RNAi-mediated down-regulation of SHATTERPROOF gene in transgenic oilseed rape

Oilseed rape is one of the important oil plants. Pod shattering is one of the problems in oilseed rape production especially in regions with dry conditions. One of the important genes in Brassica pod opening is SHATTERPROOF1 (SHP1). Down-regulation of BnSHP1 expression by RNAi can increase resistance to pod shattering. A 470 bp of the BnSHP1 cDNA sequence constructed in an RNAi-silencing vector was transferred to oilseed rape cv. SLM046. Molecular analysis of T2 transgenic plants by RT-PCR and Real-time PCR showed that expression of the BnSHP alleles was highly decreased in comparison with control plants. Morphologically, transgenic plants were normal and produced seeds at greenhouse conditions. At ripening, stage pods failed to shatter, and a finger pressure was needed for pod opening.


Introduction
Oilseed rape (Brassica napus L.) is the third most important oilseed crop in the world (Basalma 2008). Seeds have about 40-48 % oil with a high amount oleic acid and low linolenic acid suitable for frying applications and cooking. Dehiscence of pods causes significant yield loss (Raman et al. 2011). Ordinary yield losses are in the range of 10-25 % (Price et al. 1996). Seed losses have been reported as much as 50 % of the expected yield when adverse climatic conditions delayed harvesting (Macleod 1981;Child and Evans 1989). The process of pod shatter begins with degradation and separation of cell walls along a layer of few cells, termed the dehiscence zone (Meakin and Roberts 1990). Resistance to shattering is an important and necessary trait for oilseed rape improvement (Kadkol 2009). Attempts to solve this problem by interspecific hybridization using related species such as B. nigra, B. juncea and B. rapa have been faced with some difficulties as other undesirable traits will be integrated too (Prakash and Chopra 1990;Kadkol 2009). In Arabidopsis, which is in the same family of brassicaceae, several genes including the ALCATRAZ (ALC), INDEHISCENT (IND), SHAT-TERPROOF1 (SHP1) and SHATTERPROOF2 (SHP2) and FRUITFUL (FUL) have been shown to be involved in pod dehiscence (Raman et al. 2011). Genes for a number of hydrolytic enzymes, such as endopolygalacturonases, have also roles in dehiscence (Petersen et al. 1996). In Arabidopsis, SHP genes are specifically expressed in flowers with strong expression in the outer replum (Savidge et al. 1995;Flanagan et al. 1996). SHP gene also has mainly effect in the ripening of strawberries (Daminato et al. 2013). In B. napus, three BnSHP alleles (BnSHP1, BnSHP2a and BnSHP2b) have been identified (Tan et al. 2009). BnSHP1 and BnSHP2 show 80 % identity at nucleotide sequence. The expression of BnSHP2a and BnSHP2b (Two alleles of BnSHP gene which differ only in downstream sequences) are mainly in root, floral buds and pods, and most strongly in floral buds (Tan et al. 2009). It is suggested that less severe phenotype of indehiscence will be better, and the SHP, IND and ALC genes are ideal candidates for research and application in breeding new lines suitable for mechanized harvest (Liljegren et al. 2000;Tan et al. 2009). Recent advances about the role of MADSbox genes in dehiscence zone development have been reviewed (Ferrándiz and Fourquin 2014). In this study, we report the effect of the silencing cassette on expression of BnSHP alleles in transgenic oilseed rape plants using RNAi approach.

Materials and methods
Nucleic acid isolation DNA was isolated from 100 mg leaf tissues using the procedure of Dellaporta et al. (1983). For RNA isolation, total RNA was extracted from floral buds using RNeasy Mini Kit (Qiagene Co.). The quantity and quality of RNA samples were checked using nano spectrophotometry and agarose gel electrophoresis. First-strand cDNA was synthesized using 2 lg of total RNA with iScript Select cDNA synthesis kit (Bio-rad Co.) in a 20-ll reaction using oligo-dT's according to manufacturer's instructions.

Construction of RNAi cassette
A 470-bp fragment of the BnSHP1 cDNA (Accession, AY036062) without MADS-box region was amplified by PCR using specific primers; F: 5 0 -ATACTAGTGGCGCGC CCCGTTAACCCTCCACTG-3 0 and R: 5 0 -GCCTTAATT AAATTTAAATTTGAAGAGGAGGTTGGTC-3 0 containing restriction enzyme digestion sites for Asc1, Aws1, Spe1 and Pac1 (underlined), for cloning the sense and antisense fragments in the above sites in pGSA1252 behind the CaMV35S promoter. The RNAi cassette was removed with Pst1 digestion and sub-cloned in the Pst1 site in pCAM-BIA3301 to make pCAMRNAi.

