Optimization of nitrogen fertilization improves rice quality by affecting the structure and physicochemical properties of starch at high yield levels

ZHOU Tian-yang ,Ll Zhi-kang ,LI En-peng ,WANG Wei-lu,YUAN Li-min,ZHANG Hao,LIU Lijun,WANG Zhi-qin,GU Jun-fei,,YANG Jian-chang

1 Jiangsu Key Laboratory of Crop Genetics and Physiology/Jiangsu Key Laboratory of Crop Cultivation and Physiology/Agricultural College,Yangzhou University,Yangzhou 225009,P.R.China

2 Jiangsu Co-Innovation Center for Modern Production Technology of Grain Crops,Yangzhou University,Yangzhou 225009,P.R.China

3 Joint International Research Laboratory of Agriculture and Agri-Product Safety,Ministry of Education of China,Yangzhou University,Yangzhou 225009,P.R.China

Abstract A major challenge in modern rice production is to achieve the dual goals of high yield and good quality with low environmental costs. This study was designed to determine whether optimized nitrogen (N) fertilization could fulfill these multiple goals. In two-year experiments,two high yielding ‘super’ rice cultivars were grown with different N fertilization management regimes,including zero N input,local farmers’ practice (LFP) with heavy N inputs,and optimized N fertilization (ONF). In ONF,by reducing N input,increasing planting density,and optimizing the ratio of urea application at different stages,N use efficiency and the physicochemical and textural properties of milled rice were improved at higher yield levels. Compared with LFP,yield and partial factor productivity of applied N (PFP) under ONF were increased (on average) by 1.70 and 13.06%,respectively. ONF increased starch and amylose content,and significantly decreased protein content. The contents of the short chains of A chain (degree of polymerization (DP) 6-12) and B1 chain (DP 13-25)of amylopectin were significantly increased under ONF,which resulted in a decrease in the stability of rice starch crystals.ONF increased viscosity values and improved the thermodynamic properties of starch,which resulted in better eating and cooking quality of the rice. Thus,ONF could substantially compensate the negative effects caused by N fertilizer and achieve the multiple goals of higher grain quality and nitrogen use efficiency (NUE) at high yield levels. These results will be useful for applications of high quality rice production at high yield levels.

Keywords:nitrogen fertilization,rice starch,starch structure,physicochemical properties,chain length distribution

1.lntroduction

Rice is the main source of calories for more than 50%of the world’s population (http://faostat.fao.org/). The growing population has been applying great pressures for increasing the grain supply. Rice production is projected to increase by 30% to meet the demand of the extra population by 2050 compared to that in 2005-2007 (Alexandratos and Bruinsma 2012). National and international breeding projects have been launched with the goal of developing ‘super’ rice or ‘super’ hybrid rice with superior high yields to breaking the current yield ceiling (Zhang 2007). However,these ‘super’ rice cultivars are dependent on extra heavy nitrogen (N) inputs (Guet al.2017). The large amount of chemically synthesized N fertilizer inputs have resulted in serious problems in rice production,i.e.,deterioration of rice quality,instability of grain yield,and environmental pollution (Chenet al.2014;Guet al.2015,2017). Presently,the improvement of grain quality and achieving sustainable agriculture are becoming priorities because of economic development and high living standards. It is a great challenge to increase both rice production and grain quality in a sustainable way (Zhang 2007).

Scientists have made great advances in synergistically improving rice production and N use efficiency (NUE)by improving N fertilization (Penget al.2010;Yang 2015). For example,site-specific N management(SSNM) has been widely adopted in rice grown areas,which could effectively increase yield and NUE. In the study of Penget al.(2010),SSNM could reduce the N rate by 32%,but increased rice production by 5%,when comparing SSNM with traditional N management. Jianget al.(2004) and Zhanget al.(2018) also indicated that by optimizing the timing and doses of N fertilization,rice yield and N use efficiency could be improved simultaneously. Nevertheless,little is known regarding how to improve grain quality by optimizing N fertilization.The development of N fertilization practices that drive rice production with high yield,high N use efficiency and high quality is essential for the development of economic progress and the improvement of living standards.

