The effect of OsPAP10c on the growth and nutrient accumulation in rice
Previously, we have shown that Pi starvation specifically induced the expression of OsPAP10c in the roots only, which encodes the major secreted ACP isoform in rice. Overexpression of OsPAP10c by its promoter significantly improved rice growth under controlled conditions (Deng et al., 2020). To further test the effect of OsPAP10c on the yield in the field, OsPAP10c mutant and overexpression plants were planted in a long-term field experimental site in Jingzhou (30°37′N, 112°05′E). The experimental fields were applied with P fertilizer (+ P) according to local farmers' habits or without P fertilizer (-P), respectively. The yields of the OsPAP10c overexpression line were significantly increased under both + P and -P fertilizer conditions (Fig. S1). In contrast, the mutation of OsPAP10c significantly decreased the yield under + P conditions only (Fig. S1). These results indicated that the expression level of OsPAP10c significantly regulated rice yields under both + P and -P fertilizers in the field.
To elucidate the effect of OsPAP10c on the rhizosphere bacterial community, we collected paddy soils of the long-term field experimental site to grow the wild type (WT), pap10c, and OE-PAP10c plants in pots with and without P fertilizer according to the field experiments. The results showed that the genotypes significantly influenced shoot height and dry weight across several developmental stages (Table 1 and Fig. S2). Under -P fertilizer conditions, the average plant height and dry weight of OsPAP10c overexpression plants were higher than those of the WT at all four developmental stages, with plant height reaching a significant level at tillering stage and dry weight reaching a significant level at both tillering and elongation stages (Table 1). In contrast, the dry weight of the OsPAP10c mutant was significantly lower than that of WT during the ripening stage under -P fertilizer conditions and elongation and filling stages under + P fertilizer conditions (Table 1). In contrast to the influence of genotypes, P fertilizer alone has little effect on rice growth under pot-based experimental conditions. However, there is significant interaction between genotypes and P fertilizer treatments on plant dry weight at elongation and ripening stages (Table 1). The total nitrogen (N), phosphorus (P), and potassium (K) contents were measured to determine the nutrient uptake ability across different developmental stages. Under + P fertilizer conditions, there are no significant differences in the N, P, and K uptake ability between WT, OsPAP10c mutant, and overexpression plants at all four developmental stages (Table 2). Under -P fertilizer conditions, overexpression of OsPAP10c significantly increased P content at tillering and elongation stages, while mutation of OsPAP10c significantly decreased P content at the ripening stage (Table 2). Interestingly, overexpression of OsPAP10c also increased N uptake content at tillering and elongation stages under -P fertilizer conditions (Table 2).
Table 1
Shoot height and dry weight of the wild-type (WT), OsPAP10c mutant (pap10c) and overexpression (OE) plants with and without P fertilizer along the four developmental stages
P level | G-Type | Shoot height(cm) | Dry weight(g) |
Tillering | Elongation | Filling | Ripening | Tillering | Elongation | Filling | Ripening |
+P | WT | 46.88 ± 8.19a | 72.50 ± 8.88a | 66.05 ± 4.09b | 70.67 ± 4.93a | 0.45 ± 0.02a | 3.45 ± 0.60a | 5.76 ± 0.59a | 6.51 ± 0.37a |
pap10c | 51.72 ± 3.57a | 64.60 ± 6.18a | 69.83 ± 6.99ab | 66.50 ± 4.28a | 0.57 ± 0.00a | 1.83 ± 0.86b | 3.26 ± 0.75b | 7.48 ± 0.74a |
OE | 52.08 ± 8.28a | 65.50 ± 8.30a | 76.17 ± 5.65a | 70.00 ± 7.64a | 0.58 ± 0.18a | 2.58 ± 0.64ab | 3.98 ± 1.19b | 7.43 ± 0.68a |
