Wheat cultivar (Triticum aestivum L.) Huaimai 0360 (highly PHS-resistant cultivar), was used for this study. For gene cloning and expression experiments, 100 surface-sterilized seeds were germinated on moist filter paper in growth chambers at 25˚C under 12-h light/12-h dark conditions and transplanted into pots in a naturally lit glasshouse with standard irrigation and fertilization until mature. During this period, developing seeds were collected at 15, 20, 25, and 30 days after pollination (DAP). For germination, mature seeds were surface-sterilized and then imbibed on water moistened filter paper in Petri dishes in a temperature-controlled cultivation chamber (darkness at 25˚C). 3 - 5 seeds were collected at 0, 6, 12, 24, 36, and 48 h after imbibition commenced. All the collected plant samples were snap-frozen in liquid nitrogen immediately after harvesting and stored at −80˚C until use. Each treatment was repeated thrice.

For VIGS experiments, wheat cultivar Huaimai 0360 was grown in the agricultural and experimental field of Henan Agricultural University (Zhengzhou, China) during the wheat-growing season. The plot area was 15 m² (length 5.0 m × width 3.0 m) and was watered and fertilized as needed throughout the growth period.

At anthesis, an arch shed (length 5.0 m × width 3.0 m × height 1.5 m) was built directly above the wheat plants according to the method described by Liu et al. [36]. One hundred wheat spikes in the same heading stage and of similar size were labeled. During 1 day before and after BSMV inoculation, the PVC plastic film was covered with soil to prevent water loss from affecting the spread of the virus. Developing grains in the middle of BSMV-infected spikes were collected at 21, 28, 35, and 45 days post inoculation (dpi) at 10:00 a.m., snap-frozen in liquid nitrogen and then stored at −80˚C for transcription analysis. At maturity, spikes of BSMV-infected wheat plants were harvested and stored at room temperature until use. Uniform spikes at the anthesis stage were used for BSMV-VIGS inoculation.

Genomic DNA was isolated from seed at 15~30 DAP using the CTAB method. Gene specific primers TaFUS3-F (5’-TCC TCC GCC TTG ACC TCC T-3’) and TaFUS3-R (5’-CAA GGC TGG TGA CTC TGA ACT-3’) were designed based on the sequence of TaFUS3 [37] using Primer 5.0 software. The PCR was performed in an Applied Biosystems-2720 thermal cycler in total volumes of 50 μL, including 5 μL of 10 × PCR buffer, 100 mM dNTP, 0.4 μM of each primer, 2.5 unit of Taq DNA polymerase (TIANGEN, Beijing), and 100 ng of template DNA. The PCR cycling conditions comprised an initial cycle at 95˚C for 3 min followed by 35 cycles of 94˚C for 30 s, annealing at 58˚C for 1 min and 72˚C for 2 min, with a final extension at 72˚C for 10 min. PCR products were separated by electrophoresis in 1.0% agarose gels. Targeted fragments of expected size were recovered and cloned into the pMD18-T vector. Selected 20 positive clones were sequenced by Henan Shangya Technology Co., Ltd. Sequence alignment was performed by using DNAMAN software. Chromosome locations of cloned sequences were identified at urgi (https://urgi.versailles.inra.fr/blast/blast.php).

For VIGS experiments, a γRNA-based BSMV vector was utilized to silence FUS3 gene in wheat as described previously [38]. Since wheat is hexaploid, there are usually three copies of genes. To make TaFUS3 gene silenced completely, a 247 bp cDNA fragment from the wheat FUS3 gene that contains Nhe I sites was amplified with specific primers designed in the conserved regions of the three copies by Primer 5.0 software. The forward and reverse primers were 5’-GCG GGA GAT GAT CTA GTT GC-3’ and 5’-AAT CAC TGA ATG GGT CGA AGA-3’, respectively. Then, the amplified PCR products and the BSMV-γ empty vector were individually digested with restriction enzyme Nhe I. The digestion results were verified by agarose gel electrophoresis. The correct target fragments were ligated into the BSMV-γ empty vector and sequenced. The inserted clone in reverse was selected as the recombinant vector BSMV: TaFUS3 and subsequently used for gene silencing. For identifying wheat plants with the VIGS response, the BSMV: GFP was used as an infection control.

In vitro transcription of viral RNAs was carried out as described by Feng et al. [38]. The plasmids of BSMV-α, BSMV-β, BSMV-γ-derivative clones (BSMV-TaFUS3 and BSMV-GFP) were digested with restriction enzymes Mlu I and Spe I. Then, the linearized products were examined by electrophoresis. The correct linearized products were purified, and transcription was performed using a RiboMAX^TM Large Scale RNA Production Systems-T7 kit (Promega, USA) following the manufacturer’s protocol. Transcripts of each of the BSMV plasmids were mixed in a 1:1:1 ratio and 22.5 volumes of FES buffer [39] were added to the transcript mixture. This mixture was then applied to the young wheat spikes by rub inoculation. 15 µL of BSMV: TaFUS3 or BSMV: GFP transcript mixture was used for each spike. The details of the BSMV inoculation and wheat plant growth conditions were described by Liu et al. [36].

