The Impacts of Nitrogen Accumulation, Translocation, and Photosynthesis on Simultaneous Improvements in the Grain Yield and Gluten Quality of Dryland Wheat

Abstract

The effects of nitrogen (N) accumulation and translocation on photosynthesis have been widely reported, while the impacts of N accumulation, translocation, and photosynthesis on simultaneous improvements in the grain yield and gluten quality of dryland wheat still remain unclear. For this reason, the relationships between photosynthesis, N use efficiency (NUE), and related traits and grain yield, gluten quality, and the related traits of 11 representative wheat genotypes in the 2018–2021 cropping years were examined. The results show that the grain weights per spike accounted for 58.7% and 42.4% of genetic variations in the grain yield and grain protein contents, respectively. Meanwhile, N accumulation at the maturity stage caused a 49.5% genetic variation in the grain protein contents. The gluten index of MY26 and NM101 significantly decreased with a decrease in the grain number per spike in the 2018–2019 cropping season. The precipitation reduced by 53.8% in the 2019–2020 cropping season, resulting in a gluten index that increased by 13.0%. CY25 and NM101 showed high gluten quality without reducing the yield due to the high net photosynthetic rate, big grain size, large leaf area index, and high grain number per spike, respectively. Our results indicated that increasing the grain weight per spike and individual N accumulation at the maturity stage via genotype selection for a big grain size, large leaf area index, and high net photosynthetic rate simultaneously improved the grain yield and gluten quality of dryland wheat. Moreover, the effect of the genotype–environment interaction on the gluten index was related to the N translocation regulated by canopy senescence, and N translocation was affected by the source N supply associated with the net photosynthetic rate and sink N demands in relation to the grain number per spike under dryland soil conditions.

1. Introduction

Wheat is one of the three major food crops in China, and increasing wheat grain yield can meet the growing demands for grain production and ensure national food security [1]. In the meantime, agricultural development goals and requirements of “stabilizing grain, increasing production and improving quality, increasing efficiency” have been clearly proposed in the “14th Five-Year Plan” structural reform in China, which requires simultaneous improvements in wheat grain yield and quality.

Wheat quality can be subdivided into appearance quality, commodity quality, nutritional quality, and processing quality, and this study mainly focuses on processing quality [2]. Gluten protein is a central component that determines the processing quality of wheat [3], and wet gluten contents and gluten index are two important factors that affect wheat gluten quality [4]. The former is mainly related to grain protein contents [5] and the latter is mainly affected by genotypes [6]. The combined effects of grain protein contents, wet gluten contents, and the gluten index on gluten quality were assessed by the sedimentation values in this study [7]. However, the relative impacts of these three factors to improve gluten quality under dryland soil conditions require further attention.

Although the significantly negative correlation between grain yield and grain protein content has been widely reported [8,9], previous studies on grain protein deviation (GPD) have identified ideal genotypes with both high yield and grain protein contents [10,11]. Moreover, increasing the gluten index is another effective way of alleviating the contradiction between a high grain yield and low grain protein contents, as well as improving gluten quality [12]. However, this negative correlation between yield and grain protein content was affected by soil water and the N environment [13], and the gluten index was also affected by genotype–environment interactions [6]. Thus, the genotype selection for high yield and high quality in the dryland soil environment also needs to be studied in more detail.

Post-anthesis photosynthesis is a critical process in wheat grain development and yield formation [14]. However, the grain number per unit area has been determined before the flowering stage [15], which is mainly related to the involvement of N accumulation before the flowering stage in plant canopy establishment and floret development [16]. During the grain filling stage, N partitioning in photosynthetic and non-photosynthetic organs, post-anthesis N accumulation, and translocation determines the kernel weight and final yield by affecting the effective photosynthetic area, photosynthetic duration, and photosynthetic assimilation capacity [17]. These studies have also confirmed that photosynthesis is a complex physiological process, which is affected by N accumulation, partitioning, and translocation [18,19].

N is transported to grains for protein synthesis and gluten formation during the later filling stage [20]. Grain protein contents are a function of grain yield, N accumulation at the maturity stage, and the N harvest index (NHI) [21]. Previous studies have explored the effects of N accumulation at the anthesis [22], N translocation [23], and post-anthesis N accumulation [24] stages on simultaneous improvements in wheat yield and grain protein contents, but no unified conclusion has been reached. Furthermore, the effects of N accumulation, partitioning, and translocation on the gluten index are also unclear.

Consequently, N simultaneously affects wheat grain yield and gluten quality, which is related to N accumulation, partitioning, and translocation [17,25]. However, critically agronomic, physiological, and NUE-related traits of simultaneous improvements in wheat yield and gluten quality under dryland soil conditions also need to be studied in more detail. The objectives of this study were to (i) examine the relative impact of yield components and gluten quality-related traits on genetic variations in grain yield and sedimentation values, respectively; (ii) identify critically agronomic and NUE-related traits of wheat genotypes with both high yield and good gluten quality; and (iii) reveal the relationships between N accumulation, translocation, photosynthesis, and simultaneous improvements in the grain yield and gluten quality of dryland wheat.

