Genome-Wide Association Studies for Wheat Height Under Different Nitrogen Conditions

Abstract

Lodging causes a reduction in wheat (Triticum aestivum L.) yield and quality. A shorter plant height (PH) can reduce the incidence of lodging. The overuse of nitrogen promotes excessive vegetative growth, leads to taller plants, and increases lodging risk. Here, we utilized genome-wide association studies (GWASs) to explore the genetic basis of PH and the nitrogen effect index (NEI), a parameter to estimate the responses of PH under varying nitrogen conditions, using 21,201 SNP markers from the Illumina Wheat 90K SNP array. A total of 191 wheat varieties from Yellow and Huai Valley regions of China, as well as other global regions, were analyzed across two growing seasons under four nitrogen treatments, namely N₀ (0 kg/ha), N₁₅₀ (150 kg/ha), N₂₁₀ (210 kg/ha), and N₂₇₀ (270 kg/ha). GWAS results showed that 30 genetic markers were associated with PH, explaining phenotypic variance from 5.92% to 13.69%. Additionally, nine significant loci were associated with the NEI. Notably, markers on chromosomes 1A and 6B were linked to both PH and the NEI, which were insensitive to low- and high-nitrogen fertilizers. In addition, the PH of the three cultivars (Zhoumai16, Zhoumai13, and Bima1) showed little variation in four nitrogen fertilizer levels. This study identified key genetic markers associated with wheat PH and the NEI, providing insights for optimizing nitrogen use in wheat breeding.

1. Introduction

Wheat (Triticum aestivum L.) is one of the most important staple crops globally, providing calories and protein for approximately two-fifths of the global population. According to the Food and Agriculture Organization (FAO), wheat accounts for approximately 30% of global cereal consumption [1]. Plant height (PH) is one of the most critical agronomic traits that directly affects the yield and stability of wheat production. Therefore, the identification of wheat lines with reduced PH while ensuring high and stable yields has long been a key research objective for breeders [2]. In general, an appropriate PH can significantly improve grain yield and quality by reducing lodging, increasing the number of grains per spike, and improving the harvest index [3]. The utilization of dwarfing or semi-dwarfing genes has been a crucial strategy for breeders to develop semi-dwarf green revolution varieties (GRVs) of wheat [4]. Since the modern agricultural revolution, the extensive application of nitrogen fertilizers has become a conventional practice to increase wheat yield. However, this has led to issues such as low fertilizer use efficiency and environmental pollution [5,6,7,8]. The Green Revolution genes, Rht1 and Rht2, also known as “dwarfing genes”, are two key semi-dwarfing genes in wheat, designated as Rht-B1b and Rht-D1b, and are located on different chromosomes. These genes encode DELLA proteins that lack the DELLA domain, preventing their degradation and leading to protein accumulation within cells. This accumulation inhibits gibberellin (GA) from promoting plant growth, thereby reducing PH. The semi-dwarf traits conferred by Rht1 and Rht2 not only improve lodging resistance but also diminish the plant’s responsiveness to nitrogen. [9]. Therefore, identifying new quantitative trait loci (QTL) and genes that affect PH and nitrogen responses under varying nitrogen levels is crucial for breeding nitrogen-efficient wheat varieties. To date, 26 dwarfing genes have been cataloged, and 332 QTL, 270 GWAS loci, and 83 genes for PH have been identified in wheat from 2003 to 2022 [10]. Previous studies have demonstrated that varying levels of nitrogen fertilization significantly impact wheat PH [11]. However, few studies have focused on the influence of nitrogen on PH specifically.

Genome-wide association studies (GWASs) are effective methods to understand the relationship between phenotypic variations and genetic polymorphisms. GWASs have been successfully applied to identifying QTL for yield and its components [12,13], resistance to abiotic stress [14], disease [15], and grain quality [16] in wheat. Many important economic and agronomic traits in wheat, such as quality, yield-related traits, resistance, and nutrient use efficiency, are quantitative traits that are controlled by multiple genes. Due to the continuous variation typically exhibited by quantitative traits, their susceptibility to environmental influence, the complexity of their genetic basis, and the lack of clear genotype–phenotype correlations, genetic research on nitrogen use efficiency-related traits has progressed relatively slowly. QTL analysis and GWASs are essential tools for studying these complex traits. In this study, 191 varieties of wheat from the Yellow and Huai Valley Regions of China and abroad were used as the research subjects. PH was investigated under four nitrogen levels and a GWAS was performed on PH and the NEI, utilizing the 90 K SNP array developed by Illumina. The objective of this study was to identify the genetic loci associated with PH and the NEI so as to provide a reference for nitrogen-efficient wheat breeding. The findings from this research hold substantial significance for the genetic enhancement of wheat PH and the advancement of nitrogen-efficient molecular breeding strategies.

