Mapping of Candidate Genes for Nitrogen Uptake and Utilization in Japonica Rice at Seedling Stage

Nitrogen (N) is one of the essential nutrients for the growth and development of crops. The adequate application of N not only increases the yield of crops but also improves the quality of agricultural products, but the excessive application of N can cause many adverse effects on ecology and the environment. In this study, genome-wide association analysis (GWAS) was performed under low- and high-N conditions based on 788,396 SNPs and phenotypic traits relevant to N uptake and utilization (N content and N accumulation). A total of 75 QTLs were obtained using GWAS, which contained 811 genes. Of 811 genes, 281 genes showed different haplotypes, and 40 genes had significant phenotypic differences among different haplotypes. Of these 40 genes, 5 differentially expressed genes (Os01g0159250, Os02g0618200, Os02g0618400, Os02g0630300, and Os06g0619000) were finally identified as the more valuable candidate genes based on the transcriptome data sequenced from Longjing31 (low-N-tolerant variety) and Songjing 10 (low-N-sensitive variety) under low- and high-N treatments. These new findings enrich the genetic resources for N uptake and utilization in rice, as well as lay a theoretical foundation for improving the efficiency of N uptake and utilization in rice.


Introduction
Rice originated in tropical and subtropical regions and is one of the most important food crops in the world [1].Since the beginning of the 20th century, the continuous increase in fertilizer, especially nitrogen (N) fertilizer, has made a great contribution to the improvement of crop yields worldwide [2].The moderate application of N fertilizer can increase yields and improve economic efficiency, but farmers tend to apply excess N fertilizer in crop production [3,4].This tendency has led to significant reductions in N uptake and utilization and also caused a number of ecological problems [5], as well as the aggravation of crop failure, pests, and disease [6,7].Therefore, reducing the amount of N fertilizer, improving the uptake and utilization of N fertilizer, and ensuring high yields have become the main issue that rice breeders need to study.
In recent years, there have been plenty of studies on N uptake and utilization in plants [8][9][10][11].The N-efficient genotype in cotton, CCRI-69, showed higher N concentration and N accumulation at both N levels [12].TabZIP60 is a negative regulator of wheat growth and N utilization.The overexpression of TabZIP60 can inhibit wheat growth, while reducing TabZIP60 expression by RNAi interference can improve wheat yield and N utilization efficiency (NUE) partially; the expression of NADH-dependent glutamate synthase

Plant Material and Genotyping
The panel of the natural population is composed of 295 japonica rice germplasms collected mainly from China, Korea, Japan, and Russia.The DNA of 295 japonica rice varieties was extracted, and high-throughput sequencing was performed using an Illumina HiSeq 2000.SNPs were detected using the GATK4.2.1.0software toolkit [30], and a total of 788,396 SNPs were obtained for GWAS analysis by removing SNPs with the lowest allele frequencies, below 0.05, and deletion rates greater than 20% in the initially obtained population SNPs [31].The genetic structure of the population was calculated using AD-MIXTURE1.3.0 software [32].The genetic structure of this population, among others, was demonstrated in a pre-publication article from our lab [31].

