*Article* **Fine Mapping of** *qWCR4***, a Rice Chalkiness QTL Affecting Yield and Quality**

**Huan Shi, Yun Zhu, Peng Yun, Guangming Lou, Lu Wang, Yipei Wang, Guanjun Gao \*, Qinglu Zhang, Xianghua Li and Yuqing He \***

> National Key Laboratory of Crop Genetic Improvement, Hubei Hongshan Laboratory, Huazhong Agricultural University, Wuhan 430070, China; huanshi1023@gmail.com (H.S.); zy9000722@gmail.com (Y.Z.); pengyun0106@gmail.com (P.Y.); louguangming@mail.hzau.edu.cn (G.L.); wanglu001222@gmail.com (L.W.); wangyipei3@gmail.com (Y.W.); qingluzhang@mail.hzau.edu.cn (Q.Z.); xhli@mail.hzau.edu.cn (X.L.) **\*** Correspondence: gaojun8199@webmail.hzau.edu.cn (G.G.); yqhe@mail.hzau.edu.cn (Y.H.)

**Abstract:** Rice (*Oryza sativa* L.) chalkiness greatly reduces the rice quality and the commercial value. In this study, *qWCR4*, a previously reported quantitative trait locus (QTL) of white-core rate (WCR), was confirmed by a BC5F<sup>2</sup> segregation population and further fine mapped to a 35.26 kb region. In the *qWCR4* region, *LOC\_Os04g50060* and *LOC\_Os04g50070* showed significant differences in expression level in endosperm between two NILs, whereas four other genes had no expression. Starch granules in the central endosperm of chalky grains from NIL(J23B) with higher WCR exhibited a typically round and loosely packed morphology. NIL(J23B) with higher WCR accompanied a higher seed filling speed. Moreover, *qWCR4J23B* (*qWCR4* allele in J23B) increased WCR, grain numbers per plant, seed setting rate, grain width, and thousand-grain weight, contributing to a superior yield per plant. All in all, our research results not only lay a foundation for map-based cloning of *qWCR4* but also provide new genetic resources for rice yield and quality breeding.

**Keywords:** rice; white-core rate; fine-mapping; quantitative RT-PCR; yield; and quality

**Citation:** Shi, H.; Zhu, Y.; Yun, P.; Lou, G.; Wang, L.; Wang, Y.; Gao, G.; Zhang, Q.; Li, X.; He, Y. Fine Mapping of *qWCR4*, a Rice Chalkiness QTL Affecting Yield and Quality. *Agronomy* **2022**, *12*, 706. https://doi.org/10.3390/ agronomy12030706

Academic Editor: HongWei Cai

Received: 7 February 2022 Accepted: 10 March 2022 Published: 14 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

### **1. Introduction**

Rice (*Oryza sativa* L.) is a primary food crop that has greatly contributed to solving the world food security crisis. The rice yield has been improved dramatically after the first and second rice green revolutions in the past several decades [1]. In recent years, more and more attention has focused on the quality of rice concerning the improvement of people's living standards. Rice chalkiness, as an extremely unsatisfactory quality trait in rice sales and consumption, not only affects grain appearance but also has negative effects on rice processing and cooking characteristics [2].

The chalkiness of rice is characterized by the opaque part of the grain, which can be divided into white-belly (ventral side), white-core (center), and white-back (dorsal side) according to the opaque position [3]. The opaque endosperm is often associated with changes in the morphology and arrangement of starch granules, characterized by round and loosely packed starch granules [4–7]. Many studies have shown that chalkiness affects heading yield and milling yield by increasing the incidence of grain breakage [4,5,8]. The chalky grains exhibited lower protein content, and the palatability of cooked rice showed a linear decrease with increasing chalky rice proportion in the sensory evaluation [6]. Moreover, the chalkiness of kernels is often accompanied by low amylose content, which influences the sensory quality [9]. Thus, reducing the chalkiness rate has become one of the important goals in high-quality rice breeding.