Starch makes up 85-90% of rice,which is mainly composed of amylose and amylopectin (Konget al.2015). Starch is the key factor determining various aspects of grain quality,of which cooking and eating qualities are the most strongly influenced. Amylose mainly consists of a few long branches of α-1,4-linked glucose molecules and has a smaller molecular weight(Gilbertet al.2013). The content of amylose significantly influencsd rice texture. Based on amylose content,rice is divided into categories of waxy (0?2%),very low(3?9%),low (10?19%),intermediate (20?25%) and high(＞25%). In contrast,amylopectin makes up 65?85% of the matter in the starch granules,and it is highly branched with short α-1,4 chains linked by α-1,6 bonds. Key starch properties include granule size,amylose and amylopectin contents,branch chain length distribution of amylopectin,etc.,and these structural changes result in changes in the relative crystallinity and stability of the starch,which affect the physical and chemical properties of the starch,and ultimately alter the cooking and gelatinization properties of the starch (Ovando-Martínezet al.2011).The biosynthesis of starch in rice grains requires the synergy of several enzymes,including ADP glucose pyrophosphorylase (AGPase),granule-bound starch synthase (GBSS),soluble starch synthase (SS),starch branching enzyme (BE),and starch debranching enzyme(DBE). For example,GBSS is involved in the synthesis of amylose and extra long unit chains of amylopectin.Recent studies have indicated that enzymes SSI and BEII are involved in the biosynthesis of short branch chain amylopectin,whereas SSII,SSSIII,and BEI play a critical role during the elongation of long chain amylopectin(degree of polymerization (DP)＞25) (Nakamura 2002;Fujitaet al.2007).

Rice starch structural and physicochemical characteristics,and concomitant properties,are affected by the growth environment and management conditions(Zhanget al.2017;Tanget al.2019). Excessive N fertilizer application increased grain yield and protein content,but deteriorated cooking and eating qualities,by increasing gelatinization temperature and gelatinization enthalpy,while decreasing peak viscosity,hot viscosity,and breakdown value (Guet al.2015;Zhouet al.2020).However,applying N fertilizer at the flowering stage could improve starch structural and physicochemical properties when facing abotic stresses,such as high temperature(Tanget al.2019). Crop management of N fertilizer has been widely adopted in rice production systems to ensure higher grain production and N use efficiency,but their effects on starch synthesis and structure,and the relationship between starch fine structure and rice quality,are not well understood. However,they are crucial factors for developing optimized N fertilization to improve grain quality at a high yield level with lower environmental costs.

This experiment tested two high yielding ‘super’rice varieties,which are widely grown in Yangtze River Basin area,and multiple agronomic practices of N fertilization,i.e.,reducing N rate,optimizing the ratio of urea application at different stages,and increasing plant density. The structural and physicochemical responses of starch to N fertilization were studied. The purposes of this study are to investigate whether and how much optimized N fertilization could improve grain quality and N use efficiency at high yield levels. We hope this study will be helpful for building a sustainable N fertilization system of rice production for high quality at a high yield level to meet the demand of the higher food supply with good quality for humans in the coming decades.

2.Materials and methods

2.1.Rice growing and cultivation

The experiments were carried out at an experiment station of Yangzhou University,Jiangsu Province,China(32°30′N,119°25′E) during the rice growing seasons(May-October) of 2018 and 2019. The soil in the experiment was a sandy loam (typic Fluvaquent,Etisol)with 19.3 g kg-1organic matter,105.1 mg kg-1alkalihydrolyzable N,23.5 mg kg-1Olsen-P,and 95.7 mg kg-1exchangeable K. The pH value was 6.3. A hybrid‘super’ rice variety,Yongyou 2640 (Y2640),and an inbred‘super’ rice variety,Wuyunjing 24 (W24),were planted in this study. Various agronomic practices of N fertilization included reducing the N input,increasing planting density,and increasing topdressing of urea at spikelet differentiation. Across the two years,seedlings were sown on the 15 May and transplanted on the 15 June.Rice seedlings were transplanted at two seedlings per hill.Weeds,insects and diseases were controlled following local farmers’ high yielding practices.

2.2.Crop management practice

There were four treatments with two cultivars. The experimental setup is a complete randomized block design. There were three plots for each treatment×cultivar combination. Plot size was 5 m×6 m. For basal fertilizer,phosphorus (30 kg ha-1as single superphosphate)and potassium (40 kg ha-1as KCl) were applied before transplanting. Details of each N fertilization treatment are described in Table 1. Briefly,in the blank control,there was zero N fertilizer applied (0N). In the localfarmers’ practice (LFP) treatment,the general agronomic management practices in main rice production of Yangtze River Basin area were adopted. In LFP,300 kg N ha-1was applied,and the field was continuously flooded.There were 30 cm between rows and 13.3 cm between each hill within a row. The depth of tillage was 10 cm.Two treatments of optimized N fertilization (ONF) were introduced. On the basis of LFP,in ONF1,the total N application was reduced to 270 kg N ha-1(10%less),in which the basal N fertilizer was reduced,and the topdressing of urea at spikelet differentiation was increased to make bettter use of indigenous soil N and ensure grain development in the later growth stage (Penget al.2010). In ONF2,planting density was increased by 25% compared with that in ONF1 to improve N recovery efficiency.

