A novel computational approach to the silencing of Sugarcane Bacilliform Guadeloupe A Virus determines potential host-derived MicroRNAs in sugarcane (Saccharum officinarum L.)

Introduction

Sugarcane bacilliform virus (SCBV, genus: Badnavirus, family Caulimoviridae) is a circular, non-enveloped bacilliform, monopartite DNA virus that harbors about 7.5 kb double-stranded (ds) DNA with three open reading frames (ORFs) (Sun et al., 2016). SCBV was first reported in the sugarcane cultivar (B34104) in the South American continent (Autrey et al., 1992) and disseminated into many sugarcane (Saccharum officinarum L.)-growing regions around the globe (Autrey et al., 1992). SCBV is implicated in yield losses in sugarcane cultivars. SCBV infection may lead to significant losses in biomass production in susceptible sugarcane cultivars (Lockhart & Autrey, 2000). SCBV infect sugarcane cultivars utilized for sugar and bioenergy production and constitute a major threat to the global exchange of sugarcane germplasm (Borah et al., 2013). Variable follicle symptoms were observed in some sugarcane varieties showing severe chlorotic mottling, stunted growth and pronounced freckle (Lockhart, Ireyt & Comstock, 1996; Viswanathan, Alexander & Garg, 1996). SCBV has wide host range. It comprises members of plant families Poaceae (sugarcane, johnsongrass, rice, sorghum, Panicum maximum, Rottboellia exaltata and Brachiaraia sp.) and Musaceae (bananas) (Borah et al., 2013; Lockhart & Autrey, 2000; Lockhart, Ireyt & Comstock, 1996). SCBVs are highly endemic in S. officinarum germplasm. These viruses can also infect commercial sugarcane (Saccharum interspecific hybrids), S. barberi, S. robustum, S. spontaneum and S. sinensis (Lockhart, Ireyt & Comstock, 1996). SCBV is transmitted naturally by sap-sucking insect vector mealybug species (Saccharicoccus sacchari and Dysmicoccus boninsis) through vegetative cuttings. Agrobacterium-mediated inoculation is used to disseminate SCBV infection experimentally (Bouhida, Lockhart & Olszewski, 1993; Lockhart, Ireyt & Comstock, 1996; Viswanathan, Alexander & Garg, 1996). Previous studies have revealed that SCBV is a heterogeneous badnavirus species. SCBV exist in nature as a complex of viruses or distinct strains and genetic variants. Currently, four distinct SCBV species have been classified by the International Committee on Taxonomy of Viruses (ICTV). These are Sugarcane Bacilliform Guadeloupe A Virus (SCBGAV), Sugarcane Bacilliform Guadeloupe D Virus (SCBGDV) Sugarcane bacilliform MO virus (SCBMOV-MOR) and Sugarcane bacilliform IM virus (SCBIMV-QLD) originating from Guadeloupe, Morocco and Australia, respectively (Adams & Carstens, 2012; Adams et al., 2016; Bouhida, Lockhart & Olszewski, 1993; Geijskes et al., 2002; Muller et al., 2011).

SCBGAV was first identified as a sugarcane R570 hybrid variety in Guadeloupe in 2011 (Muller et al., 2011). SCBGAV was classified as a new species in the genus badnavirus under the SCBV-A group as assigned by ICTV (Adams et al., 2016). The disease symptoms appear in the form of chlorosis, leaf freckle and mottling. Some infected plants were also observed to exhibit no symptoms. Recently in China, SCBV-infection resulted in reduced sucrose content, juice, stalk weigh, purity and gravity in sugarcane plants (Ahmad et al., 2019).

SCBGAV is a mealybug-transmitted monopartite badnavirus that infects sugarcane and is composed of a 7,444 bp circular, dsDNA genome containing three ORFs on the positive strand that replicate by a virus-encoded reverse transcriptase (RT). The two small proteins of sizes (176–185 and 122–135 amino acids), were encoded by the ORF1 and ORF2 respectively. The ORF3 was a key component of SCBGAV genome, encoded by a polyprotein (1,786–1,933 amino acids) (Muller et al., 2011).