2. Materials and Methods

2.1. Experimental Site and Design

A field experiment was conducted for three years from 2018 to 2020 at the Renshou Experimental Station (29°51′ N, 104°12′ E) in southwestern China. Soil at the experimental site was classified as lithomorphic purple soil, according to FAO taxonomy. The physicochemical properties of soil (at a depth of 0–20 cm) before the sowing stage during the three cropping seasons are shown in Figure 1.

The climatic data of the region from the sowing stage to the harvest were selected from a field microclimate weather station. The monthly precipitation, cumulative precipitation, and monthly mean temperature during each wheat growing season are shown in Figure 2. The cumulative precipitation in 2019–2020 (69.3 mm) was significantly lower than that in 2018–2019 (149.9 mm) and 2020–2021 (146.2 mm). Monthly precipitation during December and May in 2019–2020 differed from that in 2018–2019 and 2020–2021. Compared to the other two cropping seasons, the 2019–2020 cropping year was considered a typical drought year. Low temperatures before jointing may have had an unfavorable effect on wheat growth, particularly in 2020–2021.

2.2. Experimental Material, Design, and Treatments

Based on our previous 2-year field experiment, 11 typical wheat genotypes, including landraces, advanced landraces, and modern genotypes, were selected from 32 wheat (Triticum aestivum L.) genotypes released in southwest China between 1965 and 2017 through the hierarchical cluster analysis of yield, yield components, and gluten quality-related traits (Figure 3). All the selected wheat genotypes are registered and widely grown (>100,000 ha per year) in southwest China, and typically have agronomic traits, such as “keep green” traits, erect flag leaves, large flag leaves, high plant height, etc. Basic information of 11 winter wheat genotypes is shown in

Table 1. Basic information of 11 winter wheat genotypes.

Genotype		Breeding Groups	Dwarfing Genes	Grain Yield (t ha−1)	Wet Gluten Content (%)	Gluten Index (%)	Harves Index	Yield and Gluten Quality Property
G1	SW482	L	Rht 8	3.2	31.7	88	0.46	Low yield and low wet gluten content
G2	CW39	AD	Rht 8	3.2	36.0	97	0.39	Low yield and high wet gluten content
G3	CW1247	L	Rht 8	3.4	28.7	98	0.48	Low yield and low wet gluten content
G4	CW66	AD	Rht 8	3.4	30.0	67	0.43	Low yield and low wet gluten content
G5	NW101	L	Rht D1b	2.9	40.4	54	0.34	Low yield and high wet gluten content
G6	MY26	AD	Rht 8	3.3	37.3	73	0.46	Low yield and high wet gluten content
G7	CW81	L	Rht 8	3.1	30.3	84	0.47	Low yield and low wet gluten content
G8	CB25	S × AD	Rht 8	4.1	38.2	52	0.47	High yield and high wet gluten content
G9	MW51	AD	Rht D1b	3.9	27.4	87	0.47	High yield and low wet gluten content
G10	CW104	S × AD	Rht 8	3.9	30.2	89	0.49	High yield and low wet gluten content
G11	CW34	S × AD	Rht 8	4.2	28.3	35	0.49	High yield and low wet gluten content

Note: G, genotype; L, landrace; AD, advanced landrace; S × AD, synthetic hexaploid wheat × advanced landrace. All data were the mean values of 11 wheat genotypes in 2016–2017 and 2017–2018 cropping seasons. The gluten quality property here was mainly divided by wet gluten content.

.

A randomized complete block design with three replicates was applied in all wheat-growing seasons. Each plot had a length of 4 m and a width of 3 m. Following conventional cultivation practices in the growing regions, the previous crop in each growing season was soybean. Winter wheat was sown at the end of October. Row spacing of 0.2 m was required in order to yield an initial seedling population of 180 seedlings m⁻². Before the sowing stage, 150 kg ha⁻¹ N (at a 6:4 ratio of sowing–jointing), 75 kg ha⁻¹ of P₂O₅, and 75 kg ha⁻¹ of K₂O were applied. Irrigation was not applied. Straw mulching based on no tillage was applied to conserve the soil moisture for wheat growth. Commercial herbicides, pesticides, and fungicides were preventively applied at monthly intervals after tillering to avoid any yield loss. Other management endeavors followed local practices.

2.3. Measurements and Calculations

2.3.1. Yield, Yield Components, and Related Traits

The initial seedling number, maximum seedling number, and fertile spikes were measured in the middle row (4 m² areas) at the seedling (GS21), jointing (GS31), and maturity (GS92) stages [26]. The maximum seedling number and fertile spikes were calculated as the ratios of the maximum seedling number to the initial seedling number and the fertile spike number to the initial seedling number, respectively.

Grain yield was determined in a harvested area of 4 m² for each plot. Fertile spikes were counted in a harvest area of 4 m² and the grain number per spike was counted as 30 spikes in each of the three trials. The grains from each plot were air-dried, weighed, and held for moisture determination using a DMC-700 Digital Moisture Tester (Seedburo, Chicago, IL, USA). The thousand grain weight was measured at a kernel moisture content of 13.5%. The grains from each plot were dried and ground into powder and passed through a 0.15 mm screen. The harvest index was calculated as the ratio of grain yield to above-ground dry matter (DM) yield at the maturity stage.

