Spring wheat cultivar Ningchun 16, which was previously demonstrated to be infectivity by the BSMV (Ma et al., 2012), was used in this experiment. Wheat seeds were germinated in a controlled growth chamber (25–27°C), and then wheat plants were grown in a green house with a light period of 16 h/day (regulated with supplementary light), a day/night temperature regime of 20–25°C/15–18°C, and 65–75% relative humidity, and watered as needed. The wheat plants were exposed to a temperature of 4°C for 14 days to achieve complete vernalization at the two-leaf stage.

The BSMV vectors, which contain α, β, and γ constructs, used in this experiment were kindly provided by Dr. Huang Li at Montana State University, Bozeman, USA. The schematic organization of the BSMV genomes was shown in Figure 1A. The γ construct was modified to include a polymerase chain reaction (PCR)-ready cloning site (TA-cloning site, TACS) following the protocol described previously (Ma et al., 2012). The BSMV γ RNA construct carrying a 185-bp fragment of the barley (Hordeum vulgare L.) phytoene desaturase gene (PDS) with the gene in antisense orientation was also provided by Dr. Huang Li (Figure 1B). The recombinant BSMV vector containing the PDS fragment (named as BSMV:PDS4as) can efficiently induce PDS silencing in wheat leaves (Scofield et al., 2005). We utilized the BSMV:PDS4as as a positive control for testing virus-induced expression of an amiRNA against the mRNA of wheat PDS gene (amiR-PDS). For simplicity, BSMV:PDS was used to stand for the BSMV: PDS4as vector, and BSMV:00 for the BSMV with no insert (empty vector) in our experiments.

The BSMV vector carrying a fragment of precursor sequences of wheat miR156 (pre-miR156, 350 bp in length) (Figure 1B), designed as BSMV:miR156, was constructed for miR156 overexpressing in wheat plants. The wheat pre-miR156 sequence available on miRBase³ was predicted based on the sequence of primary transcript of miR156 (pri-miR156) (Genbank accession CL902915) in wheat (Dryanova et al., 2008). A primer pair P7/P8 was designed according to the sequence of the pri-miR156 for amplifying the pre-miR156 sequences by PCR, with the genome DNA of Ningchun 16 as a template. The resulting PCR products were purified and directly inserted into the BSMV γ construct by T-A cloning in either the sense (BSMV: miR156-F) or antisense (BSMV: miR156-R) orientation.

The BSMV vector containing an amiR-PDS precursor, named as BSMV:amiR-PDS, was generated for amiRNA expression in wheat. To obtain the BSMV:amiR-PDS, a Web-based tool Web MicroRNA Designer⁴ was used to design a 21-mer amiR-PDS sequence, and the resulting sequence is 5′-UAAUCUGUUUAGAGGAAUCAG-3′. To generate the amiR-PDS precursor sequence, we conducted PCR amplification according to the strategies previously described (Tang et al., 2010; Gu et al., 2014). In brief, four specific oligonucleotide sequences (amiR-PDS-I, amiR-PDS-II, amiR-PDS-III, and ami-PDS-IV) suggested by Web MicroRNA Designer were used to amplify the designed amiR-PDS along with primer pair P9/P10, using the endogenous miR319a precusor of Arabidopsis (EU549293) as backbone. While, the full-length miR319a precursor was amplified from Arabidopsis. The detailed protocol for generating amiR-PDS precursor can be found in Supplementary Method S1. The resulting amiR-PDS precursor was inserted into BSMV γ construct by T-A cloning.

The BSMV vectors utilized to silence endogenous particular miRNAs were constructed to carry STTM of the certain miRNA, which is composed of two short sequences mimicking small RNA target sites separated by a linker of an 48-nucleotide RNA spacer (totally 153 bp in length) (Figure 1B), and designed as BSMV:STTM, including BSMV:STTM156/156 and BSMV:STTM166/166 which harbors two copies of short tandem target mimic of miR156 and miR166, respectively (Figures 3A, 4A). The STTM156/156 and STTM166/166 sequences were designed abiding by the rules for STTM designing reported previously (Yan et al., 2012), and synthesized by Soagon Biotech (Shanghai). Each of these two STTM modules was separately inserted into TACS of the BSMV γ construct. This was done by PCR amplification of the synthesized STTM sequences with a pair of universal PCR primer (the forward primer P11: 5′-GTTGTGTGGAATGTATGGAGC-3′ and the reverse primer P12: 5′-GCTGTAATCACACTGGCTCA-3′) complementary to the 5′- and 3′-end of the sequences of the STTM modules designed in advance, and the resulting PCR products were purified and directly inserted by T-A cloning. The sequences of all the primer pairs used in this study were listed in Supplementary Table S1.