Rice chalkiness is affected by genetic and environmental factors. Rice endosperm filling is a continuous process that typically lasts for more than a month. In this process, abiotic stresses, especially high temperature, could accelerate seed filling and led to chalkiness formation. Under high temperature conditions, chalkiness is usually trigged

7

within 9–16 days after pollination [10]. Reporters have pointed out that warmer weather during the filling stage reduces grain appearance and palatability due to chalkiness in rice grains [8,11,12]. Chen et al. [13] found that high temperature could cause the increase of amylase and the decrease of GBSS, SBEs, and BEIIb activities, and these changes are the direct or indirect reasons for the chalkiness. Therefore, high temperature is the main abiotic stress leading to the increase of rice chalkiness. Genetic background is another determinant of rice chalkiness. Grain chalkiness is a complex quantitative trait, and more than one hundred quantitative trait loci (QTLs) for rice chalkiness have been reported on all 12 chromosomes [10–12,14–17]. In addition, rice grain shape is another determinant of rice chalkiness. Changes in grain size are usually accompanied by changes in chalkiness [18,19]. Floury mutants are valuable genetic resources for dissecting the underlying mechanisms of chalkiness formation. So far, many genes responsible for floury phenotype have been confirmed in directly influenced starch and protein synthesis, and presented an extremely chalky phenotype [20–25]. However, only a few QTLs for chalkiness were successfully fine-mapped or cloned [26–28]. Among them, *Chalk5* [26] encodes a vacuolar H<sup>+</sup> -translocating pyrophosphatase. The elevated expression of *Chalk5* increased the chalkiness degree of endosperm, which may be due to the disruption of the pH dynamic balance of the endosperm transport system during seed development, thus affecting the biogenesis of protein bodies, and accompanied by the increase of vesicle structure, thus forming gaps between endosperm storage substances and resulting in the generation of chalky grains. *WCR1* [28] encodes an F-box protein that negatively regulates grain chalkiness. *WCR1* promotes the elimination of excess ROS, maintains redox homeostasis, and delays programmed cell death in starch endosperm by up-regulating the transcription of *MT2b*. Hence, the discovery of more chalkiness-related genes will be helpful to the revelation of the chalkiness genetic mechanism.

In this study, a stable QTL for white-core rate in rice, *qWCR4*, was confirmed and fine mapped to a 35.26 kb region. RT-qPCR and parental sequence analysis results showed that *LOC\_Os04g50060* and *LOC\_Os04g50070* could be the causal genes for *qWCR4*. NIL(J23B) with higher WCR was accompanied by faster seed filling speed, bigger grain size, higher yield per plant, and amylose content. In summary, the knowledge will pave the way for improving rice yield and quality.

#### **2. Materials and Methods**

#### *2.1. Plant Materials and Field Experiment*

In a previous study, an RIL population consisting of 184 lines was developed from a cross between an *indica* cultivar J23B (the recurrent parent) and a *japonica* cultivar BL130 (the donor parent) [17]. *qWCR4* was repeatedly detected to affect the white-core rate in two environments by QTL mapping. To find the underlying gene controlling WCR in the *qWCR4* interval, we selected one F<sup>7</sup> generation RIL containing the BL130 alleles of *qWCR4* to backcross with the recipient parent J23B. Five times backcrosses were performed for *qWCR4*, and the two flanking markers of the QTL mapping interval were used for hybrids screening for authenticity. After obtaining the BC5F<sup>1</sup> hybrids, plants with a heterozygous fragment of the QTL mapping interval were retained and self-pollinated to obtain the BC5F<sup>2</sup> population defined as the NIL-F<sup>2</sup> population. A total of 192 individuals were planted in 16 rows of 12 plants each for QTL validation.

For fine mapping *qWCR4*, 14 recombinants between RM241 and RM255 were firstly screened from a BC5F<sup>2</sup> population which derived from self-pollinated a BC5F<sup>1</sup> plant with a heterozygous *qWCR4* segment. After the *qWCR4* interval was narrowed to between markers M24 and M25, a total of 25 recombinants in the M24–M25 region were screened from a BC5F<sup>3</sup> population which derived from a self-pollinated BC5F<sup>2</sup> plant with a heterozygous *qWCR4* segment. All recombinant progeny lines of BC5F<sup>3</sup> and BC5F<sup>4</sup> generations were used to compare the WCR effect between homologous with J23B and BL130 alleles and each recombinant progeny line was planted in 8 rows with 12 plants per row.

The BC5F2, BC5F3, and BC5F<sup>4</sup> populations of *qWCR4* were planted under natural field conditions at the experimental station of Huazhong Agricultural University in Wuhan, Hubei Province in 2014, 2016, and 2018, respectively. The 30-days-old progeny seedlings of each recombinant were transplanted into 8 rows and 12 seedlings each row. The transplanting distance between individual plants of single-row in the field was 16.5 cm, and the row spacing was 26.4 cm. Field management followed local practices. Ten plants were harvested in the middle of each row for traits measurement.