Phenotypic Data
Two hundred plump seeds from each of the 295 japonica rice germplasms were treated in a 48 • C oven for 48 h to break dormancy.Then, the seeds were sterilized with a 0.01% sodium hypochlorite solution for 30 min, washed with sterile water three times to remove the sterilization solution residue from the seed surface, placed in an incubator at 31 • C, and incubated in the dark for 2 days.After germination, 64 germinating seeds of each variety with a uniform shoot length were selected and divided equally into two parts.One part would be treated with low N, and the other part would be treated with high N.The germinated seeds were placed in 96-well PCR plates, and one seed was placed in each well.The PCR plates with germinated seeds were then placed in plastic boxes and transferred to a greenhouse to grow seedlings (23.8 • C/22.4 • C, 10 h during the day/14 h at night).Two N fertilization treatments, low N (8 ppm) and high N (40 ppm) [33,34], were set up using NH 4 NO 3 as the N source.All germinated seeds were first incubated under high N for 5 days (64 grains per variety).Then, 32 seeds of each variety continued to grow in a high-N solution for 21 days, and the other 32 seeds per variety were cultured in a low-N solution for the same 21 days [33].The hydroponic nutrient solution was prepared by referring to the conventional formula specified by IRRI.The nutrient solution was replaced every 7 days.After 21 days of incubation under the two different N treatments, the relevant indexes were measured.
A total of 295 seedlings of japonica rice obtained after low-and high-N treatments were randomly taken from three plants each.The leaves and roots of the rice seedlings obtained after low-and high-N treatments were cut off, washed with ultrapure water, killed in an oven at 105 • C for 30 min, and dried in a constant-temperature oven at 80 • C for 12 h to a constant weight, and then the dry matter weight of the leaves and roots was weighed.The dried samples were put into 2 mL centrifuge tubes.Three to four 3 mm steel balls were added to each tube.Then, the samples were ground into powder using a sample-grinding machine.About 0.05 g of powder was weighed.The weight of the weighed sample was recorded accurately.The N content of the sample was determined using an elemental analyzer (Primacs SNC 100-IC-E), and N accumulation was calculated.Three replicates were set up for the above treatments, and the traits' relative values were calculated.
Relative character value = Low-N treatment character value/High-N treatment character value.

GWAS
GWAS was performed using the MLM method in Tassel 5.0 software [35].The GWAS of the phenotypes and SNP genotypes was performed based on the population structure (Q matrix) and the calculated genetic relationship between any two individuals (Kinship matrix).The threshold for SNPs significantly associated with the traits was set at p < 5.46 × 10 −6 , determined by genetic type 1 error calculator (GEC; http://statgenpro.psychiatry.hku.hk/gec/, accessed on 10 May 2023), which calculates the effective number of independent markers [36].Manhattan plots and Q-Q plots were drawn for the association analysis results using the CMplot package in the R language.If two or more significant SNPs were located in the same linkage disequilibrium interval, these SNPs were considered the same QTL, and the SNP with the lowest p value was defined as the lead SNP of this QTL.To identify the lead SNP with the lowest p value, redundant SNPs were filtered at the minimum distance interval.The annotation information of genes within the QTL interval was checked against the Ensembl genome database.

Linkage Disequilibrium (LD) Analysis
The paired R2 values between any two SNPs within the ±2 Mb interval of the leading SNP were calculated using LDBlockShow [37].The average of the top 10% paired R2 values in each interval was calculated and recorded as the background value of LD attenuation.The value obtained by adding 0.2 to the background value was considered the attenuation interval of LD.

Haplotype Analysis
The location interval of each gene was identified according to the China Rice Data Center (https://www.ricedata.cn/,accessed on 23 June 2023).Non-synonymous mutant SNPs within the exonic region of the gene were extracted using the RiceSNP SeekDatabase website (https://snp-seek.irri.org/,accessed on 23 June 2023) and used to identify the haplotype together with SNPs within the range of the first 2 kb of the start codon.

