Zhihao Xu,Meimei Yu,Youwei Yin,Cenwen Zhu,Wen Ji,Changquan Zhang,Qianfeng Li,Honggen Zhang, Shuzu Tang, Hengxiu Yu*, Qiaoquan Liu*

Jiangsu Key Laboratory of Crop Genetics and Physiology,Key Laboratory of Plant Functional Genomics of the Ministry of Education,Jiangsu Key Laboratory of Crop Genomics and Molecular Breeding,Jiangsu Co-Innovation Center for Modern Production Technology of Grain Crops,College of Agriculture,Yangzhou University,Yangzhou 225009,Jiangsu,China

Keywords:Rice (Oryza sativa L.)SSSII-2 RNAi Starch quality SMF transgenic lines

ABSTRACT The amylose content(AC)of rice endosperm starch varies from 0 to 35%, and is associated with rice cooking and eating quality. Soft rice has low AC, generally between 6% and 15%,and its eating quality is high whether it is consumed hot or cold.However,the appearance quality of current soft rice cultivars needs to be improved, especially opaque endosperm.Conventional genetic engineering has improved some agronomic traits of soft rice varieties,but not endosperm appearance. In the present study,a RNAi construct of the soluble starch synthase II-2(SSSII-2)and the hygromycin phosphotransferase(HPT)gene were introduced into an elite japonica rice variety, Kangtiaowuyunjing (KWY8) by co-transformation. Several selectable marker-free (SMF) transgenic lines were obtained, and SSSII-2 expression was significantly downregulated in selected transgenic lines, resulting in lower AC of the endosperm. The physicochemical properties of the transgenic rice kernels, including gel consistency(GC)and rapid visco analyzer(RVA)profile,differed significantly from those of wild-type rice and were similar to those of a soft rice variety, Nanjing 46 (NJ46). These findings indicate that the cooking, eating, and processing qualities of transgenic rice are comparable to those of NJ46. However, the transgenic rice endosperm retained a transparent appearance under low-moisture conditions. Thus, SMF SSSII-2 RNAi rice provides a resource for breeding soft rice with transparent endosperm.

1. Introduction

Rice (Oryza sativa L.) provides food for more than half of the world's population. Developing high-yielding and highquality rice cultivars is therefore important. Starch is a major component of rice endosperm, accounting for 80%-90% of dry rice seed content. Thus, starch properties play a dominant role in determining rice grain appearance and cooking and eating qualities [1-4]. Amylose and amylopectin are the two major components of starch. The amylose content (AC) of rice endosperm generally ranges from 0 to 35%, depending on the cultivar, with 20%-35% in indica rice and 15%-22% in japonica rice [5-7]. Cooked rice with high AC is usually flaky, dry, and hard, and has separate kernels, so that high AC is usually associated with poor cooking and eating quality [2,7], whereas an extremely low AC such as＜2% results in glutinous rice that is sticky after cooking.Thus, breeding rice varieties with a suitable AC is needed to improve rice cooking and eating quality. Soft rice is a special type with a relatively low AC, between 6% and 15%[8]. The quality of soft rice is between those of glutinous and non-glutinous rice and is favored by many consumers.However, the appearance quality of current soft rice varieties needs to be improved, owing largely to the opaqueness of endosperm [8-10]. Conventional genetic engineering has improved various agronomic traits of soft rice varieties, but has had little effect on the appearance of opaque endosperm.

Amylose is a polymer of glucose linked via α-1,4 bonds,whereas amylopectin has a highly branched structure of short α-1,4 chains linked by α-1,6 bonds [11]. Amylose synthesis in rice is catalyzed mainly by granule-bound starch synthase I(GBSSI), encoded by the Waxy (Wx) gene [11,12]. Amylopectin is catalyzed by soluble starch synthase(SSS),starch branching enzyme (SBE), and starch debranching enzyme (DBE) [13-15].SSS can be divided into four subfamilies (SSSI-IV) with eight isoforms (SSSI, SSSIIa, SSSIIb, SSSIIc, SSSIIIa, SSSIIIb, SSSIVa,and SSSIVb). SSSIIa is expressed in rice endosperm, SSSIIc expression is low in endosperm, and SSSIIb is expressed mainly in leaves,with lowest expression in endosperm.SSSIIb is also known as SSSII-2 [16-25]. Although the exact role of SSSII-2 in the regulation of rice starch synthesis is unclear,downregulation of SSSII-2 expression can result in lower AC[26].