#### *2.2. Phenotyping*

Plant height was measured in the field, and harvested panicles from each plant were air-dried and stored at room temperature for 3 months before further phenotyping of other traits. Effective tiller number, panicle length, primary branches, secondary branches, spikelet number per plant, and filled grain number per plant were measured for each plant before threshing. The setting rate was calculated by filled grains per plant divided by spikelet number per plant. Grain length, grain width, thousand-grain weight, and yield per plant were measured after threshing. A total of 60 plants in each near-isogenic line were used for the phenotyping of these traits.

More than 100 seeds from each normal mature plant were randomly selected to be processed into milled rice, and the seed white-core rate of milled rice was investigated through visual observation. Milled rice powder was used to determine the contents of amylose and four kinds of storage proteins (albumin, globulin, prolamin, and glutelin). The method for determining amylose content and 4 storage protein contents was mentioned previously [29,30].

#### *2.3. Genotyping and Sequence Analysis*

The parent varieties BL130 and J23B were sequenced by the illumine HiSeq2000 (Illumina, San Diego, CA, USA), and the sequencing data were compared and assembled according to the rice reference genome (Rice Genome Annotation Project, http://rice.uga. edu/, accessed on 6 February 2022) [31]. All mapping primers were designed by primer premier (version 6.0, PREMIER Biosoft, San Francisco, CA, USA) software with reference to the sequencing data of two parents. According to the cetyltrimethylammonium bromide (CTAB) method, genomic DNA was extracted from leaves [32]. The initial program of PCR amplification was 94 ◦C for 5 min, then 32 cycles of 30 s at 94 ◦C, 30 s at 55 ◦C, 30 s at 72 ◦C, and finally 5 min at 72 ◦C. The PCR products were identified by sequencing or 4% non-denaturing polyacrylamide gel electrophoresis (PAGE). DNA bands on PAGE gel were displayed by silver nitrate staining and NaOH-formaldehyde solution. Polymorphic simple sequence repeat (SSR) or insertion/deletion (Indel) markers in the QTL interval were used to identify individual genotypes. Relevant primer information can be found in Table S1.

High-quality sequencing bam documents of parents were opened and analyzed in Integrative Genomics Viewer (IGV) software (Broad Institute, Cambridge, MA, USA) [33], and all variations of candidate genes were confirmed by sequencing. Primers used to identify sequence consistency of parents and NILs were listed in Table S1, and the corresponding results were shown in Figure S3.

#### *2.4. RNA Extraction, Reverse Transcription, and qRT-PCR*

Expression levels of genes in the M24-M25 interval were examined between NILs of *qWCR4*, which were derived from two homologous progeny of SH60 in the BC5F<sup>4</sup> population. RNA extraction kit (TRIzol, Invitrogen, Carlsbad, CA, USA) was used for total RNA extraction from different endosperm stages. The first strand of cDNA was synthesized in 20 µL M-MLV reverse transcriptase reaction system (containing 2 µg RNA and 200 U M-MLV reverse transcriptase (Promega, Madison, WI, USA)). Real-time PCR was carried out using Bio-Rad T100TM real-time PCR system (Bio-Rad, Hercules, CA, USA) with the SYBR Green I mix (TaKaRa, Shiga, Japan) on the QuantStudio6Flex instrument (Applied

Biosystems, Carlsbad, CA, USA). All biological tests were repeated at least three times, with three technical repeats, and the rice *Actin1* gene was used as the internal reference. Relevant primers of this analysis are given in Table S2.

#### *2.5. Scanning Electron Microscopy (SEM)*

The rice grains of two near-isogenic lines (NILs) were cut in the middle section and gilded under the vacuum condition. The morphology of starch granules in the center of NILs' endosperm was observed by a scanning electron microscope (JSM-6390LV, Jeol Ltd., Tokyo, Japan) under 10 kV acceleration voltage and 30 nm spot size. Scanning electron microscope analyses were performed with at least three biological repeats. All procedures were carried out in accordance with the manufacturer's instructions.

#### *2.6. Statistical Analysis*

In the NIL-F<sup>2</sup> population of *qWCR4*, the linkage map of the RM241–RM255 interval was constructed using Mapmaker/Exp3.0 software with the Kosambi mapping function [34], and the *qWCR4* effect detection on WCR was performed by composite interval mapping (CIM) method using Windows QTL cartographer 2.5 software [35]. In terms of fine mapping population, we performed progeny tests to evaluate the WCR effect in recombinant progeny lines. The WCR difference between plants with two different homozygous genotypes was compared by *t*-test. If the *p*-value was less than 0.05, the candidate gene was considered to be located in the heterozygous fragment, otherwise it was considered to be located in the homozygous fragment. Differences in agronomy traits of NILs were analyzed by *t*-test.