RNA-seq
Two comparison groups (H71 vs. L71 and H284 vs. L284) were set up.Specifically, 71 is Longjing 31 (a low-N-tolerant rice variety), 284 is Songjing 10 (a low-N-sensitive variety), and H and L represent high N and low N, respectively.Firstly, sixteen seedlings each of Longjing 31 and Songjing 10 were cultured under the high-N treatment (40 ppm) for 5 days; then, eight seedlings were kept in the same solution for 7 days, while the other eight seedlings were transplanted into the low-N treatment (8 ppm) for 7 days.After 12 days, 12 root samples (three replicates of each treatment) were collected.Total RNA was extracted from the 12 samples using the TransZol Up RNA Kit (TransGen Biotech, Beijing, China).Complementary DNA was synthesized from total RNA using the HiFiScript cDNA Synthesis Kit (CWBio, Beijing, China).An Illumina library was constructed according to the manufacturer's instructions (Illumina, San Diego, CA, USA).High-throughput RNA sequencing was performed using the Illumina HiSeq 2500 platform.HISAT v2.1.0was adopted to construct the index and to map clean reads to the reference genome.The rice reference genome data used was Os-Nipponbare-Reference-IRGSP-1.0.The gene alignment and FPKM (Fragments Per Kilobase of transcript per Million fragments mapped) were calculated by using featureCounts v1.6.2 [38].p < 0.05 and |log 2 FC| > 0.585 were adopted as the thresholds to identify the differentially expressed genes between any two comparative groups using edgeR v3.24.3 [39].Other specific experimental methods and transcriptomic data are shown in previously published articles from our laboratory [33].

Candidate Gene Prediction
All genes within the LD attenuation interval of every lead SNP in the GWAS results were integrated for candidate gene prediction; the phenotypic data of the 295 japonica rice varieties were combined with the SNP data to analyze all genes by haplotype, and the genes with significant phenotypic differences among different haplotypes among these genes were identified as candidate genes.Then, the genes with significant phenotypic differences among different haplotypes and |log 2 FC| > 0.585, p < 0.05 in the low-N-tolerant variety, Longjing 31, were screened as more valuable candidate genes by combining the gene expression from RNA-seq.

Quantitative Real-Time PCR
A qRT-PCR assay was performed using the same total RNA used for the transcriptome data analysis of the two comparison groups mentioned above (H71 vs. L71 and H284 vs. L284).The first-strand cDNA (10 µL) was synthesized according to the instructions for the PrimeScript™ RT Master Mix (Takara Biomedical Technology (Beijing) Co., Ltd., Beijing, China).Primer5.0 was used to design specific primers for this assay (Table S1).The BlazeTaqTM SYBR Green qPCR Master Mix 2.0 (GeneCopoeia, Guangzhou, China) was used, and reactions were run on Roche Lightcycler 96 real-time PCR equipment in accordance with the manufacturer's instructions (Roche Medical Instruments, Basel, Switzerland).Each sample had three technical replicates, and Actian1 was used as an internal control.Finally, the expression of 5 candidate genes was calculated using the 2 −∆∆CT method.

Data Analysis
The density distribution of each trait was plotted using "ggplot2" in the R4.2.3.The phenotypic data were analyzed using IBM SPSS Statistics 25.0 software (SPSS Inc., Chicago, IL, USA) for correlation analysis and descriptive statistics, including the mean, extreme deviation, and coefficient of variation.

Phenotypic Data Analysis
The coefficients of variation of leaf N content (YN), leaf N accumulation (YNAA), root N concentration (RN), and root N accumulation (RNAA) of 295 japonica rice germplasm seedlings ranged from 20.76% to 46.25% and 17.43% to 35.20% under the high-and low-N treatments, respectively (Table S2).The coefficients of variation of the relative value of root N accumulation (RNAAR) and the relative value of leaf N content (YNR) were the largest and smallest under the low-and high-N treatments, respectively.The coefficients of variation of RN and RNAA under the low-N treatment were larger than those under the high-N treatment, while YN and YNAA were not significantly different between high-and low-N treatments.The phenotypic values of all traits showed abundant variation and were approximately normally distributed (Figure 1).
Under the low-N condition, all traits showed very significant positive correlations, with correlation coefficients ranging from 0.239 to 0.797, except for the significant positive correlation between YNAA and RN.Under the high-N condition, YN and YNAA, RN and RNAA, and YNAA and RNAA showed very significant positive correlations, with correlation coefficients of 0.338, 0.647, and 0.532, respectively.In relative values, YN and YNAA, RN and RNAA, and YNAA and RNAA showed very significant positive correlations, of which the correlation coefficients are 0.534, 0.861, and 0.287, respectively.There were very significant positive correlations between YN and YNAA, RN and RNAA, and YNAA and RNAA, whose correlation coefficients are 0.534, 0.861, and 0.287, respectively (Table S3).