2. Materials and Methods

2.1. Plant Materials and Field Trials

The natural population consisted of 191 wheat varieties, including 148 from the Yellow and Huai Valley Regions of China and 32 foreign varieties suitable for this region (Table S1). These varieties were planted during the 2022–2023 (E1) and 2023–2024 growing seasons (E2) in Qingdao, China (120°26′ E, 36°28′ N). The experiment included four nitrogen levels (0, 150, 210, 270 kg/ha) using urea (46% N) as the nitrogen source, with the higher nitrogen levels (150, 210, 270 kg/ha) top-dressed at the jointing stage in combination with irrigation. Phosphate (P₂O₅: 12%) and potassium (K₂O: 50%) fertilizers were applied as basal fertilizers at a rate of 90 kg/ha. Each variety was planted in two rows, with 40 seeds per row, a row length of 1.2 m, and a row spacing of 20 cm. The experiment was conducted with two replicates. Field management practices were carried out according to standard protocols and PH was measured after heading was completed. For each accession, the edge plants were removed and the PH of five randomly selected plants was measured. The PH was defined as the length from the base of the plant above the ground to the top of the main stem spike (excluding the awn). The nitrogen effect index (NEI) was calculated using the formula NEI = Xi/CKi, where Xi represents the two-year average PH under different nitrogen conditions and CKi represents the two-year average PH under the control condition (N₀) (Table S2).

2.2. DNA Extraction, Genotyping, and Data Processing

Leaves were collected from five healthy seedlings of each wheat variety and placed into 2 mL centrifuge tubes containing stainless steel beads. Samples were rapidly frozen in liquid nitrogen for subsequent use. Genomic DNA was extracted using a modified cetyltrimethylammonium bromide (CTAB) method [17]. DNA concentration and quality were measured using a NanoDrop2000c microvolume spectrophotometer (http://tools.thermofisher.com/, accessed on 2 July 2024), and the DNA quality was further checked by agarose gel electrophoresis. Genotyping of 191 varieties in the natural population was performed using a 90 K SNP array [18]. Chip detection and genotyping were performed by CapitalBio Technology (http://www.capitalbiotech.com, accessed on 15 July 2024). Marker quality control (QC) was conducted in TASSEL v5.0, with the following steps: (1) heterozygous genotypes with missing data were excluded [19]; (2) markers with a missing data rate greater than 20% were filtered out; and (3) markers with a minor allele frequency (MAF) of less than 5% were excluded.

2.3. Genotyping of Rht1, Rht2, Rht8, and Rht24 by Functional Markers

The Rht1 functional marker for the detection of Rht-B1b and Rht-B1a was designed (unpublished data). The Rht2 functional marker for the detection of Rht-D1b and Rht-D1a was referred to in previous studies [20]. Assays were tested in 96-well formats with 10 μL reactions (50–100 ng/μL DNA, 5 μL of 1 × KASP master mixture, and 0.5 μL of primer mixture at concentrations of 1 μM, 1 μM, and 3 μM, respectively). PCR cycling was performed based on the following protocol: 95 °C for 10 min, ten touchdown cycles (95 °C for 20 s; touchdown at 65 °C initially and decreasing by −1 °C per cycle for 60 s), followed by 34 cycles (95 °C for 20 s; 57 °C for 60 s).

For Rht8, a two-round PCR approach was adopted using Rht8-specific primers, followed by restriction enzyme digestion of the second-round PCR products to identify the Rht8 genotype [21]. Rht24-specific primers were used for PCR amplification and the resulting products were analyzed via restriction enzyme digestion to determine the genotype [22] (Tables S3 and S4).

2.4. Linkage Disequilibrium and Population Structure Analysis

A total of 21201 filtered high-quality single nucleotide polymorphism (SNP) markers were used for a population structure analysis (Table S5). The population structure was analyzed using STRUCTURE v2.3.4. The kinship matrix (K) and PCA matrix (K = 3) were calculated using TASSEL v5.0. A data visualization of PCA was obtained using the Scatterplot3d package in R software (vision 4.2.1). The extent of linkage disequilibrium (LD) between markers was measured using the r² statistic [23]. LD estimation was performed using the sliding window method provided in the LD analysis module of TASSEL v5.0. A scatter plot of the LD decay was generated by plotting the LD parameter r² against the physical distance between the markers. The LD decay window size was set to 100, meaning that LD was calculated for 100 pairs of markers both upstream and downstream of each marker. LD decay was analyzed at both whole-genome and sub-genome levels.