Table 1 Details of crop management practices in different treatments for ‘super’ rice cultivars of Yongyou 2640 (Y2640)and Wuyunjing 24 (W24)1)

2.3.Starch isolation

According to methods of Tranet al.(2011),the starch samples were extracted using the neutral protease to remove protein,to ensure that the starch structure was less degraded. All starch samples were freeze-dried and stored in a desiccator until use.

2.4.Measurement of amylose content (ACC) and protein content

The method of Zhouet al.(2020) was followed to measure the iodine absorption spectrum of starch using the Lambda 650 UV/VIS spectrophotometer (PerkinElmer,überlingen,Germany),scanned from 400 to 900 nm,to obtain the spectrum of iodine absorption. The iodine blue value of rice flour samples was measured at 680 nm and the ACC was calculated from the iodine blue value of starch standard samples prepared from corn amylopectin and potato amylose at 620 nm. For the content of protein,the N content of rice grain was determined by micro-Kjeldahl digestion,and then multiplied by 5.95.

2.5.Morphology,structure,and pasting properties of starch

For the granule morphology and granule size analysis of starch,all starch samples that had been freeze-dried and passed through a 200-mesh sieve were evenly scattered. The starch samples were coated with gold using the nanoEM (Quantum Design,USA) and then observed in environmental scanning electron microscopy(XL-30 ESEM,Philips,Amsterdam,Netherlands) to obtain images. A Malvern Master-Sizer 2000 (Malvern Instruments,Ltd.Worcestershire,UK) laser diffraction analyzer with a Fourier cell (0.02-2 000 μm) and the Hydro SM Small Volume Sample Dispersion Unit (SVSDU)were used to measure the granule size of starch.

To measure the fine structure of starch,methods for analyzing the structure of whole starch molecules and chain length distribution (CLD) of debranched amylopectin were referred to Wuet al.(2014) and Zhouet al.(2020). An Agilent 1100 Series Size Exclusion Chromatography (SEC) System (Agilent Technologies,Waldbronn,Germany) was used to measured CLD,and a PA-800 Plus System equipped a solid-state laser-induced fluorescence (Beckman Coulter,Brea,CA,USA) detector was used to measured FACE.

To measure the stability of starch crystals,an RU200R X-ray diffractometer (Rigaku,Tokyo,Japan) at 40 mA and 40 kV with Cu-Kαfiltered radiation (λ=0.154 nm) was used to measure the crystallinity of starch according to the method of Warrenet al.(2016),and a Bruker Nexus 670 spectrometer with a DTGS detector equipped with an ATR single reflectance cell (Nexus 670,Nicolet,Madison,WI,USA) was used to measure the Fourier transform infrared spectrum (Sevenouet al.2002).

Using the method of Zhouet al.(2020),the starch pasting properties were measured with a rapid viscosity analyzer (Model 3D,Newport Scientific,Australia) and the gelatinization properties of rice starch were determined by differential scanning calorimetry (DSC 8500,Perkin Elmer,USA).

2.6.Typical texture profile analysis (TPA)

The method of Lyonet al.(2000) was used to prepare rice samples and perform the instrumental texture profile analysis. For TPA,we used bench-top TA-XT2 Texture Analyzers (Texture Technologies Corp.,Scarsdale,NY,USA).

2.7.Statistical analysis

The data shown in all the Tables and Figures represent the means of the triplicate replicates. Analysis of variance(ANOVA) was carried out to assess the statistically significant differences,and Tukey’s test was used to test the significance of differences between treatments atP＜0.05. Pearson correlation analysis and principal component analysis (PCA) were also carried out for determining the relationships between different variables.

3.Results

3.1.Grain yield and NUE

Yield,yield components,and NUE (i.e.,the partial factor productivity of applied N) of 2018 and 2019 are presented in Table 2. The same tendency was observed for both cultivars in both years. The lowest grain yield was observed in the 0N treatment,and the highest grain yield was in ONF2. Optimized nitrogen fertilization increased the yield by 5.55% for ONF2 when compared with LFP. There was no significant difference in yields between treatments of LFP and ONF1. For yield components,the panicle number per area contributed the most to the increase in grain yield. When compared with LFP,the optimized N fertilization of reducing N rates,optimizing of proportions of urea application at different growth stages,and increasing plant densities,had greatly improved NUE from ONF1 to ONF2 (8.0-17.8%). Besides,the optimized nitrogen fertilization brought farmer extra economic returns (Appendix A).

