EBV-infected cells. In EBV-infected LCL, the proportion of EBV miRNAs was only 7.15% in exosomes compared with 15.8% in cells. The higher proportion of KSHV miRNAs in exosomes than in cells suggests that there is a selective mechanism by which KSHV miRNAs are sorted into the exosomes. As shown in Tables 2 and 3, KSHV miRNAs identified abundantly in the exosomes frequently contained an EXOmotif within their sequences. Two of them, CGCC and TGCG, showing more than fourfold change of exosome/cell, were found only within KSHV-encoded miRNAs, but not within host-encoded miRNAs, in the exosomes among the top 20 miRNAs found in this location. In a previous study, it was demonstrated that CCCT and CCCG motifs  Table 4. were under the control of sumoylated hnRNPA2B1 in mammalian cells, for sorting to the exosomes [12]. It is assumed that KSHV-encoded miRNAs use a special sorting mechanism other than that for host-encoded miRNAs. An EXOmotif "CCCT" did not show an enrichment in exosomes of LCL (Fig 9). Moreover, Tables 2 and 3 shows that some miRNAs in exosome have no EXOmotif in their sequences. It has been reported that short nucleotide sequences such as miRNA guides the transport of RNAs to different subcellular compartments by various mechanisms [26][27][28]. For example, the terminal motif of miR-29b causes the nuclear enrichment of this miRNA [27]. Our observations on virus miRNAs in exosomes suggested the presence of unknown mechanism other than EXOmotif-dependent enrichment of miRNAs in exosomes.
NGS data clearly demonstrated that mature miRNAs expressed in both cells and exosomes contained nonexact mature miRNAs at a high rate. It should be noted that such nonexact forms of miRNA with the addition or deletion of nucleotides at the 3' end can theoretically not be detected by stem-looped real-time PCR [29]. To date, only exact mature miRNAs have been detected in quantitative studies using stem-looped real-time PCR [30]. Indeed, real-time PCR demonstrated a similar profile of miRNAs to that of exact mature miRNAs by NGS in part (Figs 6 and 7). However, there is a large difference between the two because reads of nonexact mature miRNAs were more abundant than those of exact mature miRNAs for most of the KSHV-encoded miRNAs. NGS also revealed that the exosomes contained nonexact mature miRNAs more frequently than cells. This was observed in all of the cells that we tested, suggesting the presence of a mechanism sorting nonexact mature miRNAs to the exosomes. This also suggests that exosomes might have a function to exclude nonexact mature miRNAs from cells and to concentrate mature and functional miRNAs in cells. On the other hand, nonexact

Conclusions
In this study, using NGS, we revealed the profile of virus-and host-encoded miRNAs in the exosomes released from KSHV-or EBV-infected lymphoma cell lines. The exosomes from KSHV-infected cells contained KSHV-encoded miRNAs at a high rate. Exosomes contained nonexact mature miRNAs more frequently than cells. EXOmotifs identified as nucleotide motifs that effectively promoted the loading of miRNAs into exosomes were frequently found within the sequences of KSHV-encoded miRNAs. These findings suggest that specific virusencoded miRNAs are sorted via EXOmotifs and accumulate in the exosomes of KSHV-infected cells. Exosomes have received considerable attention in recent years as they have been suggested to have potential uses as biomarkers, vaccines, and vehicles for gene therapy [31][32][33][34]. Information on the profiles of virus-and host-encoded miRNAs and their differences between exosomes and cells should be useful for establishing these potential applications.