-P | WT | 52.23 ± 3.50b | 68.78 ± 6.42a | 70.67 ± 4.75ab | 70.67 ± 3.20a | 0.55 ± 0.06b | 2.07 ± 0.21b | 4.48 ± 0.97a | 8.19 ± 1.21a |
pap10c | 48.93 ± 2.28b | 63.52 ± 6.52a | 62.50 ± 2.59b | 71.67 ± 4.13a | 0.62 ± 0.10b | 1.90 ± 0.22b | 4.69 ± 0.60a | 6.25 ± 0.89b |
OE | 59.37 ± 2.86a | 69.53 ± 7.61a | 78.83 ± 12.58a | 73.50 ± 6.50a | 0.87 ± 0.02a | 2.98 ± 0.69a | 4.78 ± 1.89a | 8.81 ± 0.39a |
Significance | | | | | | | | |
G | | * | * | *** | ns | ** | * | ns | * |
P | | ns | ns | ns | ns | ** | ns | ns | ns |
G×P | | ns | ns | ns | ns | ns | * | ns | ** |
Note: ns = not significant, * = significant at the 5% level, ** = significant at the 1% level, *** = significant at the 0.1% level. The data represents mean ± SD. Numbers with different letters differ significantly at the 5% level by Duncan’s significant difference. + P, applying P fertilizer; - P, not applying P fertilizer. |
Table 2
N, P, and K uptake in the shoots of the wild-type (WT), OsPAP10c mutant (pap10c) and overexpression (OE) plants with and without P fertilizer along the four developmental stages
P level | G-Type | N uptake (mg/plant) | P uptake (mg/plant) | K uptake (mg/plant) |
Tillering | Elongation | Filling | Ripening | Tillering | Elongation | Filling | Ripening | Tillering | Elongation | Filling | Ripening |
+P | WT | 14.10 ± 1.08a | 55.13 ± 4.70a | 109.08 ± 5.55a | 113.89 ± 6.57a | 1.05 ± 0.05a | 6.03 ± 0.85a | 7.63 ± 0.86a | 14.37 ± 1.88a | 15.81 ± 0.71a | 51.36 ± 3.04a | 61.19 ± 3.55a | 164.74 ± 12.31a |
pap10c | 19.34 ± 0.16a | 43.80 ± 17.78a | 81.36 ± 16.86a | 128.37 ± 23.52a | 1.37 ± 0.12a | 4.30 ± 1.83a | 6.52 ± 0.84a | 17.60 ± 3.54a | 16.96 ± 1.43a | 35.28 ± 15.59a | 45.73 ± 4.9a | 180.59 ± 25.01a |
OE | 18.12 ± 5.97a | 56.75 ± 5.29a | 86.56 ± 32.07a | 141.32 ± 22.22a | 1.35 ± 0.49a | 5.01 ± 0.55a | 6.94 ± 2.74a | 18.07 ± 1.17a | 17.70 ± 4.69a | 45.11 ± 6.65a | 49.85 ± 16.48a | 183.15 ± 2.02a |
-P | WT | 16.14 ± 2.68b | 44.27 ± 7.82b | 105.99 ± 17.24a | 132.18 ± 10.84a | 1.45 ± 0.23b | 4.39 ± 0.43b | 8.37 ± 1.95a | 19.48 ± 2.29a | 17.69 ± 3.47a | 42.64 ± 1.95b | 55.77 ± 12.78a | 223.39 ± 33.98a |
pap10c | 18.46 ± 2.02b | 42.40 ± 3.63b | 88.03 ± 8.43a | 129.46 ± 20.16a | 1.62 ± 0.22ab | 4.04 ± 0.51b | 7.12 ± 0.05a | 15.81 ± 1.97b | 17.85 ± 4.81a | 35.02 ± 2.71c | 44.94 ± 2.98a | 184.47 ± 29.00a |
OE | 23.30 ± 0.79a | 74.34 ± 10.51a | 98.62 ± 20.00a | 153.29 ± 14.14a | 2.13 ± 0.32a | 6.32 ± 0.23a | 7.34 ± 1.51a | 17.11 ± 0.51ab | 22.16 ± 2.98a | 59.74 ± 0.31a | 51.25 ± 14.02a | 219.47 ± 30.59a |
Significance | | | | | | | | | | | | |
G | | ** | ** | ns | ns | * | * | ns | ns | ns | ** | ns | ns |
P | | ns | ns | ns | ns | *** | ns | ns | ns | ns | ns | ns | * |
G×P | | ns | * | ns | ns | ns | * | ns | * | ns | ns | ns | ns |
Note: ns = not significant, * = significant at the 5% level, ** = significant at the 1% level, *** = significant at the 0.1% level. The data represents mean ± SD. Numbers with different letters differ significantly at the 5% level by Duncan’s significant difference. + P, applying P fertilizer; - P, not applying P fertilizer. |
Rhizosphere soil properties and enzyme activity
Application of P fertilizer significantly increased the total P content of rhizosphere soils at tillering and filling stages (Table 3). The average Olsen-P was higher under + P fertilizer conditions than under -P fertilizer conditions, significantly increasing at tillering stage (Table 3). P fertilizer treatment also significantly increased the Po at the tillering stage but decreased the Po at the ripening stage (Table 3). P fertilizer reduced soil pH at all developmental stages (Table S1). Although OsPAP10c effectively mineralized Po and increased Pi content in hydroponic solutions (Deng et al., 2020), we did not observe