Genomics-Assisted Improvement on Blast Resistance and Low Cadmium Accumulation in An Elite Rice Variety ‘Xiangwanxian 13’

1. Introduction

Rice (Oryza sativa L.) is one of the most important crops with food supply for more than half of the world's population [1]. To ensure food security, efforts have still been made for a sustainable increase in rice yield in the last decade [2]. However, high-yield cultivation of rice varieties has caused excessive use of pesticides due to lack of resistance to pests and diseases, which increased input costs and damaged the ecological environment [3,4,5]. In addition, industrial emissions, agro-chemical use, and anthropogenic activities have resulted in excessive heavy metals such as cadmium (Cd) in rice paddies, posing a serious threat to food safety [6,7,8]. Therefore, it is necessary to develop rice varieties with superior resistance and tolerance to major biotic and abiotic stresses.

Rice blast, caused by pathogenic fungus Magnaporthe oryzae (M. oryzae), is one of the major diseases of rice resulting in yield loss from 30% to 50% in large rice planting areas under favorable conditions [3]. Great efforts have been devoted to evaluating and characterizing blast resistance of rice germplasm resources. To date, more than 100 blast resistance (R) genes have been identified and mapped on rice genome, 38 of them have been cloned and characterized [9,10]. Most of the genes are distributed in the gene clusters on chromosomes 6, 11 and 12, and very few of them were introgressed into popular rice varieties [9]. The high variable nature of the pathogen M. oryzae leads to frequent emergence of new virulent races resulting in loss of resistance within 3-5 years of rice cultivation. Among the blast resistance genes, Pi1 is an allele at the Pik locus mapped at the end of chromosome 11, conferring dominant and durable resistance against geographically diverse M. oryzae isolates [11,12]. Likewise, gene Pi2 is an allele at the Piz locus located on chromosome 6 close to the centromere, and also confers dominant and broad-spectrum resistance to various M. oryzae isolates [13,14,15,16]. Currently, several attempts had been made to deploy Pi1 and Pi2 in rice breeding [17,18,19,20], along with enhanced resistance to rice blast under high pressure in epidemic fields.

Among abiotic stresses, Cd contamination of soil poses a serious threat to sustainable agriculture and food safety. As the main source of Cd intake in human dietary [21], rice generally poses a higher risk of Cd exposure compared to other crops [22,23]. To reduce the effects of Cd toxicity on human health, a maximum Cd concentration of 0.4 mg/kg in polished rice is allowed internationally, and even a stricter limit of 0.2 mg/kg was established in China [24,25]. However, cases of excessive Cd accumulation in rice grain have occurred frequently in the past decade [26,27,28,29]. Some measures have been made to reduce the risk of excessive Cd accumulation in rice, such as soil-Cd remediation, fertilizer control, irrigation management, and growing low-Cd-accumulating varieties [25]. Studies have shown that genetic improvement and development of low-Cd rice varieties are cost-effective in agricultural practice [22,30]. Currently, the physiological and genetic mechanisms of Cd accumulation in rice have been well studied, some quantitative trait loci (QTL) and genes associated with variation in grain Cd accumulation have been identified [31,32], and four Cd-QTL hotspots in genomic regions have been mapped on chromosomes 2, 3, 7, and 8 [33,34,35,36,37,38,39,40,41,42,43,44,45,46,47]. With a series of the associated QTLs and genes being found, genetic improvements in rice specifically targeting at reducing grain Cd accumulation have become a top priority for rice breeders. Among the known cases, the natural resistance associated macrophage protein gene OsNramp5, located on Chromosome 7 and encoding a key transporter for Cd uptake in rice, is an ideal target gene to be manipulated to reduce Cd uptake [48,49,50,51]. Additionally, OsHMA3, encoding the P-type heavy metal ATPase, is another target gene that has been manipulated to sequester Cd into vacuoles of root cells, which subsequently reduced Cd transportation [52,53]. The use of OsNramp5 and OsHMA3 to generate low-Cd rice varieties were more effective when compared with genetic modification of genes such as OsNramp1, OsZIP5, OsZIP9, OsZIP7, and OsLCT1 [47].