2.5. Genome-Wide Association Study (GWAS)

GWASs for PH-related traits were conducted using the mixed linear model (MLM) in TASSEL v5.0 combined with population structure (Q) and kinship (K) [24]. The significance of the markers was adjusted for multiple comparisons using the Bonferroni–Holm method [25]. Based on the quality control of SNP markers and the population size in this study, we set a significance threshold of −log10(P) ≥ 3, equivalent to a p-value ≤ 1 × 10⁻³, as the threshold for significant markers. This threshold has also been applied in GWASs for other complex traits in wheat [26,27]. The R package CMplot (https://cran.r-project.org/web/packages/CMplot/, accessed on 7 May 2024) was used to generate Q-Q plots and Manhattan plots based on the results.

Based on the significant SNP markers identified through a GWAS analysis, the chromosomal location and physical position were referenced using IWGSC Ref Seq v2.1 (https://plants.ensembl.org/index.html/, accessed on 24 May 2024) [28]. Subsequently, the upstream and downstream 100bp of the SNP were extracted. Comparing the gene lists and annotations within a 1 Mb region flanking both the left and right sides of the SNP, candidate genes potentially associated with PH and the NEI were identified.

3. Results and Analysis

3.1. Phenotypic Assessment

In the 2022–2023 (E1) and 2023–2024 (E2) growing seasons, the phenotypic evaluation of PH was conducted on 191 wheat varieties under four different nitrogen fertilizer levels. The results showed that PH exhibited notable differences under varying fertilization conditions (

Table 1. Phenotypic variation in plant height.

Traits	Treatments a	Mean (cm)	SD b	CV (%)	Min (cm)	Max (cm)
PH	E1N0	73.38	11.53	15.72	55.22	126.98
	E1N150	80.66	13.45	16.67	57.36	135.02
	E1N210	81.65	12.55	15.37	56.83	130.70
	E1N270	77.79	13.32	17.12	52.53	127.28
	E2N0	74.43	13.08	17.56	55.97	136.92
	E2N150	84.18	14.48	17.19	60.58	136.75
	E2N210	85.38	14.81	17.34	56.50	132.17
	E2N270	81.66	12.82	15.69	57.25	125.50

^a E1 and E2 represent 2022–2023 and 2023–2024 growing seasons, respectively; N₀, N₁₅₀, N₂₁₀, and N₂₇₀ represent four nitrogen levels. ^b SD represent the standard deviation of plant height under different nitrogen levels during the E1 and E2 growing seasons.

). The lowest average PH was observed in the E1N₀ treatment group, which received no nitrogen fertilizer, with an average height of only 73.38 cm. In contrast, PH significantly increased in the E1N₂₁₀ and E2N₂₁₀ treatment groups, reaching 81.65 ± 12.55 cm and 85.38 ± 14.81 cm, respectively. Additionally, the average PH decreased slightly in the E1N₂₇₀ treatment group, suggesting that excessive fertilization may negatively affect PH growth. The standard deviation (SD) and coefficient of variation (CV) were used to measure the degree of dispersion of PH data. The E2N₂₁₀ treatment group exhibited the lowest variation coefficient. In summary, the appropriate application of nitrogen fertilizer can effectively enhance PH growth, and controlling the amount of fertilizer is crucial to avoid potential negative effects.

A correlation analysis was conducted on the PH of 191 wheat varieties over all tested environments. The results showed that the PH distribution across different nitrogen levels followed a skewed distribution, suggesting the complex genetic architecture of PH. Furthermore, a highly significant correlation (p < 0.001) was observed between different environments or treatments (Figure 1), indicating the significant roles of genetic factors on PH.

3.2. Analysis of NEI

The NEI average of two environments was used to evaluate the response of 191 varieties to different nitrogen application rates. Based on the distribution pattern observed in the two-year data and the natural distribution tendency of the dataset, the NEI was divided into four intervals (≤0.95, 0.95–1.05, 1.05–1.15, and ≥1.15), and the frequency distribution of the NEI in each interval was analyzed (

Table 2. Distribution of NEI.

