A New Strategy for Speci c Qualities of Waxy Rice Breeding by Analyzing Physicochemical Properties Between Five Waxy Mutants and Corresponding Their Wild Types

Yuhao Fu Rice Research Institute of Sichuan Agricultural University Tingting Luo Rice Research Institute of Sichuan Agricultural University Zizhong Zhu Rice Research Institute of Sichuan Agricultural University Yiping Liu Rice Research Institute of Sichuan Agricultural University Xu Liu Rice Research Institute of Sichuan Agricultural University Baoli Zhang Rice Research Institute of Sichuan Agricultural University Rui Liu Rice Research Institute of Sichuan Agricultural University Jun Zhu (  Zhujun987@126.com ) Rice Research Institute of Sichuan Agricultural University


Introduction
Waxy rice is widely used in food processing, industry, medicine, and cosmetics because of its unique functional characteristics (Bao et al. 2004, Chun et al. 2010, Lee et al. 2009, Puchongkavarin et al. 2005. It contents 80% starch of the dry weight, and almost all of the starch is amylopectin with only a small amount or no amylose (Bean et al. 1984). Some studies suggest that the different waxy rice The Waxy gene (Wx) is located on rice chromosome 6 and encodes GBSSI, which mainly controls amylose synthesis in endosperm and directly affects rice quality (Sano 1984, Wang et al. 1995. Numerous allelic variations of Wx have been found, namely, Wx lv , Wx a , Wx in , Wx b , Wx op , Wx mp , and wx, and different alleles contain varying AAC in the endosperm and has affected people preferences ( Liu et al. 2005, SATOH andOMURA 1981). For instance, in recent years, the Wx gene-edited by the CRISPR/Cas9 system obtained a wx mutantwith signi cantly reduced AAC without changing other agronomic traits ). However, the correlation of physicochemical properties between wx and WT before and after Wx editing remains unclear, which leads to uncertainty in waxy rice breeding.
The amylose content (AC), chain-length (CL) distribution of amylopectin and amylopectin ne structure have been proved to be vital indicators in uencing waxy rice's gelatinization, retrogradation, and rheology (Huang and Lai 2014, Jane et al. 1999, Vandeputte et al. 2003a, Villareal et al. 1997. For instance, waxy rice starch with amylose as the donor has increased in the degree of crystallinity, solubility and paste clarity, gelatinization temperature and enthalpy (Guo et al. 2019). The starch retrogradation is not accessible to retrograde with the increase of amylose content. (Sasaki et al. 2000). On the other hand, the higher branch density of amylopectin increases the gelatinization enthalpy (ΔH) and solubility but decreases the viscosity of waxy rice starch (Ren et al. 2017, Sorndech et al. 2015. Similarly, the higher proportion of the long chain of amylopectin accelerates starch's retrogradation (Karim et al. 2000). Reassociating long chains during retrogradation will improve gels' rigidity (Singh et al. 2012). These studies revealed the vital role of amylopectin ne structure and AC in the physicochemical properties of waxy rice. Therefore, breeding waxy rice varieties with different amylopectin ne structures enriches waxy rice germplasm resources and expands its application prospects.
This study systematically analyzed the relationship between physicochemical properties of wx mutants and WT. The result shows that the quality of wx mutant is highly correlated with WT, which can predict wx mutant properties before breeding. Based on these ndings, we enriched the knowledge of waxy rice breeding and contributed to the directed cultivation of special-purpose waxy rice.

Plant materials and growth conditions
We carefully selected two elite rice varieties with different genetic backgrounds (Rice Research Institute of Sichuan Agricultural University), mainly planted in eastern Asia, central China, and western China. Japonica waxy NY1 and indica waxy NY2 were used as controls. Then, we enlarged the study to ve elite rice varieties to nd a potential link between wx and WT. The CRISPR/Cas9-targeted genome editing tool, was constructed as previously described (Feng et al. 2013). The primer sequences used to construct the vector are displayed in Table S1. Unless indicated, all rice lines were grown in paddy elds in Chengdu, China, during normal rice-growing seasons.

Gene cloning
The primers used for Wx sequences, Wx allele genotype, target sequences and reporter gene detection are listed in Table S1.

