Identification of Genetic Loci Affecting Flag Leaf Chlorophyll in Wheat Grown under Different Water Regimes

Chlorophyll content of the flag leaf is an important trait for drought resistance in wheat under drought stress. Understanding the regulatory mechanism of flag leaf chlorophyll content could accelerate breeding for drought resistance. In this study, we constructed a recombinant inbred line (RIL) population from a cross of drought-sensitive variety DH118 and drought-resistant variety Jinmai 919, and analyzed the chlorophyll contents of flag leaves in six experimental locations/years using the Wheat90K single-nucleotide polymorphism array. A total of 29 quantitative trait loci (QTLs) controlling flag leaf chlorophyll were detected with contributions to phenotypic variation ranging from 4.67 to 23.25%. Twelve QTLs were detected under irrigated conditions and 18 were detected under dryland (drought) conditions. Most of the QTLs detected under the different water regimes were different. Four major QTLs (Qchl.saw-3B.2, Qchl.saw-5A.2, Qchl.saw-5A.3, and Qchl.saw-5B.2) were detected in the RIL population. Qchl.saw-3B.2, possibly more suitable for marker-assisted selection of genotypes adapted to irrigated conditions, was validated by a tightly linked kompetitive allele specific PCR (KASP) marker in a doubled haploid population derived from a different cross. Qchl.saw-5A.3, a novel stably expressed QTL, was detected in the dryland environments and explained up to 23.25% of the phenotypic variation, and has potential for marker-assisted breeding of genotypes adapted to dryland conditions. The stable and major QTLs identified here add valuable information for understanding the genetic mechanism underlying chlorophyll content and provide a basis for molecular marker–assisted breeding.


