Effects of the potassium application rate on lipid synthesis and eating quality of two rice cultivars

CHEN Guang-yi,PENG Li-gong,Ll Cong-mei,TU Yun-biao,LAN Yan,WU Chao-yue,DUAN Qiang,ZHANG Qiu-qiu,YANG Hong,Ll Tian

Crop Ecophysiology and Cultivation Key Laboratory of Sichuan Province,College of Agronomy,Sichuan Agricultural University,Chengdu 611130,P.R.China

Abstract Lipid content has an important effect on rice eating quality,but the effects of fertilizer application rate on the lipid synthesis and eating quality of rice are not well understood. Potassium (K) has a strong influence on rice quality and the requirement for K fertilizer in rice is greater than for nitrogen (N) and phosphorus (P) fertilizers. To investigate the effects of K fertilizer on the lipid synthesis and eating quality of rice,we used Nanjing 9108 (NJ9108,japonica) and IR72 (indica)rice as experimental materials and four K levels: K0 (0 kg ha-1),K1 (90 kg ha-1),K2 (135 kg ha-1) and K3 (180 kg ha-1).The results showed that the lipid content,free fatty acid (FFA) content,unsaturated fatty acid (UFA) content,malonyl-CoA(MCA) content,phosphatidic acid (PA) content,lipid synthesis-related enzyme activities and eating quality first increased and then decreased with increasing K in both cultivars. The maximum values were obtained under K2. However,the saturated fatty acid (SFA) content showed the opposite trend. No significant differences were found in pyruvate(PYR) content among the K treatments. The protein and oxaloacetic acid (OAA) contents and phosphoenolpyruvate carboxylase (PEPCase) activity of NJ9108 first decreased and then increased with increasing K,and the minimum values were obtained under K2;while IR72 showed the opposite trend and the maximum values were obtained under K1.Overall,increasing K optimized the fatty acid components and increased the lipid content and eating quality of rice by enhancing lipid synthesis-related enzyme activities and regulating substrate competition for lipid and protein synthesis.The optimal K application rate for lipid synthesis,eating quality and grain yield was 135 kg ha-1 for both cultivars.

Keywords: rice,potassium application rate,lipid content,lipid synthesis-related enzyme,fatty acid components,eating quality

1.lntroduction

Rice (OryzasativaL.) is the main food crop for nearly 50% of the world’s population. China is the largest rice producer and consumer in the world,and accounts for 30% of global production. Over 65% of people in China depend on rice as the staple food (Ronget al.2021;Li 2021). With increasing economic development and improvements in people’s standards of living,the demand for rice of higher quality is also increasing.

Lipids are one of the three main nutritional components of rice,with brown rice and milled rice having 3 and 0.8% lipid contents,respectively (Kimet al.2015). Although the lipid content in rice is low,most of the fatty acid constituents are high-quality unsaturated fatty acids (UFAs) that have high nutritional value and health-promoting functions in the human body (Liuet al.2013;Sharifet al.2014). Studies have shown that most high-quality rice cultivars share the same characteristics in terms of a relatively high lipid content (Yuet al.2007;Jianget al.2016). The lipid content in rice has a positive correlation with rice cooking and eating qualities,and fatty acids also have a certain effect on rice taste (Yoonet al.2008;Xu 2017). We can significantly improve the eating quality of rice by increasing the content of UFAs(Yoonet al.2009,2012).

