Nitrogen application affects maize grain filling by regulating grain water relations

WU Ya-wei,ZHAO Bo,Ll Xiao-long,LlU Qin-lin,FENG Dong-ju,LAN Tian-qiong,KONG Fan-lei,Ll Qiang,YUAN Ji-chao,3

1 College of Agriculture,Sichuan Agricultural University,Chengdu 611130,P.R.China

2 Chongqing Key Laboratory of Economic Plant Biotechnology/Collaborative Innovation Center of Special Plant Industry in Chongqing/Institute of Special Plants,Chongqing University of Arts and Sciences,Chongqing 402160,P.R.China

3 Crop Ecophysiology and Cultivation Key Laboratory of Sichuan Province,Chengdu 611130,P.R.China

Abstract Grain water relations play an important role in grain filling in maize.The study aimed to gain a clear understanding of the changes in grain dry weight and water relations in maize grains by using hybrids with contrasting nitrogen efficiencies under differing nitrogen levels.The objectives were:1) to understand the changes in dry matter and percent moisture content (MC) during grain development in response to different nitrogen application rates and 2) to determine whether nitrogen application affects grain filling by regulating grain water relations.Two maize hybrids,high N-efficient Zhenghong 311(ZH311) and low N-efficient Xianyu 508 (XY508),were grown in the field under four levels of N fertilizer:0,150,300,and 450 kg N ha-1 during three growing seasons.Dry weight,percent MC and water content (WC) of basal-middle and apical grains were investigated.The difference in the maximum WC and filling duration of basal-middle and apical grains in maize ears resulted in a significant difference in final grain weight.Grain position markedly influenced grain drying down;specifically,the drying down rate of apical grains was faster than that of basal-middle grains.Genotype and grain position both influenced the impact of nitrogen application rate on grain filling and drying down.Nitrogen rate determined the maximum grain WC and percent MC loss rate in the middle and the late grain-filling stages,thus affecting final grain weight.The use of high N-efficient hybrids,combined with the reduction of nitrogen application rate,can coordinate basalmiddle and apical grain drying down to ensure yield.This management strategy could lead to a win-win situation in which the maximum maize yield,efficient mechanical harvest and environmental safety are all achieved.

Keywords:maize,grain filling,grain drying down,nitrogen,maximum water content

1.lntroduction

China comprises one of the largest maize producing areas in the world.China’s maize planting area is more than 37 million hectares and the total production is more than 215 million tons (Houet al.2020).Furthermore,maize production accounts for approximately 55% of the overall annual crop production in the country;however,it needs to grow at an annual rate of more than 2.4% to ensure food security (Shiferawet al.2011;Rayet al.2013).Increasing nitrogen (N) application has been the most simple and effective measure to increase maize yield (Ladhaet al.2016;Mueller and Vyn 2016).However,in China,N fertilizers are typically applied at levels that are higher than the uptake demand of crops,not only leading to the decline of maize yield and quality,but also causing high production costs,environmental pollution,and decreasing plant N efficiency(Guet al.2016;Liuet al.2016;Duet al.2021).From the 1980s (13.2 kg N ha-1) to the 2000s (21.1 kg N ha-1),the annual accumulation of N increased by about 8 kg ha-1.At present,the average soil N surplus of maize agricultural system in China is 72 kg N ha-1(Liuet al.2013;Chenet al.2014).In recent years,China has gradually achieved zero growth in chemical fertilizers and pesticides,mainly due to the government’s large investment in comprehensive technical research that has reduced chemical fertilizer use,and improved fertilizer utilization rate.

As a solution,the use of high N-efficient hybrids (i.e.,those with higher N absorption and utilization) has proved to be an effective measure to decrease N application rates,ameliorate damage to the environment,and at the same time prevent food crises by ensuring a steady growth of maize yield (Chenet al.2015;Maet al.2015;Li Qet al.2017).

The development of the maize grain follows a predictable pattern whereby a gradual continuous increase in dry weight is concomitant with an initial increase in water content (WC) followed by its gradual decrease,implying a parallel continuous gradual decrease in percent moisture content (MC) from the initial stages of grain growth and development (Borrás and Westgate 2006).The time-course of the increase in grain dry weight shows a sigmoidal pattern,described as a“slow-fast-slow”growth trend (Renet al.2013).The early stage in grain development is characterized by the rapid increase in fresh weight,as active cell division and differentiation drive the initial increase in WC,while dry matter accumulation occurs at a low pace.Then,in the middle stage,rapid dry matter accumulation in the grain is observed,accompanied by further but lower increase in WC.Soon after,the maximum grain volume is attained and then the grain begins to dry down.Finally,during the late stages in grain development dry matter accumulation slows down,the grain continues to dry down until it reaches physiological maturity and dry matter accumulation peaks (Hillson and Penny 1965;Westgate 1986;Borrás 2003).

