**3. Discussion**

Many studies on the influence of P application on the growth, photosynthesis, nitrogen fixation, and yield of soybean and other crop species have been reported [17,22–26]. Most of these studies were carried out with crop cultivars that have "standard" (non-mutant) levels of seed phytic acid. In contrast relatively few such studies have been conducted with low-phytate variants of crops species [10,20] including one with low-phytate soybean lines [20] which will be discussed below. The proportion of phytate P to total P concentration in the seed of normal soybean cultivars is about 70–80%, but that of the low-phytate lines that we bred is about 30–35%. It has been reported that many low-phytate mutants impact plant growth and productivity and have decreased yield as compared with sibling normal-phytate/wild-type lines [27–29]. However in the case of soybean lines developed from crosses using the same low-phytate parent used here, CX1834 [14], previous studies demonstrated that while yields of low-phytate progeny were 85% to 90% of sibling normal-phytate lines, the yield loss was largely due to negative impacts on germination and emergence that led to reduced stand establishment, rather than impacts on plant growth and productivity subsequent to emergence and germination [30,31]. Therefore, in this study, we further investigated whether in the low-phytate lines that we bred, there was an e ffect on the growth, photosynthesis, nitrogen fixation, seed yield, and seed quality subsequent to germination and emergence. In addition, it has been reported that the phytate concentration of the seed of normal-phytate cultivars is increased by P application [17,19], and that there is a highly positive correlation between total P and phytate in seeds [17]. Therefore, we investigated whether the phytate concentration of seed increases with a higher P application level even in a low-phytate line.

Highly significant e ffects of the P application level on the plant growth, leaf photosynthesis, biological nitrogen fixation of nodules, and yield of seed were observed in this study. With higher application of P, the whole plant dry weight was markedly higher in Experiment 1 (Figure 1). However, while an increase in whole plant dry weight in the P 150 treatment compared to the P 50 treatment was observed, there were no di fferences in whole plant dry weight between the low-phytate line and normal-phytate cultivar in either P treatment. These results sugges<sup>t</sup> that the decrease of phytate in the seeds of the low-phytate line did not substantially a ffect the growth of plants subsequent to germination at any P application level.

At the low P level, the photosynthetic rate and dinitrogen fixation activity of both the low-phytate line and normal phytate cultivar were less than those of the higher P application level (Table 1). There are many reports that photosynthesis increases with P application. Wang et al. [26] reported that the activity of Rubisco, which is a key enzyme of photosynthesis, and the protein concentration of leaves were higher in a high P application than in a low P application. Furthermore, Singh and Reddy [32] also reported that Rubisco activity, RUBP regeneration, and maximum quantum yield related to the photochemical system were decreased under a P-deficient condition. Dinitrogen fixation activity was also increased by P application. It is assumed that the increase in dinitrogen fixing activity by P application is due to an increase in nodule weight (Table 1). In any case, photosynthesis and dinitrogen fixation were lower in the low P level, but the reduction was not significantly di fferent between the low-phytate line and normal-phytate cultivar. These results indicate that both seed phytate concentration and low-phytate genotype, at least those low-phytate lines derived from crosses using the CX-1834 source used here, have little impact on photosynthesis and dinitrogen fixation in soybeans at the growth stages following germination.

In this study, yield and yield components of both the low-phytate line and normal-phytate cultivar were significantly a ffected by P application. As expected, the yield of both the low-phytate line and normal-phytate cultivar was higher in the high P level than in the low P level. However, the low-phytate line had significantly higher seed yields than the normal-phytate cultivar in all the treatments (Figure 2 and Table 1). These di fferences were mainly attributed to the number of productive pods and 100 seed weight (Table 1).

Since comparisons here were made between low-phytate soybean lines obtained from the cross of cv. Tanbakuro and the low-phytate line CX1834 [14], and the unrelated normal-phytate cultivars Enrei and Akimaro, and not between the low-phytate line and a sibling normal-phytate/wild-type line derived from the same Tanbakuro by CX1834 cross, we cannot attribute any observed di fferences in yield or any trait to allelic di fferences in genes conditioning the low-phytate trait. We can only conclude that on a single-plant basis and not taking into account any e ffect on germination and emergence, yield and other traits of these specific low-phytate soybean progeny was observed to be equal to or better than either of these unrelated normal-phytate cultivars currently grown in Japan.

A previous study of several low-phytate soybean lines grown under varying levels of P supply in hydroponics reported that two such lines were more productive in terms of seed yield at given level of nutrient P than were the non-mutant, normal-phytic acid parental line, even when these lines were grown under a very growth-limiting low-nutrient P level [20]. However, one other low-phytate soybean line evaluated in this earlier study displayed reduced seed yield when grown under limiting nutrient P. Therefore one can conclude from the present and earlier studies that in soybean the impact of the low-phytate trait on growth and yield when grown under varying levels of nutrient P is variable, depending on genotype and level of nutrient P, but that some low-phytate soybean lines appear to be as productive if not more productive than normal-phytate lines.