3.2.The contents of amylose,amylopectin and protein of rice

For all the treatments,the protein content was the lowest,and the amylose and amylopectin contents were the highest in the 0N treatment. The application of N fertilizer significantly increased the protein content,but decreased the amylose and amylopectin contents in both years (Table 3). However,compared with LFP,ONF1-2 with the optimized N fertilization gradually increased the starch content but decreased the protein content,of which ONF2 had the lowest protein content and the highest starch content. Although the amount of topdressing of N fertilizer was increased in ONF1 relative to LFP,the total amount of N input was decreased by 10% (from 300 to270 kg N ha-1),which could be the reason for the slight decrease of protein content (on average 3.4% less) in ONF1 compared with LFP. These results indicated that optimized N fertilization could increase the synthesis of starch and reduce the protein content.

Table 2 Yield,yield components and nitrogen use efficiency for ‘super’ rice cultivars of Yongyou 2640 (Y2640) and Wuyunjing 24 (W24)

Table 3 Effects of different nitrogen fertilization treatments on rice amylose,amylopectin,total starch and protein content for ‘super’rice cultivars of Yongyou 2640 (Y2640) and Wuyunjing 24 (W24)

3.3.Starch granule morphology and size distribution

Starch granules exhibited irregular polygons in all the treatments of both cultivars (Fig.1). N fertilization significantly influenced the morphology of starch granules in rice varities Y2640 and W24. The surface of the starch granule was smooth in the zero N input,but the surface of starch granules was uneven and pitted,and with more small granules adhering to large granules in LFP. When comparing ONF1 and ONF2 with LFP,we find the number of large starch particles is gradually increasing,and the dents on the surface of starch granules were decreased.The surface of the starch granules of ONF2 was the smoothest and most flat. These results showed that optimized N fertilization could improve the morphology of starch granules.

Fig.1 Effects of different nitrogen fertilization treatments on the morphology of rice starch granules for ‘super’ rice cultivars of Yongyou 2640 (Y2640) and Wuyunjing 24 (W24). A-D,Y2640 under 0N (zero N fertilizer applied),LFP (local farmers’practice),ONF1 (optimized N fertilization 1),ONF2 (optimized N fertilization 2),respectively; a-d,W24 under 0N,LFP,ONF1,and ONF2,respectively. EHT,extra-high tens;WD,working distance;Msg,magfifications.

The effects of N fertilization on the particle sizes of rice starch are shown in Fig.2. Granule size distribution was divided into three groups (based on diameter) of‘a(chǎn)’ (＞15 μm),‘b’ (5?15 μm) and ‘c’ (＜5 μm). A unimodal distribution with a peak at～0.6 μm was observed for the number of starch granules. Although the distribution patterns of starch granules were not influenced by treatments,there were still differences between the treatments in number,volume and surface area. The number,volume and surface area of small particles of starch granules in groups ‘b’ and ‘c’ were significantly increased under high N inputs,when comparing LFP with 0N. Comparing ONF1 and ONF2 with LFP,the result indicates that although the N fertilizer decreased the size of starch granules,the optimization of N fertilization could increase the starch granule sizes.

Fig.2 Effects of different nitrogen fertilization treatments on the granule size distribution of rice starch for ‘super’ rice cultivars of Yongyou 2640 (Y2640,A-C) and Wuyunjing 24 (W24,a-c). 0N,zero N fertilizer applied;LFP,local farmers’ practice;ONF1,optimized N fertilization 1;ONF2,optimized N fertilization 2.

3.4.Chain length distribution of whole starch and amylopectin

Typical SEC weight molecular size distributions,ωbr(logRh) (the hydrodynamic volume (the separation parameter for size-exclusion chromatography)),the hydrodynamic volume (the separation parameter for size-exclusion chromatography),of whole (fully branched) starch and amylopectin branch chains are shown in Fig.3. Typical bimodal distributions were obtained for the full starch chain length pattern,in which one peak represented amylose (Rh(the equivalent hydrodynamic radius,nm) up to 100 nm) and the other peak represented amylopectin(Rhbetween 100 and 4 000 nm). From Table 4,the averageRhfor amylose,Rhat amylose peak maximum,averageRhfor amylopectin,Rhat amylopectin peak maximum,and averageRhfor total starch significantly decreased at first,when comparing LFP with 0N;but then these values gradually increased after optimized N fertilization of ONF1 and ONF2. The peak of the hydrodynamic radius curve first decreased,but then increased gradually from ONF1 to ONF2. Both cultivars showed these same consistent trends. This result is consistent with the results of starch granule morphology and granule size distribution. Application of N fertilizer decreased the sizes of starch granules,while optimized N fertilization practices could compensate for these effects,and increase starch granule size.