Grain filling is determined by the interaction between deposition of reserve materials and decreasing cellular WC,whereby the accumulation of reserves,such as starch,replaces water until a percent MC threshold is reached.The value of 35% was previously reported and is widely accepted as the MC of grain at physiological maturity(Borrás and Westgate 2006;Salaet al.2007a;Liet al.2020).As the grain drying down,its metabolic capacity becomes increasingly limited in the late stages of grain filling (Westgate 1994;Borrás 2003;Borrás and Westgate 2006).Thus,grain-filling duration after maximum WC has been reached is mainly determined by the rate of water loss and the rate of dry matter accumulation (Borrás and Westgate 2006).Ultimately,final grain weight is closely related to grain volume.Therefore,when the grain reaches its maximum volume,its potential final weight has been established (Borrás and Westgate 2006).As maximum grain volume is determined by the maximum WC reached(Martinez-Carrasco and Thorne 1979;Saini and Westgate 1999;Borrás 2003),the maximum WC is an independent index to estimate the sink size of the grain.

Due to differences in sink activity (Kato 2004;Yanget al.2006) and assimilate supply (Yanget al.2003,2004),there are significant differences in the degree of grain filling and final grain weight at different positions on the maize ear,thus giving rise to the basal-middle grains of ear and the apical grains.The rate and duration of grain filling affect the extent of grain storage capacity,which in turn determine maize yield and quality (Liet al.2013;Shenet al.2017).Grain yield of maize hybrids with differing maturity can be improved by prolonging the active grain-filling period,effective grain-filling time and grain-filling duration in the middle and late stages,and enhancing the mean rate of grain filling at the early stage (Wanget al.2014).The duration of linear grain filling is related to the environmental conditions during the grain-filling stage for a given hybrid(Chenet al.2013).For example,grain-filling duration can be significantly shorten by drought (Barlowet al.1980;Brookset al.1982),severe pathogen infection (Pepleret al.2006),or defoliation (Echarteet al.2006;Salaet al.2007b).Some researchers have studied the relationship between N content and grain-filling characteristics.Adequate N application effectively optimized grain filling (Qiet al.2016;Weiet al.2018).Within a certain range,N application increased the duration and rate of grain filling,thereby improving maize yield (Fernandez and Ciampitti 2019).Some researchers have found that applying N during spikelet differentiation significantly increased sink strength,grain-filling rate and grain weight of weak spikelets at the heading stage in rice(Fuet al.2011).Grain drying down was also affected by the amount of N application;thus,a low N stress accelerated maize grain drying down after physiological maturity (Hickset al.1976;Qiaoet al.2017).Previous studies have focused on the effect of N fertilizer on grain filling;however,there are few reports on the effect of N fertilizer on WC and percent MC of basal-middle and apical maize grains and the relationship between N fertilizer and grain filling.Furthermore,studies on these differences in hybrids with contrasting N efficiencies are scarce.

In this study,the response of dry matter,percent MC and WC were evaluated at each grain position and under different N levels in maize hybrids that differ in N efficiency,by using the corresponding accumulated temperature-effect model.Our objectives were:1) to understand the changes in dry matter and percent MC during grain development and the response of maize hybrids differing in N efficiency to different N application rates;and 2) to determine whether N application rate affects the grain-filling process by regulating grain water relations;especially,to assess the effect of N application rate on the maximum WC,and the relationship between changes in percent MC and grain-filling rate,and grain-filling duration.

2.Materials and methods

2.1.Growth conditions and plant materials

Field experiments were conducted in 2017,2018 and 2019 in Sichuan Zhongjiang Experimental Station of Sichuan Agricultural University (Sichuan Province,China;31.03°N,104.68°E),and the meteorological data from maize sowing to harvest are shown in Fig.1.The soil of the experimental site is purple clay,and the soil nutrient content of the top soil layer (0-20 cm)is shown in Table 1.This site generally experiences a subtropical monsoon climate.

Table 1 Soil properties of the top soil layer (0-20 cm) in the experiment

Fig.1 Daily minimum temperature (Tmin),the maximum temperature (Tmax),the mean temperature (Tmean),and precipitation during maize growing seasons in 2017 (A),2018 (B) and 2019 (C) at Zhongjiang Experimental Station,Sichuan Agricultural University,China.

Two hybrids,Zhenghong 311 (ZH311) and Xianyu 508(XY508),were used.ZH311 was provided by Sichuan Agricultural University Zhenghong Biotechnology Co.,Ltd.and XY508 was provided by Tieling Pioneer Technology Co.,Ltd.,China.The two hybrids are popularized in Southwest China,and their growth period is similar (approximately 120 days).ZH311 and XY508 were categorized as N-efficient and N-inefficient,respectively,according to their nitrogen use efficiency and nitrogen responsiveness (Li Qet al.2017;Wuet al.2019).