INTRODUCTION
Chlorophyll is the key element for photosynthesis, which captures light energy to drive electron transfer to its reaction center. Chlorophyll content is positively correlated with photosynthetic efficiency (Avenson et al., 2005), directly affecting the accumulation of photosynthates (Guo et al., 2008;Zhang et al., 2009a). Under abiotic stress situations such as drought, high temperature, salinization, and heavy metal presence, genotypes with higher chlorophyll content maintain higher photosynthetic capacity that helps to maintain higher yield achievement (Vijayalakshmi et al., 2010;Kumar et al., 2012;Ilyas et al., 2014;Talukder et al., 2014;Awlachew et al., 2016;Gupta et al., 2020;Bhoite et al., 2021;Borjigin et al., 2021). Photosynthetic activity in the flag leaves of wheat contributes about 50% to the grain yield (Verma et al., 2004;Zhu et al., 2016). Drought stress at the grain-filling stage is a common occurrence in wheat crops. This leads to accelerated degradation of chlorophyll in photosynthetic organs such as leaves, reduced photosynthetic rate, and decreased photosynthetic efficiency (Yang B. et al., 2016), hence lower fixation and assimilation of CO 2 (Yang D. et al., 2016) leading to restricted dry matter accumulation and grain development (Farooq et al., 2014). Therefore, the chlorophyll content in flag leaves is regarded as an indicator of drought resistance in wheat under drought stress (Farooq et al., 2014;Barakat et al., 2015). Molecular studies on the genetic regulation of flag leaf chlorophyll content are therefore of considerable significance for maintaining and improving yield potential under drought stress conditions. Synthesis and degradation of chlorophyll is a complex biological process, which not only involves many genes and cellular metabolic pathways, but is also readily affected by internal and external environments. Quantitative trait locus (QTL) analysis and gene cloning following construction of high-density genetic linkage maps is an effective way to study the genetics of chlorophyll (Verma et al., 2004;Thomas and Ougham, 2014;Rasheed et al., 2020). In rice, more than 900 QTLs affecting chlorophyll content have been identified by QTL mapping (Ye, 2016). More than 120 leaf color-related genes have been cloned , among which 14 were involved in chlorophyll synthesis. These included OsCAO1 encoding a chlorophyll oxygenase (Yang Y. et al., 2016); OsCHLH, OsCHLD, and OsCHLI encoding subunits of a magnesium-chelating enzyme (Jung et al., 2003;Zhou et al., 2012;Zhang et al., 2015); and YGL1 encoding a chlorophyll synthase (Liu et al., 2016). In addition, eight genes related to stay green were cloned in rice, including a DYE1-encoded light capture complex I subunit (Yamatani et al., 2018), EF8 encoding a HAP3 subunit of the HAP complex (Feng et al., 2014), and SGR that is involved in decomposition of chlorophyll (Morita et al., 2009). Some of these cloned genes have been successfully applied to rice breeding. Chen et al. (2020) found that overexpression of chloroplast gene D1 increased rice biomass by 20.6-22.9% and yield of transgenic rice by 8.1-21.0% under field conditions. Thus, identification of major QTLs/genes related to chlorophyll synthesis and degradation in grain crops could have application in wheat breeding.
As chlorophyll content is affected by water availability and environmental conditions, there are few stably expressed major QTLs (Yang D. et al., 2016). Most studies involved widely dispersed SSR markers and there are no reports on the application of QTL for chlorophyll content in wheat breeding. A few major stay green QTLs have been finemapped (Li et al., 2018;Wang et al., 2020a;Gupta et al., 2020). For example, the F 2 population involving early senescence mutant M114 with significantly reduced chlorophyll content in flag leaves was analyzed by BSR-Seq, and the els1 gene was located in the WGGB303-WGGB305 marker interval of 2BS, with 1.5 cM genetic distance (Li et al., 2018). Wang et al. (2020b) analyzed the inheritance of F 2 population constructed with premature senescence mutant LF2099 and Chinese Spring, and mapped the els2 gene to the marker interval of 2BIP09-2BIP14 on 2BL, and its genetic distance was 0.95 cM. There is no report on map-based cloning of genes regulating wheat chlorophyll content.
Genes Tackx4, Tabas1-B1, and TaPPH-7A contributing to chlorophyll content in wheat were identified by homologous cloning in wheat. Chang et al. (2015) cloned the Tackx4 allele encoding a cytokinin oxidase on chromosome 3A and validated it using a Jing411 × Hongmangchun 21 RIL population. A major QTL co-segregating with Tackx4 contributed 8.9-20.1% to chlorophyll content in four environments. Zhu et al. (2016) cloned Tabas1-B1 encoding 2-cys peroxiredoxin BAS1 on chromosome 2B and identified a major co-segregating QTL that contributed 9.0-19.2% of the variation in chlorophyll content in three environments. Wang et al. (2019) cloned TaPPH-7A encoding a pheophorbide hydrolase on chromosome 7A and found that it was closely related to the chlorophyll content of flag leaves in plants grown under drought stress. However, none of these genes was validated by transgenesis. Clearly, synthesis and degradation of chlorophyll is a complex biological process involving many genes, but currently only a few major QTLs and genes related to chlorophyll content have been reported in wheat. Thus, different research approaches and populations to map QTL are of value for a better understanding of the genetics of chlorophyll content.
In this study, the chlorophyll content of flag leaves was analyzed by QTL analysis of a DH118 × Jinmai 919 RIL population grown in six environments with different moisture conditions and validated in a Jinchun 7 × Jinmai 919 DH population to 1) identify stable QTLs that regulate chlorophyll content in flag leaves and 2) study the effects of contrasting moisture availability on the QTLs with the objective of obtaining markers for wheat breeding.

Plant Materials and Plot Design
The populations with Jinmai 919 as a same parent included 165 F 10 RILs from cross DH118 × Jinmai 919 (DJ) and 168 doubled haploid (DH) lines from Jinchun 7 × Jinmai 919 (JJ). DH118, a high-yielding variety selected for irrigated conditions by the Institute of Wheat Research, Shanxi Agricultural University, has dark green leaves and high chlorophyll content ( Figure 1A). Jinmai 919 developed by the Institute of Wheat Research, Shanxi Agricultural University, has strong drought resistance, light green leaves, and good stay green characteristics ( Figure 1A). Bred by the Institute of Maize Research, Shanxi Academy of Agricultural Sciences, Jinchun 7 is also a high-yielding variety for irrigated conditions. The DJ population was used for QTL mapping, and the JJ population was used for validating QTLs identified in the mapping population.

