Advances and Challenges in the Breeding of Salt-Tolerant Rice

Soil salinization and a degraded ecological environment are challenging agricultural productivity and food security. Rice (Oryza sativa), the staple food of much of the world’s population, is categorized as a salt-susceptible crop. Improving the salt tolerance of rice would increase the potential of saline-alkali land and ensure food security. Salt tolerance is a complex quantitative trait. Biotechnological efforts to improve the salt tolerance of rice hinge on a detailed understanding of the molecular mechanisms underlying salt stress tolerance. In this review, we summarize progress in the breeding of salt-tolerant rice and in the mapping and cloning of genes and quantitative trait loci (QTLs) associated with salt tolerance in rice. Furthermore, we describe biotechnological tools that can be used to cultivate salt-tolerant rice, providing a reference for efforts aimed at rapidly and precisely cultivating salt-tolerance rice varieties.


Introduction
Plants grow in dynamic environments and frequently experience various abiotic stresses, such as drought, high salinity, cold, and heat [1]. Salt stress is one of the most severe environmental stresses. The effects of salt stress on plants include osmotic stress, ionic toxicity, and nutritional deficiencies and eventually lead to growth inhibition and crop yield losses [2,3]. Soil salinization is mainly caused by poor-quality drainage and irrigation systems, climate change (which leads to sea level rise), and drought [4]. Soil salinity is a global problem that affects more than 20% of cultivated land, including half of all irrigated areas, and this percentage is expected to increase [5]. Therefore, improving the salt tolerance of crops would not only lead to the effective use of saline-alkali land, but also support sustainable agriculture and alleviate the world food crisis.
Rice (Oryza sativa) is an important staple food crop worldwide. As the global population continues to rise, rice production also needs to increase. However, global rice production is threatened by climate change [6]. Rice is considered to be a salt-susceptible species [7], and its salt tolerance depends on growth stage, organ type, and genotype [8][9][10]. Generally, the seedling and reproductive stages are more susceptible to salinity than the vegetative stage, roots are more sensitive than other organs [8], and japonica rice is more sensitive than indica rice [9]. Salinity stress suppresses photosynthesis and growth, leading to biomass loss, as well as partial sterility, which ultimately results in reductions in rice yield [11,12]. Therefore, the breeding of salt-tolerant rice cultivars is considered to be one of the most economic options to assure food security.
As a semi-aquatic plant, rice lives in a water-saturated environment during most of its life cycle. This environment has led to many distinct adaptations rice-for example, the ethylene response phenotype of rice is different from that of other species [13]. In Arabidopsis thaliana, ethylene plays a also causes physiological and biochemical changes, such as the inhibition of photosynthesis; decreased water content; altered metabolism; increased Na transport to the shoot; and decreased K, Zn, and P uptake [12,25,26].
Since rice is a staple food, grain quality is an important driver of marketability. Rice grain is composed mainly of carbohydrates, predominantly starches. The determinants of rice grain quality are grain texture, taste, and visual attributes, which are further determined by the composition and structure of the starch molecules [27]. The effect of salt stress on rice grain starch depends on the salt concentration and the rice genotype. Specifically, salinity treatment leads to a decrease in starch content when the salt concentration is higher than 5-dS/m 2 electrical conductivity, in both salt-tolerant and salt-susceptible cultivars of rice [28,29]. However, grain starch increased in the Nipponbar cultivar when low concentrations of salt (2 or 4-dS/m 2 electrical conductivity) were applied at the anther appearance and seedling stages [30], suggesting that the response of rice to salt stress is complex and highly context-dependent.