QTL Localization
A total of 75 SNPs that were significantly associated with traits related to N uptake and utilization in rice seedlings were localized (Table S4 and Figure 2).Among them, three SNPs located in the first and eighth linkage groups were significantly associated with

Candidate Gene Mining
The 75 QTL intervals contained 811 genes (Table S5).Among these genes, 232, 39, 8, and 2 genes had two, three, four, and five haplotypes among 295 japonica rice germplasms, respectively.Among the 281 genes with different haplotypes, 40 genes that had different

Candidate Gene Mining
The 75 QTL intervals contained 811 genes (Table S5).Among these genes, 232, 39, 8, and 2 genes had two, three, four, and five haplotypes among 295 japonica rice germplasms, respectively.Among the 281 genes with different haplotypes, 40 genes that had different haplotypes showed significant phenotypic differences in different traits among the 295 japonica rice germplasms.Therefore, it is hypothesized that these 40 genes might be associated with N uptake and utilization.
A GO enrichment analysis of these 40 genes with significant haplotype differences showed that they could be enriched in 39 functional groups, including 21 biological processes, 3 cellular components, and 15 molecular functions.Circumnutation, regulation of endocytosis, multicellular organismal movement, N,N-dimethylaniline monooxygenase activity, NADP binding, flavin adenine dinucleotide binding, and Violaxanthin de-epoxidase activity were significantly enriched (Figure 3A).A KEGG enrichment analysis of these 40 candidate genes was also performed, and the results showed that a total of six pathways were enriched (Figure 3B).
Genes 2024, 15, 327 9 of 18 haplotypes showed significant phenotypic differences in different traits among the 295 japonica rice germplasms.Therefore, it is hypothesized that these 40 genes might be associated with N uptake and utilization.
A GO enrichment analysis of these 40 genes with significant haplotype differences showed that they could be enriched in 39 functional groups, including 21 biological processes, 3 cellular components, and 15 molecular functions.Circumnutation, regulation of endocytosis, multicellular organismal movement, N,N-dimethylaniline monooxygenase activity, NADP binding, flavin adenine dinucleotide binding, and Violaxanthin de-epoxidase activity were significantly enriched (Figure 3A).A KEGG enrichment analysis of these 40 candidate genes was also performed, and the results showed that a total of six pathways were enriched (Figure 3B).

Expression Analysis of Candidate Genes
Transcriptome analysis was performed to elaborate on the expression differences of these 40 genes (Table S6).There were four significantly down-regulated genes and one significantly up-regulated gene in Longjing 31 (low-N-tolerant variety) and two significantly up-regulated genes in Songjing 10 (low-N-sensitive variety) between the treatment (low N) and control (high N).Finally, five genes, Os01g0159250, Os02g0618200, Os02g0618400, Os02g0630300, and Os06g0619000, which were differentially expressed in the low-N-tolerant variety, were selected as the most valuable candidate genes in this study (Table 1).The expression levels of these five candidate genes in Longjing 31 and Songjing 10 under low-and high-N treatments were verified by qRT-PCR.The qRT-PCR results of these five candidate genes were consistent with the expression trends of RNAseq data (Figure 4).The linkage disequilibria of these six candidate genes are illustrated in Figures 5A,D, 6A and 7A, respectively.

Expression Analysis of Candidate Genes
Transcriptome analysis was performed to elaborate on the expression differences of these 40 genes (Table S6).There were four significantly down-regulated genes and one significantly up-regulated gene in Longjing 31 (low-N-tolerant variety) and two significantly up-regulated genes in Songjing 10 (low-N-sensitive variety) between the treatment (low N) and control (high N).Finally, five genes, Os01g0159250, Os02g0618200, Os02g0618400, Os02g0630300, and Os06g0619000, which were differentially expressed in the low-N-tolerant variety, were selected as the most valuable candidate genes in this study (Table 1).The expression levels of these five candidate genes in Longjing 31 and Songjing 10 under low-and high-N treatments were verified by qRT-PCR.The qRT-PCR results of these five candidate genes were consistent with the expression trends of RNA-seq data (Figure 4).The linkage disequilibria of these six candidate genes are illustrated in Figures 5A,D, 6A and 7A, respectively.