### *2.7. Evolution Analisis*

The evolutionary history was inferred using the Neighbor-Joining method [36]. The optimal tree is shown. The tree is drawn to scale, with branch lengths in the same units as the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Poisson correction method [37] and are in the units of the number of amino acid substitutions per site. The proportion of sites where at least 1 unambiguous base is present in at least 1 sequence for each descendent clade is shown next to each internal node in the tree. This analysis involved 13 amino acid sequences. All ambiguous positions were removed for each sequence pair (pairwise deletion option). There was a total of 830 positions in the final dataset. Evolutionary analyses were conducted in MEGA11 [38].

#### **3. Results**

#### *3.1. Genetic Validation of qWCR4*

In a previous study, *qWCR4* was located between markers RM241 and RM255 on chromosome 4 using a RIL population derived from a cross between BL130 and J23B [17]. To validate the genetic effect of *qWCR4*, a BC5F<sup>1</sup> plant that carried a heterozygous *qWCR4* region was self-pollinated to obtain the BC5F<sup>2</sup> population defined as the NIL-F<sup>2</sup> population. Compared with BL130, J23B showed less white belly rate, higher white-core rate, and smaller grain size (Figure S1). Similarly, NIL with homologous J23B allele (NIL(J23B)) had a significantly higher WCR than that of NIL with homologous BL130 allele (NIL(BL130)) in the NIL-F<sup>2</sup> population of *qWCR4* (Figure 1a). We also investigated the starch granules morphology in the endosperm center between two homologous NILs. As shown in Figure 1b, starch granules in the endosperm center of white-core grains from NIL(J23B) were small, round, and loosely packed, which were notably different from NIL(BL130) with polyhedral and densely packed starch granule morphology (Figure 1b). Additionally, *qWCR4* showed an incompletely dominant effect (Figure 1c), and the phenotypic variation of WCR in the BC5F<sup>2</sup> population showed continuous bimodal distribution (Figure 1d). *qWCR4* was mapped to the region between RM241 and RM255 and explained 21.1% of the phenotypic variance (Table 1).

phenotypic variance (Table 1).

**Figure 1.** Genetic validation of QTL effects of *qWCR7* in BC5F2 population. (**a**) Chalkiness performance of NIL(BL130) (left) and NIL(J23B) (right). (**b**) SEM analysis of nonchalky (left) and chalky (lower right) grains of NIL(BL130) and NIL(J23B). (**c**) WCR difference analysis among three genotypes. (**d**) Distribution of WCR performance variation. **Figure 1.** Genetic validation of QTL effects of *qWCR7* in BC5F<sup>2</sup> population. (**a**) Chalkiness performance of NIL(BL130) (left) and NIL(J23B) (right). (**b**) SEM analysis of nonchalky (left) and chalky (lower right) grains of NIL(BL130) and NIL(J23B). (**c**) WCR difference analysis among three genotypes. (**d**) Distribution of WCR performance variation.

a significantly higher WCR than that of NIL with homologous BL130 allele (NIL(BL130)) in the NIL-F2 population of *qWCR4* (Figure 1a), . We also investigated the starch granules morphology in the endosperm center between two homologous NILs. As shown in Figure 1b, starch granules in the endosperm center of white-core grains from NIL(J23B) were small, round, and loosely packed, which were notably different from NIL(BL130) with polyhedral and densely packed starch granule morphology (Figure 1b). Additionally, *qWCR4* showed an incompletely dominant effect (Figure 1c), and the phenotypic variation of WCR in the BC5F2 population showed continuous bimodal distribution (Figure 1d). *qWCR4* was mapped to the region between RM241 and RM255 and explained 21.1% of the



We made a two-step mapping strategy to improve the efficiency of mapping *qWCR4*. Firstly, 14 recombinants between markers RM241 and RM255 were identified from a BC5F2 WCR, white core rate. QTL, quantitative trait locus. LOD, logarithms of odds. A, additive effect, the negative value means that J23B allele increases the trait value. V, variance, phenotypic variation explained by the QTL.

population consisting of 1000 individuals, and the corresponding progeny population of each recombinant was used for the progeny test. To genotype all recombinants, 7 indel markers in this region were developed based on the sequence variation between two par-These results indicated that *qWCR4* was the genetic factor responsible for WCR variation, and the BL130-derived allele decreased the endosperm WCR.