Sugarcane (Saccharum officinarum L.) has inherited an active immunity that is composed of microRNAs (miRNAs) to combat infection. The miRNAs are a class of 21–23 nucleotide-long, endogenous, noncoding RNAs that govern gene regulation and expression at the post-transcriptional level. In plants, this mechanism occurs through the process of mRNA cleavage or translational repression. They are synthesized after the processing of hairpin miRNA precursors using an RNase-III-like enzyme (Dicer) to control gene expression (Brodersen & Voinnet, 2006).

RNA silencing, through miRNAs present in the host plant, thus imparts natural immunity and resistance to the host against foreign genetic elements including plant viruses (Carbonell & Daròs, 2019; Qu, Ye & Fang, 2007).

Viral miRNA degradation is the simplest molecular approach to combat viral infection in plants. Successful demonstration about changes in the nucleotide transcripts of mature miRNAs does not affect their biogenesis and the application of modified miRNAs to target Potyviruses in Arabidopsis was reported first time (Niu et al., 2006). Several studies have been successful in generating resistance against plant viruses using amiRNA-technology in crop plants such as tomato (Chen et al., 2019; Zhang et al., 2011) and cotton (Akmal, Baig & Khan, 2017; Ali et al., 2013). Most recently, amiRNA-mediated gene silencing has been demonstrated against Potexvirus and Tobamovirus successfully (Petchthai, Le Yee & Wong, 2018).

miRNA: mRNA target identification, validation and interactions are the basis for discerning the important functions of miRNAs in the greater perspective of regulatory network leading to different biotic developments. Computational algorithms for identification and analysis of miRNA targets enable the process of narrowing down possible target binding sites for experimental validation at transgenic level. In-silico strategies model how miRNAs target specific mRNAs and a number of algorithms are accessible, each with a diverse approach to miRNA target prediction and identification. Although, it might be valuable to have access to a variety of tools and algorithms with different capabilities, the user is challenged with a significant option in determining which algorithm/tool to apply (Peterson et al., 2014).

The development of artificial microRNAs (amiRNAs)-mediated gene silencing method offers a highly precise, targeted approach that gives several advantages over RNAi-mediated gene silencing strategy. Design, construction and validation of the amiRNAs depend upon the assembly of an endogenous precursor miRNA, which is substituted with a selected miRNA nucleotide transcript complementary to the target sequence (Niu et al., 2006; Ossowski, Schwab & Weigel, 2008).

The objective of the present research work is to predict sugarcane-encoded miRNAs that have the potential to develop resistance against badnaviruses, especially the SCBGAV. The major research objective was designed to identify the potential host-derived miRNAs in the SCBGAV genome and to screen the most promising miRNAs to understand the complex host-virus interactions. The novel amiRNA silencing method has been adopted for the first time to explore as a source of creating resistance to a monopartite badnavirus. The predicted miRNAs may provide help for the generation of SCBGAV-resistant sugarcane plants through genetic engineering in the future.

Materials & Methods

Results

Sugarcane miRNAs (predicted by a consensus of algorithms) for SCBGAV silencing

Among all the predicted miRNAs for SCBGAV silencing, only four miRNAs (sof-miR167 (a, b), sof-miR396 and ssp-miR528) were predicted by all of the algorithms used (Fig. 5, Table 2 and Table S4). We have verified the efficacy of these miRNA targets against SCBGAV following suppression of RNAi-mediated badnavirus control through translational inhibition or cleavage of viral mRNA. Moreover, six miRNAs (sof-miR167 (a, b) at locus 5846 and 1310, sof-miR168a at locus 5506, sof-miR396 at locus 3394, ssp-miR444a at locus 775 and ssp-miR1128 at locus 6148) were predicted at the common locus by at least two of the algorithms used (Fig. 6). Out of 28 sugarcane miRNAs, only one miRNA (sof-miR396) was predicted by at least three algorithms used to have binding site at the same locus position 3394 (Table 3).

