Development of Self-Compatible B. rapa by RNAi-Mediated S Locus Gene Silencing

The self-incompatibility (SI) system is genetically controlled by a single polymorphic locus known as the S-locus in the Brassicaceae. Pollen rejection occurs when the stigma and pollen share the same S-haplotype. Recognition of S-haplotype specificity has recently been shown to involve at least two S-locus genes, S-receptor kinase (SRK) and S-locus protein 11 or S locus Cysteine-rich (SP11/SCR) protein. Here, we examined the function of S60, one SP11/SCR allele of B. rapa cv. Osome, using a RNAi-mediated gene silencing approach. The transgenic RNAi lines were highly self-compatible, and this trait was stable in subsequent generations, even after crossing with other commercial lines. These findings also suggested that the resultant self-compatibility could be transferred to commercial cultivars with the desired performances in B. rapa.


Introduction
In flowering plants, self-incompatibility (SI) is a genetic system that promotes outcrossing by rejecting self-related pollen [1][2][3]. In Brassica, recognition of pollen is controlled by S haplotypes (designated S 1 , S 2 ,... Sn), each of which consists of the pollen determinant gene, the S locus protein 11 (SP11, also called SCR) [4][5][6], and the pistil determinant gene, the S receptor kinase (SRK) [7][8][9]. Genes of SP11/SCR (SP11 hereafter) and SRK are closely linked each other at the S locus [10], and the S locus contain multiple alleles [11][12][13][14]. The self-incompatibility response occurs when the pollen and pistil share the same S haplotype. SP11 interacts with SRK of the same S haplotype and activates its kinase domain [15][16][17]. This activation is believed to elicit a signaling cascade within the stigmatic papilla cell that leads to rejection of self-pollen [18,19]. SP11 is expressed in the anther tapetum, a sporophytic tissue. Therefore, the SI phenotype in pollen is determined by the dominance relationships between the two S haplotypes carried by the plant. Based on these relationships, the S haplotypes in Brassica have been broadly classified into two groups: pollen-dominant S haplotypes (class I) and pollen-recessive S haplotypes (class II) [14,[20][21][22]. Pollen-dominant S haplotypes (such as S 8 , S 9 , S 12 and S 52 in Brassica rapa) are generally codominant, and are almost always dominant over recessive S haplotypes (such as S 44 , S 60 , S 40 and S 29 ) [14,21]. In S heterozygotes with dominant and recessive S haplotypes, the expression of the recessive SP11 allele is silenced as a result of tapetum-specific de novo cytosine methylation in its promoter region immediately before the initiation of SP11 transcription [23,24]. Additionally, several types of cysteine-rich peptides/ polypeptides (CRPs) are expressed specifically in flowers and seeds, where they play reproductive regulatory roles [25]. For instance, SP11, with eight conserved cysteines belonging to a subclass of defensin-like proteins, is involved in the inhibition of self-pollen germination and pollen tube growth [15]. Defensin-like LUREs act as attractants, guiding pollen tubes to the embryo sac [26,27].
SI is one of the most important obstructions to B. rapa seed production and for that self-compatible (SC) cultivar is very important in case of commercial cultivation. Manipulation of S locus genes is one of the most recognized way so far to convert SI into SC in Brassica. RNA interference (RNAi)-mediated suppression of stylar 120 kDa glycoprotein (120K), a non-S-factor, results in breakdown of the capability of the pistil to reject self-pollen, suggesting that it is required for gametophytic self-incompatibility (GSI) function [28]. In plants, RNAi can be used to regulate endogenous genes [29], and by utilizing a partial sequence of an endogenous gene in the inverted repeat regions of the silencing construct, high-level silencing of the target gene expression can also be achieved. The various RNAi techniques each have advantages and disadvantages with respect to how persistent their effect is and the range of plants to which they can be applied. For example, bombardment can be applied to any plant, but produces only transient effects. Alternatively, transformation with ihpRNAexpressing transgenes provides stable and heritable gene silencing. ihpRNA transgenes have been shown to be very effective for a wide range of target genes in various plant species [30][31][32][33][34][35], indicating that the RNAi mechanism is probably conserved in all plant species.
Here, we report for the first time the use of RNAi gene-silencing constructs to achieve silencing of the S locus SP11 gene S 60 resulting in the development of a self-compatible B. rapa transgenic line. Furthermore, we demonstrate the stable inheritance of these phenotypes in progeny derived by either selfing or intercrossing and assess the performance of these lines.