#### ents (Figure 2a). The progeny test was undertaken to determine the *qWCR4* genotype of *3.2. Fine Mapping of qWCR4*

each recombinant. According to the progeny testing results of recombinant lines ZY03, We made a two-step mapping strategy to improve the efficiency of mapping *qWCR4*. Firstly, 14 recombinants between markers RM241 and RM255 were identified from a BC5F<sup>2</sup> population consisting of 1000 individuals, and the corresponding progeny population of each recombinant was used for the progeny test. To genotype all recombinants, 7 indel markers in this region were developed based on the sequence variation between two parents (Figure 2a). The progeny test was undertaken to determine the *qWCR4* genotype of each recombinant. According to the progeny testing results of recombinant lines ZY03, ZY20, and ZY27, which have similar crossover intervals between M24 and M25, we narrowed the *qWCR4* to a 536 kb region between M24 and M25 (Figure 2a).

11

**Figure 2.** Fine mapping of *qWCR4***.** (**a**) The firs*t-*step fine mapping of *qWCR4*. (**b**) The second-step fine mapping of *qWCR4*. (**c**) Schematic representation of candidate gene in *qWCR4* region. Black, white, and grey blocks represent the genotypes of homozygous BL130, homozygous J23B, and heterozygote, respectively. Genotypes and phenotypes of recombinants, each of which was confirmed by progeny test, the A and B of the progeny test meant no WCR difference between two homologous plants in recombinant progeny, while the H of the progeny test meant significant WCR difference. After the 536 kb region was obtained from the first step, a BC5F3 population consist-**Figure 2.** Fine mapping of *qWCR4***.** (**a**) The first-step fine mapping of *qWCR4*. (**b**) The second-step fine mapping of *qWCR4*. (**c**) Schematic representation of candidate gene in *qWCR4* region. Black, white, and grey blocks represent the genotypes of homozygous BL130, homozygous J23B, and heterozygote, respectively. Genotypes and phenotypes of recombinants, each of which was confirmed by progeny test, the A and B of the progeny test meant no WCR difference between two homologous plants in recombinant progeny, while the H of the progeny test meant significant WCR difference.

ZY20, and ZY27, which have similar crossover intervals between M24 and M25, we nar-

rowed the *qWCR4* to a 536 kb region between M24 and M25 (Figure 2a).

ing of 2500 individuals was used to screen new recombinants for the second-step fine mapping. A total of 25 recombinants between M24 and M25 were identified in this population, and 3 indel markers in this region were developed and used to genotype these recombinants. As shown in Figure 2b, progeny testing results of 4 recombinants that crossover between markers S2 and S4 indicated that the candidate gene underlying for WCR in the *qWCR4* region should be in the 35.26 kb region between these two markers. *3.3. Expression and Sequence Analysis of Candidate Genes*  After the 536 kb region was obtained from the first step, a BC5F<sup>3</sup> population consisting of 2500 individuals was used to screen new recombinants for the second-step fine mapping. A total of 25 recombinants between M24 and M25 were identified in this population, and 3 indel markers in this region were developed and used to genotype these recombinants. As shown in Figure 2b, progeny testing results of 4 recombinants that crossover between markers S2 and S4 indicated that the candidate gene underlying for WCR in the *qWCR4* region should be in the 35.26 kb region between these two markers.

#### According to the annotation information of the *japonica* variety Nipponbare in RGAP website (Rice Genome Annotation Project, http://rice.uga.edu/, accessed on 6 February *3.3. Expression and Sequence Analysis of Candidate Genes*

2022), the target 35.26 kb region of *qWCR4* contains six annotated genes (*LOC\_Os04g50040*, According to the annotation information of the *japonica* variety Nipponbare in RGAP website (Rice Genome Annotation Project, http://rice.uga.edu/, accessed on 6 February 2022), the target 35.26 kb region of *qWCR4* contains six annotated genes (*LOC\_Os04g50040*, *LOC\_Os04g50050*, *LOC\_Os04g50060*, *LOC\_Os04g50070*, *LOC\_Os04g50080,* and *LOC\_Os04g50090*) (Figure 2c). Among these genes, *LOC\_Os04g50060*, *LOC\_Os04g50070,* and *LOC\_Os04g50090* encode a GRAS family transcription factor domain-containing protein, a ZOS4-13-C2H2 zinc finger protein, and a putative helix-loop-helix DNA-binding protein, respectively. The remaining genes encode two expressed proteins (*LOC\_Os04g50040* and *LOC\_Os04g50080*) and one putative retrotransposon protein (*LOC\_Os04g50050*). According to the RNA-Seq

data in the RGAP website, the two expressed proteins had no expression in endosperm or seed, and retrotransposon was usually not considered a functional gene. Therefore, we paid attention to three functional genes expressed in the endosperm.