2.3.2. NUE-Related Traits

Thirty wheat plants were collected consecutively from the middle row of each plot at the anthesis (GS65) and physiological maturity stages to measure the N content and NUE-related traits. The plant samples were separated into stems, green leaves, yellow leaves, spikes (glumes combined with rachilla, excluding grains), and grains. The segmented organ was oven-dried at 105 °C for 30 min and then dried to a constant weight at 80 °C for at least 72 h. The N content of the organs was determined using the Kjeldahl method [27]. Soil N supply was the sum of the potential N supply from soil and fertilizers. The potential N supply from the soil at the sowing stage was estimated using the initial available N and bulk densities. The total above-ground N accumulation was calculated as the product of the above-ground dry matter and corresponding N content. NUE is defined as the average grain yield produced per kg of N supply, and can be subdivided into N uptake efficiency (NUpE) and N utilization efficiency (NUtE). NUpE was calculated by dividing the total above-ground N accumulation at harvest by the soil N supply (Equation (1)). NUtE is expressed as the amount of grain yield per kilogram of the total above-ground N accumulation (Equation (2)). NHI was determined as the ratio of N accumulation in grains to the total above-ground N accumulation at the maturity stage (Equation (3)). N partitioning was determined as the ratio of N accumulation in a specific organ (i.e., stems, green leaves, and spikes) to the total above-ground N accumulation at the anthesis stage. Post-anthesis N uptake was the difference between the above-ground N accumulation at the anthesis and maturity stages. N translocation was calculated as the difference between the above-ground N accumulation at the anthesis stage and the N accumulation in vegetative organs at the maturity stage (including stems, leaves, and spikes). The transported DM and N were the DM and N lost from the vegetative organs, which were assumed to be transported to the developing organs.

NUpE (kg kg⁻¹) = total above-ground N accumulation/(N fertilizer + soil mineral N),

(1)

NUtE (kg kg⁻¹) = grain yield/total above-ground N accumulation,

(2)

NHI = N accumulation in grains/total above-ground N accumulation at the maturity stage,

(3)

N translocation amount = the total above-ground N accumulation at the anthesis stage − total above-ground N accumulation at the maturity stage (excluding N grains),

(4)

N translocation efficiency = N translocation amount/total above-ground N at the anthesis stage.

(5)

2.3.3. Gluten Quality-Related Traits

Grain protein contents were measured as the grain N content × 5.7. Wheat grains were stored in a closed space for a month after harvest. Wheat flour was placed in the ground and stored for a week. The wet gluten content and gluten index were measured in line with the gluten index method [6], using a Perten Glutomatic 2200 Instrument and a Perten 2015 Centrifuge (Perten Instruments AB, Ha¨gersten, Sweden). The sedimentation value of the wheat flour was measured using the Zeleny method [28].

2.3.4. Photosynthesis and Dry Matter Accumulation and Translocation

Flag leaf photosynthesis-related traits were measured at the beginning of the flowering stage. The area of flag leaf was measured with a specific leaf weight method. Photosynthetic gas exchange parameters, such as the net photosynthetic rate (Pn), transpiration rate (Tr), and stomatal conductance (Ci) of wheat flag leaf, were measured using LI-COR 6800 at 9:00–11:00 a.m. on typical sunny days. The water use efficiency (WUE) of flag leaf was calculated as the ratio of Pn to Ti.

Ten plants in each plot were sampled every other five days after the flowering stage. The above-ground dry matter was divided into stems (including leaf sheaths), flag leaves, green leaves (excluding flag leaves), yellow leaves, spikes (excluding grains), and grains. The grain, stem dry matter accumulation, and stem dry matter translocation dynamics were calculated.

2.4. Statistical Analysis

All data were analyzed using the mean values of three replicates. ANOVA was used, setting the genotype as the fixed effect and the year as the random effect. Statistical comparisons were considered significant at p < 0.05 or p < 0.01. Genotype plus genotype × environment (GGE) biplots and principal component analysis (PCA) were used to evaluate the mean yield and yield stability of the 11 wheat genotypes, respectively. The impact of key indicators on genetic variations was evaluated using dominance analysis.

3. Results

3.1. Combined Variance Analysis of Yield, Gluten Quality, and NUE-Related Traits

Most of the traits in this study were strongly affected by the year, although the grain yield, sedimentation value, wet gluten contents, gluten index, and N accumulation at the maturity stage were mainly regulated by the genotype (

Table 2. Combined variance analysis of grain yield, gluten quality, nitrogen efficiency and related traits in the three cropping seasons.