Measurements
Before Wx rice and rice starch physicochemical property determination, the seeds of a single Wx mutant with stable inheritance in T5 plants were dried at 37 °C for two weeks. The dried seeds were shelled, polished, and milled by a pearling rice mill, and nally screened through a 74-micron mesh. Starch was prepared as the method of Precha Atsawanan et al. (2018).

Quanti cation of starch and amylose
Before measurement, the rice our was balanced in a constant temperature and humidity cabinet for seven days. The starch content was determined according to the method of Smith and Zeeman (2006), and the amylose content of rice was determined using a colourimetric assay ).
Standard curves were plotted with standard samples of rice amylose and amylopectin (China National Rice Research Institute, Zhejiang, China).

Quanti cation of gel consistency and total protein
The gel consistency (GC) of grains was evaluated as described previously (Cagampang et al. 1973).
According to the Comin amylose assay procedure, the total protein content of grain was assayed by the Kjeldahl method.

Waxy rice starch thermal properties
The thermal properties of waxy rice starch were measured using differential scanning calorimetry (DSC Q2000, TA Instruments Ltd. Crawley UK) (Yang et al. 2018). The onset temperature (To), peak temperature (Tp), conclusion temperature (Tc), and gelatinization enthalpy (ΔH) were recorded with Universal Analysis 2000 (TA Instruments Ltd. Crawley UK).

Amylopectin chain length distribution
The chain-length distribution of amylopectin was measured using high-performance anion-exchange chromatography equipped with a pulsed amperometric detector (HPAE-PAD) according to the method of Kowittaya and Lumdubwong (2014).

Date analysis
All experiments were independently repeated three times. The results were expressed as the mean ± standard deviations. Data processing was analyzed with SPSS 25.0, and signi cance was de ned as P < 0.05 (* or lowercase letters) and P < 0.01 (** or capital letters). Snapgene was used to compare splicing and analysis carefully.

Molecular identi cation
In this study, the two cultivated rice varieties with the different major Wx alleles (Table S2) were selected rstly, QLD (Wx a ) and YSZ (Wx b ), which are widely grown in the rice region of Southwest China. We designed a CRISPR/Cas9 (Fig. 1a) construction accurately targeting the second exon (87-109 bp) of the Wx gene with the expectation to generate a null mutation (Fig. 1b). Based on this vector, the results suggest four major mutant types of QLD (Fig. 1c) and ve major mutant types of YSZ ( Fig. 1d) were obtained. This experiment observed high mutagenesis e ciency in 80% T0 transformants mutating in QLD plants and 82.35% YSZ (Table S2). Finally, our results again proved this method did not change the main agronomic traits (Fig. S1).

The quality of wx
We further identi ed two single-base homozygous mutations, YSZwx1 and QLDwx1, and performed quality analysis on them (Table S3). The results show no signi cant difference in the major grain quality of the two wx seeds except that the colour of endosperm changed into milky white (Fig. 1e) and the kernel weight (Fig. S2). The cooking and eating quality of YSZwx1 rice was soft but not sticky, while the QLDwx1 rice was very sticky (Fig. 1f), and both their gel consistency was a signi cant increase (Fig. 1g).
This study observed that the AAC signi cantly decreased to 1.16% and 2.36% (Fig. 2a) (reduced by 91.01% and 89.80%), leading to increased gel consistency (Fig. 2b). The other qualities were also experimented. For example, the total protein was increased 26.01% and 6.37%, respectively (Fig. 2c), but the total starch showed no regularity, increasing by 8.11% and decreasing by 6.19% compared to WT (Fig.   2d). The above result suggests that the other rice quality would be signi cantly changed by editing the Wx gene, not only the AAC and wx mutant does not always have the same quality trend. Breeders can cultivate the different qualities of waxy rice by this method. Differential scanning calorimetry (DSC) is widely used to study grain gelatinization temperature (GT), and the GT is usually represented by peak temperature (T p ) (Sasaki et al. 2000). The GT of rice is generally divided into three types: low (<70℃), intermediate (70-74℃) or high (>74℃) (Kongseree and Juliano 1972). In this experiment, the wx mutants showed a slightly higher GT range (To, Tp and Tc) than the WT (Fig. 2e). The GT ranges of WT starch was 66.08℃ (YSZwx1) and 73.46℃ (QLDwx1) (onset temperature, To) to 76.03℃ and 82.06℃ (conclusion temperature, Tc) with an enthalpy (ΔH) of 11.6 J/g and12.55 J/g (Table S4). The results suggest that To, Tp, Tc and ΔH from wx mutants were increased, which lead to them being gelatinized later compared with WT, but they still belonged to the same GT types as the WT. Here, we propose a hypothesis that editing the Wx gene does not change the GT, and subsequent experiments were to verify this hypothesis with knocked other rice varieties' Wx gene.