Production of transgenic plants
Agrobacterium tumefaciens strain AGL0 containing the plasmid pCAMBIA3301 was used for transformation. Cotyledon explants of rapeseed cv. SLM046 were inoculated and co-cultivated with Agrobacterium inoculum on MS medium containing 1 mg/l 2,4-D and 4.5 mg/l BAP. After co-cultivation, cotyledonary explants were transferred to MS selection medium, containing 4.5 mg/l BAP and 4 mg/l phosphinothricine , 400 mg/l cefotaxime and 300 mg/l carbeniciline. The regenerated plants were analyzed by histochemical GUS assay according to the method reported by Jefferson et al. (1987).
The rooted transgenic plants were transferred into a mixture of peat and perlite (1:1, v/v), and they were grown in the greenhouse conditions. At five-leaf stage, the plants were incubated at 4°C for 8 weeks to vernalize, and then they were moved to 25°C for 16 h in light and 8 h in dark till maturity. The presence of transgene in T1 transgenic plants was confirmed by amplification of BnSHP sense, antisense cassette (F: 5 0 -AATACTAGTGGCGCGCCCCG TTAACCC TCCTACTG-3 0 , R: 5 0 -GCCTTAATTAAATT TAAATTTGAAGAGGAGGTTGGTC-3 0 , underlined part is a tail segment) and bar (F: 5 0 -ATCTCGGTGACGGG CAGGAC-3 0 , R: 5 0 -CGCAGGACCCGCAGGAGTG-3 0 ) by PCR. A T 2 transgenic line (cultured seeds from T1 line) was used for gene expression evaluation.
To study the expression of BnSHP alleles (SHP1, SHP2a and SHP2-b), pod samples were taken from three transgenic plants of a T 2 line and one non-transgenic plants (two replications from each plant) for RNA extraction. RT-and Real-time PCR was conducted with two specific primer pairs: P14 F: 5 0 -TGAACTAGTCCATGGAGATCTTCTTCTC ATGATCAGTCGCAGCATT-3 0 , P14 R: 5 0 -AGCTTAA TTAAATTTAAATTTAAACAAGTTGAAGAGGAGGT TGG-3 0 (producing a 151-bp fragment from three alleles, underlined part is a tail segment); P19 F: 5 0 -GAACAA GGCGCGAGATTGAATCC-3 0 and P19 R: 5 0 -GATCATG AGAAGAAGACAGACCGG-3 0 (producing a 94-bp fragment from SHP1). The GAPDH gene-specific primers: F: 5 0 -AGAGCCGCTTCCTTCAACATCATT-3 0 and R: 5 0 -TG GGCACACGGAAGGACATACC-3 0 (producing a 112-bp fragment) were used as a reference gene. The relative gene expression data were analyzed using the 2 -DDCT method as described previously (Schmittgen and Livak 2008). The amount of BnSHP gene expression in transgenic and nontransgenic plants was calculated using the threshold data by 2 -DDCT method (Pfaffl et al. 2002), and the data were statistically analyzed, and the graphs were drawn by Bio-Rad software package.

Construction of RNAi cassette and transgenic plant production
To construct the RNAi cassette, a 506-bp fragment (containing a 470-bp sequence downstream the MADS-box region of the BnSHP1 cDNA) was amplified and cloned in the sense and antisense direction in either side of a GUS intron in plasmid vector pGAS1252 (Fig. 1). The cassette was then sub-cloned in the Pst1 site in pCAMBIA3301, and the recombinant plasmid was designated as pCAMR-NAi. The RNAi cassette was transferred into rapeseed, and putative transgenic plants were regenerated on medium containing phosphinotricin (Fig 2). Transformation efficiency of 3.7 % was obtained in SLM046 cultivar. Histochemical GUS assay in transgenic plants was done using Jefferson's method (Jefferson et al. 1987), and blue leaf samples were observed in transgenic plants (Fig. 3). PCR analysis on T0 putative transgenic and wild-type plants showed the insertion of 501-bp antisense segment only in putative transgenic plants (Fig. 4). Seeds of putative T1 transgenic plants were sown in soil in pots, and growing plants at 5-6 leaf stage were subjected to 1.5 % basta herbicide. Three herbicide resistant lines were selected, and the insertion of 1,000-bp fragment of the GUS gene was confirmed by PCR (Fig. 5).  The effect of silencing cassette on gene expression was evaluated using T2 transgenic lines. The transgenic plants were grown in greenhouse conditions and showed normal vegetative and reproductive characteristics compared with control wild-type plants. The presence of the transgene was confirmed by amplification of a 400 bp of the bar gene in transgenic plants (Fig. 6).
Expression of the BnSHP gene was analyzed using RT-PCR. A sharp 100-bp band for BnSHP1 and BnSHP2 gene expression was observed in wild-type plants, while transgenic plants showed a weak band as was expected (Fig. 5). The primer P14 amplifies all three BnSHP alleles (SHP1, SHP2-a and SHP2-b), and P19 amplifies BnSHP1. A very slight difference in gene expression was observed in transgenic plants. Expression level for housekeeping GAPDH gene as control was the same in both transgenic and control plant, indicating that expression of BnSHP alleles was decreased in transgenic plants as judged with the GAPDH gene expression (Fig. 7).