Preparation of RNAi constructs
The coding sequence of S 60 containing 285 bp nucleotides was placed upstream and downstream of the Gus gene encoding the bglucuronidase fragment in opposite directions using the methods described by Chuang and Meyerowitz [36]. This Gus gene with the sense and antisense coding sequence of S 60 was placed in the SalI and SacI site between the SSH and nopaline synthase terminator of binary vector pBI101 constructed from pBI121 [37]. The SSH fragment was then removed and the S 60 promoter of 489 bp was placed in the KpnI and SalI site. The resulting RNAi construct was denoted as S60-SP11RNAi ( Figure 1). This construct was then introduced into Agrobacterium tumefaciens strain EHA105.

Plant transformation and regeneration
The hypocotyl transformation protocol developed for B. rapa [38,39] was followed in this study, with some modifications. Briefly, seeds of B. rapa cv. Osome were surface-sterilized by washing with 70% ethanol for 2 min, 1% sodium hypochlorite for 15 min and double distilled water for 3-4 times. Seeds were germinated and grown in 0.16 MS medium in a culture room maintained at 22-24uC with a 16 h light/8 h dark photoperiod at a light intensity of 4500-5500 lux. Hypocotyls were excised from 6 to 7-day-old seedlings, cut into segments 2-4 mm in length, and placed onto MS-1 medium and pre-cultured for 24 h under indirect continuous light. Explants were then immersed in a suspension of 1xl0 8 bacteria/ml for 30 min with shaking at 40 rpm, then returned to feeder plates. After two days of cocultivation with Agrobacterium, explants were transferred to B5-1 medium supplemented with 500 mg/l carbenicillin and kept at 24uC under continuous light at 7500 lux intensity for 3-7 days, then transferred to B5-BZ shoot regeneration medium. These explants were cultured for seven days in B5-BZ medium supplemented with 500 mg/l carbenicillin, 10 mg/l Hyg, and 10 mg/l AgNO 3 , followed by 14 days in medium supplemented with 500 mg/l carbenicillin and 20 mg/l Hyg, 14 days in medium supplemented with 500 mg/l carbenicillin and 30 mg/l Hyg and finally, 14 days in medium supplemented with 500 mg/l carbenicillin and 30 mg/l Hyg. The cultures were then transferred onto B5-0 shoot maturation medium supplemented with 500 mg/1 carbenicillin and 50 mg/l Hyg. Two weeks later, shoots were trimmed to contain 2-3 nodes and then placed on B5 root induction medium supplemented with 2 mg/l IBA, 500 mg/l carbenicillin and 50 mg/l Hyg. Roots developed on some of the shoots after two weeks. Shoots that had not rooted were re-cut at the base and placed back onto the medium for another 2-4 weeks.

DNA extraction and PCR analysis
The putative T 0 , T 1 and BC 5 F 2 plants were analyzed by PCR to confirm the presence of transgenes. To accomplish this, plant genomic DNA was isolated from the leaves of each line using the DNeasy Plant Mini Kit (Qiagen, USA). The primers specific for the S60-SP11RNAi cassette (59-GGC ATA TGA AGC TTG TCG ACA TGA TTT AAC TTT GCA ACAG -39; 59-CTG CAG GAG CTC GCG GCC GCA TGA TTT AAC TTT GCA ACAG -39) and for NPTII (59-CAA GAT GGA TTG CAC GCA GG -39; 59-GAA GAA CTC GTC AAG AAG GCG -39) were used to identify transgenic plants. PCR reactions were carried out in a 20 l mixture at 94uC for 5 min, then subjected to 35 cycles of amplification at 94uC for 1 min, 55uC for 1 min, and 72uC for 2 min. PCR products were visualized by electrophoresis on 1% agarose gel.

