The field experiment was conducted at the experimental farm of Jiangsu Key Laboratory of Crop Cultivation and Physiology, Yangzhou University, Jiangsu Province, China (32° 30′ N, 119° 25′ E) during the rice-growing seasons of 2020 and 2021. The previous crop was wheat, and the soil texture was sandy loam consisting of 21.3 g kg^-1 organic matter, 1.32 g kg^-1 total nitrogen, 33.5 mg kg^-1 available phosphorus and 89.7 mg kg^-1 available potassium.

Two japonica rice, Nanjing 9108 (NJ9108) and Huaidao 5 (HD5) were used as materials. A two-factor randomized block design was adopted in this experiment. The first factor was two wheat straw returning treatments. One was wheat straw non-returning (NS), and the other was wheat straw full returning (WS). In the WS treatment, the wheat straw was cut into 5 cm long pieces after the previous wheat crop was harvested and then rototilled into the top 20 cm soil layer one week before rice transplanting. The second factor was two N application ratio treatment. The total N application during the rice growth period was 270 kg ha^-1, which was the common N rate for the local rice farmers(Yang et al, 2022; Liu et al., 2022). One treatment was the local farmers’ fertilizer practice (LFP), and the ratios of basal fertilizer, tiller fertilizer and panicle fertilizer to the total N rate were 50%, 10% and 40%. The other was increasing basal fertilizer rate (IBF), and the ratios of basal fertilizer, tiller fertilizer and panicle fertilizer to the total N rate were 70%, 10% and 20%. There were four treatment combinations, i.e., NS-LFP, NS-IBF, WS-LFP and WS-IBF.

Across both years, seedlings were sown on May 15-16 in 2020 and 2021 raised in a seedbed, and manually transplanted on June 12-13 at a hill spacing of 13.5cm× 25 cm. The plot area was 20 m² arranged in random blocks and repeated three times. N (urea) fertilizer was broadcasted at pre-transplanting (1 day before transplanting), tillering (7 days after transplanting) and panicle initiation (the first appearance of a differentiated apex). Before transplanting, P₂O₅ and K₂O were broadcasted as basal fertilizers at 40.5 kg ha⁻¹ and 125 kg ha⁻¹, respectively.

Experimental data were analyzed using SPSS version 28, and the mean value was tested by least significant difference (LSD), P=0.05. Plots were made using Origin 2021.

Two-year experiments showed that WS-IBF significantly increased the number of panicles and grains per panicle, resulting in a substantial increase in yield but a greater decrease in seed-setting rate (Table 1). The yield of WS-IBF increased by 6.7–9.0% and 10.3–12.2% compared with WS-LFP and NS-LFP treatments. Compared with NS-LFP treatment, NS-IBF treatment promoted the occurrence of rice tillers in the early stage (Table S1) and increased the number of panicles, but significantly reduced the percentage of productive tiller, number of grains per panicle, grain weight and seed-setting rate, resulting in a yield decrease of 8.3–12.7%. Based on conventional nitrogen proportional treatment (NS-LFP), wheat straw returning (WS-LFP) reduced the number of panicles, but would significantly increase the number of grains per panicle and grain weight, resulting in a slight increase in yield, but the difference between the two was not significant.

WS-IBF treatment had the highest root and shoot dry matter weight at all stages except the mid-tillering stage (Table 2). Under the same N fertilizer ratio, WS treatments reduced root and shoot dry matter by 2.9–17.9% and 2.2–16.4% at the mid-tillering stage, respectively, compared with NS treatments, but significantly increased dry matter weight from the panicle initiation stage to the maturity stage. Under the IBF ratio, WS-IBF treatment could dramatically improve root and shoot dry matter weight at the maturity stage. Under the LFP ratio, WS-LFP treatment had no significant increase in dry matter weight compared with NS-LFP treatment, which was consistent with the yield performance.

Under the condition of WS, WS-IBF treatment promoted dry matter weight in the whole rice growth period compared with WS-LFP treatment, while under the NS, NS-IBF treatment only increased dry matter weight from the mid-tillering stage to the panicle-initiation stage. Compared with NS-LFP treatment, NS-IBF treatment reduced root and shoot dry matter weight by 19.8–29.7% and 5.6–7.4% at the maturity stage, respectively. In contrast, WS-IBF treatment increased root and shoot dry matter weight at the maturity stage by 6.8–18.5% and 5.3–6.2%, respectively, compared with WS-LFP treatment.

There was no significant difference in soil TN content among varieties at each period, but the difference between treatments was extremely significant (Table S2). IBF treatments increased soil TN content at the mid-tillering stage, while LFP treatments increased soil TN content at the middle and late stages. Compared with NS treatments, WS treatments significantly increased soil TN content from panicle-initiation stage to maturity stage.

As shown in Figure 1, among the treatments, WS-IBF treatment had the most significant improvement in the taste quality of the two varieties (Figure 1A), which increased the taste value by 1.1–5.7% compared with the other three treatments. The stickiness (Figure 1C) of WS treatments were higher than those of NS treatments under the same N fertilizer ratio, and the taste value and stickiness gradually increased with increasing basal N fertilizer ratio under the same straw treatment. The hardness (Figure 1B) showed the opposite change.

Under the same N fertilizer ratio, WS treatments had no significant effect on hot viscosity (Figure 1E), final viscosity (Figure 1F) and pasting temperature (Figure 1I) but significantly reduced the setback value (Figure 1H), while the peak viscosity (Figure 1D) and breakdown value (Figure 1G) were increased by 5.8–14.6% and 2.6–17.4%, respectively. Under the NS and WS treatments, the peak viscosity, hot viscosity, final viscosity and breakdown value gradually increased with the increase of the ratio of basal N fertilizer; in contrast, the setback value and pasting temperature decreased.

WS treatments reduced the amylose content (Figure 1J) and gel consistency (Figure 1L) of grain and significantly increased the ratio of amylopectin content to amylose content (Figure 1K) by 8.3–24.8%. Analysis-of-variance of F-values (Table S3) indicated that all the quality measurements showed significant difference (P<0.01) among the cultivars and treatments. Correlation analysis (Figure 2) showed that the amylose content had a significant negative correlation with taste value and stickiness and a significant positive correlation with hardness. The ratio of amylopectin content to amylose content was significantly positively correlated with taste value and stickiness and significantly negatively correlated with hardness.

WS treatments significantly increased the content of protein and its components under the same N fertilizer ratio compared with NS treatments (Table 3). Under the NS and WS treatments, protein content gradually decreased with the increase of basal N fertilizer ratio. Still, it increased the percentage of glutenin to prolamin by 0.9–3.0% for the two varieties. The protein, albumin, globulin, prolamin and glutenin content of NJ9108 were generally lower than that of HD5, but the ratio of high-quality glutenin to prolamin, which is not easily absorbed by the human body, was higher than HD5. Correlation analysis (Figure 2) indicated that protein content was significantly negatively correlated with taste value, the ratio of glutenin content to prolamin content had a significant positive correlation with taste value and stickiness, and a significant negative correlation with hardness.

There is a significant difference in the 2-AP content of milled rice between the two varieties (Figure 3). The 2-AP content of fragrant rice variety NJ9108 is between 18.4–27.7 ng g^-1 and the content detected for non-fragrant rice variety HD5 is between 0.4–0.5 ng g^-1. Under the same N fertilizer ratio, WS-LFP treatment increased the 2-AP content of NJ9108 by 12.0–17.8% compared with NS-LFP treatment, while WS-IBF increased by 24.5–28.8% compared with NS-IBF treatment, which indicated that WS treatments had a more significant increase in 2-AP content under the IBF fertilizer ratio. Under NS and WS treatments, the 2-AP content in milled rice was decreased by increasing the basal N fertilizer ratio, and there was a significant difference among different N fertilizer ratios under the condition of NS treatments. Compared with NS-LFP treatment, NS-IBF treatment reduced the 2-AP content of NJ9108 by 12.0–12.4%, while WS-IBF treatment only reduced 2.6–3.7% compared with WS-LFP treatment.

With the increase of days after heading, the leaf proline content of NJ9108 and HD5 increased first and then decreased, reaching the highest value at 30 days after heading, and the grain proline content gradually decreased (Figure 4). The variation range of proline content of HD5 was smaller than that of NJ9108. Compared with NS treatments, WS treatments increased the proline content in the leaves and grains of the two varieties, and increasing the basal N fertilizer ratio would reduce the proline content in leaves and grains. However, there was no significant difference in grain proline content between WS-LFP treatment and WS-IBF treatment after 40 days of heading, while NS-IBF treatment decreased more than NS-LFP treatment.

The activities of proline dehydrogenase (PDH) (Figure 5) and ornithine aminotransferase (OAT) (Figure 6) in leaves of the two varieties generally showed a downward trend from the heading stage. Except for the leave OAT activity of NJ9108, which increased slightly at 30 days after heading, the enzyme activities of the other treatments all decreased significantly from 0 to 30 days after heading. On the other hand, the enzyme activity changed slightly from 30 days after heading to maturity and increased slightly in some treatments. The activities of the two enzymes in grains increased gradually with the advancement of the growth period, and the changes tended to be flat after 30 days of heading. Under the same N fertilizer ratio, the enzyme activities of WS treatments were significantly higher than that of NS treatments. At maturity, WS treatments increased grain PDH enzyme and OAT enzyme activities by 14.9–34.7% and 8.6–19.1% compared with NS treatments. Under NS and WS treatments, increasing basal N fertilizer ratio reduced the enzyme activity, and NS-IBF treatment reduced the PDH enzyme and OAT enzyme activity of the two varieties by 16.4–20.1% and 9.7–10.9% at maturity compared with NS-LFP treatment. However, compared with WS-LFP treatment, WS-IBF treatment only reduced the activities of these two enzymes by 6.8–7.8% and 5.1–5.4%, and the difference was slight.

Correlation analysis (Table 4) showed that 2-AP content in milled rice at maturity was significantly positively correlated with proline content in grains and leaves 30 days after heading to maturity; It also had a significant positive correlation with the grain proline dehydrogenase activity 20 days after heading to maturity, and a significant positive correlation with the leave proline dehydrogenase activity at 40 days after heading; In addition, the 2-AP content in milled rice at maturity was also significantly positively correlated with grain ornithine aminotransferase activity 30 days after heading to maturity, and significantly positively correlated with leave ornithine aminotransferase activity 30-40 days after heading. The results showed that wheat straw returning was beneficial to increase the activities of proline dehydrogenase and ornithine aminotransferase in grains at the middle and late stage of grain filling, and increasing the proline content, the precursor of 2-AP synthesis, thus promoting 2-AP synthesis in grains.

Correlation analysis (Figure 7) showed that the soil TN content at heading stage was significantly correlated with PDH enzyme activity from 30 days after heading to maturity, and OAT enzyme activity from 40 days after heading and maturity; The content of soil TN at maturity was significantly correlated with the activities of PDH and OAT enzyme 40 days after heading and maturity.

Straw returning is an effective way to improve soil properties and fertility. Many studies have shown that straw returning can improve the soil nutrient content and promote the uptake of soil nitrogen by roots, thus increasing the rice yield, and the yield increases with the increase of straw returning years (Wang et al., 2020; Cui et al., 2022). However, some studies held that straw returning has no significant effect on rice yield and even has a tendency to reduce it, which is related to soil texture and tillage methods (Zhu et al., 2011; Zhang et al., 2016). Kong et al. (2021) believed that the effect of straw returning on yield is closely related to the ratio between basal tiller fertilizer and panicle fertilizer. Straw is rich in a large number of carbon elements, and its C/N ratio is much higher than the suitable range of soil microorganisms. In addition, microorganisms will consume a certain amount of nitrogen during the decomposition of straw, which will lead to the reduction of nitrogen in the soil and fail to meet the normal growth of rice in the early stage (Tang et al., 2020). Pei et al. (2015) and Liu et al. (2017) deemed that wheat straw returning can reduce the occurrence of tillers in the early stage of rice and thus affect the yield, but significantly improve the growth indexes in the middle and late stages of rice, the results of this experiment were consistent with this. The panicle number of WS-LFP treatment was reduced compared with NS-LFP treatment under conventional N fertilizer ratio, while the yield increased slightly by a significant increase in the number of grains per panicle and the grain weight, but the difference between the two was not significant. Huang et al. (2013) and Tian et al. (2022) suggested that increasing the amount of N fertilizer used in the early stage under straw returning can reduce the inhibitory effect of straw returning on rice tillering so as to increase the number of panicles, and at the same time, the nitrogen released after straw decomposition delays the decline of leaf photosynthetic rate in the late growth stage of rice, which significantly improves the grain quality of superior and inferior grains, and eventually increases the yield. This study also showed that under wheat straw returning, appropriately increasing the basal N fertilizer ratio significantly enhanced the soil TN content thus augmenting the number of panicles and the grains per panicle, which greatly increased the yield. We also found that the panicle fertilizer rate in the WS-LFP was higher than that of WS-IBF, but the grain per panicle in the WS-LFP was lower than the WS-IBF, this is primarily due to IBF promoted straw decomposition and accelerated N release. We found that at the panicle-initiation stage, the increase rate of root dry weight of WS-IBF was higher than that of shoot dry weight compared with WS-LFP (Table 2). Liu et al. (2020) reported that increasing root dry weight at the panicle-initiation stage is conducive to promoting the spikelet differentiation and ultimately increase the grain per panicle.

Most studies have shown that the dry matter accumulation of rice in the early growth stage after straw returning to the field is smaller than that of the straw non-returning treatment (Pei et al., 2015; Li et al., 2017). This is mainly due to the accumulation of a large number of harmful substances, such as soil organic acids caused by straw returning; meanwhile, the need for nitrogen fixation by soil microorganisms to decompose the straw resulting in a phenomenon of competition for nitrogen between microorganisms and plants, so it will inhibit the accumulation of dry matter in the early stage of rice (Takakai et al., 2018; Zhang Y. J. et al., 2022). The results of this experiment are basically consistent with Pei et al. (2015) and Li et al. (2017). However, under the treatment with a higher basal N fertilizer rate, there was no significant difference in dry matter weight at the early growth stage between the WS treatments and the NS treatments, indicating that increasing the basal N fertilizer ratio under straw returning can alleviate the phenomenon of insufficient nitrogen supply in the early stage and reduce the inhibition effect of straw mulching (Yan et al., 2018). With the deepening of straw decomposition and nitrogen release in the later stage, the dry matter weights of all WS treatments were significantly higher than that of the NS treatments at the panicle-initiation stage, heading stage and maturity stage. Therefore, increasing the basal N fertilizer rate under wheat straw returning can effectively balance the population growth in the early and later stages of rice, significantly increase the dry matter weight of the rice population in the middle and later stages and ensure the coordinated growth of the number of panicles, the number of grains per panicle, and the grain weight, so as to achieve high yield.

Wu et al. (2022) concluded that wheat straw returning can significantly reduce the amylose content of rice, increase the gel consistency and improve the cooking quality of rice, and the results of this experiment are generally consistent with this. In this experiment, we found that under the LFP ratio, although wheat straw returning was beneficial to improve the stickiness and the ratio of amylopectin content to amylose content, meanwhile, reduce the hardness and amylose content of the grain, it also reduced the gel consistency and taste value, which tended to deteriorate the taste quality of rice, this is basically the same as the research results of Liu et al. (2007) and Hou et al. (2015). The effect of the N application ratio on rice cooking and eating quality was also inconsistent. Ju et al. (2018) showed that postponing nitrogen application can significantly improve the gel consistency of japonica rice, while Cheng et al. (2021) concluded that increasing N fertilizer at the later stage will increase the amylose content of rice, decrease the branching and crystallinity of amylopectin, and also the gel consistency. Hu et al. (2017) found that increasing basal N fertilizer lengthens the gel consistency, and the gel consistency of base tiller fertilizer: panicle fertilizer between 7:3 and 9:1 is significantly greater than that of 6:4. The starch RVA profile is also an important indicator of the cooking and eating quality of rice. Varieties with relatively high peak viscosity, breakdown values and low setback values are generally considered to have better taste quality. Most studies have shown that the peak viscosity and breakdown value of straw-returning treatment are higher than those of straw-non-returning treatment, while the setback value decreased (Ge et al., 2013; Chen et al., 2017), which can improve the palatability of rice. The results of this study are basically consistent with those of previous studies. Meanwhile, it was also found that the peak viscosity, hot viscosity, final viscosity and breakdown values of the WS-IBF treatment were higher than those of the other treatments, while the setback value was the opposite.

Grain protein is a key nutrient that determines rice nutrition and taste quality. In this study, the protein content of HD5 was higher than that of NJ9108 under all treatments, but the taste value of HD5 was significantly lower than that of NJ9108. Studies have shown that straw returning enhances the photosynthetic material production capacity and transport capacity of rice populations during the grain-filling period, and significantly increases the protein content of rice (Chen et al., 2014). Increasing the application of panicle and grain fertilizer can improve the activity of key enzymes of nitrogen metabolism in the rice grain filling stage, which is conducive to the degradation and transportation of leaf protein into the grain, increasing the content of protein and its components (Pan et al., 2010; Duan et al., 2013), thus reducing the rice taste quality. In this study, we found that WS-LFP treatment could reduce rice taste quality, mainly due to the significant increase in protein content. In contrast, WS-IBF treatment could reduce the content of protein and its components in grain, increase the ratio of glutelin to prolamin, and improve the taste quality and nutritional quality.

Numerous studies have shown that the aroma of rice is not only controlled by genotype but also affected by soil, water, light, temperature and fertilizer (Goufo et al., 2010). There are few reports on the effects of straw returning on rice aroma, but it is generally believed that straw returning can enhance soil nitrogen levels, increase organic matter content and improve the soil environment, while soil with high organic matter and nitrogen is conducive to improving the activities of proline dehydrogenase and ornithine transaminase in rice grains, maintaining a high level of proline content in leaves and grains, thus promoting the increase of 2-AP content in grains at maturity (Yang et al., 2012), the results of this study are basically consistent with them. It has also been suggested that the decomposition of wheat straw in a flooded state causes the soil to behave in a strongly reducing state, resulting in the accumulation of hydrogen sulfide, organic acids, phenols and other substances, which have a stress effect on the growth of rice roots (Wang et al., 2015). As an osmoregulatory substance, proline can produce an active adaptive reaction in response to the stress and promote its synthesis and accumulation in plants to cope with stress damage (Liang et al., 2013). At the same time, straw decomposition releases a large amount of potassium, which is an important solute for cellular osmoregulation, and the increase in potassium promotes the activity of OAT, a key enzyme in the ornithine pathway of proline anabolism, further increasing the content of proline, a precursor substance for 2-AP synthesis (Cao and Wei, 2010; Zhang Z. et al., 2022). Correlation analysis showed that the content of 2-AP in grains at maturity was significantly or extremely significantly positively correlated with the precursor substances and related enzyme activities of 2-AP synthesis in grains at the middle and late filling stages. The effect of the N application ratio on rice aroma is also inconsistent. The research results of Mo et al. (2018) showed that increasing nitrogen fertilizer application at the booting stage can significantly increase the content of proline and 2-AP in plants, while Ren et al. (2017) believed that increasing nitrogen application at the tillering stage can promote leaf proline accumulation at early stage and increase the amount of proline translocated from leaves to grains at late growth stage of rice, so as to increase the 2-AP content in grains at maturity stage. On the contrary, some studies suggested that nitrogen application has adverse effects on the aroma of rice, and nitrogen content was negatively correlated with grain aroma content (Ganajaxi and Hegde, 2007; Jiang et al., 2015). In this study, under the NS and WS treatments, the 2-AP content, synthetic precursor substances and related enzyme activities decreased with the increase of basal nitrogen fertilizer rate. Under the condition of NS, the 2-AP content of grains at maturity stage in the NS-IBF treatment of NJ9108 decreased significantly, averaging 12.2% less than that in the NS-LFP treatment in two years, while under the condition of WS, WS-IBF treatment only decreased 3.2% compared with WS-LFP treatment, with no significant difference.