The lipid content and fatty acid components in rice are affected not only by genetics but also by environmental factors,cultivation methods and storage conditions(Baud and Lepiniec 2010). The application of fertilizers is easier to control than other methods in field production,so modifying fertilizer applications could be a simple way to regulate lipid synthesis in rice. Previous studies have shown that fertilizer has a strong effect on the crop plant lipid content. N,P,K and calcium (Ca) fertilizers could increase the lipid,protein and free fatty acid (FFA) contents of peanut kernels. However,the application of excessive K fertilizer (450 kg ha-1) could reduce the lipid content of peanut kernels and the application of Ca fertilizer could increase the ratio of oleic acid and linoleic acid (Zhouet al.2006). The fertilizing measures to achieve oil contents over 21.5% for soybean were N,0.07-0.13 g kg-1;P2O5,0.11-0.37 g kg-1;and K2O,0.04-0.10 g kg-1(Ninget al.2007). The component products of amino acids and fatty acids in maize showed no significant differences with the maximum in response to the fertilizer application conditions of compound fertilizer at 562.5 kg ha-1and urea at 750 kg ha-1. The level of total amino acids was 1 082.38 kg ha-1,while SFAs was 164.54 kg ha-1,and UFAs was 933.64 kg ha-1(Huanget al.2013). Adding N fertilizer as top dressing in the late growth stage of rice reduced the lipid content in rice grain,whereas P and K fertilizer could increase the lipid content in rice grain (Zhuet al.2006). Applying N fertilizer can increase the activities of key enzymes involved in lipid synthesis in rice,such as acetyl-CoA carboxylase(ACCase),fatty acid synthase (FAS),glycerol 3-phosphate dehydrogenase (3-GPD),phosphatidyl phosphatase(PPase) and glycerol 3-phosphate acyhransferase (GPAT)(Tuet al.2020). The above findings indicate that the fertilizer application amount has a positive regulatory effect on the activities of key enzymes of lipid metabolism in crop plants,and thus it has a significant effect on lipid synthesis and accumulation. However,studies on the effects of fertilizer on lipid synthesis are limited compared with those on the starch and proteins in rice,especially in response to K levels (Guet al.2011).

K is an essential macronutrient that is involved in many physiological processes in plant cells,such as enzyme activation,membrane potential maintenance,and osmotic regulation (Vijayakumaret al.2021). It also influences crop yields and quality by regulating photosynthesis,assimilate transport,and carbohydrate metabolism (Pettigrew 2010). The reasonable application of K fertilizer can promote the growth and development of rice,improve the number of available tillers and the photosynthetic rate of leaves,and promote the accumulation,transformation and transportation of photosynthetic products,which have significant impacts on rice yield and quality (Maet al.2022). However,the effects of the K application rate on rice lipid synthesis and eating quality have rarely been reported. Thus,an experiment was conducted involving two rice cultivars and four K application levels. The objectives of this study were to determine how K application regulates lipid synthesis and the fatty acid components in rice and to provide theoretical and practical guidance for the improvement of rice eating quality.

2.Materials and methods

2.1.Experimental site

Field experiments were conducted during the growing seasons of 2019 and 2020 at the research farm of Sichuan Agricultural University,Chongzhou City,Sichuan Province,China (30°56′N,103°65′E). Prior to the establishment of the field experiment,soil samples from the topsoil layer (0.20 m) were analyzed. The clay soil had the following nutrient contents in samples from 2019 and 2020,respecitvely: 2.04 and 1.91 g kg-1total nitrogen(N) (Kjeldahl method,UDK-169,ITA);17.50 and 18.91 mg kg-1available phosphorus (P) (Mo-Sb colorimetry after digestion with H2SO4and HClO4);19.20 and 21.50 g kg-1organic matter (K2Cr2O7-volumetric method);60.24 and 58.31 mg kg-1available K (flame spectrometry after NH4OAc extraction);and pH 6.02 and 5.93 (tested in samples containing a 1:2.5 ratio of soil to water).

2.2.Experimental design

The experiment was conducted in a randomized block pattern with four treatments: K0,K1,K2 and K3 at 0,90,135 and 180 kg ha-1,respectively,and N fertilizer (135 kg ha-1) was used as basal manure and top dressing at a 3:7 ratio. A total of 24 treatments were performed with three repetitions. Two rice cultivars with significant differences in lipid content were used as the test materials. Nanjing 9108 (NJ9108) (Jiangsu Academy of Agricultural Sciences,lipid content in milled rice: 32.24 mg g-1) is ajaponicarice cultivar and IR72 (International Rice Research Institute,lipid content in milled rice: 29.31 mg g-1) is anindicarice cultivar,and both have high yield and good quality. Seeds were sown on 16 April 2019 and 12 April 2020,and the seedlings were transplanted on 26 May 2019 and 16 May 2020,respectively. The heading dates of NJ9108 were 8 August 2019 and 20 July 2020,and the maturity dates were 27 September 2019 and 14 September 2020,respectively. The heading dates of IR72 were 14 August 2019 and 21 July 2020,and the maturity dates were 24 September 2019 and 11 September 2020,respectively. The area of each test plot was 5.0 m×5.0 m,and the transplant density was 25 cm×20 cm with two seedlings per hill. Urea(N,46.4%) was used as the N source,superphosphate(P2O5,12.0%) was used as the P source,and potassium chloride (K2O,60.0%) was used as the K source. Basal N (94.5 kg ha-1),P (67.5 kg ha-1) and K (at the treatment rates specified above) were applied to the soil 1 d before transplanting. For the fertilizer treatments,ridges with plastic film were used for separation,and protection lines were established between the treatment blocks to ensure the isolation of the experimental plots. Field management,including the prevention and control of pests and weeds,was conducted according to the local cultural practices.