Indicators	Mean Values			F-Value
	2018–2019	2019–2020	2020–2021	G	Y	G × Y
Grain yield (t ha−1)	4.6 a	4.9 a	4.9 a	85 **	35 **	20 **
Fertile spikes (no plant−1)	1.4 a	1.2 b	1.1 c	37 **	199 **	14 **
Grain number (no spike−1)	38 c	43 b	50 a	114 **	628 **	20 **
Thousand-grain weight (g)	49 b	52 a	49 b	79 **	94 **	13 **
Grain weight per spike (g spike−1)	1.9 c	2.3 b	2.5 a	119 **	561 **	25 **
Sedimentation value (mL)	24.8 a	23.7 ab	22.6 b	137 **	19 **	8 **
Grain protein contents (%)	11.7 a	11.1 b	10.3 c	146 **	246 **	16 **
Wet gluten contents (%)	23.4 a	23.6 a	23.5 a	164 **	0 ns	12 **
Gluten index (%)	74 b	87 a	76 b	1008 **	392 **	90 **
N accumulation at anthesis (10−2 g shoot−1)	4.3 b	4.9 a	3.4 c	98 **	578 **	20 **
Post-anthesis N accumulation (10−2 g shoot−1)	0.5 b	0.5 b	1.9 a	35 **	1175 **	13 **
N accumulation at maturity (10−2 g shoot−1)	4.9 b	5.4 a	5.3 a	97 **	52 **	13 **
NUpE (kg kg−1)	0.47 a	0.47 a	0.42 b	27 **	63 **	15 **
N translocation efficiency (%)	76 b	79 a	75 b	21 **	198 **	32 **
NUtE (kg kg−1)	39 c	42 b	47 a	196 **	689 **	27 **
NHI (%)	78 c	81 b	84 a	58 **	597 **	36 **

Note: All data were the mean values of 11 wheat genotypes in each cropping season (n = 11). G, genotype; Y, year; G × Y, interaction between genotype and year. ^ns, p > 0.05; **, p < 0.01. Different lowercase letters indicate a significant difference at p < 0.05.

). Based on the average performance of 11 winter wheat genotypes, the grain number per spike and grain weight per spike in the 2018–2019 cropping season decreased by 11.6% and 24.0%, and 17.4% and 24.0%, respectively, compared with the 2019–2020 and 2020–2021 cropping seasons. The gluten index, N accumulation at the anthesis stage, and N translocation efficiency in the 2019–2020 cropping season increased by 13.0%, 14.0%, 3.0% and 11.0%, 44.1%, 4.0%, respectively, compared with the 2018–2019 and 2020–2021 cropping seasons. The post-anthesis N accumulation in the 2020–2021 cropping season increased by 280% compared with the 2018–2019 and 2019–2020 cropping seasons.

3.2. Yield and Yield-Related Traits

The impact of fertile spikes and the number of grains per spike on genetic variations in the yield of dryland wheat was strongly affected by environmental factors. In the 2018–2019 (Figure 4a,d,g), 2019–2020 (Figure 4b,e,h), and 2020–2021 (Figure 4c,f,i) cropping seasons, the fertile spikes and the number of grains per spike accounted for 49.3% and 26.0%, 54.8% and 40.6%, 58.2% and 35.1% genetic yield variations, respectively. The grain weight per spike accounted for 49.2%, 72.3%, and 54.5% genetic yield variations, with an average of 58.7%, which was higher than the fertile spike (39.2%) and thousand grain weight (17.0%) results.

The spike DM at the anthesis stage in the 2018–2019 (R² = 0.48 *) and 2019–2020 (R² = 0.83 **) cropping seasons (Figure 5a) showed significantly linear relationships with the grain number per spike (p < 0.05). However, the significantly linear relationship (p < 0.05) between the spike DM partitioning at the anthesis stage and the number of gains per spike was only observed in the 2020–2021 cropping season (Figure 5b). The grain number per spike in the 2020–2021 cropping season was significantly higher than that in the other two cropping seasons, owing to the significantly higher fruiting efficiency. However, the lower grain number per NM101 and MY26 spike in the 2018–2019 cropping season was mainly related to the lower fruiting efficiency because both had higher spike DM and (or) spike DM partitioning at the anthesis stage (Table S2).

Based on the average yield performance in the three cropping seasons, CY25, MM51, CM104, and CM34 were characterized as high-yield genotypes, while CM39 and SM482 were characterized as low-yield genotypes, due to their low grain weight per spike (Figure 6d). The high yield of CY25 and MM51 was attributed to the high kernel weight and high number of grains per spike, respectively. The yield stability of wheat was strongly affected by environmental factors, especially the low-yield stability of NM101 and MY26, which was mainly related to the large variation in the grain number per spike in the three cropping seasons. The relationships between the yield and yield components of different wheat genotypes in the three cropping seasons are shown in Figure 6 and Table S1.

3.3. Gluten Quality-Related Traits

There was a significantly linear relationship between the grain protein contents and wet gluten contents (R² = 0.46 **, Figure 7a). The grain protein contents (R² = 0.56 **, Figure 7b) showed a stronger linear relationship with the sedimentation value as compared to the wet gluten contents (R² = 0.33 **, Figure 7c). The sedimentation value was also affected by the gluten index (Figure 7d), while the effect on the sedimentation value of the gluten index less than 90% (R² = 0.62 **) was stronger than that with a sedimentation value that was higher than 90% (R² = 0.31 *).

