Increased organic fertilizer application and reduced chemical fertilizer application affect the soil properties and bacterial communities of grape rhizosphere soil

Increasing organic fertilizer application can improve the sustainability of soil productivity, but the effects of increased organic fertilizer application with reduced chemical fertilizer application over different time periods on chemical properties and bacterial community of grape rhizosphere soil in an arid region are not clear. In this study, three years of fixed-point field tests were used to compare the effects of various fertilization treatments on the soil properties and bacterial community in the grape rhizosphere. The results showed that (1) T1 and T2 significantly increased SOM, AN, AP and AK contents in grape rhizosphere soil. TN, TP and TK contents in grape leaves of T2 were the highest of those in five fertilization treatments. (2) The abundances of Proteobacteria and Bacteroidetes phyla and especially of Arthrobacter, Pseudomonas, Nitrosopira and Bacillus genera were higher in T2 than in the other samples. (3) SOM, AP and AN contents in soil were the main factors affecting soil bacterial community and mineral element contents in grape leaves and roots according to an RDA analysis. In summary, the application of organic fertilizer with reduced chemical fertilizer for two years had the greatest impact on the soil properties and bacterial community of the grape rhizosphere soil.


Results
Effects of different fertilizer treatments on grape rhizosphere soil propertie. Table 1 shows that the SOM, AN, AP and AK contents in the rhizosphere soil were the highest in T2 and the lowest in CK of the five different fertilizer treatments on day 15 after anthesis. The SOM and AK contents in the soils treated with increased organic fertilizer and reduced chemical fertilizer (T1, T2, and T3) were higher than those of CK and T0. The content of AN in the rhizosphere soil of T0 was significantly lower than that of T2 (P < 0.05); however, there was no significant difference in AN content between T3 and T0 (P > 0.05). The AP content in the rhizosphere soil of T0 was significantly lower than that of T2 and T3 by 5.86% and 2.79%, respectively (P < 0.05), but the difference in AP content between T0 and T1 was not significant (P > 0.05). The pH in the rhizosphere soils treated with organic fertilizer and chemical fertilizer (T1, T2, and T3) were significantly lower than that in CK and T0 (P < 0.05). The soil conductivity in T2 and T3 were significantly lower than that in CK and T0 (P < 0.05), and soil conductivity in T2 decreased by 31.10% and 24.24% compared with CK and T0, respectively. There was no significant difference in soil conductivity between T1 and T0 (P > 0.05) ( Table 1). The SOM, AN, AP and AK contents in the rhizosphere soil were the highest in M-T1 and the lowest in M-CK of the five different fertilizer treatments on day 75 after anthesis. The SOM, AN and AP contents in M-T1 soil were 8.40%, 5.67% and 5.82% higher than those in M-T2 soil, respectively. The soil pH in M-T2 decreased by 1.73% and 1.48% compared with that in M-CK and M-T0. Soil pH in M-T1 was lower than that in M-CK and M-T0 (P > 0.05). The soil conductivity in M-T2 Effects of different fertilizer treatments on the total N, total P and total K of grape roots and leaves. On day 15 after anthesis, the TN, TP and TK contents of grape roots were the highest of the five different fertilizer treatments in T2 and the lowest in CK ( Table 2). The content of TN in the grape roots of T2 decreased by 73.70% and 10.50% compared with that in CK and T0, respectively. The content of TP in the grape roots of T2 was significantly higher than that of CK and T0 (P < 0.05), and the TP content of T0 was higher than that of T1 and T3 (P > 0.05). The TK content in the grape roots of T2 increased by 145.83% and 37.21% compared with that of CK and T0, and the TK content of T0 was significantly lower than that of T1 and T3 (P < 0.05). The contents of TN, TP and TK in grape leaves were the highest of five treatments in T2 and the lowest in CK. The contents of TN, TP and TK in grape leaves of T2 were significantly higher than those of CK and T0 (P < 0.05). The contents of TN and TP in the grape leaves of T0 were significantly higher than those of T1 and T3 (P < 0.05). The TK content in the grape leaves of T0 was significantly higher than that of T1 and T3 (P < 0.05). On day 75 after anthesis, the TN, TP and TK contents in the grape roots of M-CK were the lowest in all treatments, whereas the TN content in grape roots treated with increased organic fertilizer and reduced chemical fertilizer (M-T1, M-T2, and M-T3) were higher than that of M-T0 (P > 0.05), and the difference between M-T0 and M-T3 was not significant (P > 0.05). The TP content in the grape roots of M-T2 was significantly higher than that of M-CK and M-T0 by 460.00% and 16.00%, respectively. The TP content of M-T1 was significantly higher than that of M-T0 (P < 0.05), but there was no significant difference in TP content between M-T3 and M-T0 (P > 0.05). The TK contents in grape roots treated with increased organic fertilizer and reduced chemical fertilizer (M-T1, M-T2, and M-T3) were significantly higher than those of M-CK and M-T0 (P < 0.05), and there was no significant difference in TK content among M-T1, M-T2 and M-T3. The contents of TN, TP and TK in grape leaves were the highest of five different fertilizer treatments in M-T2 and the lowest in M-CK. The TN content in the grape leaves of M-T2 was significantly higher than those of M-CK and M-T0 (P < 0.05). The TP contents in grape leaves of M-T2 and M-T3 were significantly higher than those of M-CK and M-T0 (P < 0.05), and the difference in leaves TP content between M-T1 and M-T0 was not significant (P > 0.05). The TK content in the grape leaves of M-T2 was significantly higher than those of M-CK and M-T0 by 29.32% and 10.53%, respectively (P < 0.05) ( Table 2).