Traits a	Frequency
	NEI ≤ 0.95	0.95 < NEI ≤ 1.05	1.05 < NEI ≤ 1.15	NEI < 1.15
N150NEI	0.52%	9.95%	61.26%	28.27%
N210NEI	1.57%	24.08%	63.35%	10.99%
N270NEI	8.38%	60.73%	29.32%	1.57%

^a NEI, nitrogen effect index; N₁₅₀, N₂₁₀, and N₂₇₀ represent 150, 210, and 270 kg/ha of nitrogen fertilizers.

). At the nitrogen application rate of 150 kg/ha, the NEI distribution exhibited a high frequency within the 1.05–1.15 interval. Specifically, the frequency of NEI values in the ≤0.95 interval was 0.52%, while it reached 9.95% in the 0.95–1.05 interval. This indicates that approximately 9.95% of wheat cultivars demonstrated a moderate response to nitrogen at a 150 kg/ha application rate. Additionally, 61.26% of wheat cultivars had NEI values within the 1.05–1.15 interval, indicating that the nitrogen application rate of 150 kg/ha moderately enhanced wheat growth and increased PH. Notably, 28.27% of samples had NEI values ≥1.15, suggesting a high nitrogen response among a considerable portion of the cultivars at this rate. As the nitrogen application rate increased to 210 kg/ha, the NEI distribution shifted, showing a substantial increase in frequency within the 0.95–1.05 interval. At this rate, 24.08% of wheat cultivars had NEI values within the 0.95–1.05 interval. This suggests that under the application rate of 210 kg/ha, wheat plants exhibited stronger adaptability and growth potential, with a significant enhancement in nitrogen use efficiency. Notably, 63.35% of wheat cultivars had NEI values in the 1.05–1.15 interval, which indicates that the majority of cultivars reached an optimal nitrogen response at this rate. The frequency of samples with NEI values ≥1.15 was 10.99%, indicating a smaller portion of cultivars responded with high nitrogen efficiency at this application rate.

However, when the nitrogen application rate reached 270 kg/ha, significant changes in the NEI distribution were observed. The frequency of NEI values in the ≤0.95 interval increased to 8.38%, while the frequency in the 0.95–1.05 interval surged to 60.73%. This suggests that the nitrogen response capacity decreased compared to the application rate of 210 kg/ha, possibly due to physiological disturbances or growth imbalances caused by excessive nitrogen application. Further analysis revealed that at the nitrogen application rate of 270 kg/ha, the frequency of NEI values in the 1.05–1.15 interval was only 29.32%, and only 1.57% of samples were found in the ≥1.15 interval. This result indicates that excessive nitrogen application may inhibit wheat growth, potentially leading to nitrogen loss and inefficiency.

3.3. Sensitivity Evaluation of Varieties to Nitrogen Fertilizer

Based on the calculation of the NEI, varieties with NEI values significantly greater than 1.20 under high nitrogen levels (N₂₁₀ and N₂₇₀) showed substantial increases in PH with an increase in nitrogen levels, indicating that these materials are nitrogen-sensitive and classified as high-NEI materials. In contrast, materials with NEI values maintained at the level of 1.00 were categorized as low-NEI materials. Additionally, we analyzed the distribution and trends (Figure 2) of NEI values for each material under the N₁₅₀, N₂₁₀, and N₂₇₀ conditions, identifying a total of seven high-NEI materials and three low-NEI materials (

Table 3. Comparison of varieties with PH and low NEIs.

Groups	Varieties	NEI a	N0PH b (cm)	N150PH b (cm)	N210PH b (cm)	N270PH b (cm)
High NEI	Linmai4	1.20	62.02	73.63	78.17	74.69
	Linmai2	1.21	68.98	82.37	88.67	83.13
	Shan512	1.22	66.82	76.93	81.56	81.23
	Fengkang2	1.22	81.30	100.38	100.07	99.44
	Lankao24	1.24	56.82	70.31	71.70	70.32
	Lumai23	1.25	66.25	76.22	84.54	82.74
	Darius	1.26	78.00	92.83	98.57	98.01
Low NEI	Zhoumai16	0.95	70.71	70.15	67.07	69.51
	Zhoumai13	0.95	59.53	61.89	56.67	56.89
	Bima1	0.97	131.95	135.89	128.64	126.35

^a NEI, nitrogen effect index; ^b N₀PH, N₁₅₀PH, N₂₁₀PH, and N₂₇₀PH represent nitrogen levels 0, 150, 210, and 270 kg/ha of plant height.