In the two closely related food legume species soybean and common bean, there are two copies of the gene perturbed in the mutant studied here, a member of the MRP gene family that encodes an ABC transporter specific to phytic acid transport [16,33,34]. The low-phytate trait in the original low-phytate soybean parental line CX1834 line [14] that was used as the trait donor in breeding the low-phytate line studied here [21], was found to be conditioned by mutations in both copies of this particular MRP gene in the soybean genome [33]. In contrast, in the common bean low-phytate mutants that are conditioned by mutations in its MRP ortholog, the low-phytate trait is the result of a mutation in only one of the two MRP duplications, and there is little or no negative impact on germination, emergence or subsequent plant growth, performance or yield [15,16]. It was hypothesized that this lack of negative impact or pleiotropic e ffect is due to a bu ffering e ffect resulting from functional complementation provided by the second wild-type duplication of the gene [16,34]. Therefore it is possible that when transferring the trait from CX1834 into the parental background cv. Tanbakuro used to breed the of the low-phytate line studied here, a similar bu ffering e ffect was obtained, either provided by variant alleles of one of the two MRP orthologs, or perhaps by other components of the Tanbakuro genome. However the most likely explanation for the good growth and performance of the low-phytate soybean lines studied here is that problems with performance of this specific genotype are probably due to negative impacts on germination and emergence, rather than to negative impacts on subsequent growth and performance [29–31].

Highly significant e ffects of higher P application on the total P and phytate P concentrations of the seeds were found in both Experiment 1 (Figure 3A) and Experiment 2 (Figure 3B). While total P concentration was higher in the higher level of P application there was little di fference in total P between the low-phytate line and normal-phytate cultivar. However the substantial e ffects of P fertilization on phytate P and inorganic P concentrations di ffered between low-phytate lines and normal-phytate cultivars. For example, in Experiment 1 the phytate P concentration in the normal-phytate cultivar was 4.09 mg g<sup>−</sup><sup>1</sup> at the 50 P application level and 5.26 mg g<sup>−</sup><sup>1</sup> at the 150 P level. However, in the low-phytate line, phytate P concentrations were similar at both P fertilization levels, about 2.0 mg g<sup>−</sup><sup>1</sup> dry weight. Similar results were obtained in Experiment 2 (Figure 3B). In the low-phytate line, inorganic P increases occurred instead of increasing phytate P in response to increasing nutrient P supply. As an outcome of this, the proportion of phytic acid P to total P concentration was reduced with increased P fertilization in the low-phytate lines but not greatly a ffected by the level of P application in the normal phytate lines (Figure 3). These results support the finding of previous studies that indicate that the accumulation and synthesis of phytic acid in seed is highly dependent on the soil P level [35]. A highly significant positive relationship between P supply and phytic acid P in seeds has been observed in soybean [17] and maize [19]. The results of this study agree with that of Oltmans et al. [36] who reported that the

mean total P of a similar low-phytate line and its sibling normal-phytate lines was not significantly different, but the mean phytate P, inorganic P, and other P concentrations were significantly different.

Lipids and protein are important factors in seed quality. There are some reports that the lipid and protein concentrations of seed were increased or decreased by P fertilization, and the influence an lipids and proteins by P application was different depending on the study. For example, in soybean, Abbasi et al. [37] reported that P application increased the oil (lipid) and protein concentrations, but Yi et al. [38] reported that the protein concentration was increased, and oil concentration was decreased with increased P application. Krueger et al. [39] reported that oil (lipid) and protein concentrations were not affected by P fertilization. Bethlenfalvay et al. [40] also found that soybean seed lipid and protein concentrations were not significantly correlated, and there was a highly significant negative correlation between seed P and lipid concentration. In this study, the lipid and protein concentration of seed in both the low-phytate line and normal-phytate cultivar were significantly higher in the higher P application level (Table 3). From this study and previous reports, the influence of P application on the protein and lipid concentration of soybean seed may be different depending on the P application level, soil moisture condition, and characteristics of the cultivar used in the study. In any case, from this study, even when the low-phytate line and normal-phytate cultivar were cultivated under the same P application condition, there was no difference in the lipid and protein concentration of the seed.

The differences in seed K, Ca, and Mg concentrations observed here could be attributed to the differences in seed phytate concentration that result either from differences in P fertilization or low-phytate genotype (Table 3). A similar result was found by Toda et al. [41]: since Ca and Mg ions form salts with phytic acid and as a result diminish its buffering effect, unbound phosphate groups of phytic acid would result from the low concentration of Ca and Mg. In contrast to K, Ca, and Mg, microelements, such as Zn, Fe, Mn, and Cu in the both low-phytate line and normal-phytate cultivar were less in the higher P application level. It is well known that the concentration of these microelements is affected by P fertilization. Naeem et al. [23] reported that the Zn concentration of seed in wheat was decreased by P application and Sánchez-Rodríguez et al. [42] reported the same in wheat and barley. Zhang et al. (2012) [43] also reported that P application significantly decreased Zn and Cu, but had no effect on the Fe and Mn concentrations of wheat seed. Zhang et al. (2017) [44] also reported that total P to Zn, Cu, Fe, and Mn molar ratios in maize seed were increased, which indicates that P application in maize may affect the bioavailability of these minerals. In this study, Zn, Fe, and Mn levels and the phytic acid:mineral molar rations for Zn, Fe and Mn were not greatly affected by differing P fertilization in either the low-phytate line or normal-phytate cultivar (Table 3). However the higher level of phytic acid in the normal-phytate cultivar as compared with the low-phytate line when grown on low or high fertilizer P resulted in much higher (2.5 to 3.5-fold) phytic acid:mineral molar ratios. From this result, it can be suggested that the low-phytate line maintained high bioavailability of these microelements in comparison to the normal-phytate cultivar regardless of P application, and will likely have positive effects on the nutrition of monogastric animals.