2.3.Sampling

After flowering,approximately 300 panicles were selected on the same day and tagged in each plot.After full heading,40 tagged panicles were sampled from each plot every 6 d at 10:00 a.m. The collected panicles were divided into three groups. Twenty tagged panicles were dried at 80°C,after which the brown rice was crushed and sieved through a 100-mesh screen for measurements of the lipid and protein contents.Another 10 tagged panicles were used to determine the phosphatidic acid (PA),malonyl-CoA (MCA),pyruvate(PYR) and oxaloacetic acid (OAA) contents,and the remaining 10 tagged panicles were placed in liquid N for 3 min and then stored at -80°C for enzymatic analysis.At harvest,10 plants from each plot were sampled randomly and allowed to dry naturally in the sun to assess the rapid visco-analyzer (RVA) value,eating quality and fatty acid components after the material was stored at room temperature for 3 mon.

2.4.Measurements and methods

Lipid contentLipids in the rice grains were obtained by ultrasound-assisted extraction (UAE) according to a previously reported method with a slight modification(Tuet al.2020). UAE was carried out with an ultrasonic cleaner (CPX3800H-C,Emerson Electric Co.,St.Louis,MO,USA). Samples (2.000 g each) were weighed and extracted with 50 mL of hexane in a centrifuge tube and then mixed thoroughly using a Vortex Genie (G560E,Scientific Industries Inc.,Bohemia,NY,USA). The tube was immersed in an ultrasonic cleaner bath,and the lipids were extracted with the appropriate sonication power(110 W),duration (37 min) and temperature (42°C). The supernatant was transferred to a centrifuge tube and centrifuged (5430R,Eppendorf AG,Hamburg,Germany)for 10 min at 7 000 r min-1. The supernatant was then transferred to a conical flask (weighed and recorded as m1beforehand) and evaporated using an electric hot plate(ML-3-4,Beijing Zhongxing Weiye Century Instrument Co.,Ltd.,Beijing,China),and the sample was ultimately dried and weighed (recorded as m2). The lipid content (m3)in the rice grain was then calculated using the following formula: m3=m2-m1.

FFA contentThe FFA content in the UAE samples was measured according to a previously reported method (Tuet al.2020). The extracted lipids and 100 mL of diethyl ether/95% ethanol (v/v,1:1) were mixed together,and three drops of phenolphthalein indicator were added to the mixture. Then,each mixture was titrated with 0.1 mol L-1KOH solution until the color changed to red,and the volume of KOH solution used was recorded. The FFA content in rice grain was calculated using the following formula: ω=V×C×282/m.

where ω is the FFA content (mg g-1),V is the volume of KOH (mL),C is the concentration of KOH (mol L-1),282 is the molar mass of oleic acid (g mol-1),and m is the lipid content in the rice grain (g).

Fatty acid componentsThe fatty acid components in the UAE samples were measured by using a gas chromatograph-mass spectrometer (GC-MS) according to a previously reported method (Xuet al.2016). The extracted lipids and 1 mL of petroleum ether/benzene(v/v,1:1) were mixed together and then flushed with 1 mL of petroleum ether/benzene. All the liquid phases were collected and transferred into a 10-mL volumetric flask. The samples were neutralized with 0.4 mol L-1KOH/carbinol,mixed thoroughly,incubated at room temperature for 10 min,and then brought to volume by the addition of water,followed by mixing. The organic layer was collected for analysisviaGC-MS (Agilent 7890-5975C,Agilent Technologies Co.,Ltd.,Santa Clara,CA,USA) and the NIST Mass Spectral Database (National Institute of Standards and Technology,Gaithersburg,MD,USA). The GC-MS conditions and components were as follows: carrier gas (He),approximately 1 mL min-1;oven temperature: initially maintained at 80°C for 3 min,increased at 10°C min-1to 260°C,and then held for 15 min;injected sample volume,1 μL;injection,split 100:1;and mass ranges,20-500 m/z. Fatty acids were identified by comparisons with fatty acid standards. The masses of fatty acid methyl esters were quantified as percentages of the total methyl ester peak area.