2.2.Experimental design

A two-factor randomized block experimental design with three replicates was used.Treatments in experiments consisted of the factorial combination of (i) two hybrids(N-efficient hybrid ZH311;N-inefficient hybrid XY508) and(ii) four N levels (0,150,300,and 450 kg ha-1).Eight treatments and three repilicates were performed.Plots were 7.5 m×6.0 m (45 m2in total).The gap between high N and low N was increased to better observe the effects of different N application rates.The N source was urea(N concentration=46%),which was applied both as basalfertilizer (50%) at the sowing stage and supplementary fertilizer (50%) at the big trumpet stage (V12).In all treatments,600 kg ha-1of superphosphate and 150 kg ha-1of potassium chloride were applied as base fertilizer.

Maize was planted on 30 March 2017,6 April 2018 and 24 March 2019,which represented normal planting dates in the region.Ridge and mulch sowing were used (remove film in V12).Cultivars were planted in wide (1.1 m) and narrow(0.5 m) rows,and 52 500 plants ha-1were used.Weeds,insects and diseases were adequately controlled during the growing period;however,no irrigation was provided because this is a typical rain fed agricultural region (irrigation was applied after seeding only).Other crop management measurements were carried out according to the maize requirements for high yield.

2.3.Sampling and measurements

One hundred representative pest-and disease-free maize plants were selected per plot before the silking stage.Between 5 and 50 d after silking,five ears were taken from each of the selected plants every 5 d.The ears were divided into basal-middle and apical grains (Fig.2).

Fig.2 Division of basal-middle and apical grains.

Incomplete grains were discarded and the remaining full grains were thoroughly mixed,at the initial grain-filling stage,especially on the 5 and 10 d,due to the larger MC in the grains.The grains were damaged greatly in the whole separation process.One hundred fresh grains were randomly sampled and weighed.Sampled grain lots for each treatment were weighed six times,wrapped,dried at 105°C for 30 min,dried again at 80°C to a constant weight,and reweighed.The grain MC was measured by the difference between the fresh and dry weights.The silking dates of ZH311 and XY508 were respectively June 15 and June 10 in 2017,June 11 and June 6 in 2018,and June 7 and June 2 in 2019.

2.4.Calculations

Grain-filling modelTo determine changes in grain filling while avoiding the influences of environmental conditions,this study investigated the relationship between the thermal time after silking (≥10°C accumulated temperature) and grain filling.The Richards equation (Richards 1959) based on the nonlinear least square method was used.It used thermal time after silking (T) as the independent variable and the dry weight (W) of 100 grains as the dependent variable(Zhanget al.2017):

whereA,B,K,andNare the model parameters.Ais the theoretical maximum dry weight of 100 grains,Bis the initial value parameter,Kis the growth rate parameter,Nis the shape parameter.

Its first-order derivative is the grain-filling rate,V(g/100-grain (°C d)):

The secondary parameters describing the grain-filling characteristics are derived from Zhuet al.(1988):

Time at the maximum grain-filling rate:

Maximum grain-filling rate:

Average grain-filling rate:

Active grain-filling stage:

The grain-filling process in maize is divided into the initial(T1),middle (T2-T1) and late (T3-T2) filling stages.The second derivative of the grain-filling rate equation forTis calculated and set as 0.The two inflection point coordinatesT1 andT2 are obtained as follows:

It is assumed that when grain filling reaches 99% of the theoretical maximum dry weight of 100 grains (A),it arrives at the actual end of filling (T3):

ValuesT1,T2 andT3 were substituted into eq.(1) to calculate the corresponding 100-grain weights (W1,W2 andW3).The grain-filling volume at each stage (W1,W2-W1 andW3-W2),the corresponding grain-filling rates (V1 (W1/T1),V2 (W2-W1)/(T2-T1) andV3 (W3-W2)/(T3-T2)) were calculated.

Grain drying down modelA logistic equation was applied using the thermal time after silking (t) as the independent variable and the grain percent MC (Wt) as the dependent variable:

whereA1,A2,X0,andPare the model parameters.A1 is the highest grain percent MC andA2 is the lowest grain percent MC.In the early stage of grain formation,the percent MC is～90% (Ashmore 2015).Until 12 d after pollination,the grain percent MC is 80-90% (Jing 2014).Therefore,A1 was set to 90% in the present study.The grain breakage rate at harvest is the lowest when the grain percent MC is 20-23% (Hall and Johnson 1970).In the present study,A2 was set to 20%.

The drying down rate equation (G (% (°C d)-1)) can be obtained by calculating the first-order derivative of equation(10):

Accumulated temperature at the maximum drying down rate:

Maximum drying down rate:

The inflection pointst1 andt2 are obtained by calculating the secondary derivative of the drying down rate equation fortand setting it to 0:

The accumulated temperaturet3 is set when the grain percent MC reaches 28%.The latter is adapted to the mechanized harvesting standard (Hall and Johnson 1970;Chowdhury and Buchele 1978),and it is consistent with the grain percent MC of maize at physiological maturity (Li L Let al.2017).The drying down process is divided into the initial (t1),middle (t2-t1),and late (t3-t1) drying down stages.The percent MC oft1,t2 andt3 areWt1,Wt2 andWt3,respectively.The drying down rate at each stage is:

A parabolic model was used to relate percent grain water content (WC) to thermal time after silking (TT) (Salaet al.2007b;Chenet al.2013).