Understanding the mechanism of starch biosynthesis and the development of modern molecular biology techniques has made it possible to regulate rice quality by manipulating the expression of starch synthesis genes,and transgenic lines with differing AC levels have been developed [26,27]. In the present study, SSSII-2 was subjected to RNA interference (RNAi) by introducing its RNAi construct together with the hygromycin phosphotransferase (HPT) gene into an elite japonica rice variety using a mini-twin T-DNA binary vector in an Agrobacterium strain. We got selectable marker-free (SMF) transgenic rice lines. Their kernels showed a transparent appearance and shared starch physicochemical and RVA properties with current soft rice cultivars.

2. Materials and methods

2.1. Plant materials

Two leading japonica rice varieties suitable for cultivation in Jiangsu province, China were employed: Kangtiaowuyungen 8 (KWY8) and the soft rice variety Nanjing 46 (NJ46). They were grown in an experimental field at Yangzhou University,and immature embryos at 12-15 days after pollination(DAP)were isolated from KWY8 for tissue culture and subsequent transformation experiments.

2.2. Vector construction

To generate RNAi-inducing constructs, a specific cDNA segment spanning 580 bp of the rice SSSII-2 gene was amplified using primers SSSII-2-F (5′-ACGGATCCAAGGCTGATCATGTTGAG-3′) and SSSII-2-R (5′-CTCTAG AAACACAAAATCAACAC-3′) with a cDNA template derived from the total RNA of a japonica rice variety,Nipponbare.For construction of hairpin constructs,the DNA fragment obtained above was cloned into the pGEMT vector(Promega, Madison, WI, USA) and transferred sequentially into the BamH I/Sal I and Bgl II/Xho I sites of the p1022RNAi vector modified from pBSK (Biovector Science Lab, Inc.,Beijing, China). The 100-bp DNA fragment of the first intron of the rice Glutelin gene was inserted into the p1022RNAi vector between BamH I/Sal I and Bgl II/Xho I. The hairpin construct was further cloned into one T-DNA region of the mini-twin T-DNA binary vector pSB1300, which is modified from pCAMBIA1300.There are two separate T-DNA regions in the pSB1300 vector. One contains a selectable marker, the HPT gene, which is driven by the Cauliflower mosaic virus(CaMV) 35S promoter, and the other harbors the maize Ubiquitin (Ubi) promoter. The hairpin construct for the rice SSSII-2 gene was cloned into the latter,resulting in the minitwin T-DNA binary vector pSBSSSII-2 (Fig. 1). The plasmid pSBSSSII-2 was introduced into an Agrobacterium tumefaciens strain EHA105 and used for transformation. All constructs were confirmed by DNA sequencing.

2.3. Rice transformation and selection of SMF transgenic plants

The procedure for Agrobacterium-mediated rice transformation and plant regeneration was performed following Liu et al. [28]. Calli derived from immature embryos were infected with an Agrobacterium strain harboring the mini-twin T-DNA vector pSBSSSII-2 (Fig. 1). Stable transgenic cells were screened on medium with hygromycin. Transgenic plants(T0generation) were transferred to soil and grown to maturity in a greenhouse. Co-transformants harboring both the HPT marker and the SSSII-2 RNAi construct were selected by PCR analysis. T1and T2seedlings were analyzed by PCR to detect the presence of the HPT marker and the SSSII-2 RNAi construct and SMF transgenic plants were identified. The T3to T6seedlings were further analyzed by PCR to confirm absence of the HPT gene in the SMF lines.The T6and T7generation seeds were subjected to starch quality analysis.

2.4.PCR and quantitative reverse transcription PCR(qRT-PCR)analysis

Fig.1- The T-DNA structure of the binary vector pSBSSSII-2.CaMV35S,promoter of CaMV 35S gene; NOS,terminator of the Napoline synthase gene;HPT,Hygromycin phosphotransferase gene;LB1 and RB1,left and right borders of the first T-DNA region;LB2 and RB2,left and right borders of the second T-DNA region.