Phenotypic Evaluation and Data Analysis
Ten plants flowering on the same day were randomly selected from each line at 10 days after flowering. The chlorophyll contents of flag leaves were measured using a SPAD-502 (Konica-Minolta, Japan) chlorophyll meter at 7:00 to 10:00 h. Each leaf was measured three times-at the base, mid-region, and tip-and the average value was used for analysis (Yang B. et al., 2016). Average values were also determined for each environment.

High-Density Genetic Linkage Map Construction and QTL Mapping
DNA was extracted from all RILs and DH lines and respective parents using the CTAB method (Vijayalakshmi et al., 2010). The RIL population was genotyped with the Infinium wheat SNP 90K iSelect assay (Illumina Inc., San Diego, CA, USA) developed by the International Wheat SNP Consortium . IciMapping v4.0 (https://www.isbreeding.net) was used to construct a high-density genetic linkage map (Li et al., 2021). SNP markers with no recombination were placed into a single bin using the "BIN" function in IciMapping V4.0. The final markers were chosen with a minimum percentage of missing data and sorted into different groups with LOD thresholds ≤8 by the "Grouping" function in JoinMap 4.0 (Li et al., 2021).
The QTLs were detected using WinQTLCart version 2.5 (https://brcwebportal.cos.ncsu.edu/qtlcart/WQTLCart.htm) based on the composite interval mapping method. QTLs were proclaimed significant at logarithm of odds (LOD) scores >2.5. The QTL contributing more than 10% to phenotypic variation in a certain environment (including BLUP) and detected in three environments (including BLUP) was considered as a stable and major QTL. QTLs less than 1 cM apart or sharing common flanking markers were treated as a single locus. The QTLs were named according to McCouch et al. (1997). The closest marker sequences flanking QTL were compared with the Chinese Spring reference genome sequence in the wheat multiomics website database (http://wheatomics.sdau.edu.cn/jbrowse-1.12. 3-release/?data=Chinese_Spring1.0) to determine the physical locations of the QTL.

Marker Development and Validation of Major QTLs
To develop kompetitive allele specific PCR (KASP) tags from the peak marker SNP sequence of the major QTLs, two specific primers (F1/F2) and a universal primer (R) were designed for each SNP. An F1 tail that could bind to induce FAM fluorescence and an F2 tail that could bind to induce HEX fluorescence were added to the specific sequences. KASP primers were designed by Polymarker (http://www.polymarker.info/) and synthesized by Beijing Jiacheng Biotechnology Co., Ltd. (Supplementary Table  S1). The developed KASP markers were used in PCR to detect previously identified QTLs in the JJ population as a means of validation. Following genotyping, the validation population was divided into two groups and differences in chlorophyll content of flag leaves between the groups were assessed by t-tests in SAS V8.0.

Gene Prediction Within QTL
Genes within the target region of major QTL were obtained using the genome browser (JBrowse) on the Triticeae Multi-omics website (http://wheatomics.sdau.edu.cn/). The GO (gene ontology) database and R package cluster profiler were applied for functional annotation and enrichment analysis of genes in the QTL regions. Identification of orthologs in wheat and rice was conducted using the Triticeae-Gene Tribe website (http://wheat. cau.edu.cn/TGT/). The expVIP public database (http://www. wheat-expression.com/) was used to search for expression data of genes in eight tissues and organs, perform log2 conversion processing, and analyze the expression patterns of candidate genes.

Analysis of Phenotypic Data
The chlorophyll contents of flag leaves of DH118 and Jinmai 919 ranged from 52.16 to 60.22 and 48.80 to 58.42, respectively, across the six environments. The chlorophyll content of DH118 was consistently higher than that of Jinmai 919, and the difference was significant in E1 and E6 (p <0.05), and highly significant in E2 and E5 (p <0.01) ( Table 1). The correlation of chlorophyll contents among different environments for the RIL population was highly significant (p <0.01), and correlation coefficients ranged from 0.303 to 0.711 (Supplementary Table S2). The H 2 of chlorophyll content was 0.90, indicating that chlorophyll content was largely determined by genetic factors. Principal component analysis showed that environmental factors had considerable influence on phenotypic values, and drought stress increases the phenotypic variation ( Figure 1B). Chlorophyll content of the RIL population was mostly between the two parents under E2, E3, E4, E5, and E6 environments, showing a continuous distribution. Bidirectional transgressive segregation was also observed in chlorophyll content among the RIL population under E1 condition ( Table 1).