Protein contentThe protein content was measured based on the total N content of head rice with a conversion index of 5.95viathe Kjeldahl method.

Metabolites related to lipid synthesis and enzymatic activity assaysSubstances produced during lipid synthesis and the relevant enzymatic activities were measured according to a previously reported method (Tuet al.2020). Based on the double antibody sandwich method,the optical density of each sample was measured at 450 nm using an ELIASA Kit,and then the concentrations of metabolites and enzymatic activities in the sample were calculated according to the standard curve.

RVA valueA 3.00-g sample and 25.0 mL distilled water were added to a test tube. Pasting properties were measured using an RVA device (3-D,Newport Scientific,Sydney,Australia) and analyzed with Thermal Cycle for Windows (TWC) Software. Viscosity values were measured in a rapid viscosity analyzer unit (RVU).

Eating qualityThe sensory properties of the cooked rice were measured using a rice sensory analyzer (STA 1B,Satake Asia Co.,Ltd.,Tokyo,Japan). Milled rice (30.0 g)was washed in a stainless steel container and then transferred into a 50-mL aluminum box containing 40 mL water. The milled rice was cooked in a multifunctional timing food steamer (GF-339,Goodway Electrical Enterprise Ltd.,Hong Kong,China). After cooking,the sensory properties of the cooked rice were determined.Cooked rice texture properties were measured using a rice texture analyzer (RHS 1A,Satake Asia Co.,Ltd.,Tokyo,Japan).

Yield and yield componentsRice was harvested at the maturity stage and the yield in each experimental plot was recorded after measuring moisture content and removing impurities. Grain yield was adjusted to a moisture content of 14%. The number of effective tillers per hill was determined before harvest using 30 plants per plot.A total of 10 selected plants were separated into single tillers according to the marked date,and were used to measure 1 000-grain weight,seed-setting rate,and filled grain number per panicle.

2.5.Statistical analysis

Data were analyzed using analysis of variance (ANOVA),and means were compared based on the least significant difference (LSD) test at the 0.05 probability level using SPSS 23.0 (Statistical Product and Service Solutions Inc.,Chicago,IL,USA). Origin Pro 2020 (OriginLab,Northampton,MA,USA) was used to draw the figures.The differences in the main indicators are shown in Table 1. Variance analysis showed that the results of key enzymes activities and metabolites related to lipid synthesis showed the same trends in both 2019 and 2020. Therefore,we only show the results for 2019 in the Results section.

Table 1 Analysis of variance for lipid contents,key enzymes activities and taste values of Nanjing 9108 (NJ9108) and IR721)

3.Results

3.1.Lipid accumulation

During the grain-filling process,the lipid content in rice grains showed a tendency to decrease gradually(Fig.1). The lipid contents of NJ9108 and IR72 first rose and then declined as the K application level increased. The highest lipid contents were observed under K2. In 2019,the lipid contents of both NJ9108 and IR72 tended to be in the order of K2＞K3＞K1＞K0 at 36 d after flowering,and the lipid contents of both cultivars under the K2 treatment were significantlyhigher than those under the K0 and K1 treatments.Compared with that under K0,the lipid contents of NJ9108 under K3,K2 and K1 increased by 6.5,9.2 and 1.5%,respectively,and the comparable lipid contents of IR72 increased by 10.4,16.1,and 6.4%,respectively.In 2020,the lipid content of NJ9108 followed the trend of K2＞K3＞K0＞K1 at 36 d after flowering,while IR72 followed the trend of K2＞K1＞K3＞K0. The lipid contents of both cultivars under the K2 treatment were significantly higher than those under the other treatments. Compared with that under K0,the lipid contents of NJ9108 under K3,K2 and K1 increased by 1.4,4.2 and -0.7%,respectively,and the comparable lipid contents of IR72 increased by 1.0,6.8,and 1.0%,respectively. Taken together,these results indicated that the appropriate application of K could increase the lipid content in rice grains. Both NJ9108 and IR72 achieved their highest lipid contents in response to the K application level of 135 kg ha-1.

Fig. 1 Effects of K application rate on lipid accumulation of Nanjing 9108 (NJ9108) and IR72 in 2019 and 2020. K0,K1,K2 and K3 refer to the different fertilizer treatments (0,90,135,and 180 kg ha-1,respectively). DW,dry weight. Bars mean SD (n=3).Different lower case letters indicate that the physical and chemical properties of both cultivars are significantly different with the different treatments (P＜0.05,LSD method).