).

3.4. Population Structure, Kinship, and LD Decay

After filtering, 21,201 polymorphic SNPs were employed for GWAS analysis (Table S6). Among polymorphic SNP markers, 39.65%, 49.97%, and 10.38% were from the A, B, and D genomes, respectively. Chromosome 1B had the most SNP markers (2025), whereas chromosome 4D possessed the least (98). The total markers spanned a physical distance of 14,064.8 Mb, with an average marker density of 0.66 Mb per marker. The average genetic diversity and polymorphism information content (PIC) for the whole genome were 0.34 and 0.27, respectively, the average genetic diversities for A, B, and D genomes were 0.34, 0.35, and 0.32, and the average PIC was 0.28, 0.28, and 0.26 (Figure 3a), respectively. A population structure analysis of 191 wheat varieties was performed using Structure software(vision 2.3.4) with 21,201 SNP markers evenly covering the entire genome. The natural logarithm of the maximum likelihood value, Ln P (D), continued to increase with the K value, without showing a clear inflection point. Therefore, the optimal number of subgroups (K) was determined based on the ΔK value [29]. As shown in the figure, ΔK reached its maximum at K = 3, indicating that the optimal number of subgroups was three (Figure 3b). This means that 191 varieties could be divided into three subgroups (Figure 3c) (Table S7). The results of the neighbor-joining (NJ) tree (Figure 3d) and principal component analysis (PCA) (Figure 3e) also indicated that the natural population could be divided into three subgroups, and the subgrouping results were consistent with those of the structural analysis. A kinship analysis of 191 wheat varieties was conducted using TASSEL v5.0 (Figure 3f), and the results showed a certain degree of kinship among these varieties, with a few varieties showing closer relationships. Although the overall kinship among the varieties was weak, this relationship might have led to false positive results in the GWAS analysis and should be considered during the analysis. Linkage disequilibrium (LD) decay across the entire genome and the three sub-genomes was also analyzed for the 191 varieties (Figure 3g). When r² = 0.5, the LD decay distance across the entire genome for the 191 varieties was 1.7 Mb; at the sub-genome level, the LD decay distances for the A, B, and D genomes were 1.5 Mb, 2.1 Mb, and 2.5 Mb, respectively, indicating that the D genome has the highest degree of linkage disequilibrium, requiring relatively fewer markers for an association analysis.

3.5. GWAS Analysis of PH and NEI

We conducted a GWAS for PH and the NEI using 21,201 SNP markers from the Illumina Wheat 90K SNP array (

Table 4. Significant SNPs associated with PH and NEI.