Effects of different fertilizer treatments on bacterial richness and diversity in grape rhizosphere soil. On day 15 after anthesis, the soil bacterial richness (estimated by Chao1) of the increased organic fertilizer and reduced chemical fertilizer treatments (T1, T2, and T3) were significantly greater than that of CK and T0 (P < 0.05), but there was no significant difference in bacterial richness among T1, T2 and T3 (P > 0.05). The Chao1 of T1 was higher than that of CK and T0 by 19.26% and 15.27%, respectively. On day 75 after anthesis, the Chao1 of M-T2 was significantly (22.67% and 24.90%) higher than that of M-CK and M-T0 (P < 0.05) (Fig. 1A). The soil bacterial diversity (estimated by the Shannon index) of T2 and T3 were significantly higher than that of CK, by 0.59% and 0.78%, respectively. The Shannon index of T0 was significantly higher than that of T1. The Shannon index of M-T2 was significantly higher than that of M-CK and M-T0, by 1.69% and 1.78%, respectively (P < 0.05) (Fig. 1B).
Effects of different fertilizer treatments on bacterial communities at the phylum level in grape rhizosphere soil. Phyla with an average abundance greater than 1% in at least one soil sample were defined as dominant and are assessed in the following analysis. The soil samples had similar bacterial compositions, but some variations were observed in the distribution of each phylum. Proteobacteria, Actinobacteria, Acidobacteria, Gemmatimonadetes and Chloroflexi were the dominant phyla, demonstrating >10% relative abundance in at least one rhizosphere soil sample. Another five predominant phyla were Planctomycetes, Bacteroidetes,  www.nature.com/scientificreports www.nature.com/scientificreports/ Nitrospirae, Verrucomicrobia and Latescibacteria on days 15 and 75 after anthesis (Fig. 2). The relative abundance of Chloroflexi in T0 was significantly lower than that of CK on days 15 and 75 after anthesis (P < 0.05), and the relative abundance of Chloroflexi in soils treated with increased organic fertilizer and reduced chemical fertilizer (T1, T2, and T3) was significantly lower than that in CK and T0 (P < 0.05) ( Table 3). On day 15 after anthesis, the relative abundances of Gemmatimonadetes and Bacteroidetes were the highest in T2 and the lowest in CK in the five different fertilizer treatments. The relative abundances of Gemmatimonadetes in T2 and T3 were significantly higher than that in CK by 51.11% and 6.83%, respectively. The relative abundance of Bacteroidetes in T2 was significantly higher than that in CK, and the relative abundance of Bacteroidetes in T0 was higher than that in T1 and lower than that in T3 (P < 0.05).
On day 75 after anthesis, the relative abundance of Proteobacteria was highest in M-T2 and lowest in M-CK of the five different fertilizer treatments. The relative abundance of Actinobacteria was the lowest in M-T2 and the highest in M-CK of the five different fertilizer treatments. The relative abundance of Proteobacteria in M-T2 and M-T3 were significantly higher than that in M-CK (P < 0.05), and the relative abundance of Proteobacteria in    www.nature.com/scientificreports www.nature.com/scientificreports/ M-T0 was lower than that in M-T3 (P > 0.05). The relative abundance of Actinobacteria in M-T2 was significantly lower than that in M-CK by 30.81% (P < 0.05) ( Table 3).