Fig.3 Effects of different nitrogen fertilization treatments on the chain length of whole starch (A and B) and on the chain length of amylopectin (C and D) for cultivars Yongyou 2640 (Y2640,A and C) and Wuyunjing 24 (W24,B and D). 0N,zero N fertilizer applied;LFP,local farmers’ practice;ONF1,optimized N fertilization 1;ONF2,optimized N fertilization 2. ωbr(logRh),the hydrodynamic volume (the separation parameter for size-exclusion chromatography). DP,degree of polymerization.

A bimodal distribution of branch chains of amylopectin is shown in Fig.3. The distribution of amylopectin branch chains peaked at a DP of 12 regardless of cultivars and treatments. Amylopectin branch chains were classified into three chain types:A chain (DP 6?12),B1 chain(DP 13?25),B2 chain (DP 26?37),and B3 chain (DP＞37).With inputs of N fertilizer,the content of short branch chains of amylopectin (i.e.,A chain,B1 chain) decreased,when comparing LFP with 0N;but the content was gradually increased after optimized N fertilization practices of ONF1 and ONF2. The opposite results were observed for long branches with DP larger than 37 (B3 chain) (Table 4).

3.5.X-ray diffraction (XRD) and relative crystallinity

There were three types (A-,B-and C-type) based on XRD spectra. The X-ray diffraction patterns of rice starch showed A-type in both years of 2018 and 2019 (Fig.4;Appendix B),with strong diffraction peaks at 15° and 23° and an unresolved doublet at 17° and 18°. These results indicated that N fertilization did affect the stability of rice starch crystals,but not crystal type. In Fig.4 and Appendix B,the order of the XRD curves is LFP,ONF1,ONF2,and then 0N. The relative crystallinity represents the stability of starch crystals,where the larger the value,the higher the stability. As shown in Table 5,the application of N fertilizer significantly increased the relative crystallinity. Comparing ONF1 and ONF2 with LFP,crystallinity showed a downward trend and reached the lowest value at ONF2,which indicated that optimized cultivation practices could reduce the relative crystallinity of starch. The reason for this result may be due to the increment of the short branch chain content of amylopectin,which reduces the stability of starch crystals.

3.6.Structural order of the starch external region

From the attenuated total reflectance (ATR)-Fourier transform infrared (FTIR),we can investigate starch structure. Peaks at wave numbers 1 045 and 1 022 cm-1are responsible for crystals and amorphous regions in the starch granules,respectively (Fig.4;Table 5;Appendix B). The value of 1 045/1 022 cm-1indicates an ordered structural content of the outer region of the starch granules,while the value of 1 022/995 cm-1indicates the content of the amorphous structure. The order of the curves is LFP,ONF1,ONF2 and 0N (Fig.4). The intensity ratio of 1 045/1 022 cm-1increased with the N rate. Comparing ONF1 and ONF2 with LFP,the value of 1 045/1 022 cm-1was gradually decreased and reached its lowest value at ONF2. The intensity ratio of 1 022/995 cm-1showed the opposite trend (Table 5). These results showed that the inputs of N fertilizer improved the stability of starch,and the optimization of N fertilization could reduce the stability of starch.

Fig.4 Effects of different nitrogen fertilization treatments on the X-ray diffraction patterns of rice starch (A and B) and attenuated total reflectance-Fourier transfors infrared spectra of rice starch (C and D) for cultivars Yongyou 2640 (Y2640,A and C) and Wuyunjing 24 (W24,B and D) in 2018. 0N,zero N fertilizer applied;LFP,local farmers’ practice;ONF1,optimized N fertilization 1;ONF2,optimized N fertilization 2.