Fig. 2 Effects of the K application rate on free fatty acid (FFA) accumulation in Nanjing 9108 (NJ9108) and IR72 in 2019 and 2020. K0,K1,K2 and K3 refer to the different fertilizer treatments (0,90,135 and 180 kg ha-1,respectively). DW,dry weight.Bars mean SD (n=3).

Fig. 3 Effects of K application rate on the activities of key enzymes involved in lipid synthesis of Nanjing 9108 (NJ9108) and IR72.K0,K1,K2 and K3 refer to the different fertilizer treatments (0,90,135 and 180 kg ha-1,respectively). ACCase,FAS and 3-GPD represent acetyl-CoA carboxylase,fatty acid synthase and glycerol 3-phosphate dehydrogenase,respectively. FW,fresh weight.Bars mean SD (n=3).

Fig. 4 Effects of K application rate on the activities of key enzymes involved in lipid synthesis of Nanjing 9108 (NJ9108) and IR72.K0,K1,K2 and K3 refer to the different fertilizer treatments (0,90,135 and 180 kg ha-1,respectively). PPase,GPAT and PEPCase represent phosphatidyl phosphatase,glycerol 3-phosphate acyhransferase and phosphoenolpyruvate carboxylase,respectively.FW,fresh weight. Bars mean SD (n=3).

Fig. 5 Effects of K application rate on the accumulation of metabolites involved in lipid synthesis in Nanjing 9108 (NJ9108) and IR72. K0,K1,K2 and K3 refer to the different fertilizer treatments (0,90,135 and 180 kg ha-1,respectively). PYR,MCA and PA represent pyruvate,malonyl-CoA and phosphatidic acid,respectively. FW,fresh weight. Bars mean SD (n=3).

Fig. 6 Effects of the K application rate on protein accumulation in Nanjing 9108 (NJ9108) and IR72. K0,K1,K2 and K3 refer to the different fertilizer treatments (0,90,135 and 180 kg ha-1,respectively). DW,dry weight. Bars mean SD (n=3). Different lower case letters indicate that the physical and chemical properties of both cultivars are significantly different with the different treatments (P＜0.05,LSD method).

Fig. 7 Effects of K application rate on the oxaloacetoc acid (OAA) contents of Nanjing 9108 (NJ9108) and IR72. K0,K1,K2 and K3 refer to the different fertilizer treatments (0,90,135 and 180 kg ha-1,respectively). FW,fresh weight. Bars mean SD (n=3).

3.2.FFA accumulation

During the grain-filling process,the FFA contents of both NJ9108 and IR72 decreased gradually (Fig.2). As the K application level increased,the FFA contents of NJ9108 and IR72 first rose and then declined. The highest and lowest FFA contents were observed under K2 and K0,respectively. In 2019,the FFA contents of both NJ9108 and IR72 tended to be in the order of K2＞K1＞K3＞K0 at 36 d after flowering. The FFA contents of NJ9108 under K3,K2 and K1 increased by 9.0,28.3 and 11.4%,respectively,and the comparable FFA contents of IR72 increased by 8.0,12.5 and 8.8%,respectively,when compared with that under K0. In 2020,the FFA contents of both NJ9108 and IR72 tended to be in the order of K2＞K3＞K1＞K0 at 36 d after flowering. The FFA contents of NJ9108 under K3,K2 and K1 increased by 17.2,21.6 and 4.9%,respectively,and the comparable FFA contents of IR72 increased by 16.9,31.9 and 11.5%,respectively,when compared with that under K0. Taken together,these results indicated that the appropriate application of K could increase the FFA content in rice grains. The highest FFA contents were obtained in response to the K application level of 135 kg ha-1for both NJ9108 and IR72.

3.3.Fatty acid composition

The fatty acid compositions of each rice cultivar at each K application rate are shown in Table 2. In both cultivars,the most abundant fatty acids found were oleic acid(C18:1),linoleic acid (C18:2),palmitic acid (C16:1) and stearic acid (C18:0). These four fatty acids accounted for more than 90% of the total fatty acid content in rice grain. No significant impacts were observed on fatty acid composition between the four K levels. As more K was gradually applied,the UFA content first rose and then declined,while the SFA content showed the opposite trend. The highest UFA and lowest SFA contents of the two cultivars were observed under K2. The UFA contents of both NJ9108 and IR72 under K2 were significantly higher than under K0. The SFA content of NJ9108 under K0 was significantly higher than those under the other treatments,and the SFA content of IR72 under K0 was significantly higher than that under K2. Compared to K0,the UFA contents of NJ9108 and IR72 under K2 increased by 5.05 and 0.82%,respectively. Taken together,these results suggested that the appropriate application of K could increase the UFA content but decrease the SFA content in the rice grain. The highest UFA and lowest SFA contents were obtained in response to the K application level of 135 kg ha-1for both NJ9108 and IR72.