Haplotype Analysis of Candidate Genes
Among the 295 japonica rice germplasms, two non-synonymous SNPs were present in the promoter region of Os01g0159250, forming three different haplotypes (Figure 5B).The three haplotypes showed significant phenotypic differences that existed in two traits, HYNAA and LYNAA (Figure 5C).Os02g0618200 and Os02g0618400 were located in the same QTL interval, and both of them had non-synonymous SNPs present in the CDS region.The number of non-synonymous SNPs and haplotypes of the two genes was three and two and four and three, respectively (Figure 6B,D).Four different haplotypes of Os02g0618200 showed significant phenotypic differences in HRNAA (Figure 6C), while three different haplotypes of Os02g0618400 showed significant phenotypic differences in HYNAA and HRNAA (Figure 6E).Both Os02g0630300 and Os06g0619000 had only one non-synonymous SNP in the CDS region and thus had two different haplotypes (Figures 5E and 7B).Two haplotypes of Os02g0630300 showed significant phenotypic differences in three traits, LRN, RNR, and RNAAR (Figure 7C).Two haplotypes of Os06g0619000 showed significant phenotypic differences in HYNAA, LYNAA, and LRNAA (Figure 5F).

Discussion
Among all fertilizers, N fertilizer has been used as a necessary fertilizer to increase crop yields, whereas the excessive use of N fertilizer has caused many adverse effects.For example, the excessive application of N fertilizer has led to the significant accumulation of nitrate-N content in the 2-4 cm soil layer, causing groundwater and surface water hazards, which has become a widespread global problem [43]; the natural transformation of N in the soil reduces the pH of the soil and destroys the soil health [44].Studies have shown an exponential increase in the coastal dead zones of the world's oceans and seas, and one of the important reasons for the dead zones is the runoff of fertilizers from rivers, which affects the activity of microorganisms and depletes dissolved oxygen in oxygenated water, resulting in the creation of dead zones in coastal areas [45].At the same time, excessive nitrogen fertilizer is not only harmful to the environment but also leads to an increase in internode length, a decrease in internode thickness, a decrease in bending resistance, and a significant increase in the number of collapses in rice stalks [6].The localization study of N-uptake-and utilization-related genes in this study yielded several excellent genetic resources, which lays an important foundation for improving N fertilizer application and enhancing N uptake and utilization efficiency in the future.
N content in rice plants is usually represented in terms of concentration, and usually, the concentration is used to diagnose the plant's nutritional adequacy, deficiency, or excess.Thus, the N concentration visually reflects the plant's uptake of N [46].The product of the weight of dry matter or grains and N content represents the amount of nutrient accumulation, which can be used as an indicator of N uptake and utilization and is directly related to crop yield [47]. N content and N accumulation in plants are closely related to N uptake and utilization [5].For example, in order to study the effect of biochar on N uptake and utilization in rice, the total N uptake, internal N utilization efficiency, and grain yield in six seasons were measured, and the dry weight, N content, and N accumulation of straw, rachis, and leafy and leafless spikelets were measured in the evaluation of internal N uptake and utilization efficiency in rice plants [48].In an experiment to investigate the relationship between the N use efficiency and yield of organic rice varieties, the N uptake and use efficiency of rice were assessed in terms of dry matter accumulation, N content, and N uptake (N accumulation), and the experimental conclusion was that the rice yield was significantly and positively correlated with N uptake [49].