Traits a	Environments b	Markers	Position (Mb)	R2%	Pmax	Pmin	Candidate Genes c
PH	E2N270, E2N0, E1N150, E2N150, E2N270	IWB54893	1A:1.48	6.24–7.67	2.50 × 10−4	9.07 × 10−4	1A03G0004200
	E1N0, E2N0	IWA4164	1A:26.96	6.24–8.68	1.24 × 10−4	7.70 × 10−4	1A03G0105800
	E1N0, E1N150, E1N210, E1N270, E2N210, E2N270	IWB38267	1A:45.79	6.27–10.76	1.11 × 10−5	9.76 × 10−4	1A03G0152900
	E1N0, E1N150, E1N210, E1N270, E2N0, E2N150, E2N210, E2N270	IWB35039	1A:496.31–497.52	6.63–12.54	2.62 × 10−6	4.94 × 10−4	1A03G0749100
	E1N0, E1N210, E1N270, E2N150, E2N210, E2N270	IWB6974	1A:505.52–511.10	5.95–9.40	4.60 × 10−5	9.31 × 10−4	1A03G0787900
	E1N0, E1N150, E1N210, E1N270, E2N0, E2N150	IWB51584	1A:517.40–517.48	5.97–10.28	1.80 × 10−5	9.65 × 10−4	1A03G0805700
	E1N0, E1N150, E1N210, E1N270, E2N0, E2N150, E2N210, E2N270	IWB50788	1A:539.56	7.47–13.69	8.41 × 10−7	2.24 × 10−4	1A03G0869100
	E1N150, E1N270, E2N210	IWB8040	1A:545.89–547.31	5.90–9.05	9.60 × 10−5	1.00 ×10 −3	1A03G0891600
	E1N0, E1N150, E1N270, E2N0, E2N150, E2N270	IWB61310	1A:564.30–571.78	5.92–7.92	1.47 × 10−4	9.63 × 10−4	1A03G0986800
	E1N150, E1N210, E2N0, E2N150, E2N210, E2N270	IWB15041	1B:116.58	7.23–13.49	1.14 × 10−6	3.53 × 10−4	1B03G0270900
	E2N210	IWB6406	1D:38.78–40.57	6.02–7.81	2.54 × 10−4	8.69 × 10−4	1D03G0123400
	E1N270, E2N210	IWB40766	1D:420.53–420.65	6.97–8.07	1.27 × 10−4	3.52 × 10−4	1D03G0772900
	E1N0, E1N270	IWB10193	2B:65.1	5.96–7.65	2.05 × 10−4	9.51 × 10−4	2B03G0226200
	E1N0, E1N210	IWB20703	2B:214.59–216.46	6.03–8.91	8.95 × 10−5	9.58 × 10−4	2B03G0528100
	E1N0, E2N0, E2N210	IWB4287	3A:645.09–650.42	6.32–7.94	1.42 × 10−4	9.49 × 10−4	3A03G0945100
	E1N0, E1N210, E2N0, E2N150, E2N210	IWB24136	3A:705.3	5.95–14.71	3.68 × 10−7	9.78 × 10−4	3A03G1105400
	E1N0, E1N150, E1N210, E1N270, E2N0, E2N150, E2N210, E2N270	IWB17930	3D:571.55	6.10–13.06	1.48 × 10−6	8.41 × 10−4	3D03G1029700
	E1N0, E1N150, E1N210, E1N270, E2N0, E2N150, E2N210, E2N270	IWB32654	3D:760.13	6.14–12.30	5.19 × 10−6	7.81 × 10−4	3D03G0533500
	E1N0, E1N150, E1N210, E1N270, E2N0	IWB60583	4A:48.62	6.33–7.48	3.70 × 10−4	9.09 × 10−4	4A03G0112900
	E1N0, E1N150, E1N210, E1N270, E2N0, E2N150, E2N210, E2N270	IWB59450	4A:681.66	6.64–13.12	1.59 × 10−6	4.93 × 10−4	4A03G1009500
	E1N150	Rht2	4D:18.78	5.89	9.84 × 10−4		4D03G0067100
	E1N210, E1N270	IWB61487	4D:38.28	6.01–6.59	5.07 × 10−4	9.42 × 10−4	4D03G0115900
	E1N0, E1N210	IWB6762	5A:535.13–540.28	6.00–8.78	6.96 × 10−5	9.70 × 10−4	5A03G0795600
	E1N0, E2N0	IWB5301	5B:531.15–534.05	6.17–7.10	3.22 × 10−4	7.60 × 10−4	5B03G0873000
	E2N0	IWB54819	6A:400.98	5.99	9.30 × 10−4		6A03G0594800
	E2N270	Rht24	6A:413.73	6.12	9.07 × 10−4		6A03G0611100
	E1N270, E2N0	IWA5747	6A:611.32–611.85	5.96–6.39	6.62 × 10−4	1.00 × 10−3	6A03G0604400
	E1N270, E2N210, E2N270	IWB10041	6B:696.46	6.08–8.13	1.48 × 10−4	9.43 × 10−4	6B03G1166500
	E1N150, E1N270, E2N0, E2N150	IWB55843	6D:455.25–455.89	6.08–7.88	2.01 × 10−4	9.89 × 10−4	6D03G0770600
	E1N150, E1N270	IWB8445	7A:87.30–88.98	6.25–10.12	2.33 × 10−5	7.10 × 10−4	7A03G0308700
NEI	N210	IWB71397	1A:432.63	6.83	4.32 × 10−4		1A03G0628100
	N270	IWB60140	1A:536.43	6.02–6.04	8.71 × 10−4	8.52 × 10−4	1A03G0857400
	N210	IWB43273	2B:56.78	6.16	9.65 × 10−4		2B03G0199300
	N150	IWB7420	6A:270.61–276.29	5.93–6.57	6.87 × 10−4	9.68 × 10−4	6A03G0500700
	N150	IWB59349	6A:399.83	6.21	7.55 × 10−4		6A03G0593300
	N210, N270	IWB51603	6B:696.15–696.46	5.97–7.18	3.66 × 10−4	9.34 × 10−4	6B03G1164300
	N270	IWB6263	6D:4.49	6.44	5.90 × 10−4		6D03G0023600
	N210	IWB9080	7A:73.74	7.26	2.72 × 10−4		7A03G0270000
	N210	IWB12096	7A:89.83	5.92	9.90 × 10−4		7A03G0316500

^a PH, plant height; NEI, nitrogen effect index; ^b E1, 2022–2023 growing season; E2, 2022–2024 growing season; N₀, N₁₅₀, N₂₁₀, and N₂₇₀ means four nitrogen levels; ^c candidate genes were retrieved from IWGSC Ref Seq v2.1. “TraesCS” was omitted for the candidate genes.