Total genomic DNA was extracted from rice leaves as described previously[29].The primers used to detect the presence of the HPT gene were HP1 (5′-GCTTCTGCGGGCGATTTGTGT-3′) and HP2(5′-GGTCGCGGAGGC TATGGATGC-3′),while primers INT-F(5′-CCTCGTAATCAATTGTTAGGAT-3′)and NOS-R(5′-GACCGGC AACAGGATTCAAT-3′) were used to detect the presence of the SSSII-2 RNAi construct.PCR amplifications were performed with denaturation at 95 °C for 5 min, followed by 30 cycles at 95 °C for 50 s,55 °C for 50 s,and 72 °C for 50 s,then a final extension at 72 °C for 10 min.The expected 598-and 583-bp PCR products for the HPT gene and the SSSII-2 RNAi construct, respectively,were obtained.

Total RNAs were isolated from endosperm at 12 DAP using the cold phenol method [12], and the RNA was reversetranscribed using an oligo-dT18 primer. The primers RTSSSII-2-F(5′-TACTCGCTCTGTTCTTG TGATAC-3′)and RTSSSII-2-R(5′-ATTCCGCTCGCAAGAACTGA-3′)were used for qRT-PCR analysis carried out as described above.

2.5. Flour preparation

After at least three months of storage at room temperature,rice seeds were dehusked using a rice huller (SY88-TH,Ssangyong Ltd., Incheon, Korea) and polished with a grain polisher (Kett, Tokyo, Japan). Polished rice was ground into flour in a FOSS 1093 Cyclotec Sample Mill (Foss Tecator,Hoganas,Sweden).

2.6. Measurement of physicochemical parameters and starch viscosity

The AC and gel consistency (GC) of mature seeds were measured using standard procedures as described previously. Briefly, AC was determined using a colorimetric method with iodine-potassium iodide [2], and GC was measured by gel length [30], with longer gels considered softer than shorter gels.

The paste viscosity of rice starch was measured using a Rapid Visco Analyzer (RVA) according to the American Association of Cereal Chemists Standard Method (AACC 61-02, 1995). The method specifies the use of 3 g of rice flour in 25 mL of water. An RVA-3D instrument operated using Thermocline Windows control and analysis software version 1.2 were used (Newport Scientific, Sydney, Australia). Five primary and three secondary parameters of the pasting curve were used to describe rice-paste-viscosity indices. They were peak viscosity (PKV), hot paste viscosity(HPV), cool paste viscosity (CPV), peak time (PeT), paste temperature (PaT), breakdown viscosity (BDV, PKV-HPV),setback viscosity (SBV, PV-PKV), and consistency viscosity(CSV, CPV-HPV).

2.7. Agronomic trait evaluation and statistical analysis

For evaluation of agronomic traits, transgenic lines were planted together with the wild type and NJ46 in experimental plots at Yangzhou University (Yangzhou, Jiangsu province, China). The field tests were laid out in a randomized complete block design with three replications.Each block consisted of 300 plants in 20 rows. Thirty plants were randomly selected from the middle region of each block for trait measurement. Chalkiness degree (CD) of polished kernels based on image processing was measured with an appearance detection analyzer (MRS-9600TFU2L,MICROTEK, Shanghai, China).

All data were analyzed using Student's t-test to test differences and represented as means ± standard deviation(means ± SD). Results with a corresponding probability value of P ＜0.05 were considered statistically significant.

3. Results

3.1. Production of transformants and identification of cotransformed rice plants

Calli induced from KWY8 immature seeds were subjected to transformation mediated with pSBSSSII-2 vector,resulting in 36 independent transgenic lines. To identify co-transformants,one or two plants of each independent T0transgenic line were selected for PCR analysis. Most transformants contained the HPT gene,but only some(30/36 transgenic lines;83%)harbored the SSSII-2 RNAi construct(Fig.2-A,B).

3.2. Breeding of SMF transgenic lines containing the SSSII-2 RNAi construct

If the target and HPT genes integrate at different sites in co-transformed plants, they could segregate in subsequent generations, and SMF transgenic plants could be selected.Accordingly, T1plants derived from 30 independent cotransformants were planted in a greenhouse and analyzed by PCR, and SMF transgenic plants were obtained from nine co-transformants (30%; Table S1). As shown in Fig. 2-C, D, four (numbers 2, 5, 7, and 9) of 10 tested plants were SMF transgenic plants. All selected SMF plants were allowed to self-pollinate for five generations, and homozygotes were confirmed by PCR in every generation (Figs. 2-E,F; S1).