any differences in the content of total P, Olsen-P, or Po in the rhizosphere soils among different genotypes (Table 3). ACP, alkaline phosphatase (ALP), and phytase activities were measured to further elucidate P cycling in the rhizosphere. The results showed that genotypes significantly affected the ACP activity in the rhizosphere soils at all four developmental stages. Under -P fertilizer conditions, ACP activity was significantly higher in the rhizosphere soils of OE-PAP10c plants than those of the WT at all developmental under -P fertilizer conditions (Table 4). However, the ACP activity of pap10c was similar to wild type under -P fertilizer conditions, suggesting other ACP isoforms could complement the loss of OsPAP10c at -P fertilizer conditions. Under + P fertilizer conditions, a lower ACP activity of pap10c was observed at both the elongation and filling stages. In comparison, a higher ACP activity of OE-PAP10c plants was observed at the ripening stage compared with that of WT (Table 4). P fertilizer significantly suppressed the ACP activity and interacted with the genotype at the filling stage (Table 4). In contrast to the negative effect on ACP activity, P fertilizer application significantly improved ALP activity in the rhizosphere soil at the filling and ripening stages (Table 4). Genotypes showed a limited effect on the ALP activity. For the phytase activity, both genotypes and P fertilizer significantly influenced the phytase activity at the elongation and ripening stages. Moreover, there is a significant interaction between genotypes and P fertilizer at the elongation, filling, and ripening stages for phytase activity (Table 4).
Table 3
Total P, Olsen-P, and organic P content in rhizosphere soil of the wild-type (WT), OsPAP10c mutant (pap10c) and overexpression (OE-PAP10c) plants with and without P fertilizer along the four developmental stages
P level | G-Type | Total P (g/kg) | Olsen-P (mg/kg) | Organic P (g/kg) |
Tillering | Elongation | Filling | Ripening | Tillering | Elongation | Filling | Ripening | Tillering | Elongation | Filling | Ripening |
+P | WT | 0.61± 0.05a | 0.47± 0.02a | 0.56± 0.02a | 0.49± 0.02b | 19.99± 4.30a | 13.37± 1.08a | 14.35± 1.89a | 10.53± 3.17a | 0.26± 0.02a | 0.21± 0.02a | 0.27± 0.01a | 0.20± 0.01a |
pap10c | 0.51± 0.09a | 0.51± 0.21a | 0.54± 0.02a | 0.54± 0.02a | 17.05± 3.96a | 12.97± 2.90a | 14.16± 1.17a | 9.68± 1.47a | 0.19± 0.09a | 0.23± 0.02a | 0.27± 0.02a | 0.22± 0.04a |
OE | 0.57± 0.04a | 0.47± 0.06a | 0.54± 0.04a | 0.50± 0.01ab | 16.78± 2.91a | 13.37± 2.04a | 13.30± 2.57a | 10.26± 0.91a | 0.24± 0.02a | 0.17± 0.05a | 0.26± 0.04a | 0.20± 0.02a |
-P | WT | 0.39± 0.12a | 0.52± 0.12a | 0.48± 0.02a | 0.51± 0.01a | 13.04± 1.29a | 10.77± 3.53a | 12.68± 1.77a | 7.55± 2.56a | 0.12± 0.06a | 0.26± 0.13a | 0.24± 0.02a | 0.25± 0.01a |
pap10c | 0.43± 0.02a | 0.45± 0.02a | 0.50± 0.02a | 0.51± 0.01a | 14.11± 0.53a | 12.37± 2.58a | 12.95± 0.88a | 9.42± 3.93a | 0.16± 0.03a | 0.20± 0.02a | 0.26± 0.04a | 0.25± 0.01a |
OE | 0.44± 0.03a | 0.43± 0.03a | 0.50± 0.02a | 0.51± 0.01a | 14.18± 1.14a | 12.57± 4.68a | 12.83± 0.30a | 9.02± 2.60a | 0.18± 0.02a | 0.19± 0.03a | 0.24± 0.04a | 0.24± 0.01a |
Significance | | | | | | | | | | | | |
G | | ns | ns | ns | ns | ns | ns | ns | ns | ns | ns | ns | ns |
P | | *** | ns | *** | ns | ** | ns | ns | ns | ** | ns | ns | *** |
G×P | | ns | ns | ns | ns | ns | ns | ns | ns | ns | ns | ns | ns |
Note: ns = not significant, * = significant at the 5% level, ** = significant at the 1% level, *** = significant at the 0.1% level. The data represents mean ± SD. Numbers with different letters differ significantly at the 5% level by Duncan’s significant difference. + P, applying P fertilizer; - P, not applying P fertilizer. |
Table 4