Another indicator of rice grain quality is the nutritional value of the grain. Rice grain contains a variety of minerals, such as Ca, Mg, and P, and some trace elements, such as Fe, Cu, Zn, and Mn. Rice also contains the vitamins like thiamine, riboflavin, and niacin [25]. Several studies have shown that the absorption and uptake of micro-and macro-mineral nutrients are altered under salinity stress [31,32]. Saleethong et al. measured the macronutrients and micronutrients in grains of Pokkali (a salt-tolerant variety) and KDML105 (a salt-sensitive variety) under saline conditions and found that salt stress resulted in significant reductions in macronutrient elements but an increase in Ca in brown rice grains of both cultivars. Moreover, the amounts of Mn, Cu, and Zn were higher in Pokkali than in KDML105 [31]. Verma and Neue reported that the contents of Na, Fe, and Zn increased, while those of P and Mn decreased in rice grain with increased salinity, but the contents of N, Mg, Cu, K, and Ca were not affected in the varieties used in this study [32]. In addition to affecting the mineral content, salt stress can increase the rice grain protein content and amino acid levels [25]. Although little is known about the effects of salt stress on vitamins in rice grains, salinity caused a significant reduction in vitamin contents in wheat grains [33]. In summary, these studies show that the effect of salt stress on rice grain quality is multifaceted and depends on the salt concentration and the rice variety.

Breeding of Salt-Tolerant Rice Varieties
Developing elite salt-tolerant rice varieties is considered to be the most economically viable and environmentally friendly method to effectively use saline-alkali land. As a food crop, yield is an important indicator to evaluate the merits of rice varieties. Salt-alkali tolerance in rice is defined as the ability to grow on land with a 0.3% saline-alkali concentration, with a yield of more than 4500 kg per hectare. Through years of hard work, breeders have sought out, collected, evaluated, and developed many salt-tolerant rice resources.
Sea Rice 86 (SR86) is a new cultivar domesticated from a wild strain of rice that was first found in 1986 in saline-alkaline soil submerged in sea water near the coastal region of the city of Zhanjiang in Southeast China [46]. SR86 showed a significantly higher ability to cope with high salinity than a highly salt-resistant rice variety, Yanfen 47, measured by both germination and salt inhibition rates [46]. After more than 20 years of breeding and selection, SR86 retains an extraordinary tolerance to salinity and is considered to be a strategic germplasm resource for the development of new rice varieties. SR86 is being used to investigate the mechanism of salt tolerance and effective breeding strategies.
In recent years, our laboratory has also carried out a breeding program to develop salt-tolerant rice. We collected more than 750 rice accessions, including 500 rice accessions from all over the world and 250 salt-tolerant rice varieties from domestic coastal cities, such as Tianjin, Liaoning, Shandong, and Jiangsu. We investigated the salt tolerance of these varieties in Dongying, Shandong Province (37 • 31 29" N 118 • 33 57" E), a typical saline-alkali field in China. Briefly, thirty-day-old seedlings were transplanted to a normal field or saline field (0.35% NaCl, pH 8.2) at a spacing of 20 cm × 15 cm, and all agronomic traits were performed at the maturity stage. Through comparing the agronomic traits of the normal field and saline field, we selected a series of varieties with excellent agronomic traits and salt tolerance (Table 1) (unpublished data). Using these varieties, we made more than 300 hybrid combinations, generated more than 100 salt-tolerant genetic populations, and identified more than 1000 high-yielding salt-tolerant recombinant inbred lines. "Control" indicates that the variety was grown in a normal field. "Salt" indicates that the variety was grown in a field containing 0.35% NaCl and pH 8.2 throughout its life cycle. "DYST" means Dongying Salt Tolerance.

Mapping and Cloning of Salt-Tolerant Genes/QTLs
Salt stress has two main stress effects on rice: osmotic stress and ionic stress [2,3]. Osmotic stress reduces the water uptake by roots and causes internal dehydration. It also leads to the excessive accumulation of reactive oxygen species (ROS), which damages various cellular components and macromolecules and, eventually, leads to plant death [47]. Ionic stress is caused by the excessive accumulation of Na + and Cl − in metabolically active intracellular compartments. High intracellular concentrations of Na + inhibit the uptake of other ions, which can disrupt the metabolism and potentially kill the plant [2,3]. Thus, enhancing the ability of rice to minimize and adjust to osmotic and ionic stresses is an effective way to improve the salt tolerance of rice.