3.7.Pasting properties of starch

There were significant differences in the pasting properties of starch of both cultivars under different treatments (Table 6). With the application of N fertilizer,the values of breakdown,hot viscosity,peak viscosity,and final viscosity were decreased,but the setback and pasting temperature increased for both cultivars. When comparing ONF1 and ONF2 with LFP,the opposite results were observed. RVA characteristics are important parameters which indicate the gelatinization,disintegration,swelling and gelling of starch granules. Peak viscosity indicates the ability of extension of swelling;a high breakdown viscosity value reflects low resistance to heating;and setback viscosity means the retrogradation of starch paste during cooling. In general,rice cultivars with better taste quality have larger viscosity and disintegration values,and lower digestion values.From Table 6,we can conclude that the higher amounts of N fertilizer tend to deteriorate the grain quality. However,through optimized N fertilization,the quality of rice can be restored. The higher breakdown viscosity and lower setback viscosity values at ONF2 showed that rice neither became stiff nor retrograded during cooking. With the increase of the amylopectin short branch chain content,the stability and crystallinity of the starch crystals decreased,which led to the decrease of the pasting temperature of the starch,so the starch exhibited excellent pasting properties and the rice quality was improved.

3.8.Thermal properties of starch

The gelatinization temperature (onset temperature (To),peak of gelatinization temperature (Tp) and conclusion temperature (Tc)) and gelatinization enthalpy (ΔHgel),showed significant differences among the treatments for both cultivars (Table 6). Rice with a high gelatinization temperature has been shown to have a harder endosperm and more resistance to grain breakage.ΔHgelis correlated with crystallinity,an indicator of the susceptibility to a loose crystal structure and double helix formation during gelatinization.Long chains of amylopectin play an important role in gelatinization,which would increase crystallinity,limit the water absorption and cooperation in the amorphous region,and result in higher gelatinization temperature and gelatinization enthalpy. When comparing treatments of LFP with 0N,theTo,Tp,Tc,andΔHgelwere increased. In this experiment,we also found that gelatinization temperature was positively correlated with the DP of amylopectin. The results in this study showed that N fertilizer would increase the gelatinization temperature,and the susceptibility of starch gels to retrograde and syneresis under high N fertilizer would make the resulting starch unsuitable for products requiring low-temperature storage. In contrast,the optimized N fertilization practices would improve rice starch quality in these aspects (Table 6).

Table 5 Effects of different nitrogen fertilization treatments on relative crystallinity,infrared absorption spectroscopy (IR) ratio of rice starch for ‘super’ rice cultivars of Yongyou 2640 (Y2640)and Wuyunjing 24 (W24)

3.9.Typical texture profile analysis (TPA) of cooked rice

TPA is a two-compression-cycle test,providing a sense of the chewing behavior of food products. The hardness and adhesiveness of the two rice cultivars under different treatments are presented in Fig.5. Hardness is the maximum force reached during the first compression,while adhesiveness indicates the forece required to separate the probe from the base plate after the sample has been compressed. The hardness and adhesiveness of rice decreased with the application of N fertilizer. Comparing ONF1 and ONF2 with LFP,the hardness and adhesiveness gradually increased,with the highest values observed in ONF2,which adopted optimized N fertilization practices.

Fig.5 Typical texture profile analysis (TPA) of cooked rice for ‘super’ rice cultivars of Yongyou 2640 (Y2640) and Wuyunjing 24(W24). 0N,zero N fertilizer applied;LFP,local farmers’ practice;ONF1,optimized N fertilization 1;ONF2,optimized N fertilization 2.Bars mean standard deviation from each treatment (n=3) and different letters between different treatments of the same cultivar means the values are significantly different (P＜0.05).

3.10.Relationships between starch properties

The correlations between different parameters of grain quality are presented in Fig.6. Generally,relative crystallinity was negatively correlated with breakdown,A-and B1-chain contents,peak viscosity,finial viscosity,and adhesiveness,and was positively correlated with B2-and B3-chain contents,gelatinization temperature,gelatinization enthalpy,and hardness. Adhesiveness was positively correlated with A-and B1-chain contents,peak viscosity,breakdown,and negatively correlated with gelatinization enthalpy,setback,protein content and hardness. These results indicated that besides the well-known relationships between protein content,amylose and amylopectin contents with the gelatinization properties of rice,the branch chain length of amylopectin also played an important role in this aspect. The higher content of short-chain amylopectin would result in the lower ordering and decrease the relative crystallinity,and improve the eating and cooking quality of the starch.

Fig.6 Pearson correlations for rice grain quality parameters for both ‘super’ rice cultivars under different nitrogen fertilization treatments. ＊,＊＊ and ＊＊＊ indicate the significant differences at P＜0.05,P＜0.01 and P＜0.001 levels,respectively. Am,amylose;Ap,amylopectin;Rh,the equivalent hydrodynamic radius;To,onset temperature;Tp,peak of gelatinization temperature;Tc,conclusion temperature;ΔHgel,gelatinization enthalpy;ΔHret,retrogradation enthalpy;R (%),retrogradation percentage (100%×ΔHret/ΔHgel).