Molecular breeding can not only improve target traits of recipient materials, but also develop new varieties with excellent performance by pyramiding different traits, which is more accurate and efficient than traditional breeding. Marker-assisted selection (MAS) is regarded as a highly efficient breeding method because of the rapid and precise selection of the desired genes [54]. Application of MAS has significantly improved the rice breeding process, although with some disadvantages such as linkage drag and unknown genetic background effects [19]. Furthermore, genomic technology has been now widely used for improving breeding efficiency by foreground and background selection with MAS [55]. Based on large numbers of informative markers, foreground selection chooses individuals carrying the target genes [56], while background selection is used to improve the genomic recovery of the recipient parent [57]. The kompetitive allele specific PCR (KASP) markers have considerably enhanced crop improvement and variety development with more efficient and precise genotyping [58]. In addition, single nucleotide polymorphism (SNP) arrays based on the Affymetrix and Illumina technologies have been developed and used for background selection applied in rice [59], maize [60], and other crops [61,62,63]. Recently, new SNP panels have also been developed by using Genotyping by Pinpoint Sequencing of multiplex PCR products (mGPS) and Genotyping by Pinpoint Sequencing of captured targets (cGPS) technologies (https://www.higentec.com). To date, a series of rice SNP arrays with different marker numbers and genomic coverage have been developed and used for MAS [64], such as the 44K SNP chip, RICE6K array, 50K SNP chip, Rice700K SNP array, Cornell 7K Array Infinium Rice (C7AIR), Rice3K56 SNP array, and so on [65,66,67,68,69,70]. Therefore, all these advancements in genomic technologies can provide more powerful selection of plant traits, and facilitate comprehending and improving various crops.

Xiangwanxian 13 (XWX13) is a fragrant indica rice variety with a late-maturing characteristic in China. It is highly welcomed by farmers due to the high yield and good grain quality, and has been commercially cultivated over 600,000 hectares in total. However, XWX13 is highly susceptible to rice blast, and the grain Cd concentration in high pollution paddy fields still exceeded the Cd-safe standard of 0.2 mg/kg though it was ever an emergency variety for low Cd [71], indicating a production risk and limit for large-scale commercial utilization. In the present study, an efficient genomic breeding strategy was adopted for enhancing blast resistance and lowering grain Cd accumulation in XWX13. Four major genes, namely Pi1, Pi2, OsHMA3, and OsNramp5, were precisely introduced into XWX13 from three donor parents with aid of genomics-assisted selection. Our results provide a successful breeding practice in genomics-assisted rapid and directional improvement of XWX13 for rice blast resistance and low Cd accumulation.

2. Materials and Methods

2.1. Plant Materials

Rice plant materials used in this study include the original parent Xiangwanxian 13 (XWX13, a high-quality indica rice cultivar with fragrance), blast resistance gene donor HZ02411 (Pi1+Pi2), low-Cd accumulation gene donors HZ02415 (OsHMA3) and HZ02416 (OsNramp5). The rice variety CO39 was used as susceptible control, while Gumei4 was used as resistant control.

2.2. Population Development and Breeding Selection Procedure

The improved XWX13 lines (iXWX13-1 and iXWX13-2) were developed by marker-assisted backcrossing as described in Figure 1. The late-maturing indica rice cultivar XWX13 was crossed with each of three different donors, and the foreground SNP markers were used to genotype XWX13, the corresponding donors, and the F₁ plants. The F₁ plants were then backcrossed with the recurrent parent XWX13. The BC₁F₁ plants were genotyped for foreground and background selections, and three plants with the target genes and the highest background recovery rate were selected to further backcross with XWX13. The same genotyping and selections were done for the BC₂F₁ and BC₃F₁ populations. To pyramid genes for different traits, the BC₃F₁ plants with the target genes and the highest background recovery rate were used for multiple hybridization, and double-cross F₁ population was generated. Based on continuous self-pollination, foreground and background selections, and field phenotypic screening, two improved lines iXWX13-1 (Pi1+Pi2+OsHMA3) and iXWX13-2 (Pi1+Pi2+OsNramp5) were successfully developed.

2.3. Marker-Assisted Selection and Background Analysis

SNP markers tightly linked to the genes, Pi1, Pi2, OsHMA3, and OsNramp5, were used to track the target genes in foreground selection. And 120 polymorphic SNP markers evenly distributed on 12 chromosomes were used for background selection. KASP genotyping was performed as described by He et al. [20].