Table 2 Effects of K application rate on fatty acid composition in rice grains (%)

3.4.Key enzyme activities

Acetyl-CoA carboxylase (ACCase),fatty acid synthase(FAS),glycerol 3-phosphate dehydrogenase (3-GPD),phosphatidyl phosphatase (PPase) and glycerol 3-phosphate acyhransferase (GPAT) are the key enzymes involved in lipid synthesis in rice grains. PEPCase and ACCase compete for the same PYR substrate and can catalyze the conversion of PYR into OAA and acetyl-CoA (A-CoA),respectively. The enzymatic activities of NJ9108 and IR72 first increased and then decreased during the grain filling process (Figs.3 and 4). As more K was gradually applied,the ACCase,FAS,3-GPD,PPase and GPAT activities of NJ9108 and IR72 also first increased and then decreased,and both followed the trend of K2＞K3＞K1＞K0 during the entire filling stage. As the K application level increased,the PEPCase activities in NJ9108 decreased at first and then increased,while in IR72 it showed the opposite trend (Fig.4-E and F).The activities of PEPCase in NJ9108 and IR72 at the peak stage followed the trends of K0＞K1＞K3＞K2 and K1＞K2＞K3＞K0,respectively. Overall,the results indicated that the activities of key enzymes involved in lipid synthesis in rice grains could be enhanced with appropriate K application levels. The highest activities of ACCase,FAS,3-GPD,PPase and GPAT were obtained in response to a K application level of 135 kg ha-1.

3.5.Metabolite accumulation

PYR is the initial substrate in lipid synthesis in rice grains,and MCA and PA are important metabolites in the lipid synthesis pathway. The contents of these metabolites in rice grains can indicate the activity of lipid synthesis. Duringthe entire filling stage,the PYR,MCA and PA contents of NJ9108 and IR72 first increased and then decreased(Fig.5). However,no significant differences in PYR content were found with increasing K application,whereas the MCA and PA contents increased at first and then decreased,and they followed the trend of K2＞K3＞K1＞K0. These results indicated that increases in K application levels did not enhance the PYR content but significantly increased the MCA and PA contents. The increases in MCA and PA contents might be due to increases in the activities of key enzymes involved in their synthesis,which was beneficial to lipid synthesis as a result. The highest MCA and PA contents of both cultivars were obtained in response to a K application level of 135 kg ha-1.

3.6.Protein accumulation

During the grain-filling process,the protein content of NJ9108 decreased gradually while that of IR72 first decreased and then increased (Fig.6). As more K was gradually applied,the protein and OAA contents of NJ9108 first decreased and then increased,and followed the trend of K0＞K1＞K3＞K2;while IR72 showed the opposite trend and tended to be in the order of K1＞K2＞K3＞K0 (Fig.7).The two-year test results showed that the protein content of NJ9108 at 36 d after flowering under K1 was significantly higher than under K2 and K3,and the protein contents of

IR72 under K1 and K2 were significantly higher than under K0. In 2019,compared with that under K0,the protein contents under K1,K2 and K3 decreased by 0.8,4.9 and 3.7% for NJ9108 and they increased by 2.9,2.9 and 2.6%for IR72,respectively. In 2020,compared with that under K0,the protein contents under K1,K2 and K3 decreased by 0.1,1.3 and 1.5% for NJ9108 and they increased by 1.9,1.6 and 1.4% for IR72,respectively. The OAA content represents the activity level of OAA synthesis from PYR.Correlation analysis showed that the OAA content was correlated with the protein content of NJ9108 (r=0.980＊)and IR72 (r=0.824). Taken together,these results indicated that the effects of the K application rate on the protein and OAA contents of the two different cultivars differed. The highest protein and OAA contents were obtained at the K levels of 0 kg ha-1for NJ9108 and 90 kg ha-1for IR72.