We then examined the relative expression level of these three genes in different stage endosperm by quantitative real-time PCR (Figure 3a). The result showed that *LOC\_Os04g50090* was not expressed in the endosperm. *LOC\_Os04g50060* was notably expressed at a relatively high level throughout the whole endosperm development period with higher expression in NIL(BL130). *LOC\_Os04g50070* expressed only in 5E (endosperm 5 days after pollination) with a significant difference between NILs. We also analyzed the nucleotide diversity of two parents in *LOC\_Os04g50060* and *LOC\_Os04g50070*. The genomic region used for nucleotide diversity analysis contains the promotor (~2 kb) and the entire ORF of two genes. According to the sequencing data of two parents, 22 SNP variants were found in the *LOC\_Os04g50060* region, of which 17 were in the promotor and 5 in the ORF (Figure 3b). 9 SNP variants were found in the *LOC\_Os04g50070* region, of which 8 were in the promotor and 1 in the ORF. Evolutionary analysis showed that *LOC\_Os04g50060* and *LOC\_Os04g50070* exhibited high homology with GRAS family transcription factor and zinc finger protein family, respectively (Figure S4). Taken together, both *LOC\_Os04g50060* and *LOC\_Os04g50070* could be the candidate genes for WCR in the *qWCR4* region. *Agronomy* **2022**, *12*, x FOR PEER REVIEW 8 of 14

**Figure 3.** Expression and sequence variations of two candidate genes in *qWCR4* locus. (**a**) Relative expression level analysis of three candidate genes. *LOC\_Os04g50060*, *LOC\_Os04g50070,* and *LOC\_Os04g50090* represent three causal genes underlying the *qWCR4* locus. 5E, 10E, and 15E mean endosperm 5 days, 10 days, and 15 days after pollination, respectively. \*\* means significant differences at *p* < 0.01, student's *t-*tests. (**b**) Variation analysis of *LOC\_Os04g50060*. (**c**) Variation analysis of *LOC\_Os04g50070*. Boxes filled with black and white in **b** and **c** mean coding and untranslated regions, respectively. The solid lines on the left side of boxes in **b** and **c** mean the promotor region. The dotted lines on the solid lines and boxes in **b** and **c** represent the position from the translation initiation site (ATG), + and − mean located in down-stream and up-stream of ATG. *3.4. Agronomic Traits of Two NILs*  The abnormal starch granule morphology may be caused by a developmental disorder. Researchers have proved that high filling speed decreased rice quality [39,40]. Here, **Figure 3.** Expression and sequence variations of two candidate genes in *qWCR4* locus. (**a**) Relative expression level analysis of three candidate genes. *LOC\_Os04g50060*, *LOC\_Os04g50070,* and *LOC\_Os04g50090* represent three causal genes underlying the *qWCR4* locus. 5E, 10E, and 15E mean endosperm 5 days, 10 days, and 15 days after pollination, respectively. \*\* means significant differences at *p* < 0.01, student's *t*-tests. (**b**) Variation analysis of *LOC\_Os04g50060*. (**c**) Variation analysis of *LOC\_Os04g50070*. Boxes filled with black and white in (**b**) and (**c**) mean coding and untranslated regions, respectively. The solid lines on the left side of boxes in (**b**) and (**c**) mean the promotor region. The dotted lines on the solid lines and boxes in (**b**) and (**c**) represent the position from the translation initiation site (ATG), + and − mean located in down-stream and up-stream of ATG.

we investigated the seed filling speed after pollination of two NILs. NIL(J23B) with higher WCR accompanied with a higher filling speed in the middle and later stage (from 12 days after pollination to 20 days after pollination) (Figure 4a). This result implied that quicker

#### *3.4. Agronomic Traits of Two NILs*

The abnormal starch granule morphology may be caused by a developmental disorder. Researchers have proved that high filling speed decreased rice quality [39,40]. Here, we investigated the seed filling speed after pollination of two NILs. NIL(J23B) with higher WCR accompanied with a higher filling speed in the middle and later stage (from 12 days after pollination to 20 days after pollination) (Figure 4a). This result implied that quicker endosperm filling might result in abnormal starch granule morphology. *Agronomy* **2022**, *12*, x FOR PEER REVIEW 9 of 14

**Figure 4.** Trait performance of two NILs. Differences in Seed filling rate (**a**), Filling grain number per plant (**b**), Setting rate (**c**), Grain length (**d**), Grain width (**e**), Thousand grain weight (**f**), Yield per plant (**g**), Amylose content (**h**) and protein content (**i**) were analysised between two NILs. \* and \*\* mean significant differences at *p* < 0.05 and *p* < 0.01 respectively, student's *t*-tests. **Figure 4.** Trait performance of two NILs. Differences in Seed filling rate (**a**), Filling grain number per plant (**b**), Setting rate (**c**), Grain length (**d**), Grain width (**e**), Thousand grain weight (**f**), Yield per plant (**g**), Amylose content (**h**) and protein content (**i**) were analysised between two NILs. \* and \*\* mean significant differences at *p* < 0.05 and *p* < 0.01 respectively, student's *t*-tests.