Based on the average gluten quality performance in the three cropping seasons, CM39 showed the highest sedimentation value, due to the highest wet gluten contents and a gluten index above 90%. The wet gluten contents of CY25 and MY26 increased by 5.8 and 4.4%, respectively, in comparison to the average values of that of 11 wheat genotypes. However, the gluten index of CY25 was below 60%, which lead to a decrease in the sedimentation value. CM482 and NM101 maintained a higher gluten index at a higher wet gluten content level, which helped to increase their sedimentation value. The sedimentation values of CM1247 and CM104 were significantly higher than those of other genotypes with low wet gluten content (Figure 8d). The gluten index was also strongly influenced by the interaction between the genotype and the environment. The gluten index of CY25 in the 2019–2020 cropping season and NM101 in the 2018–2019 cropping season significantly increased and decreased, respectively, compared with the other two cropping seasons (Figure 8a–c and Table S3).

3.4. Relationships between NUE and Yield as Well as Gluten Quality-Related Traits

NUpE and NUtE accounted for 51.3%, 44.4%, and 62.5% and 48.3%, 54.9%, and 36.4% genetic yield variations in the 2018–2019, 2019–2020, and 2020–2021 cropping seasons, with averages of 52.7% and 46.5%, respectively (Figure 9 and Table S4). Similarly, the grain weight per spike and N accumulation at the maturity stage accounted for 43.8%, 59.7%, and 23.7% and 45.6%, 33.4%, and 69.4% genetic variations in grain protein contents in the 2018–2019, 2019–2020, and 2020–2021 cropping seasons, with averages of 42.4% and 49.5%, respectively (Figure 10 and Table S5), which were higher than the average of NHI (6.2%).

N accumulation at the wheat maturity stage was mainly caused by N accumulation at the anthesis stage. N accumulation at the anthesis stage showed significantly linear relationships with the N accumulation rate, plant height, and leaf area index at the anthesis stage in the 2018–2019 and 2019–2020 cropping seasons, while a significantly linear relationship between N accumulation at the anthesis stage from the sowing stage to the flowering stage was only observed in the 2019–2020 cropping season (Figure 11). In this study, NM101 showed higher N accumulation at the anthesis stage, which was mainly related to its higher plant height and leaf area index at the anthesis stage. CM34 was characterized as a late flowering variety, while its lower N accumulation at the anthesis stage in the 2018–2019 cropping season was mainly related to the dilution effect resulting from higher fertile spikes (Table S6).

3.5. Critical NUE-Related Traits Associated with High Yield and High Gluten Quality

There was a negative correlation between the grain yield and sedimentation value in this study (R² = 0.35 **, Figure 12a). Some of wheat genotypes in group 2 showed both a higher grain yield and sedimentation value. The grain yield and sedimentation value of CM104 in the 2018–2019 cropping season increased by 10.9% and 2.0%, respectively, compared with the average values (Figure 12b). Similarly, CY25 and NM101 showed a higher yield and sedimentation value in the 2019–2020 (Figure 12c) and 2020–2021 (Figure 12d) cropping seasons.

In the 2019–2020 and 2020–2021 cropping seasons, the thousand grain weight and post-anthesis N accumulation values of CY25 increased by 12.2%, 46.4% and 12.8%, 28.7%, respectively, and the wet gluten contents increased by 7.2 and 6.5%, respectively, compared with the average values. The grain number per spike and N accumulation at the anthesis stage of NM101 increased by 34.2%, 24.1% and 15.0%, 13.7%, respectively, and the gluten index increased by 7.0 and 15.0%, respectively, compared with the average values. In addition, CY25 and NM101 also showed a higher grain weight per spike and N accumulation at the maturity stage (Figure 13).

3.6. Photosynthesis, Dry Matter Accumulation, and Translocation

In the 2019–2020 and 2020–2021 cropping seasons, the flag leaf area of CM34 and NM101 increased by 20.5% and 43.6% and 20.8% and 31.1%, respectively, while the flag leaf N content decreased by 0.2% and 0.3% and 0.5% and 0.4%, respectively, and Pn decreased by 5.5% and 2.7% and 4.9% and 5.4%, respectively, compared with the average values. These results show that the decreased Pn value of CM34 and NM101 was mainly associated with the decrease in N content caused by the increased flag leaf area. CY25 was characterized as a high Pn variety, and its Pn increased by 16.1% and 18.5% in the 2019–2020 and 2020–2021 cropping seasons, respectively, compared with the average values. The high Pn value of CY25 in the 2019–2020 cropping season was mainly attributed to a high WUE (

Table 3. Net photosynthetic rate related traits of flag leaf of 11 winter wheat genotypes in the 2019–2020 and 2020–2021 cropping season.