Activities of acid phosphatase, alkaline phosphatase and phytase in rhizosphere soil of the wild-type (WT), OsPAP10c mutant (pap10c) and overexpression (OE-PAP10c) plants with and without P fertilizer along the four developmental stages
P level | G-Type | ACP (µmol /g · h) | ALP (µmol / g · h) | Phytase (µmol / g · h) |
Tillering | Elongation | Filling | Ripening | Tillering | Elongation | Filling | Ripening | Tillering | Elongation | Filling | Ripening |
+P | WT | 0.77 ± 0.18ab | 2.21 ± 1.00a | 1.92 ± 1.14a | 1.51 ± 0.49b | 0.10 ± 0.03a | 0.31 ± 0.09a | 0.34 ± 0.06a | 0.36 ± 0.11a | 0.14 ± 0.01a | 0.05 ± 0.02b | 0.07 ± 0.03b | 0.06 ± 0.01b |
pap10c | 0.66 ± 0.10b | 1.16 ± 0.22b | 0.53 ± 0.19b | 1.56 ± 0.35b | 0.10 ± 0.02a | 0.16 ± 0.09b | 0.22 ± 0.02b | 0.40 ± 0.06a | 0.16 ± 0.03a | 0.07 ± 0.01a | 0.09 ± 0.01ab | 0.10 ± 0.01a |
OE | 1.04 ± 0.44a | 3.72 ± 1.27a | 1.72 ± 0.13a | 2.32 ± 0.66a | 0.11 ± 0.02a | 0.21 ± 0.10ab | 0.24 ± 0.09b | 0.40 ± 0.11a | 0.15 ± 0.03a | 0.06 ± 0.02ab | 0.10 ± 0.02a | 0.10 ± 0.02a |
-P | WT | 0.84 ± 0.30b | 0.97 ± 0.46b | 2.22 ± 1.05b | 1.28 ± 0.57b | 0.11 ± 0.01a | 0.15 ± 0.06b | 0.20 ± 0.10a | 0.28 ± 0.10a | 0.14 ± 0.02a | 0.07 ± 0.02b | 0.09 ± 0.01a | 0.13 ± 0.01a |
pap10c | 1.13 ± 0.35ab | 1.69 ± 0.24b | 2.27 ± 0.16b | 1.57 ± 0.44b | 0.12 ± 0.02a | 0.22 ± 0.04b | 0.18 ± 0.06a | 0.30 ± 0.11a | 0.11 ± 0.04a | 0.09 ± 0.02b | 0.06 ± 0.01a | 0.12 ± 0.03a |
OE | 1.32 ± 0.44a | 3.80 ± 0.88a | 3.63 ± 1.75a | 3.14 ± 1.37a | 0.13 ± 0.04a | 0.28 ± 0.07a | 0.22 ± 0.06a | 0.33 ± 0.08a | 0.16 ± 0.06a | 0.11 ± 0.04a | 0.09 ± 0.05a | 0.09 ± 0.02b |
Significance | | | | | | | | | | | | |
G | | *** | *** | ** | *** | ns | ns | * | ns | ns | * | ns | * |
P | | ns | ns | *** | ns | ns | ns | ** | * | ns | *** | ns | *** |
G×P | | ns | ns | * | ns | ns | ** | ns | ns | ns | * | * | *** |
Note: ns = not significant, * = significant at the 5% level, ** = significant at the 1% level, *** = significant at the 0.1% level. The data represents mean ± SD. Numbers with different letters differ significantly at the 5% level by Duncan’s significant difference. + P, applying P fertilizer; - P, not applying P fertilizer. |
Diversity of bacterial microbiomes across different plant developmental stages
The linear mixed model analysis suggested that plant developmental stages and fertilization significantly influenced bacterial Shannon, Chao1, and Richness indexes in the rhizosphere (Table S2). Although the genotypes alone did not significantly regulate the overall α-diversity, the interactive effects of genotype × developmental stages and genotype × fertilization × developmental stages were significant, indicating OsPAP10c may change the α-diversity of rhizosphere microbiomes in a developmental-dependent manner (Table S2). Therefore, we further analyzed the α-diversity of rhizosphere microbiomes along each of the four developmental stages. The results showed that P fertilizer decreased the α-diversity at tillering, elongation, and filling stages but increased α-diversity at ripening stages (Fig. 1). Moreover, the α-diversity was significantly increased and decreased in the rhizosphere of OsPAP10c overexpression plants under P fertilizer conditions at tillering and ripening stages, respectively (Fig. 1a, d). PERMANOVA analysis and NMDS ordinations on bacteria indicated that the plant developmental stages explained the most considerable variations in bacterial communities (Fig. S3 and Table S3). Although genotypes alone only explained approximately 2% of the variation in the bacterial community, it can explain a total of 14% of the variation when considering the cross effects (Table S3). At different developmental stages, the genotypes can explain 6.722–9.543% of the bacterial variation, suggesting that the expression of OsPAP10c significantly influenced rhizosphere bacteria composition at all development stages (Fig. 2 and Table S3). CAP analysis based on the P factors and Bray-Curtis dissimilarities of the bacterial community indicated that soil Olsen-P and total P level but not the Po level were significantly correlated with the bacterial community (Fig. 3). The bacterial community also showed a significant correlation with phytase, ALP, and ACP activities (Fig. 3).