RNA extraction and expression analysis
RNA was extracted from the anthers of T 1 and BC 5 F 2 plants using an Rneasy mini kit (Qiagen, USA). RNA was treated with RNase-free DNase (Promega, USA) to remove genomic DNA contaminants. Between 50 and 100 ng of total RNA was used to make single-strand cDNA using SuperScript III reverse transcriptase (Invitrogen, Toronto, Canada) in a 20 ml reaction with oligo (dT18) primers according to the manufacturer's instructions. RT-PCR was conducted using an AMV one step RT-PCR kit (Takara, Japan). Primers specific for S 60 (59-ATG AGA TAT GCT ACT TCT ATA TAT ACA -39; 59-TGA TTT AAC TTT GCA ACA GTA GCA -39) were used for RT-PCR, and actin primers specific for Brassica (59-ATG GCC GAG GCT GAT GAC AT -39 and 59-AGC CTC GGT AAG AAG AAC CG -39) were used as a control. PCR was conducted using 50 ng of cDNA from the anthers of respective plants as templates in master mixes composed of 20 pmol of each primer, 150 mM of each dNTP, 1.2 U of Taq polymerase, 1x Taq polymerase buffer, and double-distilled H 2 O diluted to a total volume of 20 ml in 0.5 ml PCR tubes. The samples were then subjected to the following conditions: initial denaturation at 94uC for 5 min, followed by 30 cycles of denaturation at 94uC for 30 s, annealing at 58uC for 30 s and extension at 72uC for 1 min, with a final extension for 5 min at 72uC. PCR products were visualized by electrophoresis on 1% agarose gel. Real-time quantitative PCR was performed using 1 ml of cDNA in a 25 ml reaction employing iTaq TM SYBR H Green Super-mix with ROX (California, USA). The same primers used for RT-PCR were employed for real-time PCR, while Brassica actin primers (59-CAA CCA ATC GTC TGT GAC AA -39; 59-ATG TCT TGG CCT ACC AAC AA -39) were used as a control. The conditions for real-time PCR were as follows: initial denaturation for 10 min at 95uC, followed by 40 cycles of 94uC for 30 s, 58uC for 30 s, and 72uC for 45 s. The fluorescence was measured following the last step of each cycle, and three replications were used for each sample. Amplification, detection, and data analysis were conducted using a Rotor-Gene 6000 real-time rotary analyzer (Corbett Life Science, Australia).

Self-compatibility analysis
SC was tested by observing the pollen-tube behavior and fruit set ratio. For observation of pollen-tube behavior, flower buds were covered with a bag one day before anthesis and then pollinated with pollen from the same plant, after which the samples were placed on 1% solid agar plate at room temperature for 24 h. The pollinated pistils were then separated from the buds and fixed in acetic alcohol (ethanol: acetic acid = 3:1) for 5 h at room temperature. The fixed pistils were hydrolyzed in 1 N NaOH for 2h at 60 uC, after which they were stained with decolorized aniline blue solution (0.01% aniline blue in 2% K 3 PO 4 ) for 2 h. The stained stigmas were mounted on glass slides with 50% glycerol and observed using a UV fluorescent microscope (Nikon Eclipse 80i, Japan). Three flowers were used from each plant. During flowering, the major inflorescence was bagged for self-pollination and the self-compatibility index (SCI) (self-compatibility index = the number of seeds/the number of flowers) was calculated according to Zhang et al. [40].