Cropping Season	Genotype	N Content (%)	Leaf Area (cm2)	Pn (µmol m−2 s−1)	Tr (mmol m−2 s−1)	Ci (µmol m−2 s−1)	WUE (10−3 mol mol−1)
2019–2020	SM482	3.6 de	32 e	18.4 b	2.9 de	405 a	6.5 c
	CM39	3.6 de	38 d	17.9 c	2.3 f	252 e	7.9 b
	CM1247	3.4 f	29 f	16.7 e	2.7 e	254 e	6.1 d
	CM66	3.8 b	48 b	16.2 f	4.9 a	380 b	3.3 g
	NM101	3.3 g	56 a	17.1 d	3.4 b	228 f	2.0 h
	MY26	3.8 bc	36 d	17.2 d	3.5 b	250 e	4.9 f
	CM81	3.6 e	31 ef	17.0 d	3.2 c	211 f	5.3 e
	CY25	3.7 cd	43 c	20.4 a	2.3 f	308 c	8.6 a
	MM51	3.6 de	37 d	18.6 b	2.2 f	258 e	8.5 a
	CM104	4.1 a	32 ef	17.2 d	2.2 f	285 d	8.0 b
	CM34	3.4 f	47 b	16.6 e	2.9 d	211 f	5.8 d
2020–2021	SM482	3.4 bd	36 f	24.6 a	4.7 d	220 ab	5.2 c
	CM39	3.3 d	36 ef	16.8 f	5.7 ab	162 f	2.9 f
	CM1247	3.5 b	29 h	19.3 de	3.8 e	130 g	5.1 c
	CM66	3.3 d	37 ef	21.2 c	6.0 a	205 bd	3.6 e
	NM101	2.9 e	51 a	20.2 ce	5.5 bc	159 f	3.7 e
	MY26	3.7 a	41 cd	24.5 a	4.0 e	230 a	6.1 b
	CM81	3.4 bc	38 de	19.1 e	5.3 c	193 de	3.6 e
	CY25	3.4 bc	43 c	25.3 a	5.0 d	197 ce	5.1 c
	MM51	3.4 bc	37 ef	22.6 b	4.7 d	186 e	4.8 d
	CM104	3.3 cd	33 g	20.9 c	3.0 f	164 f	6.8 a
	CM34	2.8 e	47 b	20.3 cd	5.5 bc	211 bc	3.7 e

Note: All values were measured with three replications (n = 3). Different lowercase letters indicate a significant difference at p < 0.05.

).

The maximum accumulation of stem DM and the translocation to the developing grains occurred 10~15 days after the flowering stage. The translocation contribution of DM before the flowering stage to grains in the 2019–2020 cropping season was higher than that in the 2020–2021 cropping season. By contrast, most of the grain yield was caused by post-anthesis DM accumulation through photosynthesis in the 2020–2021 cropping season. The stem DM translocation of NM101 and MM51 with a high grain number per spike in the 2019–2020 cropping season increased by 30.8% and 23.1%, respectively, compared with the average values. The grain DM accumulation of NM101 and CY25 with a high grain weight per spike in the 2019–2020 cropping season increased by 26.6% and 22.2%, respectively, compared with the average values. These results show that the impact of stem DM translocation on grains of NM101 with a high grain number per spike was greater than that of CY25 with a high thousand grain weight in the 2019–2020 cropping season. In addition, the grain DM accumulation of both MY26 and CY25 with a high thousand grain weight in the 2020–2021 cropping season increased by 14.6%, compared with the average values (

Table 4. Stem DM translocation and grain DM accumulation of 11 winter wheat genotypes in the 2019–2020 and 2020–2021 cropping seasons.

Genotype	2019–2020		2020–2021
	Stem DM Translocation (g Plant−1)	Grain DM Accumulation (g Plant−1)	Stem DM Translocation (g Plant−1)	Grain DM Accumulation (g Plant−1)
SM482	1.3 cd	2.1 f	1.0 ad	1.4 g
CM39	1.3 cd	2.3 d	0.8 de	1.7 d
CM1247	1.0 e	1.9 h	0.7 e	1.5 f
CM66	1.1 e	1.9 h	0.9 bd	1.8 c
NM101	1.7 a	2.8 b	1.0 ac	1.8 c
MY26	1.2 d	2.2 e	0.9 bd	2.0 a
CM81	1.4 c	2.1 f	0.8 de	1.6 e
CY25	1.3 cd	2.9 a	1.1 a	2.0 a
MM51	1.6 b	2.5 c	1.1 ab	1.8 c
CM104	1.1 e	2.0 g	0.9 ce	1.7 d
CM34	1.3 cd	2.5 c	1.1 a	1.9 b

Note: All values were measured with three replications (n = 3). The data of 2019–2020 and 2020–2021 cropping seasons were collected 10–40 days and 15–40 days after flowering, respectively. Different lowercase letters indicate a significant difference at p < 0.05.

).

4. Discussion

4.1. Variations in Grain Weight per Spike and Individual N Accumulation at the Maturity Stage Affect Simultaneous Improvements in the Grain Yield and Protein Contents

Previous studies have shown that the grain yield of wheat was mainly determined by the grain number per unit area [29] due to the trade-off relation between the fertile spikes and grain number per spike [30], as well as the decisive effect of genotypes on the thousand grain weight [31]. However, both fertile spikes and grain number per spike were driven by environmental factors [32], which was in agreement with our results. The results in this study also clearly indicated that the grain weight per spike accounted for a greater yield variation than fertile spikes and can be considered as a direct indicator for selecting dryland wheat with a high grain yield. This is because a large number of infertile tillers died after jointing due to a decrease in the N uptake resulting from water deficits under dryland soil conditions [33]. In fact, the high grain yield that mainly resulted from high fertile spikes was only observed on CM34 in the 2018–2019 cropping season. Therefore, most of the grain yield of dryland wheat was caused by the main stem due to the competitive advantages of main stem leaves in relation to light, water, and nutrition [34].