180 wheat seeds from both the BSMV-TaFUS3 and BSMV-GFP plants (40 dpi) with uniform size and plumpness were surface sterilized in 70% alcohol for 1 min, and then in 0.1% HgCl₂ for 5 - 10 min and washed four times with sterile distilled water. Six groups (30 grains/group) were distributed on two layers of filter paper in a 9-cm Petri dish containing 10 mL of sterilized water, and germinated at 25˚C under continuous dark for 7 days. Three of them were used to investigate the germination rate and the other three groups were used for the expression analysis of ABA/GA synthetic and metabolic genes. The number of germinated seeds (radicle protruding through the seed coat) was recorded daily. Germination ratio refers to the number of germinated seed as a proportion of the total number of seeds. For the gene expression analysis, germinated seeds were harvested at 24, 48, 72, and 96 h snap-frozen in liquid nitrogen and stored at −80˚C for further analysis.

Spike sprouting tests were performed with intact spikes (40 and 45 dpi). Twelve spikes each from the BSMV-TaFUS3 and BSMV-GFP plant groups were surface sterilized in 0.1% HgCl₂ for 10 min, and washed four times with sterilized water, vertically inserted into a foam board (length 20 cm × width 20 cm × height 3 cm), and covered by a plastic film to prevent water loss. Water was sprayed once every 6 h. After 7 days, the test spikes were taken out and immediately threshed manually to record seed germination. The calculation method of germination rate was the same as for the seed germination test. In addition, the plumule and radicle length was measured on germinated seeds. All tests were performed in a climate chamber at 25˚C (24 h darkness).

Seeds from BSMV-TaFUS3 and BSMV-GFP plants (40 dpi) were soaked in sterilized water for 4 h, and then the embryos were isolated with dissecting needles under aseptic conditions and were placed on two layers of filter paper in a 9 cm Petri dish containing 10 mL of sterilized water with 50 µM ABA at 25˚C in darkness for 48 h. RNA was extracted from embryos after 0 and 48 h of treatment. The experiments were performed in triplicate.

The RNA extraction and the first strand cDNA synthesis were described in our previous study [40]. Based on the sequences of the genes, including TaFUS3-3A, TaFUS3-3B, TaFUS3-3D, TaNCED1, TaNCED2, TaCYP707A2, TaGA20ox-D4, TaGA20ox1d, TaABI3, TaABI4, TaABI5, TaPKABA1, and β-actin, specific primers were designed using Primer 5.0 software. The primer sequences and amplified product are provided in Table 1. The qRT-PCR analysis was performed with a SYBR Green quantitative RT-PCR kit (TaKaRa, Dalian, China) on a CFX connect real-time system (Bio-RAD, USA) following the manufacturer’s instructions. The expression level of each gene was determined using the comparative Ct method. Relative quantification for each gene was calculated by the 2^−ΔΔCt method using wheat β-actin gene as an internal control [41].

Based on the known cDNA sequence of TaFUS3 gene, a pair of specific primers

TaFUS3-F and TaFUS3-R were designed in the 5’ and 3’ non-coding regions, respectively, to amplify the complete gene DNA sequence from wheat cultivar “Huaimai 0360”. An amplicon of about 3500 bp was obtained. After sequencing, three genes, containing seven exons and six introns, and with sequence lengths of 3477, 3534 and, 3501 bp was obtained (Figure 1). The URGI BLAST analysis showed that the three sequences were highly similar to the, respectively, 3A, 3B and 3D sequences from the long arm of chromosome 3 in Chinese spring. So, the three TaFUS3 homologues were named TaFUS3-3A (GenBank accession MZ408248), TaFUS3-3B (GenBank accession MZ408249) and TaFUS3-3D (GenBank accession MZ408250), respectively. The corresponding cDNA sequences of TaFUS3-3A, TaFUS3-3B and TaFUS3-3D were 1015, 1012 and 1015 bp in length, and contained 906, 909 and 906 bp open reading frames (ORF), respectively. TaFUS3-3A, TaFUS3-3B and TaFUS3-3D were predicted to encode proteins containing 301, 302 and 301 amino acid, respectively, and shared 98.57% similarity.

The expression profiles of three TaFUS3 genes in seed development and germination stages were investigated by qRT-PCR using gene-specific primer pairs TaFUS3-3A-F/R, TaFUS3-3B-F/R and TaFUS3-3D-F/R (Table 1). Expression analysis indicated that the significant differences in transcription were observed among the three homoeologous genes. TaFUS3-3B and TaFUS3-3D had similar expression patterns (Figure 2), whereas the expression of TaFUS3-3A was almost undetectable in all samples (data not shown). During grain filling stage, the transcription of TaFUS3-3B and TaFUS3-3D showed a gradual upward trend at 15 - 25 DAP, and then declined quickly at 30 DAP (Figure 2(a)). During seed imbibition, the expression of TaFUS3-3B and TaFUS3-3D showed an up-regulation trend in the 0 - 24 h period. Afterwards (36 - 48 h), the transcripts of the two genes showed gradual declined (Figure 2(b)). Additionally, the two TaFUS3 genes also shared different transcript abundances during seed development and germination process. For instance, TaFUS3-3D had higher expression level than

that of TaFUS3-3B. These results indicated that the two TaFUS3 homoeologs might play an important role in wheat seed development and germination.