Chalkiness is closely related to grain size and quality traits, such as grain length, grain width, starch, and protein content [19,26,27]. In the present study, we investigated agronomy traits of NILs to detect the effect of *qWCR4* on these traits. No significant differences were observed in plant height, effective tiller number, panicle length, primary branches number, secondary branches number, and spikelet number per plant, leading to similar plant architecture between these two NILs (Figure S2a–f). However, compared with the NIL(J23B), NIL(BL130) with lower WCR had more grain length but less filled grain number, setting rate, and grain width, leading to an obviously decreased thousand-grain weight and yield per plant (Figure 4b–g). In addition, we also examined the amylose content and four storage protein content of these two NILs. NIL(BL130) showed decreased amylose content but higher storage protein content (Figure 4h,i). The above results suggested that the *qWCR4* allele from J23B has a positive effect on yield-related traits and storage content but a negative effect on amylose content and chalkiness.

#### **4. Discussion**

Unlike waxy grain, chalky grain is attributed to the air gap between irregular starch granules, which cause scatter light, leading to an opaque phenomenon [41]. However, the opacity of waxy rice is due to the diffusion of light from micropores and hollows in starch granules [42]. The external performance and internal characteristics of rice dramatically influence consumer choice. As one of the evaluation indexes of poor-quality rice, chalkiness not only affects the appearance of rice [43–45] but also greatly affects the rice milling yield [4,5,8] and cooking and eating quality [6]. Therefore, understanding the formation mechanism of endosperm chalkiness is essential to rice yield and quality.

#### *4.1. A New QTL Was Found to Control WCR in Rice Endosperm*

Many chalkiness QTLs were identified in the past decades, but only a few QTLs/genes related to chalkiness have been fine mapped/cloned. There are two possible reasons. Firstly, chalkiness is a quantitative trait controlled by polygenes, abnormal seed filling, sugar transport, and starch synthase could lead to the chalky phenomenon. Second, the chalkiness is seriously affected by an especially high temperature environment. Therefore, researchers usually used different populations under different environments to obtain stable chalkiness QTLs [11,12,14,17,46]. In a previous study, *qWCR4* was repeatedly detected in two environments and confirmed in a NIL-F<sup>2</sup> population [17]. In this study, *qWCR4* was confirmed in a higher generation genetic population (Figure 1) and showed a higher phenotypic variation explained by the QTL (Table 1). Starch granules of chalky grains from NIL(J23B) exhibited a typically chalky morphology according to previous research (Figure 1b). Through two-step fine mapping, *qWCR4* was narrowed to a 35.26 kb region between markers S2 and S4 (Figure 2a,b). Until now, few chalkiness-related genes have been reported on chromosome 4, while *qWCR4* is a newly found QTL affecting chalkiness in this region. Therefore, we found a new QTL *qWCR4* controlling WCR in chromosome 4.

#### *4.2. LOC\_Os04g50060 and LOC\_Os04g50070 Could Be the Candidate Genes for qWCR4*

According to the annotation data on the RGAP website, six genes exist in the *qWCR4* region. Two of them encoded expressed protein not expressed in the endosperm. *LOC\_Os04g50050* encoded a putative transposon not considered a functional gene in general. *LOC\_Os04g50090* encoded an HLH DNA binding protein not expressed in endosperm. *LOC\_Os04g50060* encoded a GRAS family transcription factor with a dramatic expression change in different endosperm stages, and showed high protein homology with identified GRAS family transcription factors in other species. *LOC\_Os04g50070* encoding a C2H2 zinc finger protein with significant expression changed only in 5E, and also exhibited protein homology with a reported zinc finger protein family. Cai et al. [47] reported a GRAS family transcription factor *ZmGRAS20*, which is expressed in endosperm specifically, may function as a starch synthesis regulatory factor in rice endosperm. *ZmGRAS20* transgenic seeds exhibited altered starch granules morphology, and overexpression of *ZmGRAS20* led to a chalky region