Figure 5. Venn diagram representing common sugarcane miRNAs predicted by all algorithms.

Open in a new tab

Table 2. miRNA-target pairs were selected from miRanda analysis.

Locus position was selected by consensus analysis of at least two algorithms.

SugarcanemicroRNAs	miRNA-target pair	Locusposition	MFE*(kcal/mol)
sof-miR167(a, b)	Query: 3′ gucUAGUACGACCGUCGAAGu 5′ Ref: 5′ aaaATCAAGTT-GCAGCCTCa 3′	5846–5865	−26.60
sof-mi396	Query: 3′ guCAAGUUC-UUUCGACACCUu 5′ Ref: 5′ agGATTAGGTGATGCTGTGGAg 3′	3394–3415	−24.90
ssp-miR528	Query: 3′ gaggAGACGUACGGGGAAGGu 5′ Ref: 5′ gcgaTCCGC-CCCCCCTTCC- 3′	7426–7444	−28.0

Open in a new tab

Notes.

MFE (Minimum Free Energy)* and mode of target Inhibition** was determined by RNAhybrid.

Figure 6. Intersection plot showing sugarcane miRNAs predicted from at least two algorithms.

Open in a new tab

Table 3. Sugarcane miRNAs and their target positions in the SCBGAV genome identified by all four algorithms.

miRNA name	Position miRanda	Position RNA22	Position RNAhybrid	Position psRNATarget	MFE* miRanda	MFE** RNA22	MFE RNAhybrid	Expectation psRNATarget
sof-miR156			7104				−23.2
sof-miR159a	5282		1659	3779	−17.18		−25.4	6
sof-miR159a(1)	6739				−18.11
sof-miR159b	5282		1659	3779	−17.18		−25.4	6
sof-miR159b(1)	6739				−18.11
sof-miR159c	6739		6896	3779	−16.70		-28	6
sof-miR159d	5282		1659	3779	−17.18		−25.4	6
sof-miR159d(1)	6739				−18.11
sof-miR159e	5282		733		−15.81		−25.5
sof-miR167a	5846	1310	1304	5846	−16.08	−15.80	−26.6	7
sof-miR167b	5846	1310	1304	5846	−16.08	−15.80	−26.6	7
sof-miR168a	5506	5506	6084		−21.24	−17.70	−25.6
sof-miR168a(1)	7050	7050			−18.45	−17.10
sof-miR168a(2)		5137				−14.30
sof-miR168b	7050	7050	6937		−20.05	−19.40	−25.6
sof-miR168b(1)		5506				−13.70
sof-miR396	3394	3394	3394	5732	−19.99	−17.80	−24.9	5.5
sof-miR396(1)		4115				−14.20
sof-miR408a		1735	5136			−13.70	−27.6
sof-miR408b		1735	5136			−13.70	−27.6
sof-miR408c		1735	5136			−13.70	−27.6
sof-miR408d		1735	5136			−13.70	−27.6
sof-miR408e			6152				−28.9
ssp-miR166	3249		5572	981	−21.45		−26.7	6.5
ssp-miR166(1)	5589				−17.54
ssp-miR166(2)	5761				−21.74
ssp-miR169			752				−24.8
ssp-miR437a			978				−21.7
ssp-miR437b			6829				−21.8
ssp-miR437c			978				−24.3
ssp-miR528	7426	1310	7407	2897	−18.37	−13.70	-28	6.5
ssp-miR528(1)	7395				−18.05
ssp-miR827	5378		3010		−16.40		-24
ssp-miR444a		4103	774	775		−16.30	−27.6	6.75
ssp-miR444b		4103	774	756		−16.30	−27.6	6.0
ssp-miR444b(1)				2615				6.5
ssp-miR444b(2)				5009				6.5
ssp-miR444b(3)				5731				6.5
ssp-miR444c-3p			774	4940			−26.7	6.5
ssp-miR1128	6141		6148		−22.96		−27.8
ssp-miR1432	6070		1301		−15.48		−20.4

Open in a new tab

Notes.