Grain yieldHarvest all the remaining ears and threshed and grains were dried to determine grain yield (percent MC was adjusted to 14.0%).

2.5.Statistical analysis

Data were analyzed statistically by analysis of variance(ANOVA) procedure using SPSS v.20.0 (IBM Corp.,Armonk,NY,USA).Comparisons among different treatments were performed with Duncan’s multiple range tests.AP-value＜0.05 was considered significant.Pearson’s correlations were calculated to determine the relationship between different parameters of maize.The Richards equations were fitted and graphs were plotted in OriginPro v.9.0 (OriginLab Corp.,Northampton,MA,USA).

3.Results

3.1.Grain yield

Increasing N application significantly increased maize yield and the high N-efficient maize hybrid ZH311 had a significant yield advantage over the low N-efficient hybrid XY508 at all N fertilizer levels tested in 2017 and 2018,and at 0 and 150 kg N ha-1in 2019 (Fig.3).In 2017 and 2019,the yields of both hybrids showed increasing trends with increasing N application,with the highest yield at 450 kg ha-1.The differences in yield among 0,150 and 300 kg N ha-1were significant,whereas the difference in yield between 300 and 450 kg N ha-1was not significant (ZH311 in 2017 and 2019,XY508 in 2019);in 2017,the yield of XY508 increased significantly more with increasing N application than that of ZH311,while in 2018,the yields of both hybrids first increased and then decreased with increasing N application,with the highest yield observed at 300 kg N ha-1,and there was a significant difference between adjacent fertilization amounts.Further,the yield of ZH311 was less affected by reduced N application than that of XY508.Because of the difference of meteorology and basic soil fertility,the grain yield varies from year to year.In 2017 and 2018,the content of available N in soil was higher before maize sowing.However,the precipitation in the late growth period (July)of maize in 2018 was large,which seriously affected the growth of maize.Therefore,the grain yield was the highest in 2017 and the lowest in 2018.

Fig.3 Grain yield of maize hybrids Zhenghong 311 (ZH311) and Xianyu 508 (XY508) in 2017,2018 and 2019 under four nitrogen levels (0,150,300,and 450 kg N ha-1).Different lowercase letters within year indicate significant differences among N levels and between the two hybrids (P＜0.05).Vertical bars represent ±SE (n=24).

3.2.Effect of N fertilizer on the characteristics of grain filling

The determination coefficients (r) of the equation of maize basal-middle and apical grains were above 0.97,indicating a reasonable fit by Richards equation (Table 2).From the perspective of theoretical maximum 100-grain dry weight(A),basal-middle grains showed an obvious superiority over apical grains.TheAvalues of the basal-middle ZH311 and XY508 grains were 10.63 and 24.19% higher,respectively,than those of their apical counterparts (three-year average).N level and its interaction with maize hybrid showed significant effects on grain filling at different ear positions.TheAvalue for ZH311 and XY508 increased with increasing N (except for ZH311 in 2017),particularly at low N levels.Applying more N fertilizer optimized the filling of apical grains and increased the 100-grain weight,thus decreasing the gap with respect to basal-middle grains,especially in the high N-efficient hybrid ZH311.When N application increased from 300 to 450 kg N ha-1,the increase of 100-grain weight in the low N-efficient hybrid XY508 was significantly larger than that of ZH311,which showed almost no change,and even showed a decreasing trend.This explains the advantage shown by ZH311 to develop a higher grain weight potential at low and medium N application rates,while XY508,a low N-efficient hybrid,needs higher N levels than ZH311 to stimulate its grain weight potential.

The grain-filling characteristic parameters of basalmiddle and apical grains of different N-efficient hybrids were significantly different under different N levels (Table 3).ForD(active grain-filling stage),the basal-middle grains were 15.21% larger than apical grains.Increasing N fertilizer can prolong theD.Interestingly,based on the three-year average,the highestDof the basal-middle and apical grains of the high N-efficient hybrid ZH311 was observed at 150 and 300 kg N ha-1,respectively,whereas for the low N-efficient hybrid XY508,it was observed at 450 kg N ha-1.The basal-middle grains entered the middle filling-stage first (T1 of the basal-middle grains was smaller than that of the apical grains,with an average difference of 96.23°C d),which allowed the basal-middle grains to reach the maximum filling rate first (Tmax·vof the basal-middle grains was smaller than that of the apical grains,with an average difference of 58.40°C d).In contrast,the duration of the middle filling stage (T2-T1) in basal-middle grains was longer than in apical grains (average over 74.44°C d).