3.7.Rice eating quality

In 2019,the peak viscosity and breakdown viscosity of NJ9108 and IR72 first increased and then decreased as more K was gradually applied (Table 3),and the maximum values were observed under K2. In contrast,the final viscosity,setback viscosity and pasting temperature showed the opposite trend,and the minimum values were observed under K2. There were no significant differences in trough viscosity or peak time between the four K levels.With the exception of breakdown viscosity,all the other RVA profile characteristics of IR72 were higher than those of NJ9108. A similar trend was found in 2020 as well. Previous studies showed that higher values for peak viscosity and breakdown viscosity and lower values for final viscosity,setback viscosity and pasting temperature could lead to improved rice eating quality. In 2019,the taste value,mouthfeel and appearance of NJ9108 and IR72 rice first increased and then decreased as more K was gradually applied (Table 4),and the maximum values were observed under K2. No significant differences in hardness,stickiness or balance were observed between the four K levels. Compared to IR72,NJ9108 had a higher taste value,mouthfeel,appearance,stickiness and balance,but a lower hardness. These results indicated that the appropriate application of K fertilizer could improve rice eating quality,and the best eating qualities of both cultivars were obtained at the K level of 135 kg ha-1.

Table 3 Effects of K application rate on the rapid visco-analyzer (RVA) profile characters of rice in 2019 and 2020

Table 4 Effects of K application rate on the eating quality parameters of rice in 2019 and 2020

3.8.Yield and yield components

In 2019,as more K was gradually applied,the panicle number,filled grain number per panicle,seed-setting rate,1 000-grain weight and grain yield of NJ9108 increased significantly,while those of IR72 increased (Table 5). No significant differences were found between K2 and K3.The highest panicle number of NJ9108 was observed under K3 and the maximum values of the other indexes were observed under K2. The highest filled grain number per panicle of IR72 was observed under K3,and the maximum values of the other indexes were observed under K3. A similar trend was found in 2020 as well.These results indicated that the appropriate application of K could increase the rice yield. Both NJ9108 and IR72 achieved the highest values for filled grain number per panicle,seed setting rate,1 000-grain weight and grain yield in response to a K application level of 135 kg ha-1.

Table 5 Effects of the K application rate on rice yield and yield components in 2019 and 2020

3.9.Correlations between the main indicators

The lipid contents showed significant positive correlationswith the MCA contents and ACCase activities of both cultivars (Table 6). Taste value showed positive correlations with the lipid and UFA contents,but it showed a negative correlation with SFA content. However,the correlations of lipid contents with protein contents,and of OAA contents and PEPCase activities differed between the two cultivars. The lipid content of NJ 9018 showed significant negative correlations with the protein content,OAA content and PEPCase activity,while IR72 showed the opposite trends.

Table 6 Correlation coefficients (r,n=48) between lipid contents,key enzyme activities and taste values of Nanjing 9108 (NJ9108)(below the diagonal) and IR72 (above the diagonal)1)

4.Discussion

4.1.Effects of K application rate on the activities of key enzymes involved in lipid synthesis and lipid accumulation

The lipid content in rice grains is determined by the combined effects of the activities of several key enzymes involved in lipid synthesis,such as ACCase,FAS,3-GPD,PPase and GPAT (Caiet al.2018;Zhaoet al.2018;Zhanget al.2019). PA is the initial substrate for both lipid and protein synthesis,and PA can be converted either to MCA by ACCase as part of lipid synthesis or to OAA by PEPCase as part of protein synthesis. The activities of these enzymes are not only regulated by genetic characteristics but they are also affected by other factors,such as the amount of fertilizer applied (Guet al.2011;Tuet al.2020). Prior to this study,limited information was available on the effects of different K levels on rice lipid content. In the present study,we found that the lipid,FFA,MCA and PA contents and lipid synthesis-related enzyme activities of two cultivars followed the same trend of first increasing and then decreasing as more K was applied.The maximum values were obtained at the K level of 135 kg ha-1. However,trends of protein and OAA contents and PEPCase activities of the two cultivars differed as more K was applied. The effect equations relating lipid content with K application amount were established:y=55 925.99047+3 586.17947x-57.30029x2(R2=0.9965)(NJ9108) andy=34642.19085+2358.66257x-39.94958x2(R2=0.9127) (IR72). The optimal K application amounts,174.91 kg ha-1(NJ9108)and 172.25 kg ha-1(IR72),were calculated. Previous studies have shown that high ACCase concentrations can reduce PEPCase activity,which could promote the ACCase-catalyzed conversion of more PA in the lipid synthesis pathway (Wanget al.2013). Since there is competition between ACCase and PEPCase for the same substrate,PEPCase is the key enzyme that can balance the lipid and protein contents in rice grains(Songet al.2008). Most studies have found that protein content is closely correlated with PEPCase activity,while it is negatively correlated with lipid content (Zhang 2006;Li 2007). In this study,when the lipid content and ACCase activity of NJ9108 increased,the protein content and PEPCase activity showed the opposite trend,which agrees with the results of previous studies(Wuet al.2006). However,the changes in IR72 did not fit this trend well. Further analysis showed that PEPCase activity was correlated with the protein contents of NJ9108 (r=0.984＊) and IR72 (r=0.894)(Table 6). We speculated that the effects of the Kapplication rate on lipid and protein contents and on ACCase and PEPCase activities varied between the two different cultivars in this study. Therefore,the effects of different cultivars should be considered in further studies on K application rate.