Identification of target gene and plasmid construction
The sequences of mature SP11 proteins are highly divergent, except for the presence of conserved cysteines [41]. Class II SP11s show several similarities with class I SP11s such as, they are small, secreted proteins with conserved putative signal peptides, their mature proteins show S-haplotype-specific polymorphisms in spite of their common hydrophilic and basic properties, like other pollen coat proteins (PCPs), they have eight conserved cysteine residues [20]. All eight cysteine residues are conserved in an arrangement that is characteristic of SP11. Conservation of the eight cysteine residues suggests a common three-dimensional protein structure of SP11s that is stabilized by intramolecular disulfide bonds, similar to the defensin family of antimicrobial proteins [42]. Two other residues that are conserved among most of the class I SP11s, a glycine residue between C1 and C2 and an aromatic amino acid residue between C3 and C4 [41], also are conserved among all of the class II SP11s. Hydrophilicity analysis of the class II SP11s does not suggest a hydrophilic (surfaceexposed) structure in the C3-C4 region, which has been suggested for class I SP11s [43]. Three other regions, C1-C2, C2-C3, and C5-C6, are highly divergent and contain two or three amino acid residues that are completely variable across the four class II Shaplotypes of B. rapa [20]. B. rapa cv. Osome, a heterozygote of the S 52 and S 60 haplotype [44], was used for silencing of SP11/SCR gene based on RNA interference. S 52 is a class I and S 60 is a class II SP11 S-haplotype [9,14]. Amino acid sequence identities among class II SP11s are 62.3 to 94.6%, rather high compared with those of class I SP11s which ranges from 19.5 to 76.1% [20]. Due to having low diversity and number of alleles of class II SP11s, we considered class II SP11, S 60 as a crucial target for conversion of the SI B. rapa cv. Osome to SC. The 285 bp-length sequence of the S 60 gene was selected as the target for RNA silencing.
We constructed an RNAi binary vector designated as S60-SP11RNAi ( Figure 1) for introduction into SI B. rapa cv. Osome. The target sequence was incorporated into pBI101 in invertedrepeat orientations interrupted by GUS (1023bp). To enhance the efficiency of RNA silencing, strong S 60 promoter (2489 to 21) was used to control expression of the transgene, and a plasmid was constructed to specifically suppress expression of S 60 . It is well known that the 59-flanking region up to 2192 bp SP11 is sufficient to direct gene expression in tapetum and pollen [45]. Moreover, the vector carried plant resistant gene NPTII driven by the Nos promoter and HPT driven by the CaMV 35S promoter.

Transformation and selection of homozygous transgenic plants
Using the hypocotyl as the explants, plasmid pBI101 with a S60-SP11RNAi cassette was transformed into B. rapa mediated by A. tumefaciens EHA105. Subsequently, the NPTII-resistant plantlets were regenerated, and the insertion of silencing fragments was  confirmed by genomic DNA PCR (data not shown). Overall, 16 transgenic plants were obtained and grown to generate T 1 seeds, after which they were harvested separately. Around, 20 to 25 T 1 seeds were harvested from each transgenic line by selfing and then cultured in pots to produce T 2 seeds. For genetic segregation of the T 2 generation, 15-20 T 2 seeds from each T 1 line were tested onto 1/10 strength of MS plate with 50 mg/l hygromycin for antibiotic resistance screening. After five weeks of culture, heterozygote and homozygote events were counted on the basis of hygromycin resistance and susceptibility of plants. Among the 16 transformant lines, three lines produced transformants and non-transformants in a 3:1 (x 2 = 0.003-0.04, P = 0.88-0.93) ratio. This pattern of segregation is only possible when single copy transfer of genes occurs in the transgenic plants. The other transformants were not segregated in such a pattern in the T 2 generation, indicating the transfer of more than one copy of the target gene. Therefore, only homozygous transformant lines, self-compatible by RNAi (SR) 6-10, SR11-8 and SR18-7 were used for further study.

Suppression of S 60 gene and conversion to selfcompatible state
Real-time PCR was conducted to analyze the relative transcription level of the S 60 gene among the three aforementioned homozygotic T 2 lines. The results revealed much lower transcription levels in the transformants lines than the non-transformant line (Figure 2). Suppression levels varied from almost 2.5 to approximately 20-fold when compared with the untransformed control, with the highest suppression being observed in SR6-10. These three T 2 lines were self-pollinated to produce T 3 seeds and the compatibility of these lines was also tested by investigating the pollen tube behavior and fruit set. Unlike the self-incompatible tester line (Figure 3A), the S60-SP11RNAi lines showed numerous pollen tubes on stigma papilla cells upon self-pollination ( Figure 3B-D). The mean SCI values of the SR6-10, SR11-8 and SR18-7 T 2 lines were 6.89, 6.28 and 5.91, respectively, which were comparable to their average seed per open-pollinated flower values (7.95, 7.55 and 7.38). The mean SCI of the tester line was 0.82. Taken together, the pollen tube observations and fruit set data confirmed that the S60-SP11RNAi lines were highly selfcompatible.
The SI system is sporophytic in Brassica, and SP11s from class-II S-haplotypes (e.g. S 40 and S 60 ) exhibited strictly sporophytic expression patterns, suggesting that the expression of SP11 in the tapetal cell layer was sufficient for SI [20]. Recent biochemical studies revealed that SP11 functions as the sole ligand for its cognate SRK receptor complex. Their interaction induces the autophosphorylation of SRK, which is expected to trigger the signaling cascade that results in the rejection of self-pollen. This socalled ligand receptor complex interaction and receptor activation occurs in an S-haplotype-specific manner, and this specificity is almost certainly the basis for self-pollen recognition [5,15]. In this study, we silenced the S 60 gene in B. rapa to interfere with the socalled ligand receptor interaction and receptor activation and prevent consequent rejection of self-pollen and the resultant S60-SP11RNAi homozygous plants were SC.