of ventral endosperm with decreased starch content. Ji et al. [48] discovered an *Opaque 2* (*O2*)-*ZmGRAS11*-*ZmEXPB12* regulatory module that regulates endosperm cell expansion and endosperm filling. On the other hand, Royo et al. [49] identified two closely related C2H2-type zinc finger proteins, *ZmMRPI-1* and *ZmMRPI-2,* which interact with *ZmMRP-1* and modulate its activity on transfer cell-specific promotors. Jiang et al. [50] found *ZmZAT8*, a conserved feature of plant C2H2-type zinc finger protein, plays a positive role in regulating starch synthesis in maize endosperm and could be strongly stimulated by ABA. Taken together, both GRAS transcription factor and C2H2-type zinc finger protein could be the candidate gene control WCR in *qWCR4* interval. Sequence variation analysis between two parents showed that 22 and 9 SNPs exist in *LOC\_Os04g50060* and *LOC\_Os04g50070*, respectively. Variations in the promotor of these two genes may lead to differences in the expression levels, and ultimately affects the WCR performance of two NILs.

#### *4.3. qWCR4 Changed the Yield and Quality Performance of NILs*

Many studies have shown that high filling speed would decrease the quality of rice and increase the rice chalky rate [39,40]. Compared to NIL(BL130), NIL(J23B) showed higher seed filling speed (Figure 4a), and then more white-core grains occurred. NIL with higher WCR showed superiority in filled grain number, grain width, and TGW, which resulted in a significant advantage in yield per plant (Figure 4b–g). According to previous reports, the GRAS family transcription factor and C2H2-type zinc finger protein may also have a function in regulating grain width [51,52], and many studies have shown that the increase of grain width increases the WCR [18,19]. Therefore, whether the higher WCR in NIL(J23B) is caused by the larger grain width needs to be further studied.

Many researchers have reported that rice endosperm with higher WCR is often accompanied by lower starch [9] and protein content [4]. We further analyzed the content of amylose and storage proteins in NILs, and higher WCR was found along with higher amylose content and lower storage protein content (Figure 4h,i). It may be because *qWCR4* affects the amylose content and storage protein content through other mechanisms. The increase of grain width is accompanied by increased amylose content [53]. *qWCR4* may affect amylose and storage protein content by increasing grain width. All in all, fine-mapping *qWCR4* could provide gene resources for breeding high-yield and quality rice varieties.

#### **5. Conclusions**

Our research revealed that *qWCR4* was a genetic factor conferring WCR variation and was narrowed to a 35.26 kb region. The quantitative RT-PCR and sequence variation analysis of genes in the *qWCR4* region showed that *LOaC\_Os04g50060* and *LOC\_Os04g50070*, coding a GRAS and a C2H2 family transcription factor in the *qWCR4* region, respectively, may be candidate genes for *qWCR4*. Different *qWCR4* alleles significantly changed the agronomic traits of NILs. *qWCR4*J23B could significantly increase the seed setting rate and grain width of NIL(J23B), resulting in a higher yield per plant, but also an increase in WCR of about 9%. However, whether more chalky grains are caused by higher seed filling speed or wider grain needs further study. Finally, the results of this study also laid a foundation for map-based cloning of *qWCR4*.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/agronomy12030706/s1, Table S1: Primers used for fine mapping and NILs identification; Table S2: Primers used for quantitative RT-PCR; Figure S1: Grain characters of parents and NILs; Figure S2: Agronomic traits difference of NILs; Figure S3: Sequence consistency of NILs and parents; Figure S4: Evolutionary tree of *LOC\_Os04g50060* with GRAS family transcription factor (a) and *LOC\_Os04g50070* with zinc-finger protein family (b).

**Author Contributions:** Conceptualization, H.S., Y.H., Y.Z. and P.Y.; methodology, H.S., Y.H., Y.Z. and P.Y.; investigation, H.S., Y.Z., L.W. and Y.W.; resources, Y.H., P.Y., G.G., Q.Z. and X.L.; writing-original draft preparation, H.S.; writing-review and editing, H.S., Y.H. and G.L.; supervision, Y.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by grants from the National Natural Science Foundation of China (U21A20211), the Ministry of Science and Technology (2021YFF1000200, 2020YFD0900302), the science and technology major program of Hubei Province (2021ABA011, 2020BBB051), and the China Agriculture Research System (CARS-01-03).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The datasets generated during the current study are available from the corresponding author on reasonable request.

**Acknowledgments:** The authors extend their appreciation for the support from the National Key Laboratory of Crop Genetic Improvement, Huazhong Agricultural University.

**Conflicts of Interest:** Authors declare that there are no conflict of interest.

#### **References**