Effects of different fertilizer treatments on N and P metabolism-related bacteria in grape rhizosphere soil at the genus level. On day 15 after anthesis, the relative abundance of Nitrospira in T2 was the highest of the five different fertilizer treatments and that in CK was the lowest. The abundance of Nitrospira in T2 was significantly higher than that in CK and T0 by 91.80% and 112.65%, respectively (P < 0.05) ( Table 4). The relative abundances of Pseudomonas and Arthrobacter in T2 were the highest of the five different fertilizer treatments and that in CK was the lowest. The relative abundances of Pseudomonas in T2 increased by 277.84% and 140.83% compared with that in CK and T0, respectively. The relative abundance of Arthrobacter in T2 and T3 was significantly higher than that in CK and T0 (P < 0.05). The relative abundance of Bacillus in T2 increased by 105.74% and 74.96% compared with that in CK and T0 (P < 0.05) ( Table 4).
On day 75 after anthesis, the relative abundance of Nitrospira in M-T2 was the highest of the five different fertilizer treatments and that in M-CK was the lowest. The abundance of Nitrospira in M-T2 was significantly higher than that in M-CK and M-T0 by 25.08% and 44.93%, respectively (P < 0.05) ( Table 4). The relative abundances of Pseudomonas and Arthrobacter in M-T2 were the highest of the five different fertilizer treatments, and those in M-CK were the lowest. The relative abundance of Pseudomonas in M-T2 increased by 105.05% and 124.20% compared with that in M-CK and M-T0 (P < 0.05). The relative abundances of Arthrobacter in M-T2 and M-T3 were significantly higher than that in M-CK (P < 0.05). The relative abundance of Arthrobacter in the M-T2 was significantly higher than that in M-CK and M-T0, by 110.90% and 52.10%, respectively (P < 0.05) ( Table 4).
Relationship between grape rhizosphere soil properties, the bacterial community in grape rhizosphere soil and mineral element contents of grape roots and leaves. The associations between grape rhizosphere soil properties and the mineral element contents of grape leaves and roots were analysed by RDA. pH and conductivity were negatively correlated with TN, TP, TK, TN.R, TP.R and TK.R on days 15 and 75 after anthesis. SOM, AN, AP and AK were positively correlated with TN, TP, TK, TN.R, TP.R and TK.R on days 15 and 75 after anthesis (Fig. 3A,C). On day 15 after anthesis, axis 1 and axis 2 explained 95.15% and 3.99% of the total variation, respectively (Fig. 3A). The mineral element contents of grape leaves and roots were significantly associated with SOM, AN and AP (P < 0.05). On day 75 after anthesis, axis 1 and axis 2 respectively explained 92.71% and 6.25% of the total variation (Fig. 3C). The mineral element contents of grape leaves and roots were significantly associated with SOM and AN (P < 0.05). Figure 3B,D show that the relationships between the grape rhizosphere soil properties and the microbial community composition in the grape rhizosphere soil and grape rhizosphere soil properties were considered environmental variables. The Chao1 and Shannon index values and the abundances of Arthrobacter, Pseudomonas, Nitrosopira and Bacillus were negatively correlated with pH and conductivity on days 15 and 75 after anthesis (Fig. 3B,D). The Chao1 and Shannon index values and the abundances of Arthrobacter, Pseudomonas, Nitrosopira and Bacillus were positively correlated with SOM, AN, AP and AK. The microbial community composition was significantly correlated with SOM and AP (P < 0.05).

Discussion
In this study, the soil pH was significantly decreased by the increased organic fertilizer and reduced chemical fertilizer treatments compared to that in the typical chemical fertilizer and unfertilized treatments (Table 1). Intriguingly, increasing the amount of organic fertilizer and reducing the amount of chemical fertilizer did not stabilize or increase the soil pH but significantly decreased the soil pH, which disagreed with previous findings 26,27 . This discrepancy might be attributed to differences in the soil type, as proposed by Wei et al. 13 , who observed that organic fertilizer decreased the soil pH in alkaline soils but increased the soil pH in acidic soils. Our fertilization treatments also resulted in distinct soil compositions. The five different fertilization treatments showed significant differences in SOM content ( Table 1). The SOM content in T2 was the highest of the five fertilizer treatments on day 15 after anthesis and that in T1 was the highest of the five fertilizer treatments on day 75 www.nature.com/scientificreports www.nature.com/scientificreports/ after anthesis. This was in line with previous studies in which the application of organic fertilizer increased the soil organic carbon content 28 . Some reports have shown that organic matter treatments with high levels of C directly increased the soil organic C content 29 . The AN content in the T0, T1, T2 and T3 fertilization treatments was higher than that in CK; AP and AK showed the same trends. Another previous study showed that the long-term combined application of chemical fertilizer and organic fertilizer significantly increased the contents of TN, AN and AP in the soil 30 . Garcia-Orenes et al. found that the application of organic fertilizer increased the contents of SOM, AK and AP 31 . In this experiment, we found that on days 15 and 75 after anthesis, the increased organic fertilizer with reduced chemical fertilizer treatments for one and two years led to the contents of SOM, AK and AP being significantly higher than those of unfertilized and typical chemical fertilization.