Specific taxa of the rhizosphere soil microbiota associated with OsPAP10c in rice
To investigate the changes in microbial taxa by OsPAP10c, the relative abundance of rhizosphere soil microbiota was analyzed at the phylum level among different rice genotypes. Taxonomic classification showed that the dominant bacterial phyla enriched were consistent at different developmental stages, P fertilizer conditions, or genotypes (Fig. 4). Rhizobacteria composition under different genotypes showed significant changes in the abundance of some bacterial phyla under different developmental or P application conditions. For instance, the relative abundance of Actinobacteriota, Chloroflexi, and Proteobacteria, which account for nearly 50% of the total reads, was significantly changed by overexpression or mutation OsPAP10c under only -P fertilizer conditions at certain developmental stages (Fig. 4a, b, c, d). OsPAP10c overexpressing suppressed the relative abundance of Actinobacteriota in the rhizosphere soil at elongation, filling, and ripening stages under -P fertilizer conditions (Fig. 4b, c, d). In contrast, OsPAP10c overexpressing mainly changed the relative abundance of Bacteroidota and Firmicutes under + P fertilizer conditions (Fig. 4b, c, d). To further clarify the effect of genotypes on the composition of the rhizosphere soil microbial community, Linear discriminant analysis effect size (LEfSe) was conducted. At each indicated condition × developmental stage, 28 to 70 differentially abundant taxa, ranging from the phylum to genus, were detected with linear discriminant analysis (LDA) effect size (LEfSe) (LDA score > 3) (Fig. S4-S11). 5 to 35, 6 to 21 and 5 to 21 different taxa had greater relative abundances in the rhizosphere of WT, pap10c, and OE-PAP10c, respectively, across the four developmental stages under + P or -P fertilizer conditions (Fig. S4-S11). For example, at the elongation stage and + P fertilizer conditions, we found that the rhizosphere of WT enriched 26 taxa, such as Anaeromyxobacter, Roseiflexaceae, Comamonadaceae, Ciceribacter, Magnetospirillum, Pleomorphomonas. Seventeen taxa were enriched in the rhizosphere of pap10c, such as Gemmatimonadaceae, Defluviicoccus, Rokubacteriales, Lysobacter, Vicinamibacteraceae, and Methylomonas. 18 taxa were enriched in the rhizosphere of OE-PAP10c, such as SB_5_o_Bacteroidales, Stenotrophomonas, Xanthomonadaceae, Prolixibacteraceae, Clostridiaceae, Gracilibacteraceae, Bacteroidetes_vadinHA17. In addition, at the elongation stage and -P fertilizer conditions, 18 taxa were more abundant in the rhizosphere of WT, such as Solirubrobacterales, Roseiflexaceae, Solirubrobacteraceae, Nocardioidaceae, Caldilineaceae, Caldilineaceae, Defluviicoccus, Gaiella. 21 taxa were more abundant in the rhizosphere of OE-PAP10c, such as Christensenellaceae_R_7_group, Sporacetigenium, g_OPB41, Geobacter, WCHB1 (Prolixibacteraceae), Prolixibacteraceae, Peptostreptococcaceae, Methylobacter, Clostridiaceae, Sulfuritalea. Only 7 taxa were more abundant in the rhizosphere of OsPAP10c mutant than the other genotypes, including Bacteroidetes_vadinHA17, Methylomonadaceae, Rhizobiales, Rhodocyclaceae, and three from Bacteroidota. At the filling stage, 11 taxa were more abundant in the rhizosphere of WT, such as g_67_14_o_Solirubrobacterales, MB_A2_108 (Actinobacteriota), Nocardioidaceae, and Microtrichales. Fourteen were relatively enriched in OE-PAP10c, such as Citrifermentans, Subgroup_7 (Holophagae) Vicinamibacteraceae, Geobacteraceae, Vicinamibacteraceae, and Rokubacteriales. 21 taxa were more abundant in the rhizosphere of OsPAP10c mutant, such as Clostridium, SB_5 (Bacteroidales), OPB41 (Coriobacteriia), Intrasporangiaceae, SBR1031 (Anaerolineae), Micrococcales, Sporacetigenium, Propionibacteriales, KD4_96 (Chloroflexi).