Fig.2-PCR analysis of target and HPT genes in transgenic plants.(A)Target and(B)HPT genes in the T0 generation.Lanes 1-10,10 plants from 10 independent transgenic lines.(C)Target and(D)HPT genes in the T1 generation.Lanes 1-10,10 T1 plants from one T0 transgenic plant.(E)Target and(F)HPT genes in the T2 generation.Lanes 1-10,10 T2 plants from one T1 transgenic plant.M,DL2000 markers;P,plasmid pSBSSSII-2; W,non-transformed wild-type KWY8.

3.3. Expression of the SSSII-2 gene in SMF transgenic plants

SSSII-2 expression in transgenic lines was tested by qRTPCR.Expression levels of SSSII-2 were reduced significantly in immature seeds of five tested lines, compared with wild-type KWY8 over two years/generations, although the degree of reduction differed between transformants (Fig.3).

3.4. Quality characteristics of transgenic rice kernel starch

To investigate the effects of reduced SSSII-2 expression on rice starch quality, the AC of transgenic lines was measured.Compared with KWY8,of the five tested transgenic lines,the highest reduction was observed in line SSSII-2-13, with AC values of 11.18% and 12.10% for T6and T7generations (2015 and 2016), respectively (Table 1). These AC values were comparable to those of the soft rice NJ46 (11.00% and 12.86%in 2015 and 2016,respectively;Table 1).

The GC values indicated that transgenic lines displaying reduced AC were softer than the WT. For example, the gel lengths of the transgenic line SSSII-2-13 were 123.35 mm and 120.60 mm, respectively, in T6and T7generations, similar to that of NJ46(112.45 mm and 113.47 mm during 2015 and 2016,respectively), and much longer than the gel length of KWY8(103.20 mm and 104.60 mm during 2015 and 2016)(Table 1).

Fig.3- qRT-PCR analysis of SSSII-2 gene expression in immature endosperm of transgenic lines.(A)T6 generation.(B) T7 generation.The Actin gene was used as internal control for normalization of gene expression.Error bars represent the SDs;*P ＜0.05, **P ＜0.01.

Table 1-Physicochemical qualities of mature seeds of transgenic lines.

To further evaluate these qualities of the transgenic lines,RVA paste viscosity was measured. Most RVA paste viscosity values and RVA profile indices of transgenic lines were altered significantly, including lower setback value (SBV) and cool paste viscosity (CPV), and increased breakdown viscosity(BDV) (Table 2, Fig. 4). These RVA values of the transgenic lines reached or exceeded those of the elite soft rice cultivar NJ46. For example, the BDV of the transgenic lines was between 707.5 and 1011, considerably higher than that of KWJ8 (674.5). The BDV value of the soft rice NJ46 was 1050.5.High BDV,low SBV and CPV values are always associated with high eating quality;cooked rice is elastic,soft,and sticky,and cold rice is flexible [31]. These results indicate that the cooking, eating, and processing qualities of transgenic lines were comparable with those of NJ46, demonstrating the soft rice quality of the SSSII-2 RNAi lines.

3.5. Agronomic traits and kernel characteristics of transgenic lines

Agronomic traits of selected SMF transgenic and WT lines were measured under normal field conditions.Over the two years,the 1000-kernel weight of some transgenic lines was reducedsignificantly compared with WT.The lowest 1000-kernel weight was observed for the transgenic line SSSII-2-23 (24.91 g and 23.40 g in 2015 and 2016, respectively). Most of the other trait values were comparable to those of the WT(Table 3).

Table 2-Thermal properties of mature seeds of the transgenic lines determined by RVA.

Fig.4- RVA profiles of endosperm starch of the transgenic lines.The abscissa represents time and the ordinate represents relative viscosity value(RVU).A and B show the results in 2015 and 2016,respectively.

To assess the appearance of transgenic rice endosperm,dried mature rice kernels were polished. The transparent appearance and low chalkiness degree of mature endosperm of transgenic lines were similar to those of KWY8.By contrast,opaque and high-chalkiness endosperm was observed in NJ46(Figs. 5; S2). These findings show that we have successfully bred soft rice with desirable appearance by down regulating the SSSII-2 gene.