Salinity tolerance in rice is a polygenic trait controlled by QTLs [17]. Using mapping populations derived from crosses between salt-sensitive varieties and salt-tolerant varieties, researchers have identified a large number of QTLs (Table 2). Only a few major salt tolerance QTLs or genes have been identified by genomic methods. For example, qSKC-1 and qSNC-7 are involved in regulating K + /Na + homeostasis under salt stress and explain 48.5% and 40.1%, respectively, of the total phenotypic variance [48]. Isolation of the qSKC-1 gene by map-based cloning revealed that it encodes a member of the HKT-type transporter family [49]. The QTL Saltol, which acts to maintain shoot Na + /K + homeostasis in the salt-tolerant cultivar Pokkali, explained 43% of the variation in the seedling shoot Na + /K + ratio in a recombinant inbred line (RIL) population derived from a cross between indica varieties IR29 and Pokkali [50]. qSE3, which encodes a K + transporter gene, OsHAK21, promotes seed germination and seedling establishment under salinity stress in rice [51]. qST1 and qST3, which are located on chromosomes 1 and 3, respectively, conferred salt tolerance at the young seedling stage and explained 36.9% of the total phenotypic variance in the RIL population derived from a cross between Milyang 23 and Gihobyeo [52]. qST1.1, which plays a key role in salt tolerance in SR86, explained 62.6% of the phenotypic variance [53]. The identification of QTLs is only one step toward determining the genes associated with salinity tolerance; further research is needed to discover the major salt-tolerance genes and investigate their regulatory mechanisms in rice.  [72]. The overexpression of APX genes in rice reduces ROS levels and enhances salt tolerance [73][74][75]. The knockdown of OsVTC1-1 and OsVTC1-3, which are involved in ascorbate synthesis, increases the accumulation of ROS and decreases the salt resistance of rice [76][77][78]. Genes involved in Na + /K + homeostasis are also candidates for improving salt tolerance in rice. High-salinity stress results in altered K + /Na + ratios, which lead to metabolic changes in the plant [79]. Na + causes growth inhibition via Na + toxicity, whereas K + is essential for plant growth and development. Therefore, restricting the intracellular accumulation of toxic sodium (Na + ) is beneficial for survival under salt stress [2,3]. Members of the high-affinity K + transporter (HKT) family in rice function as Na + and Na + /K + transporters in controlling Na + accumulation, and enhancing or inhibiting the expression of HKT genes changes the intracellular Na + content and the salt tolerance of rice [80,81]. Thus, modifying ion channels is an option to improve the salt tolerance of rice.
In addition to genes related to ROS and Na + /K + homeostasis, many other genes influence the salt tolerance of rice [16]. For example, rice expansin 7 (OsEXPA7), which encodes a cell wall-loosening protein, positively regulates salt tolerance in rice by coordinating sodium transport, ROS scavenging, and cell-wall loosening [82]. Response regulator 22 (OsRR22) encodes a B-type response regulator protein that acts as a transcription factor regulating genes involved in the osmotic stress response and/or ion transport between parenchyma cells and vascular tissue cells in the root [83]. OsCYP71D8L, a cytochrome P450 monooxygenase in the CYP71 clan in rice, negatively regulates salt tolerance by affecting gibberellin and cytokinin homeostasis [84]. The dominant suppressor of KAR2 (OsDSK2a) is a ubiquitin-binding protein that mediates seedling growth and salt responses in rice through interacting with elongated uppermost internode (EUI) to affect gibberellin metabolism [85]. Although a number of salt-tolerance genes have been identified (Table 3), further information is needed about how to effectively use these genes to improve rice salt tolerance. Table 3. Genes associated with rice salt tolerance.

Biotechnology Promises to Accelerate Breeding of Salt-Tolerant Rice Cultivars
Publication of the rice reference genome and the development of next-generation sequencing (NGS) techniques provide an opportunity to explore genome-wide genetic variations and carry out genotyping in a highly efficient way [86]. The relatively low cost of sequencing enables the use of genome and transcriptome sequencing to discover a large number of sequence polymorphisms and to map some agronomic traits [87]. Genome-wide association studies (GWAS) are a powerful approach for identifying valuable natural variations in trait-associated loci, as well as allelic variations in candidate genes underlying quantitative and complex traits, including those related to growth, salt tolerance, and nutritional quality [88,89]. Compared to a traditional QTL linkage analysis, GWAS is based on high-density variations in natural populations and can detect multiple alleles at the same site [90]. Several loci associated with salt tolerance have been identified in rice based on GWAS. Kumar et al. conducted a GWAS of 12 different salt tolerance-related traits at the reproductive stage and identified 20 single-nucleotide polymorphisms (SNPs) significantly associated with the Na + /K + ratio and 44 SNPs associated with other traits observed under salt stress conditions [91]. Liu et al. used 708 rice accessions to perform a GWAS to identify the genes associated with rice salt tolerance and identified seven accessions carrying favorable haplotypes of four genes significantly associated with grain yield under salt stress. These promising candidates will provide valuable resources for salt-tolerant rice breeding [89]. Lekklar et al. conducted a GWAS for salt tolerance during rice reproduction, and more than 73% of the identified loci overlapped with the previously reported salt QTLs [92]. Yuan et al. performed a GWAS using 664 cultivated rice accessions from the 3000 Rice genomes, and twenty-one salt-tolerant QTLs and two candidate genes were identified [93]. These studies indicate that GWAS is a powerful strategy for mapping QTLs of salt tolerance in rice.