4.2.Relationship between lipid content and rice eating quality

Previous studies have reported that high quality rice is characterized by a high lipid content,rich flavor,good luster and good palatability (Guet al.2011;Jianget al.2016). Rice cultivars with high eating quality normally show higher breakdown viscosity and lower setback viscosity,as determined by RVA. Compared to rice grain starch and protein contents,the lipid content,and especially the starch lipid content in milled rice,has a significant effect on the cooking and eating quality of rice,which increases with an increase in the starch lipid content in milled rice (Zhanget al.2009;Guet al.2011). Studies have also shown that there are significant differences in lipid content betweenindicaandjaponicarice based on an analysis of 120 different rice cultivars,and the average lipid contents inindicaandjaponicaricewere found to be 2.67 and 2.30%,respectively (Chenet al.2020). Furthermore,lipids have a greater influence on rice eating quality than other quality indicators,and they only have a certain correlation with the amylose content of rice (Jianget al.2016),which suggests that increasing the lipid content and decreasing the amylose content could be beneficial for improving the cooking and eating qualities of rice.

The results of this study showed that the highest lipid content,breakdown viscosity,and rice eating quality,and the lowest setback viscosity were obtained at the K level of 135 kg ha-1,in agreement with previous studies (Jianget al.2016). UFAs accounted for approximately 75% of the total fatty acids in rice grain,and not only have high nutritional value but also are closely associated with rice cooking and eating qualities (Yoshidaet al.2010).Increasing the UFA content could significantly improve the eating quality of rice (Yoonet al.2009;Yoonet al.2012).In this study,four main fatty acids,oleic acid (C18:1),linoleic acid (C18:2),palmitic acid (C16:1) and stearic acid(C18:0),were detected in both cultivars,which accounted for more than 90% of the total fatty acid content in the rice grains. The RVA results showed that the highest peak viscosity and breakdown viscosity,as well as the best taste value,appearance and mouthfeel and the lowest final viscosity,setback viscosity and pasting temperature were all observed at the K level of 135 kg ha-1in both NJ9108 and IR72;and this treatment also resulted in the highest UFA content and lowest SFA content,in agreement with previous studies (Yoshidaet al.2010).In summary,this study found that an appropriate K application rate could improve rice lipid content and eating quality. Therefore,modifying the K application rate is a very practicable way to increase the lipid content and rice eating quality.

5.Conclusion

The appropriate application of K fertilizer can optimize the fatty acid components and increase the lipid contents by enhancing lipid synthesis-related enzyme activities,improving rice eating quality as a result. As more K was gradually applied,the lipid,FFA,UFA,MCA and PA contents,lipid synthesis-related enzyme activities and eating quality of two cultivars first increased and then decreased. The maximum values were observed at the K level of 135 kg ha-1. The effects of K application rate on the protein contents and PEPCase activities of NJ9108 and IR72 differed. K application could stimulate lipid synthesis more than protein synthesis at the K levels of 135 kg ha-1for NJ9108 and 180 kg ha-1for IR72. In summary,the optimal K application rate for both NJ9108 and IR72 in terms of lipid synthesis,eating quality and grain yield was 135 kg ha-1.

Acknowledgements

This work was supported by the Sichuan Science and Technology Program,China (2020YFH0146 and 2022YFH0029).

Declaration of competing interest

The authors declare that they have no conflict of interest.