Increasing N fertilizer could decreaseT1,allow the grain to enter the middle filling stage earlier,and prolong the duration of the middle filling stage.Concomitantly,it is worth noting that under high N level,the duration of the late filling stage was significantly longer.The high N-efficient hybrid ZH311 maintained a longer duration of the grain middle filling-stage at low and medium N levels,whereas the low N-efficient hybrid XY508 only experienced a longer duration of the middle grain-filling stage at high N levels.

Vmax,Vmeanand mean grain-filling rate of middle stage(V2) of basal-middle grains of ZH311 were lower than those of apical grains (except in 2019,although the difference in this case was not significant),whereasVmax,VmeanandV2 of basal-middle grains of XY508 were larger than the corresponding values for apical grains (over the three years of study) (Table 3).This may be the reason why the difference in 100-grain weight between basal-middle and apical grains in ZH311 was significantly less than that between basal-middle and apical grains in XY508.

Table 2 Richards model predictions of basal-middle and apical grains parameters under different N levels

In terms of cultivar,Vmax,Vmean,mean grain-filling rate of initial stage (V1),V2,and mean grain-filling rate of late stage(V3) of apical grains of ZH311 ears were greater than those of XY508 ears,which seemingly confirms that apical grains of ZH311 have an obvious grain-weight advantage over those of XY508.Conversely,theVmax,Vmean,V2,andV3 of the basal-middle grains of XY508 were greater than those of ZH311,which again confirms that the basal-middle grains of XY508 have an obvious grain-weight advantage over those of ZH311 (Table 2).Our data showed that the variation in grain growth rate was the major factor determining the differences in final grain weight between different genotypes.Based on the three-year average,Vmax,Vmean,V2,andV3 of ZH311 increased gradually with increasing N level,whereas they all decreased in XY508 with increasing N.At 0 kg N ha-1,the basal-middle grains of ZH311 and basal-middle and apical grains of XY508 showed higher filling rate,which may be the response of both genotypes 3/4 and crops in general to N deficiency.

There was a significant positive correlation (P＜0.01)between theoretical maximum 100-grain dry weight (A) and duration of active filling stage (D),middle filling stage (T2-T1) and late filling stage (T3-T2).Conversely a significant negative correlation (P＜0.01) was detected betweenAand duration of initial filling stage (T1) (Table 4).Further,N level was positively correlated with duration of active filling stage(D),middle filling stage (T2-T1) and late filling stage (T3-T2),but negatively correlated with duration of initial grainfilling stage (T1) (Fig.4).Therefore,increasing N levels shortened the duration of initial grain filling,caused grains to enter the middle filling stage earlier,and prolonged the duration of the middle and late grain-filling stages,ultimately increasing final grain weight at maturity.

Fig.4 Correlation analysis for N level and grain-filling characteristic parameters.A,relationship between nitrogen fertilizer level(N-levels) and duration of active grain-filling stage (D).B,relationship between nitrogen fertilizer level (N-levels) and grain-filling duration of initial stage (T1).C,relationship between nitrogen fertilizer level (N-levels) and grain-filling duration of middle stage(T2-T1).D,relationship between nitrogen fertilizer level (N-levels) and grain-filling duration of late stage (T3-T2).Each point in the figure represents the average value of the two hybrids over three years.

Table 4 Correlation analysis between filling characteristic parameters and ultimate growth1)

3.3.Effect of N level on the characteristics of grain drying down

With increasing post-silking accumulated temperature,grain percent MC decreased first slowly then rapidly,and gradually slowed down thereafter.This trend is consistent with the logistic equation (Fig.5).TheR2values indicating the degree of model fit laid between 0.9659 and 0.9964(Table 5).The trends depicted reliably reflect the association between grain percent MC and post-silking accumulated temperature,and accurately predict grain position,hybrid and inter-annual variations.

Fig.5 Predictive model of percent moisture content (MC) of basal-middle and apical maize grains under different N levels.Each data point represents the average of grain percent MC (measured six times).Drying down model of apical grains (A,B,E,and F)and basal-middle grains (C,D,G,and H) at different N levels.Drying down model of Xianyu 508 (XY508) (A,B,C,and D) and of Zhenghong 311 (ZH311) (E,F,G,and H).

Table 5 Logistic model parameters for basal-middle and apical grains under different N levels

The basal-middle grains first entered the middle drying down stage (t1 of the basal-middle grains was lower than that of the apical grains,with an average difference of 95.72 °C d),which allowed the basal-middle grains to reach the maximum drying down rate first (tmax·Gof the basalmiddle grains was lower than that of the apical grains,with an average difference of 58.40°C d) (Table 6).Thet2-t1andt3-t2 of basal-middle grains were larger than those of apical grains.The application of nitrogen increasedt2-t1 andt3-t2.These results showed that the increase of N fertilizer prolonged drying down during the middle and late stages of grain filling.