4. Discussion

As a major component of rice endosperm, starch quality and characteristics of starch granules affect the appearancequality of rice seeds. ＞20 genes have been identified as being involved in starch biosynthesis, and they work coordinately.Amylose synthesis in rice endosperm is regulated mainly by the Wx gene,while the synthesis of amylopectin is modulated by multiple genes [12-17,32]. Variation in the Wx gene sequence affects AC in rice endosperm, and the Wx alleles Du,Flo2,Wxop/hp,Wxmp,and Wxmqhave been used for soft rice breeding [9,10,33-38]. Soft rice varieties can be generated using these resources by conventional breeding methods such as backcrossing. However, this approach is time-consuming and labor-intensive, and there are obvious defects in current soft rice cultivars harboring these alleles, including opaque endosperm,which negatively affects appearance and thus the market value of soft rice. Genetic engineering provides an alternative method to manipulate AC by regulating expression of the starch synthesis-related genes and thereby improving rice quality [26,27]. In the present study, the expression of the SSSII-2 gene was down regulated using RNAi, resulting in lower AC and their GC and RVA values similar to those of a soft rice cultivar, NJ46. In addition,transgenic rice kernels retained a transparent appearance.Therefore, down-regulation of SSSII-2 gene expression is a good alternative way to cultivate soft rice with good appearance quality.

Table 3-Agronomic traits of the transgenic lines.

Fig.5-Kernel phenotypes of transgenic lines.Kernel phenotypes of(A)T6 and(B)T7 generation transgenic lines.a,b,c,d,e,f,and g represent kernels from KWY8,SSSII-2-5,SSSII-2-9,SSSII-2-13,SSSII-2-23,SSSII-2-16,and NJ46,respectively.

Genetic modification of starch synthesis genes may lead to pleiotropic effects, including affecting expression of other starch biosynthesis genes. For example, in SSSII-2 RNAi transgenic lines the expressions of many other genes were affected, including Wx, SSSI, SSSII-1, SSSIII-3, SSSIII-1, SSSIII-2,SSSIV-1,SBEI,SBEIIa,and SBEIIb[26].These changes resulted in modified amylopectin chain length, thermodynamic properties, and starch viscosity of the transgenic rice, beside lower AC.The regular shape and intact surface of the starch granules of SSSII-2 RNAi rice could also be one reason for its transparent appearance at low moisture contents [26]. We accordingly propose that the reduction of AC and transparent appearance in the SSSII-2 RNAi transgenic lines is caused mainly by the reduction of SSSII-2 expression, but that the expression changes of many other starch synthesis related genes may also have contributed. The reason for the transparent appearance of SSSII-2 RNAi lines awaits investigation.

Selectable marker genes, which render transgenic cells resistances to antibiotic or herbicide, are essential for selection of the relatively small number of cells in which integration of foreign DNA occurs. However, selectable marker genes are dispensable once transgenic plants are obtained. Their continued presence prevents retransformation with the same marker system, given the limited number of selectable marker genes available. In addition, some marker genes may raise safety or public concerns, such as whether the expression product of resistance marker genes could result in antibiotic resistance of plant pathogens,harming the health of humans and other animals. For these reasons, generation of SMF genetically modified (GM) plants is a target of plant biotechnology research as well as of transgenic breeding. To a large extent these concerns can be addressed by developing SMF GM plants,circumventing the biosafety issue,given that rice is a food crop [39]. Some existing transgenic crops harbor selectable marker genes, such as the soft rice previously[26] derived by downregulation of SSSII-2. Selectable marker genes also increase the time and effort required for safety evaluation preceding the commercial production of transgenic crops, given that safety evaluation is conducted caseby-case[39-47].The optimal solution to these problems is to develop GM plants with no selectable markers. The soft rice lines generated in this study are free of selectable markers,and the DNA fragment for RNAi is derived from rice. Thus,the SMF transgenic soft lines with transparent kernels developed in the present study are free of any foreign gene.Furthermore, the recipient, KWY8, is a commercial variety with high yield, quality, and disease resistance, suitable for cultivation in Jiangsu province,China.These SMF transgenic soft rice lines can be used in soft rice production after safety evaluation.

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2019.05.006.

Acknowledgments

This work was supported by the Key Transgenic Breeding Program of China(2016ZX08001006,2016ZX08001002-003),the National Key Research and Development Program(2016YFD0102000), Yangzhou City Science and Technology Plan(YZ2017059),the National Natural Science Foundation of China (31872859), Jiangsu Agricultural Science and Technology Innovation Fund (CX181001), the Priority Academic Program Development of Jiangsu Higher Education Institutions(PAPD).