*MFE: Minimum free energy measured in /Kcal/mol where *MFE represents minimum folding energy measured in Kcal/mol.

Discussion

For possible miRNA target prediction in the genome of SCBGAV, a combination of the aforementioned computational tools was used in order to filter out the false-positive results and to increase the accuracy of miRNA target prediction. miRanda was implemented to validate various parameters, from target site conservation to whole genome prediction of miRNA target genes (John et al., 2004; Riffo-Campos, Riquelme & Brebi-Mieville, 2016). Then, RNAhybrid and psRNATarget were used, both of which are strongly prescribed for plant miRNA target identification (Dai, Zhuang & Zhao, 2018; Srivastava et al., 2014; Zhang & Verbeek, 2010). RNA22 is a novel algorithm that applies an in-silico strategy that is highly divergent in comparison to other algorithms. Pattern-based recognition is the key feature of this algorithm (Min & Yoon, 2010; Miranda et al., 2006).

The results from this study suggest that SCBGAV is susceptible to targeting by consensus sugarcane-encoded miRNAs. The genome components of SCBGAV (ORF1, ORF2 and ORF3) seemed to be principally prone sites for sugarcane-miRNA regulation. While in vivo demonstration requires validating functional efficacy, the degree of complementarity between target mRNA and miRNA concludes the fate of the predicted sites. A full complementarity binding between target mRNA and miRNA sequence results in endonucleolytic cleavage and disruption. Contrary to this, partial target-site complementarity characteristically down-regulates target-gene expression by suppressing translation of target mRNA.

The present study identifies suitable sugarcane-encoded miRNAs to exhibit a stronger degree of target-site complementarities within the ORF1, ORF2 and ORF3 of SCBGAV. These predicted miRNAs may be utilized to develop effective amiRNA constructs, which could be used to enhance the sugarcane immunity to SCBGAV. Pairing multiple miRNAs with a single mRNA induces effective RNA silencing (Doench & Sharp, 2004). In the current study, we have predicted several miRNA targets which were associated with SCBGAV (ORF1, ORF2, and ORF3) genes at multiple loci. A deeper understanding of these vital ORFs involved in SCBGAV epidemic via miRNA-mediated control of gene expression would significantly assist in the development of molecular approaches to combat the dissemination of SCBGAV. The miRNA-target pairs ensuring MFE exceeding the threshold standards were predicted.

Our study is unique from previous studies in that it provides an in silico approach for comprehensive analysis of host-derived miRNAs using seven computational algorithms and advanced statistical approaches. Previous studies that have identified plant-encoded miRNA Targets in plant viruses have been limited by small numbers of miRNAs screened, small sample sizes, lack of independent validation sets, and the use of inappropriate statistical methods to present miRNA interaction data. Notably, we have validated the efficacy of the consensus seven sugarcane-encoded miRNA targets (detected two algorithms at the same locus) against SCBGAV following suppression of RNAi-mediated viral combat. These are achieved by translational inhibition or cleavage of viral miRNA.

Applying a combination of four diverse computational algorithms (miRanda, RNA22, Tapirhybrid and psRNATarget), the predicted consensus results are highly reliable and robust for assessing interaction between sugarcane miRNAs and viral genome (Oliveira et al., 2017). This miRNA-viral gene interaction was further verified by RNAhybrid (standard MFE predicted) and Circos algorithms (network visualization). False-positive miRNA-target interactions predicted by computational algorithms depend upon the mechanism of the miRNA-target recognition (Pinzón et al., 2017). In addition, MFE is one of the key factors to affect the miRNA-targets interactions for results sorting. A lower MFE value has high potential to develop miRNA-target complex (Hammell et al., 2008; Kertesz et al., 2007). We have applied a stringent cut-off −15 kcal/mol for miRanda and −20 kcal/mol for RNAhybrid to narrow down miRNA candidates.