Interestingly,the drying down rate of apical grains was significantly higher than that of basal-middle grains(Table 6).For apical XY508 grains,the drying down rates at the initial (G1),middle (G2) and late (G3) drying down stages,and the maximum (Gmax) drying down rates were 8.05,12.17,18.22,and 6.75% higher,respectively,than those for the basal-middle grains.The same parameters for the apical ZH311 grains were 12.99,14.11,7.84,and 12.40% higher,respectively,than the values for the basa-middle grains in ears of this genotype.Increasing N application reduced the grain drying down rate;however,this effect seemed to be position-and hybrid-dependent.For each 150 kg ha-1increase in N application,Gmax,G1,G2,andG3 of apical grains decreased by 2.23,1.53,2.19,and 2.34%,respectively,while those of basal-middle grains decreased by 2.68,1.87,2.77,and 3.38%,respectively.In turn,for each 150 kg ha-1increase in N application,Gmax,G1,G2,andG3 of ZH311 decreased by 1.97,1.32,2.15,and 2.51%,respectively,and the corresponding values for XY508 decreased by 2.95,2.08,2.82,and 3.21%,respectively.Therefore,the application of N fertilizer can cause the grain to enter the middle drying down earlier,but it can reduce the grain drying down rate and prolong the drying down time in the middle and late stages of the grain,thus delaying the suitable timing of grain harvest.

3.4.Synchronization of grain filling and drying down in maize

Previous studies on grain development in maize reported a significant positive correlation between the maximum WC and final grain weight (Borrás 2003;Borráset al.2009;Chenet al.2013).The present study used a parabolic model to fit the relationship between grain WC and thermal time after silking.The determination coefficient of the correlation for apical grain was greater than 0.7387 (P＜0.001),while that for basal-middle grain was greater than 0.7329 (P＜0.001)(Fig.6).Maximum WC of basal-middle grain was higher than that of apical grain at each N level.Thus,applying more N fertilizer can increase maximum WC in maize grains during their development (especially in basal-middle grains).This study studied the relationship between N level and maximum WC and found a positive correlation between them (especially in basal-middle grains).The relationship between the maximum WC and final grain weight at maturity was then established.It should be noted that apical grains showed a positive correlation between these two variables,although not significant (R2=0.6130,P=0.2171);conversely,basal-middle grain showed a highly significant and positive correlation (R2=0.9510,P＜0.001).Maximum WC differed among hybrids (Table 7).On average,maximum WC in apical grains of the high N-efficient ZH311 and the low N-efficient XY508 was similar,whereas maximum WC in basal-middle grain of ZH311 was significantly lower than in basal-middle grain of XY508.

Fig.6 The relationship between water content (WC) and thermal time after silking (≥10°C) during grain development in maize in 2017 and 2018.A,apical grain.B,basal-middle grain.The insets show the relationship between N-level and the maximum WC and the relationship between the maximum WC and ultimate growth A,which is divided into apical grain and basal-middle grain.

The correlations between some parameters of the maize grain-filling model and the drying down model were analyzed.There were significant negative correlations between the duration of middle stage of grain filling and the grain-filling rate at that same stage to the drying down rate at the middle drying down stage (P＜0.001).Additionally,there were significant negative correlations between the duration of the late grain-filling stage and the grain-filling rate at that same stage to the drying down rate at the late drying down stage (P＜0.001).This result implies that the decrease in grain-filling rate and grain drying down rate will prolong grain-filling duration in the middle and late stages of grain filling.

4.Discussion

Maize grain weight is a quantitative trait controlled by multiple genes that are frequently affected by environmental factors (Zhanget al.2013).Further,grain size,dry matter accumulation rate and weight are reportedly lower in apical grains than in basal-middle grains (Zhaoet al.2018).Dry matter accumulation during grain-filling is generally determined by the duration and rate of grain filling.In this study,the duration of the middle and late grinfilling stages of basal-middle grains was longer than that of apical grains,whereby the weight of basal-middle grains was higher.The earlier the linear phase of grain filling (i.e.,rapid grain filling) starts,the larger the final grain weight will be,whereas,the later the linear phase of grain-filling starts,the lower the final grain weight will be (Chenet al.2013).Our research showed that the middle grain-filling stage (i.e.,rapid grain-filling stage) of basal-middle grains occurred earlier than that of apical grains due to their earlier fertilization and development initiation,and thereby,to their greater sink strength.In contrast,having been fertilized later,apical grains had a lower sink strength,which reflected as a lower growth rate,a shorter linear grain-filling phase duration,and ultimately,a lower final grain weight.These findings are consistent with previous reports (Reddy and Daynardet al.1983;Cárcova and Otegui 2007;Chenet al.2013).The difference between basal-middle and apical grain-weight was significantly greater in XY508 than in ZH311 (Table 2).Thus,the lower weight of apical grains limited yield performance of XY508.Indeed,such a large variation between basalmiddle and apical grains indicates that individual grain-weight variation along the ear cannot be neglected while estimating maize yield,especially in stress-prone environments.As grainfilling duration was shorter in apical grains than in basal-middle grains,the maximum and mean grain-filling rates,and the grain-filling rate at the middle grain-filling stage of apical grains in the high N-efficient hybrid ZH311 were higher than those of the basal-middle grains,whereby the gap between basal-middle and apical grains was effectively reduced.Conversely,the grain-filling rate of the apical grains of the low N-efficient hybrid XY508 was lower than that of the basal-middle grains,which partly explained the large difference in grain weight between ear positions in XY508 (Table 4).