Linkage Map Construction
A high-density genetic linkage map for the RIL population was constructed by using Wheat90k SNP chip. The total length of the map was 5,858.63 cM with an average genetic distance of 1.65 cM, including 3,553 SNP markers and covering all 21 chromosomes (   Table S3). In addition, the average chlorophyll content of lines with only Qchl.saw-5A.3 allele in RIL population was higher than that of other lines with only one favorable allele ( Figure 2B) Figure S1). Analysis of gene expression in various tissues    identified 18 candidate genes related to chlorophyll metabolism ( Table 4). These 18 genes were divided into three categories according to their function. The first category was related to the composition of chloroplasts. TraesCS5A02G420700 related to chloroplast thylakoid membrane, and TraesCS5A02G377000 related to chloroplast membrane formation and the homologous gene TraesCS5A02G423000 of pRRFNR14 (Os03g0784700) in rice involved in the process of chloroplast composition (Aoki and Ida, 1994). The second category was related to eight new genes of chlorophyll photosynthesis, including TraesCS5A02G414400, TraesCS5A02G378700 (OsLOX 2 ), TraesCS5A02G373600, TraesCS5A02G424100, TraesCS5A02G376700, TraesCS5A02G369500, TraesCS5A02G392300, and TraesCS5B02G356300 (OsUgp1) E et al., 2015). These genes participated in photosystem I reaction center subunit III, ATP binding, metal ion binding, and transferase activity. The third kind of genes responded to drought stress by regulating photorespiration, mediating auxin response, and participating in the regulation of ABA signal transduction pathway, such as rice homologous gene GLO1, OsIAA13/OsIAA1, OsSAPK8, and OsUBC9 (Thakur et al., 2001;Zhang et al., 2012;Xu et al., 2013;E et al., 2015).
We also identified three novel genes TraesCS5A02G411200, TraesCS5A02G374500, and TraesCS5A02G426100 that responded to drought stress by redox reaction, activation of enzyme activity, and ATP binding ( Table 4).

Comparison with Previous Research Results
According to reviews by Gupta et al. (2017Gupta et al. ( , 2020, a total of 82 QTLs controlling chlorophyll content were identified in previous studies. These QTLs were distributed across all 21 chromosomes and explained 2.7-59.1% of the phenotypic variation, but most of these QTLs were different. The reasons could be due to 1) different methods of chlorophyll measurement that cause differences in phenotypic values, e.g., some studies used a spectrophotometer (Zhang et al., 2009b) and others used a chlorophyll meter, leading to differences in QTL analysis results (Bhusal et al., 2018); 2) chlorophyll content is a complex quantitative trait and genes controlling leaf chlorophyll are expressed differently at different developmental stages (Yang D. et al., 2016), and different measurement periods will inevitably lead to different identified genes; 3) due to different types of populations and molecular markers, it is not easy to compare results across different genetic backgrounds.
In this study, 29 QTLs controlling chlorophyll content in flag leaves were located on 12 chromosomes, most of which were A and B genome chromosomes with only three detected in the D genome. Similar results were reported in previous studies (Zhang et al., 2009b;Yang D. et al., 2016). We detected four stably expressed major QTLs on chromosomes 3B (Qchl.saw-3B.2), 5A (Qchl.saw-5A.2 and Qchl.saw-5A.3), and 5B (Qchl.saw-5B.2), with contribution rates of 5.28-23.25% to the variation in chlorophyll content. These QTLs still need further validation before application in marker-assisted selection (Ahmed et al., 2021).
Fourteen, seven, and nine QTLs for chlorophyll content were located on chromosomes 3B, 5A, and 5B, respectively, in previous studies ( Table 5). The three major QTLs controlling chlorophyll content of flag leaves identified in our study were consistent with results of previous studies. The major QTL Qchl.saw-3B.2 on chromosome 3B was in the interval 52.83-54.75 Mb. Kumar et al. (2010) reported a major QTL QSg.bhu-3B for flag leaf senescence in the same region, explaining 17.9% of the variation in stay green phenotypic, and Puttamadanayaka et al. (2020) reported QChl.iari_3B that controlled chlorophyll content. The QTLs in our study spanned shorter physical distances and are therefore more conducive for gene cloning. Qchl.saw-5A.2 was in the range 569.54-582.38 Mb. Puttamadanayaka et al. (2020) reported QChl.iari_5A for chlorophyll content spanned by AX-94531685 (567.52 Mb) and AX-94726381 (582.96 Mb). In the same region, Wang et al. (2017) detected three major QTLs controlling 1,000grain weight, and their adjacent markers were BS00073670_51, wsnp_Ex_c1138_2185522, and Tdurum_contig71499_211, respectively. Yang et al. (2019) cloned a TaGL3-5A allele that conferred larger grain size based on homology with rice. Many studies have confirmed the high correlation between chlorophyll content and yield-related traits (Zhang et al., 2009b;Vijayalakshmi et al., 2010). Although there was no investigation of yield-related traits in this study, we have colocated QTL/genes for chlorophyll content, 1,000-grain weight, and grain size in the same interval with previous studies and confirmed the correlation between chlorophyll content and yieldrelated traits. The major QTL Qchl.saw-5B.2 on chromosome 5B was located in the interval 536.05-536.68 Mb, which coincided with chlorophyll content QTL Qspad.acs-5B.4 spanned by Xwmc415 and Xwmc508 reported by Yang et al. (2016). Qchl.saw-5A.3 with the strongest genetic effect in our study was in the chromosome 5A interval 586. . Given no previous report of gene for chlorophyll content in this interval, Qchl.saw-5A.3 is a novel QTL.