The whole-genome SNP array RICE 1K mGPS was further used for detecting the genetic background between the recurrent parent XWX13 and the improved lines, iXWX13-1 and iXWX13-2. The SNP array RICE 1K mGPS was developed by Huazhi Biotechnology Co., Ltd., and comprises 5,456 loci in 1,048 intervals uniformly distributed on the rice genome. According to the similar methods previously reported [20], SNP array genotyping was performed and the background recovery rate was calculated.

2.4. Evaluation of Rice Blast Resistance

XWX13 and the improved lines were tested for resistance spectrum at the seedling stage during 2022-2023 in Huazhi Biotechnology Co., Ltd., Changsha, China. The 39 M. oryzae isolates were used for spraying inoculation in the greenhouse according to the method described by Liu et al. [72]. Disease reaction was assessed one week after inoculation following a standard 0-9 rating evaluation system [73,74]. The rice varieties CO39 and Gumei4 were used as the susceptible and resistant control, respectively. Disease evaluation was performed in three replications. Resistance frequency was defined as ratio of resistance-showing M. oryzae isolates to the total number of inoculated M. oryzae isolates on the rice genotype (line), and a higher frequency represents a broader resistance spectrum.

Leaf and panicle neck blast resistance of XWX13 and the improved lines were evaluated in 2022 in two natural blast nurseries, Dawei Mountain in Liuyang City and Taojiang in Yiyang City, Hunan, China. The blast resistance was assessed according to the evaluation system mentioned above [73,74].

2.5. Evaluation of Yield, Main Agronomic Traits and Grain Quality

XWX13 and the improved lines were planted in multi-site trials in 2022. Five test sites were set up, namely Changsha and Yueyang in Hunan, Yichun and Shangrao in Jiangxi, and Hezhou in Guangxi. In the field layout, each variety or line was planted in a plot of 13.34 m². The field cultivation management was carried out according to the method of China National Rice Variety Regional Trial. The heading date was recorded for each plot. At maturity stage, ten individual plants in the middle of the central row in each plot were sampled for measurements of main agronomic traits including plant height (PH, cm), panicle length (PL, cm), spikelets per panicle (SPP), filled-grain percentage (FGP, %), and 1000-grain weight (GW, g). The plot yield was measured for each variety or line at each site.

Bulked harvested seeds from each plot were used for grain quality analysis, including brown rice percentage (BRP, %), milled rice percentage (MRP, %), head rice percentage (HRP, %), chalky rice percentage (CRP, %), chalkiness degree (CD, %), rice grain length (RGL, mm), grain length/width ratio (L/W), alkali spreading value (ASV), amylose content (AC, %), gel consistency (GC, mm).

The single factor ANOVA in the software SPSS Statistics 19 was used to test the significance of variation for yield, main agronomic traits and grain quality.

2.6. Determination of the Grain Cd Concentration

The grain Cd concentration of XWX13 and the improved lines in multi-site trials were determined by using NX-100FA food heavy metal detector (NCS Testing Technology Co., Ltd., Suzhou, Jiangsu, China). Bulked harvested seeds from each plot were randomly sampled three times, and measured for the grain Cd concentration following the instrument manual. The values of grain Cd concentration given are mean of three replications. Differences of the grain Cd concentration between XWX13 and the improved lines were statistically tested by the single factor ANOVA in SPSS Statistics 19.

3. Results

3.1. Development of the Improved Lines

Following the breeding procedure shown in Figure 1, the four resistance genes were respectively transferred into the XWX13 background from HZ02411 (Pi1+Pi2), HZ02415 (OsHMA3), and HZ02416 (OsNramp5). Three single-cross BC₃F₁ populations were developed by marker-assisted foreground and background selections, and then two double-cross F₁ populations were generated by multiple hybridization. After four generations of selfing screening and identification, two improved lines iXWX13-1 (Pi1+Pi2+OsHMA3) and iXWX13-2 (Pi1+Pi2+OsNramp5) were obtained.

The two improved lines, iXWX13-1 and iXWX13-2, and the original recipient parent XWX13 were further genotyped for genetic background analysis using SNP array RICE 1K mGPS. It was found that the genetic background recovery rates of iXWX13-1 and iXWX13-2 were 94.44% and 94.63%, indicating a high genomic similarity with XWX13 (Figure 2 and Figure 3).