) (Figure 4). The results showed that under different nitrogen fertilizer treatments over two years, the significant markers affecting PH were concentrated on chromosomes 1A, 1B, 1D, 2B, 3A, 3D, 4A, 4D, 5A, 5B, 6A, 6B, 6D, and 7A. Among them, markers IWB35039 and IWB50788 on chromosome 1A (p-value up to 8.41 × 10⁻⁷) and IWB17930 and IWB32654 on chromosome 3D (p-value up to 5.19 × 10⁻⁶) were consistently associated with PH across the four nitrogen treatment environments over two years and explained a significant proportion of the variance in PH, with the highest R² value reaching 13.69%. Candidate genes TraesCS1A03G0749100, TraesCS1A03G0869100, TraesCS3D03G1029700, and TraesCS3D03G0533500 may influence overall PH by regulating cell division and elongation, as well as secondary metabolic pathways. Additionally, the Rht2 and Rht24 gene markers were also significantly associated with PH, explaining 5.89–6.12% of phenotypic variances. Under different nitrogen treatment conditions, genetic markers on chromosomes 1A, 2B, 6A, 6B, 6D, and 7A showed significant effects on the NEI. Notably, the marker IWB71397 on chromosome 1A was significantly associated with the NEI under N₂₁₀ treatment, explaining 6.83% of phenotypic variances, while the marker IWB12096 on chromosome 7A was significantly associated with the NEI under N₂₁₀ treatment, explaining 7.26% of the variation. The marker IWB51603 on chromosome 6B was significantly associated with the NEI under both N₂₁₀ and N₂₇₀ treatments, explaining 5.97–7.18% of the variation. Candidate genes TraesCS1A03G0628100, TraesCS7A03G0270000, and TraesCS6B03G1164300 may influence nitrogen use efficiency in plants by regulating the expression of genes associated with nitrogen metabolism. It is noteworthy that the results indicated that PH and the NEI were co-located in similar regions on chromosomes 1A, 2B, and 6B, suggesting these loci may influence PH by modulating nitrogen utilization.

4. Discussion

Wheat is an important staple crop in China [30] and its high and stable yield plays a crucial role in ensuring national food security. PH is a critical agronomic trait in wheat and serves as a key parameter affecting lodging resistance and grain yield [31]. A reduction in PH can sometimes correlate with a decrease in grain yield. Thus, the goal of breeding programs is to identify variations that reduce height without negatively affecting yield. Genetic studies have shown that PH is a complex trait regulated by both Mendelian genes and QTL. Previous studies have identified multiple major QTL for PH across 21 wheat chromosomes, providing a foundational understanding of the genetic mechanisms underlying PH [10]. Previous research has shown that different nitrogen levels significantly affect wheat PH [32]. Although nitrogen fertilizers increase yield, some dwarfing genes reduce wheat responsiveness to nitrogen, making yield increases highly dependent on large amounts of nitrogen fertilizer. However, excessive nitrogen fertilizer application can significantly increase PH, which underscores the importance of identifying genetic loci that balance PH and nitrogen responsiveness [33].