The results of a 3-year field experiment in this study indicate that the mean and stability of the grain yield of dryland wheat were mainly related to the grain number per spike, except for the stable and high grain yields of CY25 resulting from a high thousand grain weight. Genetic variations in the grain number per spike could be explained by the difference in the spike DM at the anthesis stage and fruiting efficiency among wheat genotypes [35], while the trade-off effect between the spike DM at the anthesis stage and fruiting efficiency was mainly determined by the environment [36]. Although the plant height and spike DM have been used as critically agronomic traits for selecting wheat genotypes with a high grain number per spike, CY51 with an erect flag leaf, and NM101 with a large flag leaf in this study showed high and low grain numbers per spike in the 2018–2019 cropping season at an equally high spike dry matter level, respectively. This result is consistent with the finding [37] that the canopy structure with erect leaves increased the grain number of wheat by improving the photosynthetic rate and radiation use efficiency during the spike growth period. The grain number per spike and fruiting efficiency significantly increased in the 2020–2021 cropping season, compared with that in the other two cropping seasons, with a significant increase in the partitioning of DM to spikes at the anthesis stage. These results confirm that the increased grain number per spike resulting from an improved fruiting efficiency may be related to an increase in the proportion of assimilates allocated to a spike [38]. Based on these findings, it is evident that improving fruiting efficiency by reducing ineffective tillers [39] and increasing light interception ability is an effective way of increasing the grain number per spike and grain yield under dryland soil conditions.

Simultaneous increases in the grain weight per spike and individual N accumulation at the maturity stage simultaneously improved the grain yield and protein contents of dryland wheat. In this study, the genetic variation in grain protein contents was mainly determined by the grain weight per spike and N accumulation during the maturity stage, rather than the NHI. These results also supported the finding that the NHI of modern wheat has already approached the maximum [29]. Therefore, increasing the grain protein contents without reducing the grain yield requires improvements in the grain weight per spike alongside increases in individual N accumulation at the maturity stage. However, the findings of previous studies on the impact of N accumulation at anthesis [22] and post-anthesis N accumulation stages [24] on GPD (higher grain protein contents at an expected grain yield) are still controversial. Our results suggest that simultaneous increases in the grain number per spike and N accumulation at the anthesis stage may lead to great dilution effects of C to N. These results also support the conclusion that source N supply cannot match the increase in sink N demands [16]. It is noteworthy that NM101 showed higher grain protein contents than MM51 at an equally high grain number per spike due to a larger leaf area index. These results indicate that the simultaneous increases in the source N supply and N storage capacity helped to decrease the above dilution effect under dryland soil conditions. Additionally, we attributed the simultaneous increases in the grain yield and protein contents of CY25 to a high thousand grain weight and post-anthesis N accumulation, which was also consistent with the conclusion that GDP was mainly related to post-anthesis N accumulation [24]. Root traits [40] and high expressions of TaNRT2 [41] helped to increase N acquisition, which may explain the critical role of increased post-anthesis N accumulation in simultaneously improving the grain yield and protein contents of dryland wheat. Physiological mechanisms [42] related to these hypotheses need to be confirmed in future studies.

4.2. The Genotype–Environment Interaction Effect on the Gluten Index Was Related to N Translocation Affected by Source N Supply and Sink N Demands

Grain protein contents showed a stronger linear relationship than wet gluten contents with sedimentation values in this study. This result suggests that genetic variations in the gluten index weakened the expected linear relationship between the wet gluten contents and sedimentation values. In addition, the results based on a great variation in the gluten index of some wheat genotypes (e.g., CY25, CY26, and NM101) among three cropping seasons confirmed the finding that the gluten index was also affected by the interaction effect between the genotype and the environment [6].

The translocation of N to the grains for the synthesis and polymerization of gluten protein occurred during the later grain filling stage [22]. However, the onset and ratio of the synthesis of glutenin and gliadin in relation to the gluten index varied among wheat genotypes and were strongly affected by the environment [43]. One possible hypothesis is that the genotype and environment factors changed grain protein components by affecting the grain N accumulation rate and N accumulation duration, respectively, resulting in variations in the gluten index [44]. Therefore, N translocation to grains may be related to the gluten index. In the current study, drought stress caused by reduced precipitation in the 2019–2020 cropping season and the extension of grain filling duration caused by increased post-anthesis N accumulation in the 2020–2021 cropping season led to variations in the gluten index of some wheat genotypes among different cropping seasons. Although a study attributed the improvement in the gluten index caused by drought to the increase in the N translocation efficiency [44], our study supported another hypothesis which stated that the gluten index was related to the onset of N translocation regulated by canopy senescence, rather than N translocation efficiency. This was because the N was efficiently transferred to grains at last [45].