The Pearson correlation analysis showed pair-wise correlations between grain quality traits,however,the overall relationships among different parameters were obscure. In Fig.7,PCA revealed the relationships between different properties of starch from various N fertilization treatments. The clustering of rice starch property parameters indicated positive correlations between them. When rice starch property parameters are located in opposite directions in the Fig.7,negative relationships among these parameters are indicated. For example,the loading plot indicated that gelatinization enthalpy and protein content were closely associated;adhesiveness and retrogradation enthalpy are tightly correlated;To,Tp,Tc,and B3-chain content,and relative crystallinity are tightly linked;and hardness and A chain content are positive associated. The negative correlations also indicated that amylose and amylopectin contents were negatively associated with relative crystallinity and gelatinization enthalpy,respectively.

Fig.7 Principal component analysis loading plot:relationships between properties of rice starch for both cultivars under different nitrogen fertilization treatments. Am,amylose;Ap,amylopectin;Rh,the equivalent hydrodynamic radius;To,onset temperature;Tp,peak of gelatinization temperature;Tc,conclusion temperature;R (%),retrogradation percentage (100%×ΔHret/ΔHgel).

4.Discussion

4.1.The effect of N fertilizer on grain yield,NUE and rice quality

N is essential for plant growth and development. Rice requires N during the vegetative growth stage to promote growth and tilllering,which in turn determines the number of panicles. In the panicle initiation stage,N contributes to the number of spikelets per panicle,and it decreases the number of degenerated spikelets and increases the hull size in the late panicle formation stage. The newly bred‘super’ rice varieties with great yield potentials are highly dependent on the N rate (Yanget al.2015;Guet al.2017;Zhuet al.2021). Farmers tend to apply excessive amounts of N fertilizer to paddy rice fields to maintain a high grain yield level. For example,in some counties of Jiangsu Province,China,the average N fertilizer rate was as high as 300 kg N ha-1(Denget al.2011). In this study,a very high grain yield level of ＞11.9 t ha-1for cultivar Y2640 was achieved at LFP,with 300 kg N ha-1input(Table 2),but at the expense of environmental risk,as most of the N applied would be lost to the environment by ammonia volatilization,denitrification,runoff,and leaching(Zhu and Chen 2002),which causes environment pollution,such as soil acidification,groundwater pollution,water eutrophication,and air pollution (Zhouet al.2016).However,agriculture today faces more challenges than ever before due to the dual goals of increasing food production and protecting the environment for sustainable agriculture (Chenet al.2014). For achieving these goals,optimized N fertilization strategies are introduced in the present study. In ONF1,N application at pre-transplanting as basal fertilizer was shifted to topdressing of N at the stage of initial spikelet differentiation for several reasons.One reason is that reducing the N rate at pre-transplating would reduce NH3volatilization and make full use of the indigenous N supply from soil (Penget al.2010). The second reason is that it is now widely recognized that more N is taken up in the middle to late growing stage of plants (Menget al.2013;Chenet al.2014). Thirdly,topdressing N fertilizer at the late growing stage would help in regulating the structure of the rice canopy and in improving radiation use efficiency (Guet al.2017).Finally,topdressing of N fertilizer at the stage of initial spikelet differentiation would help to promote spikelet differentiation and inhibit spikelet degeneration (Wanget al.2018). In this case,the N rate was further reduced by 10% to improve NUE without yield loss. In ONF2,planting density was increased by 25% to increase the size of the rice population to make full use of solar radiation and soil nutrients. On average,grain yield and partial factor productivity of applied N under ONF were increased by 1.7 and 13.06%,respectively.

N fertilizer also significantly affects rice quality. It was reported that the use of N fertilizer improves the milling and appearance qualities of rice (Douet al.2017;Zhouet al.2018). The cooking and eating quality are significantly influenced by N fertilizer.Larege amounts of nitrogen fertilizer would result in increments in protein content,which decreased adhesiveness.This could be due to the protein affecting the water absorption of rice,hence influencing the pasting properties (Martin and Fitzgerald 2002;Dinget al.2021). In starch gelatinization,the interactions between water and starch granules are,to some extent,blocked by the disulfide bonds of the protein,which influences adhesion (Martin and Fitzgerald 2002). That is why a higher protein content is always associated with hardness,higher gelatinization enthalpy,lower adhensiveness and deteriorated eating and cooking quality. Amylose content is another important determinant for cooking and eating quality. A positive relationship between amylose content and hardness,and an inverse relationship with the stickiness of cooked rice is often reported (Julianoet al.1984;Bianet al.2020). A lower amylose content would contribute to the improvement of cooking and eating quality to some extent. However,the decrease of amylose is marginal in comparing LFP with 0N,and the cooking and eating quality is deteriorated at the N rate of 300 kg N ha-1in LFP.