Real-time PCR
To analyze the relative changes in gene expression, a quantitative Real-time PCR using two specific primers for BnSHP alleles and GAPDH gene was applied (Fig. 8). The efficiency of amplification using dilution series was 95 and 94 % for the GAPDH and the BnSHP genes. A higher C T value was observed for transgenic plants than non-transgenic plants, implying that the BnSHP gene expression has been decreased in transgenic plants. The data were statistically analyzed, and the related vertical graph showed the 97 % reduction in gene expression (Fig. 9). The genetic variation in shatter resistance has been reported in Brassica species, including B. rapa L., B. juncea L., B. hirta L. and within wild relatives of Brassica (Kadkol et al. 1985;Wang et al. 2007, Bagheri et al. 2012. However, to avoid windrowing of crops on a routine basis, the level of resistance is insufficient (Raman et al. 2011). Random impact test (RIT) on 229 accessions of B. napus for shatter resistance detected only two shatter resistant lines (Wen et al. 2008). Using introgression method, shatter resistance could be improved in B. juncea (Kadkol 2009;Raman et al. 2011). However, unwanted traits could be present too as reported by Summers et al. (2003). The DK142 line (resynthesized B. napus using B. oleracea alboglabra and B. rapa chinensis) showed higher shatter resistance in all locations, but significantly lower seed yield than commercial variety (Summers et al. 2003).
Indehiscent and harder siliques were obtained by constitutive MADSB expression in winter rape lines compared with wild-type winter rape plants, and precocious seed dispersal was prevented (Chandler et al. 2005). However, for unclear reason, the transgenic summer rape lines did not show a non-opening silique phenotype. This indicated that the tender rape cultivars differences could affect the coordination of signaling events involved in fruit dehiscence. A number of genes have been shown to involve in fruit dehiscence and seed shattering. Among the reported genes, SHP, IND and ALC genes have been suggested as ideal candidates for manipulation in breeding shatter resistant lines (Liljegren et al. 2000;Tan et al. 2009). In B. napus, three SHP alleles have been reported, and due to more probable redundancy of these alleles, resistance to shattering may need control of all these alleles simultaneously (Tan et al. 2009).
In Arabidopsis, double mutant of both SHATTER-PROOF1 (SHP1) and SHATTERPROOF2 (SHP2) produced indehiscent pods (Liljegren et al. 2000). Pod shatter resistance was also observed in mutants of the INDEHIS-CENT gene (Liljegren et al. 2004;Wu et al. 2006), and the ALCATRAZ gene in Arabidopsis (Rajani and Sundaresan 2001). Sorefan et al. (2009) showed that the INDEHIS-CENT (IND) mutant produced indehiscent fruits by preventing differentiation of tissue in dehiscence zone into an abscission layer.
Over-expression of FRUITFUL (FUL), a repressor of SHP and IND, produced indehiscent siliques in Arabidopsis (Ferrandiz et al. 2000). Ectopic expression of FUL gene in B. juncea produced indehiscent fruits, but they could not be threshed in a combine harvester without seed damage (Ostergaard et al. 2006). In the present study, introduction of the BnSHP gene-silencing cassette into B. napus showed a drastic reduction in BnSHP expression. Morphologically, transgenic plants were normal and set seeds at greenhouse conditions. At ripening stage, pods failed to shatter. Further phenotypic evaluation of shatter resistance such as the cantilever test, pendulum test (Kadkol et al. 1984) and the RIT are needed to be done in produced transgenic plants. Fig. 9 The high reduction in the BnSHP gene expression in B. napus transgenic plant compared with wild-type plant using Real-time PCR analysis