Application Of Small RNAs Deep Sequencing Technology For Identifying Viruses From Woody Plants

Plant viruses are important pathogens causing great damage to crop production. Wood plants have been used for thousands of years for both fuel and as a construction material, and play an important role in social economy and maintaining ecosystem. Viruses in woody plants have a wide range of geographical distribution, and have a great potential threat to wood production and ecosystem. Due to the low virus titer, it is extremely difficulty for virus detection and discovery in woody plant using traditional methods. High-throughput sequencing technology is a milestone technology in genomic biology and has been used as a powerful tool for studies on plant pathogens, especially for plant virus identification. This technology has broken out the limitations of traditional virus detection methods, which can simultaneously detect the known and unknown RNA viruses, DNA viruses and viroids in plant samples even with low virus titer. This method has already opened up the field of large-scale, rapid diagnosis of plant viruses. In this research, in order to comprehensively understand the diversity of woody plant viruses, small RNAs deep sequencing technology is used to detect viruses infecting several woody plants in China. Our research is helpful for the discovery of new plant viruses and the prevention of viral disease epidemic in woody plants.Field samples of three woody plants were collected in Zhejiang province and used to detect viruses by high-throughput sequencing technology, the samples include Wisteria sinensis leaves showing mosaic symptom collected in Hangzhou, Nandina domestica leaves showing mosaic symptom collected in Linan, and mulberry (Morus alba L.) leaves showing mosaic and roll symptoms collected in Tongxiang. Using small RNAs deep sequencing technology, Wisteria vein mosaic virus (WVMV), Cucumber mosaic virus (CMV) and Mulberry mosaic leaf roll-associated virus (MMLRaV) were identified, respectively, in the three samples. To our knowledge, this is the first time report of CMV infecting N. domestica in China.Leaf samples showing mosiac symptom were collected from pecan(Carya illinoinensis) in Lin’an, Zhejiang province. Using small RNAs deep sequencing technology, an unknown RNA virus was identified. This novel virus is classified into the family Potyviridae, and the name Pecan mosaic-associated virus (PMaV) is tentatively proposed. The PMaV genome has 9310 nucleotides (nt) in length, with the contents of A, C, T, G of 35.35%,25.49%,18.49% and 20.66%, respectively. It has a 57-nt length 5’-UTR and a 224-nt 3’-UTR. PMaV genome possesses a typical organization characteristic of the genus Potyvirus, encoding a single open reading frame (ORF) of 3026 amino acids (aa), which then undergoes proteolytic processing to form 11 gene products:PI, P3,6K1, C1,6K2, NIa-VPg, NIb, HC-Pro, NIa-Pro, CP and PIPO. Proteolytic cleavage sites of PMaV polyprotein are predicted as follows: LYWRGF/S (P1/HC-Pro), KLYKVG/G (HC-Pro/P3), DVITLQ/S (P3/6K1), CMTFQ/S (6KI/C1), DAIRFQ/S (C1/6K2), DTITLQ/G (6K2/NIa-VPg), DEIQFE/A (NIa-Vpg/NIa-Pro), DSMKFQ/G (NIa-Pro/NIb), and AGASAQ/S (NIb/CP). Nucleotide/animo acid identity and phylogenetic analysis were carried out between PMaV and other full-sequenced potyviruses. Our results indicated that PMaV and. Lettuce mosaic virus (LMV) had the highest nucleotide sequence identity (58.9%). PMaV shared 52%-53% amino acid sequence identities with other potyvirus encoding proteins. PMaV CP had the highest sequence identity with LMV on nt level (73%) and with Turnip mosaic virus (TuMV) on aa level (68%). Phylogenetic analysis also suggested that PMaV and LMV were the most closely related species based on aa sequence identities of the ORFs. Our results showed PMaV is a new Potyvirus species based on both the criteria demarcation for a new Potyvirus species from Admas et al (complete nucleotide identity less than 76%, and the ORF amino acid identity less than 82%) and ICTV (CP amino acid identity less than 80%, and complete nucleotide identity less than 85%). The PMaV can’t be mechanically transmitted to Nicotiana benthamiana and N. tabacum. Furthermore, by using negative staining and transmission electron microscope observation of the infected pecan leaves, we could find the typical Potyvirus linear viral particles which showed non-enveloped, flexuous, filamentous, and of 750 nm in length.Lemon (Citrus limon) samples displaying stunting, leaf roll and mottle symptoms were collected from in Dehong Autonomous Prefecture, Yun’nan province. Using high-throughput sequencing and the assembly of small RNAs, we showed that the Hop stunt viroid (HSVd) was present in C. limon leaf samples. The full length of HSVd shared 100% similarity with HSVd citrus strain CC-D isolate Hop 1 (GenBank accession number:FJ716188.1). HSVd derived siRNAs (HSVd-siRNAs) in C. limon were mainly 21 nucleotides in length, and nearly equal amount of HSVd-siRNAs originated from the plus-genomic RNA strand as from the complementary strand. A bias of HSVd-siRNAs toward sequences beginning with a 5’-Guanine was observed. Furthermore, hotspot analysis showed that a large amount of HSVd-siRNAs derived from the central and variant domains of the HSVd genome. Infection and symptom development of plant viroids are related to viroid-derived siRNAs and "Cachexia expression motif" of HSVd, which is composed of 5-6 nucleotides located in the variable domain (V domain) modulates host symptoms. An algorithm was created with perl script and was used for selective extraction of the HSVd-siRNAs that allowed the first ten nucleotides at the 5’end could cover the "cachexia expression motif. The 67 HSVd-siRNAs were then used for target genes prediction by psRNATarget tools, and the five target genes with highest value, as well as their related functions in symptom development were analysed. The possibilities that these five genes might be targeted by its corresponding HSVd-siRNA were also validated by using other miRNA target prediction software RNA22. Our results suggest that the large amounts of HSVd-siRNAs from central and variant domains might be involved in interference with host gene expression and affect symptom development.Clerodendrum cyrtophyllum leaf samples showing mosaic symptom were collected in Lin’an, Zhejiang province. Using small RNAs deep sequencing, clerodendrum golden mosaic china virus (ClGMCNV) was identified in the samples. The full length of DNA-A was 2776 bp, and shared 99% similarity with the DNA-A of C1GMCNV isolate YX1 (GenBank accession number FN396962.1). The full length of DNA-B was 2729 bp, and also shared 98% similarity with the DNA-B of ClGMCNV isolate YX1 (GenBank accession number FN396963.1). Deep sequencing data were analyzed using the Velvet and Velvet-Oases methods with different parameters (k-mer value 15,17,19,21, respectively). Using the Velvet method, when K-mer was 15, the number and total length of contigs assembled was the largest, which covered 70% of the full length of CIGMCNV genome; while K-mer was 21, the total length of contigs covered only about 30% of the genome. Using the Velvet-Oases method, we assembled the contigs with different K-mer values and found that the coverage proportions of the contigs were significantly increased, with the highest number reached to 88.25%. In order to see the distribution of the vsiRNAs along CIGMCNV genome more clearly, we used visualization tool SeqMonk to carry out the whole genome scan of the vsiRNAs sequences. The results showed clearly that some regions along CIGMCNV genome did not produce any vsiRNAs.