The N translocation was affected by both the source N supply and the sink N demand under dryland soil conditions. Accelerated senescence generally occurs in a dryland soil environment due to the decreases in N availability and post-anthesis N uptake caused by water deficit [46]. The increase in N accumulation at the anthesis stage and the decrease in post-anthesis N uptake in the 2019–2020 cropping season accelerated N translocation, resulting in simultaneous increases in the N translocation efficiency and the gluten index. However, a significant decrease in the gluten index of MY26 and NM101 in the 2018–2019 cropping season was related to the reduced sink N demand resulting from a decrease in the grain number per spike. Our results may support another hypothesis: that the increased precipitation during the later grain filling stage delayed the translocation of N [47], resulting in a decrease in the gluten index.

4.3. Photosynthesis Improvements Simultaneously Improved the Grain Yield and Gluten Quality by Increasing N Accumulation or Contributing to N Translocation

Drought decreased the net photosynthetic rate of flag leaves by reducing rubissco enzyme activity and stomatal opening, resulting in a decrease in the grain yield. Thus, the net photosynthetic rate during the 2019–2020 cropping season was lower than that during the 2020–2021 cropping season. At the same time, we attributed an increase in the net photosynthetic rate during the 2020–2021 cropping season to the improved sink carbon demands resulting from the increased grain number per spike. The interaction effects between the photosynthesis, N uptake, and accumulation have been reported in previous studies [19]. Therefore, the genetic variation in grain protein contents resulting from differences in individual N accumulation at the maturity stage may be related to the factors affecting carbohydrate production, such as the leaf area, net photosynthetic rate, and “stay-green” property. In this study, we attributed an increase in the individual N accumulation at the maturity stage of NM101 and CY25 to the large leaf area index and high net photosynthetic rate, respectively. This was because the high temperature and drought under dryland soil conditions generally reduced rubisco enzyme activity and induced stomatal limitation, which resulted in a decrease in the net photosynthetic rate and a fast canopy senescence [48].

However, our results show that a large leaf area and high net photosynthetic rate differed in simultaneously improving the grain yield and gluten quality of dryland wheat, which may be related to N translocation. On the one hand, an increase in the leaf area index improved N accumulation before the flowering stage. On the other hand, the large leaf area reduced the net photosynthetic rate and increased the water stress due to the decrease in the N content per unit area and the increase in transpiration, respectively, which contributed to N translocation and helped to maintain a high gluten index. In addition, the translocation of N from the lower leaves to flag leaf delayed the senescence of the flag leaf, resulting in the extension of flag leaf photosynthesis. Based on these findings, it is evident that simultaneous increases in the grain yield and gluten quality of NM101 were related to simultaneous improvements in N accumulation at the maturity and N translocation stages resulting from the increased leaf area and grain number per spike.

By contrast, simultaneous increases in the grain yield and gluten quality of CY25 were related to the improvements in N accumulation at the maturity stage and the grain filling rate resulting from increases in the net photosynthetic rate and thousand grain weight. Previous studies have reported that increases in post-anthesis N accumulation improved the grain protein contents of wheat without decreasing the grain yield [24], which was consistent with our finding that CY25 showed significantly higher post-anthesis N accumulation than other wheat genotypes, especially in the 2019–2020 and 2020–2021 cropping seasons. However, the results from the 3-year field experiment in this study indicate that the simultaneous increases in the grain yield and gluten quality of CY25 were related to improvements in N accumulation at the maturity stage and the N accumulation rate resulting from increases in net photosynthetic rate and thousand grain weight. On the one hand, the increased carbon assimilation ability in relation to the high net photosynthetic rate provides the carbon source and energy for the N uptake of root, especially during the grain filling stage after the flowering stage. On the other hand, the increase in the grain filling rate associated with a big grain size improved the net photosynthetic rate and N accumulation rate. Based on these findings, synergy effects among the net photosynthetic rate, grain filling rate, and N accumulation rate contributed to simultaneous increases in the grain yield and wet gluten contents. The results in our study also support the findings that a single kernel weight, post-anthesis N accumulation, and N accumulation rate were determined by genotype effects [44,49].

5. Conclusions

Our findings suggest that simultaneous improvements in the grain yield and gluten quality of CY25 were related to increases in the grain size and wet gluten contents, which benefited from synergy effects in the net photosynthetic rate, the post-anthesis N accumulation rate, and the grain filling rate. By contrast, we attributed simultaneous improvements in the grain yield and gluten quality of NM101 to the increasing grain protein contents while maintaining a high gluten index, which was mainly related to increases in the leaf area index and the number of grains per spike. Moreover, our study clearly indicates that the effect of genotype–environment interaction on the gluten index was related to N translocation regulated by canopy senescence and N translocation under dryland soil conditions was affected by source N supply and sink N demands in relation to the grain number per spike. Our findings should help to simultaneously improve the dryland wheat grain yield and gluten quality by identifying desirable N-efficient genotypes in future breeding and N management studies.