Stable inheritance of the trait
To determine the stability of gene silencing in the following generations, three T 2 lines that had relatively strong gene silencing based on the reduction of the transcript abundance levels were selected for further analysis. All T 3 individuals produced from each T 2 line were checked through amplification of NPTII and S 60 primers using gDNA extracted from 4-week-old leaves to identify positive and negative progenies and all plants were found to be positive (data not shown). S 60 gene expression in these plants was also analyzed through RT-PCR analysis of three plants selected at random from each T 3 line using S 60 specific primers (Figure 4). The results showed that the gene suppression level was similar to that of T 1 transgenic plants when compared to the untransformed control.

Performance evaluation of RNAi transgenic lines
Transgenic SC B. rapa lines were subjected to performance evaluation as seed set experiments under two different conditions; use of pollinating agents (Bees and CO 2 ) and no pollinating agents. It should be noted that the commercial cultivation of different Brassica crops uses bees and CO 2 as pollinating agents, and CO 2 is not environment friendly. In this evaluation study, three homozygous RNAi transgenic lines and a non-transformant (WT) line were used. For the non-transformant plants, bees and CO 2 were used as pollinating agents, while for the transformant plants no bees or CO 2 were used (File S1). In each case, the total seed number was counted from four plants and highest average seed set was found in the SR6-10 RNAi transgenic line (Table 1). These findings indicate that the transgenic RNAi lines could perform well in fruit settings without any pollinating agents and could be used for commercial cultivation with great economic benefits and minimal environmental hazards.

Use of RNAi transgenic line as breeding material
SI is generally the main obstruction for B. rapa seed production and for that SC cultivar is crucial issue in case of commercial seed production. In this study, we developed a SC B. rapa line and then transferred this trait in a strictly SI commercial variety, B. rapa 'Seoulbechhu'. We utilized the backcross method to transfer the S60-SP11RNAi gene from the SR6-10 RNAi transgenic homozygous line into this variety and obtained 19 lines in the BC 5 F 2 generation. The insertion of silencing fragments was confirmed by genomic DNA PCR of these 19 lines using hygromycin resistant gene primers (data not shown). We then re-checked the inserts using class I and class II SRK universal primers to identify the homozygous lines for the S 60 gene introduced into these 19 lines and found no class I insert in lines no. 7, 10, 12 and 17 (File S2), indicating that these four lines are homozygous for class II S60-SP11RNAi gene. Because pollen-dominant class I S haplotypes (such as S 8 , S 9 , S 12 and S 52 in Brassica rapa) are generally codominant, they are almost always dominant over recessive class II S haplotypes (such as S 44 , S 60 , S 40 and S 29 ) [14,21]. Upon expression analysis, only these four lines showed expression of the S60-SP11RNAi gene at different levels, while no expression was observed in the other lines (File S3). The expression level of the S60-SP11RNAi gene in these four homozygous lines was compared with that of non-transformed S 60 containing plants. Line 10 showed the highest level of suppression when compared to nontransformed S 60 bearing plants, followed by lines 7, 17 and 12 ( Figure 5). The fruit setting of these four lines was also very high in this generation.
Taken together, it can be concluded that the transgenic RNAi lines were fully self-compatible, and that this trait was stable in subsequent generations, even after crossing with other commercial lines with higher performance than the non-transgenic lines.