N uptake and utilization efficiency are important research topics in rice cultivation worldwide [3,50].In recent years, a lot of important genes related to N uptake and utilization in rice have been localized and identified by GWAS, such as OsNLP4, OsNRT1.1B,OsNPF6.1,OsNAC42, OsTCP19, etc. [22][23][24]51,52].In this study, the N content and N accumulation in the roots and leaves of 295 japonica rice varieties at the seedling stage were measured under low-and high-N treatments, respectively.The above phenotypic data were used for GWAS to locate QTLs associated with N uptake and utilization in rice.The localized QTLs were mainly concentrated in linkage groups 4, 6, 8, 10, and 11, which is consistent with some previous studies [33,53,54].Transcriptome analysis has been used to study the response of many plant species to various environmental stresses [55][56][57][58][59].With the development of high-throughput sequencing technology and the continuous improve-ment of transcriptome databases, the combination of transcriptome data with traditional QTL mapping or GWAS has been widely used in the prediction of candidate genes.For example, by combining a population segregation analysis with transcriptomics, four candidate genes related to rice blast resistance were obtained [60]; by integrating GWAS and transcriptome data, one N utilization efficiency (NUE)-related candidate gene, OsNAC68, was identified in rice [61].Our previous studies combining GWAS/BSA with transcriptome data identified many salinity tolerance candidate genes (OsIRO3, OsSAP16) [31,62].In this study, we obtained 40 haplotypic differential genes within the interval localized by GWAS.By sequencing the transcriptomes of low-N-tolerant and low-N-sensitive varieties, five genes differentially expressed in low-N-tolerant varieties were identified as more valuable candidate genes.And according to the phenotypic differences between different haplotypes, five favorable alleles were identified, namely, Hap2 of Os01g0159250, Hap1 of Os02g0618200, Hap2 of Os02g0618400, Hap1 of Os02g0630300, and Hap2 of Os06g0619000.Varieties containing the above alleles showed excellent phenotypic variation.The study of these five genes in relation to N uptake and utilization in rice has not been reported.Varieties containing the above alleles exhibited excellent phenotypes, so the approach integrating GWAS and transcriptome data is an effective way to mine genes related to N uptake and utilization in rice.
Among the five most valuable candidate genes in this study, Os02g0618200, Os02g0618400, and Os02g0630300 have been cloned.Os02g0618200, whose gene symbol is OsPRR1, encodes a response regulator that is expressed in multiple tissues of rice and has a circadian rhythm of expression in leaves, roots, tiller shoots, and endosperm.The expression of OsPRR1 suppresses the expression of OsCCA1, which mediates rice tiller and spike development [41].Os02g0618400 (OsMPS) encodes an R2R3-like MYB transcription factor.OsMPS is expressed in many tissues, including the aboveground parts of seedlings, roots, pollen, vascular tissues of glumes, endosperm sheaths, and leaves, but not in the endosperm, stigma, ovary, or embryo; OsMPS can regulate adaptive growth by integrating signals between different plant hormones and the environment [42].Os02g0630300 (OsGA2ox9) encodes a gibberellin 2-oxidase.OsGA2ox9 is transcribed in a variety of tissues, including rice roots, leaves, pollen, and seeds [40].It was shown that OsGA2ox9 knockout lines germinate spikes under continuous rainy weather, whereas seed dormancy was increased in overexpressing lines, but this was restored by externally applied GA3; OsAmy (α-amylase isozyme 3D) expression was significantly increased in OsGA2ox9 knockout seeds, and the amount of glucose and sucrose in the seeds of the knockout lines was found to be increased; thus, OsGA2ox9 may regulate the transcription of α-amylase-encoding genes through GA signaling, which promotes the hydrolysis of starch to soluble sugars and inhibits ABA signaling, leading to spike germination [40].Two genes, Os01g0159250 and Os06g0619000, were not cloned.Os01g0159250 is a hypothetical protein.The expression product of Os06g0619000 is a protein containing the structural domain of NB-ARC, and proteins containing the structural domain of NB-ARC are usually associated with plant adversity responses [63][64][65].However, the responses of the five candidate genes to N have not been studied in depth.The two candidate genes that have not yet been cloned should be further investigated for their relevance to N uptake and utilization.And among the cloned genes, Os02g0618200 (which mediates tillering and spike development), Os02g0618400 (which regulates rice growth), and Os02g0630300 (which inhibits ABA signaling) may be related to N regulation, and their molecular mechanisms in N uptake and utilization should be further investigated in depth in the future.