In order to control false-positive prediction, union and intersection approach was applied. Union approach based on grouping of one or more algorithmic tools which results increase of number of true targets as well the number of false targets. It decreases specificity at the cost of increasing sensitivity. Using intersection approach relies on the mixing two or more tools which improves the specificity at the cost of decreasing sensitivity (Witkos, Koscianska & Krzyzosiak, 2011). Our analysis indicated that both the strategies showed the highest performance as shown in (Figs. 5 and 6). These miRNAs were identified and considered as the most desirable sugarcane miRNAs against the genome of SCBGAV. These predicted miRNAs were identified after setting the parameters of minimum free energy, seed pairing, target site accessibility, folding energy and pattern recognition, thus integrating all aspects of miRNA target prediction. Therefore, they are the most suitable selections for gene silencing.

Similar studies have been published previously using multiple computational algorithms to identify a best fit miRNA for a desirable target to understand plant-virus interaction. First time, genome-wide identification and comprehensive analysis was carried out to predict maize (Zea-mays)-encoded miRNA targets against maize chlorotic mottle virus (MCMV) (Iqbal et al., 2017). We have adopted the same novel computational strategy to predict the sugarcane-encode miRNA targets in the SCBGAV genome to combat badnaviruses in sugarcane.

As the amiRNAs have high specificity to the designed target gene, detrimental off-target effects can be minimized, permitting their silencing expression to be stably transmitted to future generations (Ossowski, Schwab & Weigel, 2008; Zhao et al., 2009). Furthermore, the small size of amiRNA permits for the insertion of multiple and distinct amiRNAs within a single gene expression cassette, which can then be transformed to develop transgenic plant resistant to multiple viruses simultaneously (Niu et al., 2006; Park, Zhai & Lee, 2009; Schwab et al., 2010). We have designed future work to validate this promising amiRNA-based strategy can in fact be used to develop durable SCBGAV- resistance in transgenic sugarcane.

Conclusions

The current study details an organized computational approach for the identification of host-derived miRNAs aimed at silencing viral genes affecting the sugarcane host plant by amiRNA-mediated interference targeting different genes of the SCBGAV. This study offers an enhanced means to computationally investigate the best-candidate miRNAs against badnaviruses, prior to cloning. As our approach allows a narrow-range of match-mismatch in microRNA-mRNA attachment, it screens most of the falsely predicted attachments. These predicted miRNAs were identified after setting the parameters of MFE, seed pairing, target site accessibility, folding energy and pattern recognition, thus using key features of miRNA target prediction. Most importantly, we identified six miRNAs (sof-miR167 (a, b), sof-miR168a, sof-miR396, ssp-miR444a and ssp-miR1128) that could target the SCBGAV genome. Among them, sof-miR396 was predicted as the top-ranking effective candidate, capable of targeting the vital ORF3. To further refine our results, we also identified key interaction of miRNA networks and target genes by constructing Circos graph of consensus miRNAs and their targets. This short-listed miRNA is the best candidates to be utilized in sugarcane plant transformation for the development of SCBGAV-resistant sugarcane cultivars. These results offer an evidence of idea for the construction of innovative amiRNA-based therapeutics against emerging SCBGAV badnaviruses.

Supplemental Information

Funding Statement

This study is supported by the National Natural Science Foundation of China (No. 31771865), the Sugar Crop Research System (CARS-170301) and the “Talented Young Scientist Program” (TYSP-4th Batch 2017-2018) of China Project ID (Pakistan-18-004). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Additional Information and Declarations