Effective phenotyping data (phenomics data) is a prerequisite for the discovery of genes/QTLs, association mapping, and GWAS. Despite recent advances in genomics, the lack of appropriate phenomics data limits the progress in genomics-assisted crop improvement programs. Therefore, the acquisition of high-throughput, effective, and comprehensive trait data in rice has become an acute need [94,95]. High-throughput phenotyping offers the opportunity capture phenotypically complex variations underpinning adaptation in traditional phenotypic selection or statistics-based breeding programs [94,95]. With the development of plant phenotyping platforms, we can obtain the phenomics data more effectively and cost-efficiently, and the integration of high-throughput trait phenotyping with genomics will greatly promote the genetic dissection of salt tolerance-related traits.
Genomic and transcriptomic analysis, proteomics, and metabolomics are powerful tools for identifying genes related to salt tolerance in rice. Salt stress causes a series of changes in rice, including changes in gene expression, protein content, and metabolite levels [96][97][98]. By comparing the transcriptome, proteome, and metabolome of rice under salt stress versus normal conditions, or in salt-tolerant versus salt-sensitive varieties, we can obtain many potential genes related to salt tolerance. For example, a study comparing transcriptome profiles of FL478 and IR29 found more than one thousand genes that were differentially expressed under salt stress compared to normal conditions [96]. Sun et al. analyzed the transcriptome data of a salt-tolerant rice landrace called Changmaogu and detected a large number of genes that were differentially expressed at the germination and seedling stages under salt stress. A further analysis revealed that most of the differentially expressed genes were clustered in the pathways of ABA signal transduction and carotenoid biosynthesis [97]. Peng et al. used an Isobaric Tags for Relative and Absolute Quantitation (iTRAQ)-based proteomic technique to detect proteins that become more or less abundant under salt stress and identified 332 differentially abundant proteins in seedlings of salinity-tolerant and dwarf 58 (sd58) and Kitaake [98]. Although these studies provide many candidate genes for salinity tolerance, whether these genes can be used for breeding salt-tolerant rice remains to be verified. The transgenic approach and genome editing approach are powerful tools for verifying whether genes can be used for salt-tolerant rice breeding. Using the transgenic approach and genome editing approach, ideal materials for target genes can rapidly be obtained and used to assess the salt tolerance, which would accelerate efforts to improve the salinity tolerance in rice.
Based on high-density genome-wide SNP markers detected by next-generation sequencing, the SNP marker-assisted selection method is used to accelerate the process of breeding salt-tolerant rice. Using this method, Rana et al. precisely introgressed the hitomebore salt tolerant 1 (hst1) gene from the salt-tolerant cultivar Kaijin into the high-yielding cultivar Yukinko-mai. Their offspring, YNU31-2-4, had agronomic traits similar to Yukinko-mai under normal growth conditions. Under salt stress, the yield of YNU31-2-4 was significantly higher than that of Yukinko-mai [99]; Bimpong et al. used marker-assisted selection to develop salt-tolerant rice cultivars through introgressing Saltol into the lowland cultivar, Rassi, and obtained 16 introgression lines for further African-wide testing prior to release in six West African countries [100]. Pauyawaew et al. introgressed Saltol into KDML105 by two rounds of marker-assisted backcrossing, and introgression lines with positive Saltol alleles are being tested for salinity tolerance in the salt-affected areas in the northeast of Thailand [101]. All these results suggest that SNP marker-assisted selection breeding is a fast and effective method to improve rice salt tolerance.