Zhao's research showed that the combination of chemical fertilizer and organic fertilizer significantly increased the contents of N, P and K in apple leaves 32 . Tao et al. found that the combination of chemical and organic fertilizers increased cotton plant N and P uptake and fertilizer utilization efficiency 33 . The contents of TN, TP and TK in the grape leaves and roots treated with chemical and organic fertilizers were significantly higher than those in unfertilized leaves and roots, and T2 increased those parameters the most significantly (Table 2). This indicated that applying organic fertilizer and reducing chemical fertilizer application could improve nutrient absorption in grapes, and the effect of applying organic fertilizer and reducing chemical fertilizer for two years was greatest. This may be due to the organic fertilizers improving soil aeration and water retention in the graperhizosphere, which promotes the absorption of N, K and P by the grape roots.
Increasing or decreasing soil nutrients will change the soil bacterial abundance and diversity 34 . Yuan et al. found that organic fertilizer application significantly increased soil bacterial diversity and viability 35 . Sun et al. also showed that bacterial abundance was increased by the application of chemical fertilizer alone and the combined application of chemical and organic fertilizers 34 . In this study, the bacterial diversity under increased organic fertilizer and reduced chemical fertilizer for two years was significantly higher than that in the unfertilized and typical chemical fertilizer treatments. Most of the dominant bacteria in the soil were basically the same, including approximately 10 bacterial phyla 36 . Proteobacteria, Actinobacteria, Acidobacteria, Gemmatimonadetes and Chloroflexi were the five most abundant bacterial phyla in the soil samples on days 15 and 75 after anthesis (Fig. 2). Nitrification is the driving force of soil nitrogen cycling 4 . Nitrification, which is the oxidation of ammonia (NH 3 ) to nitrite (NO 2 − ), and then to nitrate (NO 3 − ), is a key step within the N cycle. In agricultural systems, nitrification is associated with potential N losses from soils, either through the loss of NO 3 − by leaching and/or through denitrification 4 . Nitrospira is an ammonia-oxidizing bacteria that contributes to soil Figure 3. Redundancy analyses depicting the relationships among grape rhizosphere soil properties, the bacterial community in the grape rhizosphere soil and mineral element contents of grape roots and leaves. The relationships between the mineral element contents of grape roots and leaves and the grape rhizosphere soil properties (A,C). The relationships between the bacterial community in the grape rhizosphere soil and the grape rhizosphere soil properties (B,D). TN.R: total N in grape roots, TP.R: total P in grape roots, TK.R: total K in grape roots, TN: total N in grape leaves, Con: conductivity, Nir: Nitrosopira, Arthro: Arthrobacter. (2020) 10:9568 | https://doi.org/10.1038/s41598-020-66648-9 www.nature.com/scientificreports www.nature.com/scientificreports/ nitrification 37 . Phosphorus is an indispensable nutrient for plant growth 22 . Unavailable forms of phosphorus in soil account for 95% of the total soil phosphorus, and plants cannot absorb and utilize these forms directly 22 . Phosphorus-solubilizing bacteria can transform insoluble phosphorus, which is difficult for plants to absorb and utilize, into available forms and improve the soil phosphorus utilization efficiency. Several studies have demonstrated that members of the Pseudomonas and Bacillus genera were the dominant phosphate-solubilizing bacteria (PSB) in agricultural soils. Kumar and Singh reported that the inoculation of vermicompost with the phosphate solubilizing Pseudomonas striata significantly increased the available P 38 . Bacterial strains belonging to the genera Bacillus and Pseudomonas have been isolated from the rhizosphere of various plants and are reported to improve plant growth 39 . In our study, the relative abundance of Nitrospira in the increased organic fertilizer and reduced chemical fertilizer for two years treatment (T2) was significantly higher than that in the unfertilized or typical chemical fertilizer treatments (CK and T0) ( Table 4), similar to Bacillus and Pseudomonas. This indicated that grape rhizosphere soil treated with increased organic fertilizer and reduced chemical fertilizer for two years (T2) contained more available N and available P than the rhizosphere soil of the other treatments, which is beneficial for plant growth.