In vitro transcription of viral RNAs was conducted according to the procedure previously described (Scofield et al., 2005). The BSMV RNAs in vitro synthesized were rub-inoculated onto the 2nd or the 3rd fully expanded leaf of wheat at 3- or 4-leaf stage. Ten plants were infected with each of the BSMV:amiR-PDS, the BSMV:miR156, the BSMV:STTM156/156, the BSMV:STTM166/166, the BSMV:PDS, and the BSMV:00, respectively, as described previously (Scofield et al., 2005; Ma et al., 2012), the inoculation with BSMV:PDS and BSMV:00 being as a positive control, and a control, respectively. Three independent biological replicates were included in this experiment.

For analyzing the abundance of a particular mature miRNA and its target mRNA in wheat infected with each reconstructed BSMV, the 2nd leaves upper the inoculated ones were collected at 2, 4, 8, 12, 15, 20, 22, and 30 days post inoculation (dpi), with three biological replicates included. Total RNA was extracted from each of these leaf samples by using TRIzol reagent (Invitrogen, Grand Island, NY, USA) according to the instructions the manufacturer recommended, and treated with RNase–free DNase I (TaKaRa, Dalian, China). The integrity of each RNA sample was checked by electrophoresis of 1% agarose gel, and the concentration was determined with a Nanodrop ND-1000 spectrophotometer (Nano Drop Technologies, Wilmington, DE, USA). For detection of coding genes, first-strand cDNA was synthesized using 1 mg of total RNA with oligo (dT) primer and MMLV reverse transcriptase (TaKaRa, Dalian, China), and for detecting STTM mudules, first-strand cDNA was synthesized using the primer P12 specific to STTM. For testing mature miRNA, the specifically designed stem-loop reverse transcription primer for each of miRNA was used to reverse-transcribe (RT) individual miRNA, primer P13 and P14 being used for RT-PCR of miR156 and miR166, respectively, with 1 mg of total RNA, following the procedures described previously (Chen et al., 2005; Han et al., 2014). The resulting cDNA was diluted and used for further study.

To detect the abundance of the BSMV RNA and the inserts, including pre-miR156, pre-amiR-PDS, and STTM module, in the infected plants, the cDNAs from the leaf samples collected at 12 dpi were used for semiquantitative RT-PCR analysis, primer pair P1/P2 being used for detecting the γ RNA, P7/P8 for pre-miR156, P9/P10 for pre-amiR-PDS, and P11/P12 for STTM 156/156. The primer pair P15/P16 for an internal conference gene Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), which is constitutively expressed in wheat, was used for cDNA normalization (Sun et al., 2010). All the sequences of the primer pairs used were listed in Supplementary Table S1.

To determine the abundance of the interested mature miRNAs and their target mRNAs in wheat infected with BSMV:STTM156/156, BSMV:STTM166/166, BSMV:amiR-PDS, and BSMV:miR156, respectively, quantitative real-time RT-PCR (qTR-PCR) was performed with U6 (small U6 spliceosomal RNA) and GADPH as internal reference genes for cDNA normalization. Forward primer P17 and P18 specific for miR156 and miR166, respectively, were separately used with universal reverse primer P19 to detect mature miR156 or miR166 levels, and primer pair P20/P21 for U6 was used for cDNA normalization. Primer pair P22/P23 was used for miR156 target sequence (GenBank ID: CK196549)⁵, which is the homolog of SPL2 in O. sativa and Brachypodium distachyon, and P24/P25 was used for Homeobox-leucine zipper protein HOX33-like (Accession no.Ta#S52543234), which was predicted to be one of the miR166 target in wheat (Sun et al., 2014). The primer pair P15/P16 for GAPDH was used for cDNA normalization of encoding genes. Real-time RT-PCR with three technical replicates were performed on a CFX96 real-time system (BIO-RAD, USA) by using SYBR Premix Ex TaqTM II (TaKaRa, Dalian, China). The relative abundance of each miRNA and its target mRNA was calculated by the 2^-ΔΔCT method as previously reported (Livak and Schmittgen, 2001). All primers used in this experiment were listed in Supplementary Table S1.