The most current methods of genomic selection use single SNP markers to predict the genetic merits of individuals, but haplotypes may have several advantages over single markers for genomic selection. A haplotype is the combination of a series of genetic mutations that coexist on a single chromosome; it contains multiple SNPs and can better capture the linkage disequilibrium and genomic similarity in different lines [102,103]. The use of haplotypes can improve the accuracy of genomic prediction. Thus, identifying key haplotypes related to the salt tolerance of rice will provide useful genetic resources and parent materials for breeding salt-tolerant rice.

Challenges and Perspectives
Salt stress is one of the main environmental factors affecting rice growth and rice yield worldwide. Improving the salt tolerance of rice is the most direct and effective method to solve this problem. After years of collection and screening, scientists have obtained several salt-tolerant rice germplasms (Table 2), which have laid the foundation for studying the mechanism of salt tolerance in rice and developing salt-tolerant rice cultivars. This review summarizes the progress to date on the breeding of salt-tolerant rice and proposes that genomics and molecular tools for precision breeding will accelerate the development of salt-tolerant cultivars.
Rice salt tolerance is controlled by multiple genes and is a complex physiological characteristic [17]. The evaluation of rice salt tolerance is also complex. Phenotypic traits are typically used to evaluate the salt tolerance of rice varieties [18], an approach that lacks comprehensiveness and accuracy and may result in an assessment of salt tolerance that is not consistent with actual performances in the field. Therefore, a standardized system to evaluate rice salt tolerance is urgently needed. Moreover, salt tolerance is genotype-dependent. Although numerous QTLs controlling salt tolerance traits have been identified in different mapping populations (Table 2), only a few major salt tolerance genes have been isolated from QTLs [48,49]. The inconsistency and variability of QTLs in different genetic backgrounds and environments have limited their applications in breeding programs.
Using traditional breeding methods, scientists have bred a series of salt-tolerant rice cultivars, but the mechanism of salt tolerance in rice is largely unclear. How plants perceive salt stress, how these signals are translated into adaptive responses, and how multiple salt tolerance genes coordinately regulate salt tolerance in rice are all questions that need further investigation. Moreover, traditional breeding methods are time-consuming and inefficient. SNP marker-assisted selection and genetic engineering technology will greatly improve the molecular breeding process [99,104]. Therefore, efforts should be made to capture useful salt tolerance genes as possible genetic markers to introgress into elite rice varieties. In addition, it is difficult to obtain salt-tolerant varieties that can be used in field production by the introduction of a single gene or several genes. To cultivate valuable salt-tolerant rice varieties, multiple salt tolerance genes or haplotypes need to be gathered into elite rice varieties; how to quickly and effectively introduce these genes or haplotypes simultaneously is another problem that needs to be solved. With the innovation of breeding technology, many new breeding technologies, such as MutMap, knock-out/in, and genome editing, will accelerate the process of salt-tolerant rice cultivation [83,105]. Further studies should focus on cloning salt tolerance genes and elucidating their regulatory mechanisms and on investigating how multiple genes or haplotypes can be transferred at the same time with stable inheritance by their offspring.
Due to variations in saline-alkali soils and environmental conditions in different areas, the cultivation of salt-tolerant rice varieties is subject to strong geographical constraints; varieties cultivated in one place are unsuitable for planting in another. Therefore, research efforts should focus on identifying molecular markers and haplotypes associated with salt tolerance and on developing salt-tolerant rice. The main rice varieties in various regions should be transformed with genes associated with improved salt tolerance by using a combination of molecular breeding and traditional breeding approaches. The resulting core rice germplasm with a high salt tolerance and excellent agronomic traits would provide a valuable germplasm resource for breeding programs aimed at developing salt-tolerant rice.
Author Contributions: All the authors H.Q., Y.L., and R.H. discussed and created the review's outline, H.Q. wrote the manuscript, and R.H. edited the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding: This work was funded by the National Natural Science Foundation of China, grant numbers 32030079, 31801445, and 31871551, and the Agricultural Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences.

Conflicts of Interest:
The authors declare no conflict of interest.