Conclusions
A total of 75 QTLs were obtained using GWAS, which contained 811 genes.Of the 811 genes, 281 showed different haplotypes, and 40 genes had significant phenotypic differences among different haplotypes.Five genes (Os01g0159250, Os02g0618200, Os02g0618400, Os02g0630300, and Os06g0619000) that differentially expressed in the low-N-tolerant variety (Longjing 31) were finally identified as the more valuable candidate genes associated with N uptake and utilization among these 40 genes.However, the molecular mechanisms involved in the regulation of N uptake and utilization by the above candidate genes still need to be further investigated.

Figure 1 .
Figure 1.Phenotypes of 295 japonica rice varieties under high-and low-N treatments.(A) HRN (root N content under high-N treatment); LRN (root N content under low-N treatment).(B) RNR (relative value of root N content under low-and high-N treatments).(C) HRNAA (root N accumulation under high-N treatment); LRNAA (root N accumulation under low-N treatment).(D) RNAAR (relative value of root N accumulation under low-and high-N treatments).(E) HYN (leaf N content under high-N treatment); LYN (leaf N content under low-N treatment).(F) YNR (relative value of leaf N content under low-and high-N treatments).(G) HYNAA (leaf N accumulation under high-N treatment); LYNAA (leaf N accumulation under low-N treatment).(H) YNAAR (relative value of leaf N accumulation under low-and high-N treatments).Lowercase a and Lowercase b represent the significance between treatments at the 0.05 level.

Figure 1 .
Figure 1.Phenotypes of 295 japonica rice varieties under high-and low-N treatments.(A) HRN (root N content under high-N treatment); LRN (root N content under low-N treatment).(B) RNR (relative value of root N content under low-and high-N treatments).(C) HRNAA (root N accumulation under high-N treatment); LRNAA (root N accumulation under low-N treatment).(D) RNAAR (relative value of root N accumulation under low-and high-N treatments).(E) HYN (leaf N content under high-N treatment); LYN (leaf N content under low-N treatment).(F) YNR (relative value of leaf N content under low-and high-N treatments).(G) HYNAA (leaf N accumulation under high-N treatment); LYNAA (leaf N accumulation under low-N treatment).(H) YNAAR (relative value of leaf N accumulation under low-and high-N treatments).Lowercase a and Lowercase b represent the significance between treatments at the 0.05 level.

Figure 2 .
Figure 2. Manhattan plots of N content and N accumulation.(A) Leaf N content under high-and low-N treatments.(B) Root N content under high-and low-N treatments.(C) Leaf N accumulation under high-and low-N treatments.(D) Root N accumulation under high-and low-N treatments.The detailed annotations for abbreviations are shown in Figure 1.

Figure 2 .
Figure 2. Manhattan plots of N content and N accumulation.(A) Leaf N content under high-and low-N treatments.(B) Root N content under high-and low-N treatments.(C) Leaf N accumulation under high-and low-N treatments.(D) Root N accumulation under high-and low-N treatments.The detailed annotations for abbreviations are shown in Figure 1.

Figure 3 .
Figure 3. Enrichment analysis of 40 genes.(A) Top 15 categories from GO classification analysis of 40 genes.(B) KEGG functional classification and biological pathway enrichment of 40 genes.

Figure 3 .
Figure 3. Enrichment analysis of 40 genes.(A) Top 15 categories from GO classification analysis of 40 genes.(B) KEGG functional classification and biological pathway enrichment of 40 genes.