Previous studies have shown that an adequate N application rate contributes to optimize grain filling,while the difference in weight between grain positions might be minimized by N level (Shenet al.2017).Consistently,our results showed that increasing N fertilizer application rate optimized apical grain filling,thereby decreasing the weight difference between basal-middle and apical grains.Interestingly,the effect of increasing N application rate on the development of apical grains of the high N-efficient effect on the low N-efficient hybrid XY508,while the effect of increasing N on the development of basal-middle grains was the opposite (Table 2).In maize,the apical grains initiate growth 4-5 days later than basal-middle grains,and have a higher possibility of abortion prior to the onset of grain filling (Tollenaar and Daynard 1978a).In this study,the increase of N fertilizer application reduced the duration of early grain-filling,which in turn significantly reduced the risk of grain abortion,thus allowing the grain to enter the middle stage of grain-filling earlier,and prolong the duration of the middle and late grain-filling stages.The duration of grainfilling is limited by the availability of assimilate supply from the source leaves and the ability for continuous metabolism of imported assimilate in the grain (Salaet al.2007a).With the increase of N level,the coordination between source capacity and sink strength was ensured,and the supply of assimilates was sufficient,which effectively explained the extension of grain-filling duration.When the application of N fertilizer increased from 300 to 450 kg N ha-1,the increase in 100-grain weight of XY508 was significantly greater than that for ZH311,in which case little change was observed;indeed,even a slightly downward trend was detected (Table 2).The duration of the highest active filling stage in ears of ZH311 was attained at 150 and 300 kg N ha-1,at which levels this hybrid maintained a longer middle grain-filling stage;in turn,the duration of the highest active filling stage in ears of XY508 required 450 kg N ha-1for the longer middle grain-filling period to occur (Tables 3 and 4).Therefore,ZH311,a high N-efficient hybrid,showed greater advantage to develop grain weight potential at low and medium N application levels,whereas the low N-efficient hybrid such as XY508 needed a higher N level to stimulate grain weight potential.

Table 7 Maximum water content (WC) of basal-middle and apical grains of different hybrids under different N levels

Grain drying down is affected by the interaction between genotype and environment,and is mainly controlled by grain growth and development during grain filling (Cross 1991;Magariet al.1997),when starch,protein and other substances are continuously accumulating in the grain,and water is replaced and consumed,thus leading to the gradual decrease in percent MC (Liet al.2018).In the present study,the apical grains of ZH311 showed an advantage in filling rate,while in XY508,the basal-middle grains showed an advantage in filling rate,although the drying down rate of the apical grains of both hybrids was faster than that of the basal-middle grains.Nonetheless,this finding was not explained by a positive relationship between grain filling and grain drying down rates,but by the relative differences in grain volume.Because of such differences,the dry matter content in apical grains was relatively higher over the same time;therefore,the relative percent MC decreased rapidly.Increasing N application slowed down the rate of drying down and prolonged the duration of drying down in the middle and later stages,which was consistent with the pattern and extent of dry matter accumulation.N fertilizer effects on grain drying down was grain position-and hybriddependent.Specifically,the effect was higher on XY508 than on ZH311,and on the basal-middle grains than on apical grains.It is worth noting that,the grain drying process in the field can be divided into physical and physiological drying processes.Our study found that,N applications affected the physical drying processes significantly.However,the drying after physiological maturity is mainly the process of water vapor exchange between grains and the atmosphere(Cross 1991;Li L Let al.2017;Wang and Li 2017a),and whether fertilization measures have a sustained effect on it remains to be studied.