Effect of Environment on Expression of QTL for Chlorophyll Content
Synthesis and degradation of chlorophyll are complex biological processes and regulation likely differs under different water regimes (Yang D. et al., 2016). Under irrigated conditions, higher chlorophyll content could ensure fixation of more photosynthetic assimilates (Zhang et al., 2009b). Under drought stress conditions, stay green is closely related to higher yield (Verma et al., 2004;Thomas and Ougham, 2014). Drought-tolerant genotypes usually have higher chlorophyll content, and chlorophyll degrades more slowly under drought stress (Kumar et al., 2012;Lopes and Reynolds, 2012).
In this study, QTL analysis of chlorophyll content in flag leaves under irrigated and dryland (drought stressed) conditions was made using a RIL population derived from a cross between a Spanning markers were used to locate positions in the physical map if the certain markers failed to be located on the physical map. The physical locations of some markers were not available leaving the physical location as a single marker.   Table 3). The number of QTLs under drought stress was much more than that under well-watered conditions, showing that environmental stress could induce to express genes originally keeping silent under irrigated conditions to reduce plant damages from environmental stress (Yang et al., 2007;Guo et al., 2008;Vijayalakshmi et al., 2010;Christopher et al., 2018). In addition, it was not difficult to find that there were some differences in QTL mapping data between the well-watered and drought stress, which implied that there were different QTL expression patterns under different water regimes (Yang et al., 2007;Yang B. et al., 2016;Xu et al., 2017;Hassan et al., 2018;Christopher et al., 2021). It also implies that different QTLs should be used for marker-assisted breeding of wheat varieties under irrigated conditions and dryland. For example, the Qchl.saw-3B.2 detected in this study was not only confirmed to be stably expressed without the influence of genetic background, but also detected under several well-watered conditions, which may be more suitable for molecular marker-assisted selection of varieties under irrigated conditions. In addition, Kumar et al. (2012) and Hassan et al. (2018) considered that the major QTL detected under drought stress may contain genes that contribute to drought resistance and have the application potential to increase yield under drought stress. In our study, three major QTLs (Qchl.saw-5A.2, Qchl.saw-5A.3, and Qchl.saw-5B.2) were detected in drought stress environments. Qchl.saw-5A.3 could be detected in all drought stress environments (E4, E5, and E6), and the contribution rate to phenotype was 6.04-23.25% (Table 3), which may be more suitable for marker-assisted selection breeding of drought-resistant varieties. In short, this study used high-density chips for QTL mapping, and the SNP and KASP markers of four major QTLs could be applied to the next development of molecular markers under different water conditions.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
SY, LQ, and JuZ designed the experiment and wrote the article. BY and XW carried out the experiments. HW and YF analyzed the data. JiZ, BW, XZ, and CY did the field experiments. All authors contributed to the article and approved the final article to publish.