3.2. Blast Resistance of the Improved Lines

The two improved lines, iXWX13-1 and iXWX13-2, the recurrent parent (XWX13), CO39 (susceptible control), and Gumei4 (resistant control) were scored at the seedling stage by spraying inoculation with 39 M. oryzae isolates in the greenhouse (Table 1), and the typical symptoms of two strains on five varieties or lines were shown in Figure 4. The two improved lines, iXWX13-1 and iXWX13-2, and the resistant control Gumei4 exhibited broad-spectrum resistance to rice blast, with a high resistance frequency of 97.44%, 100%, and 92.31%, respectively, whereas the recipient variety XWX13 and the susceptible control CO39 showed a low resistance frequency of 17.95% (Table 1). The results verified that the enhanced blast resistance is occurring in the two improved lines (possessing Pi1+Pi2).

The increased resistance of the two improved lines (iXWX13-1and iXWX13-2) were also observed under two natural blast nurseries in comparison to the recurrent parent (XWX13). In detail, XWX13 showed 7-8 level of leaf blast (S-HS) and 15-18% of panicle neck blast infection rate (MS) in the blast nursery of Taojiang County in 2022, and thus was susceptible to rice blast (Table 2, Figure 5). While the two improved lines (possessing Pi1+Pi2) exhibited a higher level resistance reaction, both the iXWX13-1 and iXWX13-2 showed grade 2 of leaf blast (R) and 4-5% of panicle neck blast infection rate (R). In the same year, iXWX13-1 and iXWX13-2 demonstrated resistance to leaf blast (0-1 grade) in the blast nursery of Dawei Mountain, whereas XWX13 showed grade 7 of leaf blast (S) (Table 2). In short, the resistance to leaf and panicle neck blast under natural conditions were also effectively enhanced in the two improved lines (possessing Pi1+Pi2).

3.3. Yield, Agronomic Traits and Grain Quality of the Improved Lines

To test whether the traits of the two improved lines，iXWX13-1 and iXWX13-2， were identical to that of the original recipient parent XWX13, the main agronomic traits of the lines were observed in a multi-site trial in 2022. Minor differences were found for a few traits, including days to maturity (DTM), plant height (PH) and 1000-grain weight (GW) in overall. For example, the DTM was slightly longer (123 d vs. 120.4 d) for the two lines than XWX13, and the GW of iXWX13-1and iXWX13-2 were subtle higher (0.6-0.8 g) than that of XWX13 together with a decrease in plant height (2-3 cm) of iXWX13-1and iXWX13-2 (Table 3). In general, no significant difference was observed between the two improved lines and XWX13 for most of the agronomic traits. These results indicated that the main agronomic traits of the improved lines are highly similar to that of XWX13.

The grain yield of these tested lines in the multi-site trial was investigated. The average yields of XWX13, iXWX13-1, and iXWX13-2 from five test sites were 7.50 t/ha, 7.37 t/ha, and 7.51 t/ha, respectively (Table 3). Non-significant differences were found between XWX13 and the improved lines. In term of each test site, the yield of iXWX13-2 was slightly higher than that of iXWX13-1, but the difference was not statistically significant as indicated by the single factor ANOVA test (Table 3).

To profile grain quality, rice grains from the five test sites were investigated for ten important rice quality traits with significant difference only in grain shape (Table 4). For iXWX13-1 and iXWX13-2, grains were shorter (0.2-0.3 mm) than that of XWX13, thus resulting in relatively round grains. The other eight rice quality traits, BRP, MRP, HRP, CRP, CD, ASV, AC, and GC, showed statistically non-significant differences between XWX13 and the improved lines, revealing highly similar grain quality.

3.4. Grain Cd Accumulation in the Improved Lines

To evaluate the grain Cd content under Cd-polluted conditions, the grain Cd concentration of all the lines in the multi-site trial were examined and showed that Cd was hardly accumulated into the grains of iXWX13-2. Across five test sites, the grain Cd concentration in iXWX13-2 ranged from 0.009 to 0.077 mg/kg, significantly lower than that in XWX13 (0.618-1.052 mg/kg) and iXWX13-1 (0.569-0.864 mg/kg) (Figure 6). Compared with the original recipient parent XWX13, the grain Cd concentration in iXWX13-2 decreased by 90.98-98.87%. Regarding to grain Cd accumulation between XWX13 and iXWX13-1, significant differences were only detected in two sites, Shangrao and Hezhou, where the grain Cd concentration in iXWX13-1 were both relatively lower than that in XWX13 (Figure 6).