To investigate whether a BSMV-based vector can be used to block miRNA activity or to overexpress a mature miRNA, we first tested the possibility of the BSMV-derived vector causing accumulation of the inserts it carried in wheat. We inoculated the 2nd fully expanded leaves of wheat at three-leaf stage with BSMV:00, BSMV:amiR-PDS, BSMV:STTM156/156, and BSMV:miR156, respectively. Different phenotypes were observed in the 4th leaves of the infected plants at 12 dpi (Figure 2A). The leaves of the plants infected with BSMV:amiR-PDS exhibited photobleaching phenotype, those infected with BSMV:STTM156/156 displayed shrinking in the middle section of the leaves, while those infected with BSMV:miR156 did not showed any obviously symptom, presenting mosaic and chlorotic stripes similar as that of the plants infected with BSMV:00. Semiquantitative RT-PCR analysis was conducted to detected the abundance of the BSMV γ RNA, pre-amiR-PDS, STTM156/156, and pre-miR156 in the leaves exhibited in Figure 2A. The results showed that the BMSV γ RNA and the inserts were accumulated in the leaves of the infected plants at 12 dpi (Figure 2B), suggesting that a BSMV-based vector is able to overexpress the inserts it carried in wheat.

miR156 and miR166 are conserved among plant species and highly expressed throughout the growing period of wheat (Han et al., 2014). It has been demonstrated that the miR165/166 family target and repress the expression of target gene homeodomain-leucine zipper (HD-Zip) III, determining the behavior of apical dominance and the phenotype of ectopic leaf growth and development in Arabidopsis (Jung and Park, 2007; Yan et al., 2012) and in N. benthamiana (Sha et al., 2014). miR156 targets the SQUAMOSA PROMOTER BINDING PROTEIN-LIKE (SPL) transcription factor genes which are involved in the promotion of vegetative phase transitions as well as flowering (Wu and Poethig, 2006). It has been demonstrated that the functional blockage of miR156/157 triggered an early vegetative phase change and early flowering in A. thaliana (Yan et al., 2012). Therefore, we chose miR156 and miR166 as target miRNA to knock down in wheat in order easy to check whether a BSMV-derived vector can induce endogenous miRNA silencing in wheat or not. So, we constructed BSMV:STTM156/156 and BSMV:STTM166/166 (Figures 3A, 4A), and inoculated the wheat plants with these BSMV-derived vectors. First, we inoculated wheat plants onto the 2nd fully expanded leaves with BSMV:STTM156/156 and BSMV:00, respectively. The experiment included three independent biological replicates and each replicate contains ten wheat plants. As previously observed, the symptoms of wheat infected with BSMV:STTM156/156 were indistinguishable from those infected with BSMV:00 at 10 dpi or earlier. However, the wheat plants inoculated with BSMV:STTM156/156 appeared to shrink in the middle section of the 4th leaves at 12 dpi, apparently curled at 14 dpi, and further curled up into a needle-like leaves at 18 dpi, but the leaves remained green, while the needle-like part of the leaves began to wither 22 dpi later (Figure 3B). This phenotype was rarely found in the 5th leaves of these plants. No leaf-curling phenomenon was observed in the wheat infected with BSMV:00. To confirm that the observed leaf-curling phenomena were the result of miR156 silencing, the 4th leaves of the BSMV:STTM156/156 infected plants were collected at 4, 8, 12, 14, 18, and 22 dpi to measure the abundance of endogenous mature miR156 by qRT-PCR described in section “Materials and Methods.” The results indicated that the miR156 levels were significantly knocked down in BSMV:STTM156/156 infected plants from 4 to 22 dpi (Figure 3C). About 15% reduction in mature miR156 abundance was detected as early as 4 dpi, and a remarkable reduction of miR156 level was detected at 14 ∼ 18 dpi. The needle-like leaves were seen at 18 dpi, much later than the reduction of the miR156 level. It is known that the mRNA sequence (GenBank ID: CK196549) is a homolog of miR156 target gene SPL2 in B. distachyon by sequence alignment, therefore, we expect that the CK196549 is one of the miR156 target genes in wheat. Thus, we conducted qRT-PCR to analyze the level of the CK196549 in the leaves of the infected wheat. Indeed, the levels of the CK196549 were significantly higher in BSMV:STTM156/156 infected plants than in control plants from 4 to 22 dpi (Figure 3D). Contrasting to the reduction of miR156 abundance, a remarkable increasing of mRNA abundance of the miR156 target gene was detected at 14 dpi. These suggested that BSMV:STTM156/156 can induce miR156 silencing in wheat plants.

Then, we inoculated wheat plants with BSMV:STTM166/166 and BSMV:00, respectively, with three replicates included and each replicate contains 10 wheat plants, in order to further confirm the conclusion that a BSMV-based vector can induce miRNA silencing in wheat plants. Similarly, the wheat infected with BSMV:STTM166/166 and BSMV:00 did not show any distinguishable symptoms at 10 dpi or earlier. However, the plants infected with BSMV:STTM166/166 showed a slightly different phenotype from those control plants in the 4th leaves from 12 to 30 dpi. The BSMV:STTM166/166 infected plants exhibited a symptom of leaf-twist in the 4th leaves, while the control plants did not (Figure 4B). Our observation mirrored the damage of leaf primordial functions generated by the inhibition of miR166. This result is consistent with the previous report (Yan et al., 2012). Homeobox-leucine zipper protein HOX33-like (Ta#S52543234) was predicted to be one of the miR166 target in wheat (Sun et al., 2014). Analysis of qRT-PCR confirmed that the miR166 abundance was significantly knocked down in the BSMV:STTM166/166 infected plants (Figure 4C). In contrast, the mRNA level of Ta#S52543234 in the BSMV:STTM166/166 infected plants was much higher than that in control plants (Figure 4D).