Pasting properties
The wx mutants and WT pasting properties were signi cantly variable (p<0.05) (Fig. 3a). For example, the PKV, HPV, BDV, CPV, SBV, peak time and PT of QLD and QLDwx1 ranged from 2647 to 2775 cp, from 1729 to 1362 cp, from 918 to 1413 cp, from 4080 to 1757 cp, from 1333 to -1018 cp, from 6.38 s to 4.57 s and from 80.7℃ to 81.5℃, respectively (Table S5). The results show that QLDwx1 and YSZwx1 have more waxy rice pasting properties (Fig. 3b).

Chain length distribution of amylopectin
The chain length distributions of amylopectin in wx mutants and WT were distinctly different (Fig. 4a). At chain length degree of polymerization (DP) 7-20, wx mutants increases compared to WT, while at DP > 20, wx was decreased (Fig. 4b).According to the value of ΣDP≤10/ΣDP≤24 (amylopectin chain ratio, ACR value), amylopectin structure from cultivated rice can be classi ed into three types: L-type (ACR≤0.200), S-type (ACR≥0.240) and M-type (ACR 0.201-0.239) . Interestingly, even although the ACR value of the wx mutants were signi cantly lower than that of WT, they still belonged to the same amylopectin structure type. The amylopectin structure QLDwx1 and QLD belonged to the L-type, and YSZwx1 and YSZ belonged to the M-type (Fig. 4c).

Data Analysis
We applied the same method to three other rice varieties with the same material treatment to nd a potential link between wx mutants and WT (Fig. S3 and Table S6). The other three rice varieties belonged to the Wx b genotype (Table S6). The results showed that WT (Fig. 5a) with higher AAC and the AAC of corresponding its wx mutant (Fig. 5b) was also higher, and some wx mutants AAC were more than 2%, which means they may not belong to high-quality waxy rice. The same results were observed that the Tc and ΔH of wx mutants were increased, however, their GT types were not changed (Table 1). Similarly, a result showed the ACR value was altered but not beyond speci ed value such as M-type (ACR 0.201-0.239), therefore, wx and WT have the same amylopectin structure type (Fig. 5b). Further data analysis showed a direct linear correlation between ACR and GT (Fig. 5c), and the ACR value depended on GT.
Here, there are two new types be proposed to identify the structure of amylopectin, respectively, which were HGT-type (high gelatinization temperature type, ACR < 0.18) and LGE-type (low gelatinization temperature type, ACR > 0.18). Based on these ndings, we demonstrated that GT type and ACR type would not be changed by editing the Wx gene and if WT with a higher AAC, its wx mutant would be endowed a higher AAC.
In this study, we also separately analyzed the correlation between the related physicochemical properties of WT or wx mutants. Pearson's correlation analysis suggested that WT and wx mutants had dissimilar correlations between rice quality, amylopectin structure and physicochemical properties (Fig. 6). The data of WT re ected that GT was negatively correlated with DP6-12 and ACR and positively correlated with DP13-24. The AAC was positively correlated with peak time, CPV, HPV, SBV, and was negatively correlated with BDV and GC. However, the wx mutantsdata re ected that GT was negatively correlated with DP6-12, DP6-24 and ACR and positively correlated with DP13-24, peak time, CL, and DP25-100. The AAC was negatively correlated with GC and positively with peak time HPV and CPV.
The above results implied that amylopectin structure and AAC in diverse rice varieties determines different physicochemical properties. So, the quality data of wild and waxy varieties added together in statistics analyzing may not be an excellent way to nd the relationship between the physicochemical properties. Here, we suggest that if there are accurately diagnoses of the physicochemical properties of waxy rice, it is best to use glutinous rice data only.