Of all tested samples, only the grain Cd concentration in iXWX13-2 (possessing OsNramp5) in each plot was obviously lower than the limit value of 0.2 mg/kg in China National Standard for Food Safety (GB 2762-2022), revealing that iXWX13-2 is a stable Cd-safe improved line and superior to iXWX13-1 (possessing OsHMA3). Therefore, iXWX13-2 (stacking Pi1+Pi2+OsNramp5) demonstrated the enhanced blast resistance and low Cd accumulation without penalty of the yield, main agronomic traits and grain quality.

4. Discussion

Improvement Achieved by Genomic Marker-Assisted Selection

The phenotype-based selection in conventional breeding has played a great role in developing elite rice varieties in the past decades, although with a long breeding cycle and low efficiency [75,76]. Alternatively, genomic marker-assisted breeding is an innovative approach that utilizes modern molecular tools and genomic information to improve the accuracy and efficiency of plant breeding [58], and it has been successfully applied to improve biotic or abiotic stresses tolerance in rice varieties [20,47,77,78,79,80]. Also high-throughput sequencing and genotyping platforms have accelerated the process of genomic marker-assisted breeding [58]. In this study, an efficient genomic breeding strategy for the integration of blast resistance and low Cd accumulation was adopted by the genomic marker-assisted foreground and background selection. Four major genes, namely Pi1, Pi2, OsHMA3, and OsNramp5 from three donors, were precisely introduced into the genetic background of a good quality fragrant rice variety XWX13 by using genomic marker-assisted selection. Eventually two preferable improved lines iXWX13-1 (Pi1+Pi2+OsHMA3) and iXWX13-2 (Pi1+Pi2+OsNramp5) were obtained with XWX13 background recovery rates of 94.44% and 94.63%. Compared to the traditional 6-8 year breeding cycle, our strategy shortens the cycle up to 2-3 years as reported in the previous studies [19,20]. Therefore, this study demonstrated a successful breeding practice in genomics-assisted rapid improvement of XWX13 for blast resistance and low Cd accumulation.

Pyramiding Genes to Improve Blast Resistance and Cd Accumulation

Blast resistance is easily lost after a few years of cultivation due to variation and evolution of pathogen populations. For instance, Tang et al. [81] found elite indica rice varieties from South China with Pi1 were not resistant to blast, while those with Pi2 exhibited 83.3% resistance frequencies. Similarly, it was found that multi-gene pyramiding lines showed significantly higher resistance levels than monogenic lines [82]. So there is an urgent need for pyramiding multiple genes in breeding programs [83,84], and pyramiding of multiple resistance genes has proven to be an effective way for developing rice varieties with durable resistance to M. oryzae [20,55]. As found in our study, the two improved lines iXWX13-1 and iXWX13-2 carrying Pi1+Pi2 exhibited more than 97.44% blast resistance frequency under artificial inoculation at the seedling stage, indicating an obviously increased resistance spectrum compared with that of the original recipient variety XWX13. Furthermore, the two improved lines iXWX13-1 and iXWX13-2 showed a grade 2 of leaf blast (R) and 4-5% of panicle neck blast infection rate (R) in blast nursery in Taojiang County, and thus possessed the enhanced resistance to leaf and panicle neck blast. The increased resistance to leaf blast were also observed in blast nursery in Dawei Mountain. These results revealed that it is feasible to pyramid multiple broad-spectrum R genes into elite rice varieties, enhancing their resistance and durability against blast disease.