The regulation of water absorption and loss by maize grains is an important determinant of grain development(Borrás 2003;Borrás and Westgate 2006;Gambínet al.2007).Many studies have shown that final grain weight relates to the sink strength of the grain as established in the early stage of grain filling,which can maximize sink potential after ensuring a sufficient supply of assimilate in the middle and late stages of grain filling (Borrás 2003;Borráset al.2004;Chenet al.2013).Furthermore,maximum sink strength is concomitant to maximum grain volume,and consistently,water absorption during the early grain-filling stage largely accounts for final grain volume.Ultimately,maximum WC is associated with the maximum grain volume,and is positively correlated with the final grain weight.Therefore,the maximum grain WC is considered a useful proxy of potential sink strength of the grain (Gambínet al.2007;Borráset al.2009;Chenet al.2013).In our study,the maximum grain WC was increased by N fertilizer.The maximum WC of both basal-middle and apical grains was positively correlated with grain final weight under the different N levels tested;further,although it was highly significant in the case of basal-middle grains,the correlation was nonsignificant for apical grains (Fig.6).Dry matter accumulation in basal-middle grains is limited by sink strength,while the growth of apical grains is mainly limited by assimilate supply,i.e.,source capacity (Tollenaar and Daynard 1978b;Frey 1981;Hanftet al.1986).In our study,the maximum WC of basal-middle grains of ZH311 was significantly lower than that of the basal-middle grains of XY508,which explained the difference in final grain weight.Maximum WC of apical grains in the two hybrids was similar,which indicated that the sink strength of apical grains of the two hybrids was similar in the early stage,but the final grain weight of the apical grains of ZH311 was significantly greater than that of their counterparts in XY508.This showed that a limited assimilate supply was the main reason for the low weight of apical grains in XY508.Therefore,in the case of apical grains,the first goal is to increase source capacity,that is,to increase the supply of photosynthetic products,and then to expand sink strength.The duration of grain filling was affected by water loss and dry matter accumulation (Gambínet al.2007;Yeet al.2020).In this study,final grain weight was positively correlated with the duration of the middle and late grain-filling stages (Table 4),which was significantly and negatively correlated with filling rate in the middle and late stages of grain filling and with drying down rate in the middle and late stages of drying down (Fig.7).

Fig.7 Correlation analysis for grain-filling parameters and drying down parameters.A,relationship between grain-filling rate (V2)and duration (T2-T1) of the middle grain-filling stage.B,relationship between drying down rate (G2) at the middle grain-drying down stage and the duration (T2-T1) of the middle grain-filling stage.C,relationship between the grain-filling rate (V3) and the duration (T3-T2) of the late grain-filling stage.D,relationship between drying down rate (G3) at the late grain-drying down stage and the duration (T3-T2) of the late grain-filling stage.Data from basal-middle grain in 2017 and 2018.

The concept that biomass accumulation stops when grains reach a critical percent MC value underlines the importance of maintaining grain percent MC above this threshold to extend grain-filling duration (Gambínet al.2007).As described above,in the experiments reported herein,the increase of N fertilizer application slowed down the drying down rate in the middle and late stages of grain drying down(Table 6).Furthermore,the regulating effect of N fertilizer on grain weight was reflected in the absorption of water at the early stage of grain filling to increase maximum grain WC and potential sink strength.On the other hand,it reduced the rate of water loss in the middle and late stages of grain filling,thus keeping percent MC above the threshold value for an extended time,thereby ensuring a continued supply of assimilates,prolonging grain-filling duration,maximizing the realization of the sink potential established in the early stage of grain filling,and ultimately increasing final grain weight.

However,the maize mechanized grain-harvesting area in China is less than 5% and maize production costs are consistently higher than in developed countries (Li 2017;Wanget al.2019);thus,there is an urgent need to improve the mechanical grain-harvesting rate for maize.Additionally,a higher percent MC of the grain significantly reduced harvest quality (Maioranoet al.2014;Wang and Li 2017b);therefore,rapid grain drying down becomes a crucial issue in maize production and breeding (Li 2017).In actual production,the varieties with low MC at early maturity and physiological maturity are selected widely.Conversely,increasing N fertilizer application slowed down the rate of grain drying down and prolonged the duration of grain filling,which was beneficial to the increase of grain weight.While in the same production area,the WC of different hybrids was stable at physiological maturity stage (Liet al.2020).Increasing N fertilizer delayed the time from silking to physiological maturity.However,in production areas where the maize cropping season is short,the time left for grain drying down after physiological maturity is insufficient (Liet al.2018;Liuet al.2019).According to this study,ZH311,a high N-efficient maize hybrid,can guarantee an economic yield at low and medium nitrogen levels (Fig.3);thus,the selection of this high N-efficient hybrid for production will reduce N fertilizer use and subsequent environmental pollution,while enhanceing the coordination among the stages of grain filling,drying down,and mechanical harvest.

5.Conclusion

This study found that the difference in the maximum WC and filling duration between basal-middle and apical grains in maize resulted in a significant difference in final grain weight.Grain position markedly influenced grain drying down;specifically,the drying down rate recorded for apical grains was faster than that for basal-middle grains.Ntrogenmediated changes in grain water relations determined the influence of nitrogen fertilizer-input on grain filling.The use of high nitrogen-efficient maize hybrids,coupled with the reduction of nitrogen application,can coordinate basalmiddle and apical grain drying down while ensuring yield.This management strategy can achieve a win-win situation in which maize yield,mechanical harvest efficiency and environmental safety are all accomplished at once.

Acknowledgements

We gratefully acknowledge funding support from the National Key Research and Development Program of China(2018YFD0301206,2016YFD0300209,2016YFD0300307,and 2017YFD0301704).

Declaration of competing interest

The authors declare that they have no conflict of interest.