The functional profiles of rhizosphere soil microbiomes
Tax4Fun2 prediction was applied to assess how genotypes, P fertilizer, and developmental stages affect the potential P cycling functions of the bacterial community, including P-starvation response regulation, P-uptake, and transport system, and inorganic P-solubilization and organic P-mineralization genes. Adonis2 analysis indicated that the developmental stages had the most potent effect on the relative abundance of the predicted genes involved in P cycling (P < 0.001) (Table S4). P fertilizer application only significantly affected the abundance of genes involved in P starvation response regulation (Table S4). P fertilizer suppressed the predicted relative abundance of genes that regulate phosphorus starvation response at the tillering stage in all the genotypes (Fig. S12). In contrast to P fertilizer treatments and developmental stages, the effects of genotypes alone were insignificant on the predictable composition. However, the genotype × fertilization × developmental stage significantly affected the total abundance of predicted genes mainly involved in the P-uptake and transport system (Table S4). Moreover, the genotype × developmental stage interaction also significantly affected the abundance of predicted genes involved in Pi solubilization and Po mineralization (Table S4). For the P-uptake and transport system genes, their predicted abundances were induced at tillering and elongation stages but suppressed at the ripening stage by P fertilizer application in the rhizospheres of WT plants (Fig. 5). These genes maintained more stability in both the rhizospheres of OsPAP10c mutation and overexpression plants irrespective of the P fertilizer treatments (Fig. 5). The relative abundance of Pi solubilization and Po mineralization genes had a higher relative abundance in the rhizosphere soils of pap10c than that of OE-PAP10c during elongation stage under -P fertilizer conditions (Fig. 6). It became much lower in both the rhizosphere soils of mutant and overexpression plants at the filling stage under -P fertilizer conditions (Fig. 6).
Heatmap for the individual P cycling genes showed that these genes were clustered into three groups (Fig. 7). The genes in group I had a relatively higher prediction level in the early developmental stages. In contrast, the genes in group III showed an opposite trend (Fig. 7). However, the P fertilizer treatments and genotypes have little effect on the group I and III genes. Interestingly, the relative abundance of group II genes was significantly changed by all the factors, including developmental stages, P fertilizer treatments, and genotypes (Fig. 7). This group mainly contained genes that were involved in Pi solubilization and Po mineralization, with fewer P uptake and transport system genes. For example, under P fertilizer conditions, the relative abundances of some specific genes (such as phnH, phnN, phnI, etc.) of the microbial community for pap10c were much lower than WT and OE-PAP10c in the elongation stage but much higher at the ripening stage. In contrast, the predicted relative abundance of phnH, phnN, phnI, pstB, etc., in the rhizosphere of pap10c mutant, was significantly higher than WT and overexpression plants under -P fertilizer conditions at the elongation stage (Fig. 7).