Suggestion of speci c quality breeding of waxy rice
It is generally believed that the AC of waxy rice should be less than 2% (Juliano 1992). We recommend that if you want to obtain high-quality waxy rice, you should rst choose rice varieties with low AAC, otherwise, these wx varieties may not be called waxy rice. On the other hand, GT also can be selected before waxy rice breeding, breeders only have to choose a rice variety with the targeted GT. Finally, any new waxy rice variety breeders want can be obtained based on the ACR value, AAC and agronomic traits.

Discussion
New waxy rice varieties have been developed in the past few decades by hybrid breeding and mutagenesis breeding. However, these methods have the disadvantages of unclear quality, time consumption and laboriousness (Deng 1992, Olsen and Purugganan 2002, Toda 1980 (Han et al. 2018, Zeng et al. 2020. Unfortunately, these studies did not address how to obtain waxy rice with directional properties. Our study found that the quality of wx mutantcan be predicted and has similar properties inextricably linked to WT. Previous studies have shown that other starch-synthesizing enzymes also in uence rice AAC. That different alleles have different effects, such as soluble starch synthases (SSs), branching enzymes (BEs), debranching enzymes (DBEs) and isoamylase 1 (ISAI) (Man et al. 2013, Shufen et al. 2019, Umemoto et al. 1995, Zhu et al. 2020). The AAC of rice is different, even in the same Waxy gene type. In this study, WT with higher AAC, its corresponding wx mutant AAC, will be higher than other mutants. It is reported that silencing Wx gene expression regulates genes related to starch synthesis, with upregulation of granulebound starch synthase II (GBSSII) compensating for some amylose in the endosperm (Pérez et al. 2019). The above reasons may rationally explain the different AAC and chain length distributions of amylopectin of the wx mutant after editing the same Waxy gene type. Currently, however, the reasons for these discrepancies are unde ned. This nding suggests that if you want to cultivate a lower AAC wx mutant variety, you should edit a lower WT, but this mechanism needs to be further studied in the future.
In rice starch, the functional properties of amylopectin have been extensively researched. Gelatinization temperature (GT) was negatively correlated with short amylopectin chains (DP 6-12) and the proportion of chains with DP ≤ 10 to those with DP ≤ 24 (the amylopectin chain ratio, ACR value) but positively correlated with amylopectin long chains (DP > 37) (Nakamura 2002, Vandeputte et al. 2003b).However, this study indicated that GT was only negatively correlated with ACR and DP6-12 when all were wild type lines, signi cantly negatively correlated with ACR and DP6-12 and signi cantly positively correlated with DP25-100 when all were waxy rice lines. Based on these ndings that physicochemical properties are analyzed separately in different rice types is necessary, which can obtain accurate results.
Moreover, editing the Waxy gene resulted in similar ACR and GT between WT and wx and that ACR and GT were linearly correlated. We conclude that ACR is the main reason for the difference in GT that editing the Wx gene does not change the ACR type and GT type. Therefore, this makes it possible to cultivate different amylopectin ne structures and GT type waxy rice.

Availability of data and materials
All data supporting the conclusions of this article are provided within the article and its (Additional le 1: Figure S1-S4, Additional le 2: Table S1-7). The rice seeds of the landraces used in the studies are available at the Rice Research Institute of Sichuan Agricultural University, China.

Competing interests
The authors declare that they have no competing interests.   Table S1: Primers used in this study Table S2: Percentage of T0 plants with mutations in the target locus   Table S3: Waxy alleles and mature seed quality of wx mutants and corresponding WT lines    mutant and corresponding their WT: appearance of cooked rice (f), gel consistency (g). Error bars are mean ± SD (n = 3). Signi cant differences were determined by Student's t-test (*P < 0.05, **P < 0.01).  bars are mean ± SD (n = 3). Signi cant differences were determined by Student's t-test (*P < 0.05, **P < 0.01).  represents the ACR value of a rice variety. Black is onset temperature (To), bluish-black is peak temperature (Tp), red is conclusion temperature (Tc). Error bars are mean ± SD (n = 3).

Figure 6
Pearson's correlation analysis of amylopectin structure relationships of WT and wx. '×' means 'no correlation'. The stronger the red color, the more signi cant positive relationship, the stronger bluish black, the more Signi cant negative relationship.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. TableS1S7.docx