To functionally test the role of TaFUS3 gene, we used a virus-induced gene silencing (VIGS) approach in the cultivar “Huaimai 0360”. This cultivar has a relatively high seed dormancy level. A 247-bp cDNA fragment of the TaFUS3 gene was isolated. Then, a recombinant viral vector BSMV: TaFUS3 was constructed and used to inoculate the wheat spikes during the heading stage under field conditions. We collected developing seeds at 21, 28, 35, and 45 dpi and then mature seeds (45 dpi) were imbibed for 24, 48, 72, and 96 h. The qRT-PCR analysis showed that the transcript level of TaFUS3 in the BSMV: TaFUS3-silenced plants was significantly decreased (by 61%, 41%, 55%, and 25% at 21, 28, 35, and 45 dpi, respectively) compared with the control plants (Figure 3(a)). During seed germination, the transcript abundance of TaFUS3 declined by 18% - 24% in mature seeds (45 dpi) imbibed for 48 - 96 h (Figure 3(b)). These findings indicated the target gene was successfully repressed in BSMV: TaFUS3-silenced plants.

The effect of TaFUS3 knockdown on the seed germination was characterized at 40 dpi. The seed germination capacity of BSMV-TaFUS3-silenced plants was always higher than that of control plants (BSMV: GFP infected) for 7 days of imbibition (Figure 4(a)). Moreover, the germination speed was faster in BSMV:TaFUS3 than BSMV:GFP infected plants following the addition of water. As shown in Figure 4(b), the seed germination rate in TaFUS3-silenced plants reached 16.7%, 24.5%, and 27.8% on the first, second and third day of germination, which were 5 times, 3.7 times and 2.8 times of the control, respectively.

Similar results in PHS experiment were observed at 40 and 45 dpi. The data showed that the sprouting capacity was obviously higher in spikes of BSMV:TaFUS3 than BSMV:GFP plants (Figure 4(c)). Compared with the control, the germination rates in seed of BSMV-TaFUS3 wheat spike increased on average 50% (40 dpi) and 24.1% (45 dpi) (Figure 4(d)). In addition, the plumule and radicle lengths in grains of BSMV: TaFUS3 wheat spike were increased by, respectively, 60.4% and 55.8% at 40 dpi (Figure 4(e)) and 14.3% and 32.5% at 45 dpi (Figure 4(f)). These findings indicated that the sprouting speed was significantly faster in spikes of BSMV: TaFUS3 than BSMV-GFP wheat. Above results suggested that TaFUS3 silencing resulted in enhanced seed germination and PHS in wheat grains; thus, TaFUS3 may act as a negative regulator of wheat seed germination.

To define the role of TaFUS3 in regulating seed germination and PHS, qRT-PCR analyses were carried out to detect the expression levels of ABA biosynthesis genes TaNCED1, TaNCED2, GA biosynthesis genes TaGA20ox-D4, TaGA20ox1d and ABA metabolism gene TaCYP707A2 (Table 1). The result revealed that the transcription of one GA synthesis genes (TaGA20ox1d) and one ABA metabolism gene (TaCYP707A2) was evidently induced and their transcript levels in TaFUS3-silenced plants were up-regulated on average by 1.9 and 2.0 times, respectively, in the seeds imbibed for 48 hours (Figure 5). However, the expression of ABA biosynthesis genes TaNCED1 and TaNCED2 was obviously inhibited and the transcription levels of TaNCED2 in TaFUS3-silenced plants decreased by 32.86% (Figure 5), indicating that TaFUS3 might control seed dormancy and PHS -resistance by positively regulating ABA synthesis genes and negatively regulating ABA metabolism and GA synthesis genes.

The expression levels of TaFUS3 and endogenous ABI3, ABI4, ABI5, and PKABA1 (Table 1) were analyzed in embryos of the BSMV-infected plants upon ABA treatment. As shown in Figure 6, after ABA treatment, the expression of these genes was induced. Compared with the control, the relative transcript level of TaFUS3 in BSMV:TaFUS3 plants was significantly lower. In addition, the expression of TaFUS3 in TaFUS3-silenced plants significantly up-regulated the expression of PKABA1 and markedly down-regulated the expression of ABI5

(Figure 6). Although the expression of ABI3 was also up-regulated, the difference was not significant. These findings suggested that repressed expression of TaFUS3 gene reduced ABA sensitivity through negatively regulating ABI4 and PKABA1 genes and positively regulating ABI5 gene involved in the ABA signaling pathway.