Taken together, the above results suggested that the BSMV vectors carrying STTM sequences can effectively block endogenous miRNA activity in wheat.

To test whether a BSMV-derived vector can be used to over-express endogenous miRNA to silence target genes in wheat, we constructed BSMV-derived vector BSMV:miR156-F and BSMV:miR156-R, which carries pre-miR156 sequences (upper pane in Figure 5A) in the sense and antisense orientation, respectively. Ten wheat seedlings gone through vernalization were inoculated onto the 3rd fully expanded leaves with BSMV:miR156-F, BSMV:miR156-R, and BSMV:00, respectively, at 4-leaf stage. It was observed that the wheat infected with BSMV:miR156-F, BSMV:miR156-R or BSMV:00 did not show any distinguishable symptoms at 18 dpi or earlier (Figure 5B). However, qRT-PCR analysis revealed that the abundances of miR156 were greatly increased in the leaves of wheat infected with BSMV:miR156-F or BSMV:miR156-R, compared to that in control plants (Figure 5C), on the contrary, the levels of miR156 target CK196549 were largely reduced in the leaves of plants infected with BSMV:miR156-F or BSMV:miR156-R (Figure 5D). Further observation found that the wheat plants infected with BSMV:miR156-F and BSMV:miR156-R showed very different phenotypes from control plants at 20 dpi and later. The control plants exhibited normally booting, heading and flowering (Figures 6A–C), however, the wheat plants infected with either BSMV:miR156-F or BSMV:miR156-R displayed two different phenotypes: type I (plants) showed the increased tiller number (5 more than the control plants) but not booting and heading (Figures 6A,C), and type II (plants) exhibited the similar tiller number as control plants did but with very late booting and heading date (7–10 days later than the control plants) (Figure 6B), the control plants heading at 25 dpi whereas the BSMV:miR156-infected wheat heading at about 32 dpi. These observation reflected the delay of the vegetative phase transitions and heading as well as flowering caused by miR156 overexpression resulting in great reduction of the levels of the target mRNA CK196549 (Wu and Poethig, 2006). We repeated this experiment five times, with ten plants for each construct in each experiment, and similar results were obtained. Furthermore, qRT-PCR analysis of miR 156 abundance was conducted to investigate if the different phenotype caused by distinct mature miR156 abundances leading to different mRNA levels of the miR156 target in these BSMV-miR156-infected wheat. The results confirmed that the mature miR156 levels in type I plants were significantly higher than in type II plants (Figure 6D). Taken together, a BSMV-based vector is able to overexpress endogenous miRNAs to silence target genes for rapidly studying miRNA functions in wheat.

In recent years, amiRNA technique has been established by using an endogenous miRNA precursor to yield a new target-specific miRNA for gene silencing in plants (Schwab et al., 2006; Molnar et al., 2009). And amiRNA has been demonstrated to specifically and efficiently silence single or multiple target genes by a well-designed precursor sequence (Schwab et al., 2006). To examine whether a BSMV-based vector is able to express amiRNAs to silence endogenous target genes in wheat, we designed an amiRNA targeting PDS gene in wheat and inserted it into the BSMV vector to obtain BSMV:amiR-PDS (Figures 1B, 7A). We inoculated ten wheat plants onto the 3rd fully expanded leaves with BSMV:amiR-PDS (for silencing PDS), BSMV:PDS (for positive control) and BSMV: 00 (control), respectively. As expected, mosaic and chlorotic stripes were observed on the tip portions of the 4th leaves of all inoculated plants at 7 dpi. Photobleaching was observed only at the 5th leaves of the wheat infected with BSMV:amiR-PDS or BSMV:PDS during 10 ∼ 30 dpi (Figure 7B). qRT-PCR analysis further confirmed that the endogenous PDS level, similar as that in the BSMV:PDS infected plants, was largely decreased in BSMV:amiR-PDS infected plants (Figure 7C). These suggested that the mRNA abundance of endogenous PDS in the BSMV:amiR-PDS infected plants was knocked down by amiR-PDS expression, resulting in leaf photobleaching in these plants. Thus, the BSMV-derived vector is capable of expressing target-specific artificial miRNAs to silence endogenous genes in wheat.