Figure 4 .
Figure 4. Gene expression of five candidate genes in Longjing 31 and Songjing 10 under high-and low-N treatments.* denotes significant difference at the 0.05 level, ** denotes significant difference at the 0.01 level.

Figure 5 .
Figure 5. Associated region and haplotype analysis of Os01g0159250 and Os06g0619000.(A) Regional Manhattan plot and linkage disequilibrium (LD) heatmap of Os01g0159250.(B) Gene structure of Os01g0159250.(C) Haplotype analysis of Os01g0159250.(D) Regional Manhattan plot and LD heatmap of Os06g0619000.(E) Gene structure of Os06g0619000.(F) Haplotype analysis of Os06g0619000.Lowercase a and Lowercase b represent the significance between treatments at the 0.05 level.

Figure 4 .Figure 4 .
Figure 4. Gene expression of five candidate genes in Longjing 31 and Songjing 10 under high-and low-N treatments.* denotes significant difference at the 0.05 level, ** denotes significant difference at the 0.01 level.

Figure 5 .
Figure 5. Associated region and haplotype analysis of Os01g0159250 and Os06g0619000.(A) Regional Manhattan plot and linkage disequilibrium (LD) heatmap of Os01g0159250.(B) Gene structure of Os01g0159250.(C) Haplotype analysis of Os01g0159250.(D) Regional Manhattan plot and LD heatmap of Os06g0619000.(E) Gene structure of Os06g0619000.(F) Haplotype analysis of Os06g0619000.Lowercase a and Lowercase b represent the significance between treatments at the 0.05 level.

Figure 5 .
Figure 5. Associated region and haplotype analysis of Os01g0159250 and Os06g0619000.(A) Regional Manhattan plot and linkage disequilibrium (LD) heatmap of Os01g0159250.(B) Gene structure of Os01g0159250.(C) Haplotype analysis of Os01g0159250.(D) Regional Manhattan plot and LD heatmap of Os06g0619000.(E) Gene structure of Os06g0619000.(F) Haplotype analysis of Os06g0619000.Lowercase a and Lowercase b represent the significance between treatments at the 0.05 level.

Figure 6 .
Figure 6.Associated region and haplotype analysis of Os02g0618200 and Os02g0618400.(A) Regional Manhattan plot and linkage disequilibrium heatmap of Os02g0618200 and Os02g0618400.(B) Gene structure of Os02g0618200.(C) Haplotype analysis of Os02g0618200.(D) Gene structure of Os02g0618400.(E) Haplotype analysis of Os02g0618400.Lowercase a and Lowercase b represent the significance between treatments at the 0.05 level.

Figure 6 .
Figure 6.Associated region and haplotype analysis of Os02g0618200 and Os02g0618400.(A) Regional Manhattan plot and linkage disequilibrium heatmap of Os02g0618200 and Os02g0618400.(B) Gene structure of Os02g0618200.(C) Haplotype analysis of Os02g0618200.(D) Gene structure of Os02g0618400.(E) Haplotype analysis of Os02g0618400.Lowercase a and Lowercase b represent the significance between treatments at the 0.05 level.

Figure 7 .
Figure 7. Associated region and haplotype analysis of Os02g0630300.(A) Regional Manhattan plot and linkage disequilibrium heatmap of Os02g0630300.(B) Gene structure of Os02g0630300.(C) Haplotype analysis of Os02g0630300.Lowercase a and Lowercase b represent the significance between treatments at the 0.05 level.

Figure 7 .
Figure 7. Associated region and haplotype analysis of Os02g0630300.(A) Regional Manhattan plot and linkage disequilibrium heatmap of Os02g0630300.(B) Gene structure of Os02g0630300.(C) Haplotype analysis of Os02g0630300.Lowercase a and Lowercase b represent the significance between treatments at the 0.05 level.

Table 1 .
The candidate genes that were expressed differentially under the low-N condition.

Table 1 .
The candidate genes that were expressed differentially under the low-N condition.