The grain Cd accumulation poses serious safety concern for the rice industry. To date, various approaches, including water management and administration of contaminated soil additives, have been used to alleviate Cd accumulation in rice grain [85,86]. Correspondingly, some low-Cd rice varieties have been identified and developed through mass selection, marker-assisted selection, and genetic manipulation [39,47]. Hu et al. [80] confirmed that gene OsNramp5 was the most efficient in lowering Cd accumulation, and first generated low Cd accumulation restorer (R) lines by editing OsNramp5, OsLCD, and OsLCT1 in both the japonica and indica rice. In the present study, two improved lines iXWX13-1 (Pi1+Pi2+OsHMA3) and iXWX13-2 (Pi1+Pi2+OsNramp5) were obtained by genomic marker-assisted selection. The evaluation of Cd accumulation in a multi-site trial showed that the grain Cd concentration in iXWX13-2 carrying OsNramp5 decreased by 90.98-98.87% compared with that in the original recipient variety XWX13, which was far below the threshold value of 0.2 mg/kg in China (GB 2762-2022), demonstrating a non-transgenic Cd-safe improved line, and also reconfirmed the stable efficiency of OsNramp5 in lowering Cd concentration [47,80]. Here, statistically significant differences of the grain Cd concentration between XWX13 and iXWX13-1 were observed in two test sites (Shangrao and Hezhou), but not in the other three sites (Changsha, Yueyang, and Yichun), and both of them exceeded the safe threshold value of 0.2 mg/kg, indicating that the improved line iXWX13-1 carrying OsHMA3 had no obvious effect on lowering Cd accumulation. The gene OsHMA3 is not as effective in lowering Cd accumulation as previously reported [47,87,88]. As a typical quantitative trait, the variations of grain Cd accumulation in rice are determined by both the genotypes and the environmental factors that determine Cd phyto-availability in soil [28,89,90], which make it difficult to accurately evaluate the genetic effects of grain-Cd variation [91]. It is usually accepted that there is natural variation in Cd accumulation among different rice varieties [92]. Previous studies have indicated that the indica rice varieties generally accumulate more Cd comparing to the japonica rice varieties [93]. However, a consistent conclusion has not been obtained by comparing Cd accumulation in the same genetic background with different Cd-related genes. In our study, the two improved lines iXWX13-2 and iXWX13-1 have the similar genomic background of XWX13 (Figure 2 and Figure 3), and thus OsNramp5 is obviously more effective in lowering grain Cd accumulation than OsHMA3 in the same genetic background (Figure 6). The performance of the two genes in other genetic backgrounds needs to be investigated in the upcoming further studies.

Improvement of Traits Without Penalty of the Yield, Main Agronomic Traits and Grain Quality

Plants face the constant challenge of allocating limited resources to immunity and growth to ensure survival and reproduction, therefore, one major aim of crop breeding is to balance biotic resistance and abiotic tolerance with yield [94]. The genomics-assisted breeding strategy can provide more powerful selection of quantitative traits, and modify tradeoffs effectively. In this study, the characterization of the two improved lines iXWX13-1 and iXWX13-2 showed that their yield and main agronomic traits were almost identical to those of the recurrent parent XWX13, and the average values for each trait in a multi-site trail were not significantly different. In term of rice quality, eight traits (brown rice percentage, milled rice percentage, head rice percentage, chalky rice percentage, chalkiness degree, alkali spreading value, amylose content, and gel consistency) showed no statistically significant difference between XWX13 and the improved lines. In view of the potential tradeoff between resistance and yield, multiple traits should be paid attentions when screening the target plants selected based on the genomic marker-assisted foreground and background selections [95]. The successful cases for how to select superior recombinants in the breeding programs have been previously reported [79,96]. In general, when the donor line is a wild relative of the crop, linkage drag is a very common problem which can strongly limit the use of a particular gene in breeding [97]. Our study also provide an optimized strategy for overcoming linkage drag in crop breeding by using commercial variety harboring favorable genes as an improved donor instead of the wild donor, which would help in further improving the background of the pyramided lines and shorting the breeding cycle.

5. Conclusions

Two improved good-quality rice lines iXWX13-1 (Pi1+Pi2+OsHMA3) and iXWX13-2 (Pi1+Pi2+OsNramp5) were developed by using genomic breeding strategy, which is composed of genomics-assisted foreground selection and background selection, and phenotypic screening. The improved line iXWX13-1 can be used as a blast-resistant version, and the improved line iXWX13-2 can be used as a blast-resistant and low Cd accumulation version to replace XWX13. The present study is the first successful practice to apply genomic marker-assisted backcrossing and background selection to improve the blast resistance and to lower the grain Cd accumulation in XWX13 without penalty of the yield, main agronomic traits and grain quality. This proposed strategy would pave the way to integration of multiple desirable genes into the background of elite varieties in crop genetic improvement.