4.1. Tall and Dwarf Alleles of Rht2 and Rht24 Showed Distinct Effects on PH and NEI Under Varying Nitrogen Levels

Many genetic loci related to PH have been identified and 26 dwarfing genes have been cataloged. Of them, Rht2 (Rht-D1b) and Rht24 have been widely used in wheat breeding [34,35]. The introduction of semi-dwarf genes Rht1 (Rht-B1b) and Rht2 (Rht-D1b) in the 1960s contributed to the Green Revolution in wheat [4]. Rht-D1b encodes N-truncated DELLA proteins, which cannot be degraded owing to a lack of the DELLA domain, but it could repress the effect of GA on PH [36,37]. Rht24 encodes gibberellin 2-oxidase (TaGA2ox-A9), contributing to the reduction in bioactive GA in stems [35]. Here, the different alleles of Rht2 and Rht24 exhibited distinct effects on PH during growth. Notably, under low nitrogen levels (N₀ and N₁₅₀), the difference in PH between Rht2 and Rht24 genotypes was particularly pronounced, suggesting that gibberellin had a greater impact on PH under low-nitrogen conditions. As the nitrogen gradient increased (N₀, N₁₅₀, N₂₁₀, and N₂₇₀), the PH corresponding to the Rht2 and Rht24 genotypes also increased, but the increase in Rht-D1a and Rht24a (tall types) was more significant. This phenomenon is consistent with the role of nitrogen as an essential nutrient for plant growth. Nitrogen not only influences photosynthesis but also promotes protein synthesis; thus, nitrogen application generally enhances plant growth and increases PH. However, the figure also shows that under high nitrogen levels (N₂₁₀ and N₂₇₀), the PH of Rht-D1b and Rht24b (dwarf types) did not increase as much as expected. This could be due to a saturation point in the plant’s response to nitrogen, where further nitrogen application no longer significantly boosts PH, demonstrating typical diminishing marginal returns.

In addition, the NEI of the Rht2 and Rht24 alleles also showed significant differences across different nitrogen gradients (Figure 5). At lower nitrogen levels (N₁₅₀), the NEI of Hap1 was higher, whereas that of Hap2 was relatively lower, indicating that the tall type can utilize nitrogen more efficiently under low-nitrogen conditions. As the nitrogen gradient increased (N₂₁₀ and N₂₇₀), the NEI generally declined, indicating a gradual decrease in the nitrogen utilization efficiency of the plant in response to additional nitrogen. This phenomenon aligns with the saturation effect of crop nitrogen absorption and utilization, suggesting that under high-nitrogen conditions, the nitrogen use efficiency of crops no longer improves significantly. Nitrogen plays a crucial role in wheat physiology by facilitating the synthesis of proteins and chlorophyll, supporting photosynthesis, and regulating hormonal and metabolic processes [38]. However, excessive nitrogen application can diminish photosynthetic efficiency, heighten the risk of photoinhibition, and lead to nutrient imbalances that inhibit the absorption of other essential elements (such as phosphorus and potassium), ultimately restricting further wheat growth.

4.2. GWAS Analysis for PH and NEI

In the GWAS analysis of PH and the NEI for 191 varieties, we identified the markers IWB35039, IWB6974, and IWB51584 located in the 496.31–497.52 Mb, 505.52–511.10 Mb, and 517.40–517.48 Mb regions on chromosome 1A, respectively. These three markers were significantly associated with traits across multiple nitrogen gradients. Additionally, similar regions were identified by previous studies near 495.55 Mb [39], 510.98 Mb [40], and 515.74 Mb [41], further confirming the importance of genetic loci on chromosome 1A in the regulation of PH. The IWB6406 marker, located in the 38.78–40.57 Mb region on chromosome 1D, aligns with previous findings that identified a key PH regulatory locus in a similar position (34.27 Mb) on chromosome 1D. Moreover, prior research identified loci related to PH on chromosomes 1D, 3A, 4D, and 5A, consistent with our findings, indicating a common regulatory role of these regions in different wheat populations. Another study [42] also identified a region on chromosome 5B (541.42 Mb), which aligns with our results, further emphasizing the importance of chromosome 5B in the regulation of PH. Additionally, we found that PH at approximately 539 Mb on chromosome 1A, 60 Mb on chromosome 2B, and 696 Mb on chromosome 6B was significantly associated with the NEI at similar positions on chromosomes 1A, 2B, and 6B, respectively, suggesting a genetic correlation between these two traits, which may indicate that these chromosome regions regulate both PH and the NEI. This significant correlation suggests a dual effect of loci on nitrogen use and growth characteristics. This dual effect may be due to the co-regulation of gene adaptation and nitrogen response mechanisms in wheat, ensuring better adaptation in different nitrogen environments and thus optimizing wheat growth performance. In this study, we identified key genetic markers that can be used in future breeding programs to optimize PH and the NEI to achieve stable yield potential. The success of the Green Revolution was due to the introduction of Rht semi-dwarf genes and the application of high-nitrogen fertilizer, and modern agriculture needs to further balance the relationship between PH and the NEI. By selecting genotypes that exhibit a higher NEI and moderate PH at different nitrogen levels, we are able to reduce the overuse of nitrogen fertilizer while maintaining high crop yields. This strategy not only adapts to the needs of sustainable agriculture but also offers possible solutions to the future challenges of limited resources and climate change.