4.2.Optimized N fertilization improves the structure and physicochemical properties of starch,and ultimately improves rice quality at high yield levels

High grain yield necessitates the input of N fertilizer,however,it can also deteriorate the grain quality due to the increses in protein content. In this study,by reducing the N input,increasing planting density,and optimizing the ratio of urea application at different stages,the protein content in the grain is decreased,which is consistent with the work of Zhanget al.(2019).Futhermore,the changes of amylose and amylopectin contents,starch chain length distribution,amylopectin fine structure,and amylopectin chain length distribution all have an impact on the textural properties of starch. ONF significantly increases the short branch chain (A and B1 chains) content of amylopectin and decreased the long branch chain (B2 and B3 chains)content (Fig.3;Table 4). The change of chain length distribution of amylopectin influenced the stability and gelatinization properties of the starch. The changes of the crystallinity and FTIR spectrum value were mainly related to the short-range order content in the outer region of the starch granules (Warrenet al.2016;Huanget al.2020). The higher the content of short branch chains(A and B1 chains) of amylopectin,the lower the content of long branch chains,and the lower the stability and crystallinity of the starch (Tanget al.2019;Zhouet al.2020). Our research indicated that the increase of Aand B1-chain amylopectin could significantly decreased the crystallinity and the IR ratio of 1 045/1 022 cm-1,and so the stability of the starch crystals decreased (Fig.4;Table 5). Simultaneously,starch gelatinization properties were greatly influenced by chain-length distributions.Stickiness increased with the short amylopectin branch chain ratio (Liet al.2017). In starch gelatinization,amylopectin molecules with short branch chains formed a relative super-globe and cluster. The molecular entanglements within the super-globe are less tight than those between long polymer chains of amylose,which contributed to the improvement in the thermal properties(Xieet al.2009;Huanget al.2015). In this research,compared with LFP,ONF1 and ONF2 could significantly increase the viscosity value and adhesiveness,and reduce the hardness,pasting temperature,gelatinization temperature,gelatinization enthalpy and retrogradation enthalpy (Fig.5;Table 6). The changes in branch chain length distribution could be due to changes in the expression levels of starch synthesis-related genes includingOsSSI,OsBEIIb,OsSSIIa,andOsSSIIIa.SSIpreferentially synthesizes the shortest amylopectin chains of DP 6-7;BEIIbis pivotal in the formation of amylopectin A chains;SSIIahas a specific role in elongating short A and B1 chains to produce long cluster chains;andSSIIIacontributes to the synthesis of amylopectin chains of DP＞30 from intermediate chains (Fujitaet al.2006,2007;Ryooet al.2007). The changes in expression levels of starch synthesis-related genes involved in the synthesis of short branch chains (A and B1 chains) and long branch chains (B2 and B3 chains) of amylopectin between the ONF and LFP treatments need further study.

5.Conclusion

Rice is the main staple food for most human beings in the world. High rice grain yields have been achieved at the expense of large inputs of chemically synthesized N fertilizer. High inputs of nitrogen fertilizer are causing a serious environmental crisis and lower resource use efficiency. Furthermore,excessive application of N fertilizer would deteriorate rice grain quality. Is it possible to synergistically improve yield,quality and resource use efficiency by optimized nitrogen fertilization? The nitrogen application commonly used today in local farmers’ practice affected the physicochemical and textural properties of starch. The changes could due to the high rate of nitogen fertilizer increasing the protein content of rice,long branch chain content of amylopectin and relative crystallinity,and decreasing the starch granule size and the content of short branch chain of amylopectin. Optimized nitrogen fertilization(ONF1 and ONF2) could substantially compensate for the effects caused by N fertilizer and achieve all three goals of higher yield,higher grain quality,and higher N fertilizer use efficiency. Meanwhile,the economic benefits for the farmers would be significantly improved.

Acknowledgements

This study was financially supported by the National Natural Science Foundation of China (32071943 and 31872853) and the Priority Academic Program Development of Jiangsu Higher Education Institutions,China (PAPD).

Declaration of competing interest

The authors declare that they have no confilict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm