**3. Discussion**

### *3.1. E*ff*ect of P Fertilizer on the PA Content*

To develop *lpa* mutants, DNA mutations in the PA biosynthesis genes have been used. These *lpa* mutants helped reveal the mechanism of PA biosynthesis; however, *lpa* mutant rice often exhibited low plant biomass and yield compared with those of wild type [20]. Therefore, the use of *lpa* rice mutants in commercial agriculture is not very competitive. In our study, two contrasting rice cultivars selected in terms of PA content from natural variation in the rice germplasm did not show any significant di fference in the initial growth (Table 3) and yield-related traits (Table 1). P fertilization did not a ffect the panicle weight and yield in both WRC 5 and WRC 6 (Table 2). It suggested that the sensitivity of P application in terms of plant growth and yield was similar between WRC 5 and WRC 6. Thus, these contrasting lines are e ffective in analyzing the mechanism of P fertilizer response of PA synthesis without consideration of the plant yield and growth di fferences.

In rice, P is accumulated in the leaves until the heading stage, and then transported to the stem in the heading stage. Finally, P is stored in the grain at maturity [33]. According this P translocation system, we expect that rice would show a di fferent response of PA content with the time of P fertilizer application. Our study showed that the PA content increased with P fertilizer in WRC 5, irrespective of the time of P fertilizer application (Figure 1). Moreover, the PA content did not change in WRC 6 with P fertilizer. These results indicate that the e ffect of P fertilizer on the assimilation and translocation of P is not stage dependent but di ffers with the genotype.

When a higher amount of P was applied, the PA content increased in both cultivars (Figure 2a). Because the grain yield was not a ffected by higher P (P4) application, the PA content per grain (Figure 2b) or per plant (Figure 2c) also increased, indicating that PA biosynthesis was upregulated by P application. On the contrary, the PA content under the control treatment di ffered with the genotype; WRC 5 showed a significantly lower PA content than WRC 6. This result suggests that WRC 5 and WRC 6 have a di fferent responsiveness toward soil P, and WRC 6 might be sensitive to PA accumulation under lower soil P conditions. The Pi content in WRC 6 at 10 DAF was higher than that in WRC 5 (Figure 3a), and this might be due to the higher PA content in WRC 6. Interestingly, the Pi content at 10 DAF did not respond to P fertilizer application, and Pi influx in the early development stage of grain might determine the genotypic di fferences in the PA content.

### *3.2. E*ff*ect of P Fertilizer on PA Biosynthesis*

While the PA content varied in response to P application, the Pi content remained constant and was not a ffected by the P application and genotype (Figure 3a). The Pi content in the grain was found to be regulated at a constant level. While the Pi content was constant, the ratio of PA-P to Pi responded to P fertilizer, and it was associated with the PA content (Figure 3b). This suggests that the biosynthesized PA using influxed Pi determines the PA response to P application. To assess the PA biosynthesis level, we analyzed *INO1* gene expression, which we previously reported as the key determinant that explains genotypic di fferences between WRC 5 and WRC 6 [32]. The expression of *INO1* di ffered with genotype; WRC 6 responded to P fertilizer, whereas WRC 5 did not show any response to P fertilizer (Figure 4). This di fferential response to P fertilizer might be one of the key factors explaining the genotypic di fference in the PA content and the di fference in P responsiveness. Simultaneously, while *INO1* showed no response to P fertilizer, the PA content was stimulated by P fertilizer in WRC 5 (Figure 2b), suggesting there might be other factor(s), besides *INO1* gene expression, that regulate the PA biosynthesis response to P fertilizer.

## *3.3. Other Factors Regulating Grain PA Content*

The biosynthesis of PA continues until 25 DAF [32,34]; we investigated the expression level of other genes related to PA synthesis identified from the Nipponbare genome database [35]. Seven genes, namely, inositol 1, 3, 4-trisphosphate 5/6-kinase 2 (*ITPK2*), 2-phospho-glycerate kinase (*2-PGK*),

inositol-pentakisphosphate 2-kinase 1 (*IPK1*), myo-inositol kinase (*MIK*), inositol 1, 3, 4-trisphoshate 5/6-kinase *2* (*ITPK6*), inositol 1, 3, 4-triskisphosphate 5/6-kinase 1 (*ITPK1*), and *INO1* showed no response to P fertilizer treatment in Nipponbare (Figure 5). We could not compare these gene expression levels in WRC 5 and WRC 6 directly because WRC 5 and WRC 6 might have di fferent DNA sequences in the priming site of the polymerase chain reaction (PCR) and this interferes with the accurate comparison of the gene expression level. Although we should further analyze gene expression using WRC 5 and WRC 6, the results indicated that the PA content response to P fertilizer might not be regulated by PA biosynthesis in Nipponbare.

There are three potential steps that determine the grain PA content, namely, uptake of soil-Pi, translocation of Pi within the plant body, and biosynthesis and accumulation of PA in the grain. To analyze Pi uptake by the root system and Pi translocation capacity from the root to shoot, we analyzed the xylem sap in seedlings grown by hydroponic culture (Figure 6). The Pi content in xylem sap increased with P fertilizer application, but there was no di fference between WRC 5 and WRC 6. These results indicate that the Pi uptake and translocation in the seedling stage are stimulated by P fertilizer; this additional Pi might be used for additional PA biosynthesis. Contrarily, we could not find any genotypic di fference in the Pi content in xylem sap in the seedling stage. Further investigation of xylem sap collected at various growth stages will be required to asses Pi uptake and Pi translocation capacities. Recently, SPDT is reported to regulate Pi distribution between the leaves and grains [9]. A comprehensive analysis throughout Pi uptake, translocation, and grain accumulation will be important to further elucidate the e ffect of P fertilizer on PA accumulation in grains.