In our study, the RDA analysis showed that SOM and AP were the main environmental factors affecting the soil bacterial community structure in grape rhizosphere soil (Fig. 3). The microbial community composition was significantly correlated with SOM and AP in Xu's study 3 . Nie also found that the fungal community composition was significantly correlated with SOM 40 . Pan's study found that the bacterial community structure was significantly related to soil properties, such as SOM and pH 41 . Gu et al. found that the available N content in soil was positively correlated with the total N content in Baren Apricot leaves 42 . We found that the mineral element contents of the grape leaves and roots were significantly positively correlated with SOM, AN and AP in soil (P < 0.05). In summary, fertilization changed the chemical properties of the grape rhizosphere soil, which in turn affected the soil bacterial community structure, and ultimately affected the mineral element contents of the grape leaves and roots.

conclusions
In conclusion, this study showed that the different fertilization treatments resulted in different soil properties, bacterial diversity conditions and TN, TP and TK contents in grape roots and leaves. The T2 treatment had the highest AN, AP and SOM contents of all the treatments on day 15 after anthesis. The contents of TN, TP and TK in grape leaves and roots treated with increased organic fertilizer and reduced chemical fertilizer for two years (T2) were the highest of the five fertilizer treatments on days 15 and 75 after anthesis. The fertilizer treatments changed the abundance and diversity of bacteria and the relative abundances of Nitrospira, Pseudomonas, Arthrobacter and Bacillus in the rhizosphere soil of 'Summer Black' grapes. The relative abundances of Nitrospira, Pseudomonas, Arthrobacter and Bacillus were most significantly increased by increased organic fertilizer and reduced chemical fertilizer for two years. The SOM, AP and AN contents in soil were the main factors affecting the soil bacterial community and the mineral element content in grape leaves and roots according to the RDA analysis.

Materials and Methods
Experimental site and soil description. The field plot experiment was conducted at the Shihezi Agricultural Science Research Institute of Xinjiang Uygur Autonomous Region for three consecutive years from 2016 to 2018 (45°19′N, 86°03′E). The average annual temperature was between 6.5 and 7.2 °C, the annual precipitation was 225 mm, and the annual evaporation was 1342 mm. The soil was a sandy loam soil. The basic chemical properties of the 0.20-0.40 m soil were as follows: pH, 8.02; soil organic matter (SOM), 31.23 g/kg; total nitrogen, 135.32 mg/kg; alkali-hydrolysed nitrogen (AN), 51.11 mg/kg; total phosphorus (TP), 89.93 mg/kg; available phosphorus (AP), 32.95 mg/kg; total potassium (TK), 369.56 mg/kg; available potassium (AK), 130.35 mg/kg; and conductivity, 0.221 us/cm. Experimental design. The experiment used 7-year-old 'Summer Black' table grape plants as the research material. The grapes are planted in a north-south orientation, with a line spacing of 3 m and a plant spacing of 0.7 m. The grape trellis is a V-shaped scaffold, placed between the vines. The study used a randomized block design with three replicates of five treatments. Fifteen plots were used, and there were 7 well-grown, healthy grapes plants in each plot. The five treatments were (1) no fertilization for three consecutive years (CK), (2) typical chemical fertilization for three consecutive years (T0), (3) increased organic fertilizer and reduced chemical fertilizer for one year (T1), (4) increased organic fertilizer and reduced chemical fertilizer for two years (T2) and (5) increased organic fertilizer and reduced chemical fertilizer for three years (T3). The nutrient content of the fermented and decomposed cow manure was 2.48% (total nitrogen), 1.79% (total phosphorus) and 2.50% (total potassium). Organic fertilizer (cow manure) was applied in spring at depths of 25 cm and 30 cm from the main trunk of the grape plant(unilateral fertilization). The chemical fertilizers (drip application) used were urea (N: 46%), monoammonium phosphate (N: 12%, P 2 O 5 : 60%), and potassium sulfate (K 2 O: 50%) applied during the growing season. The chemical fertilizer was applied with water-fertilizer integrated device with a flow rate of 2.6 L/h and a working pressure of 0.05-0.1 MPa. The drip irrigation belt was located at 30 cm to the north and south sides of the grapes, just above the area where the organic fertilizer was applied. During the growing season, the fertilizer application times and field management practices for the five treatments were the same as those for conventional grapes 43 . The specific fertilizer dosage and application times of each treatment are shown in Table 5.
Soil and plant sampling. The rhizosphere soil samples (20-40 cm underground) were collected with an auger on days 15 and 75 after anthesis (pre-swelling and mature stage). Soil samples were collected using the five-point sampling method. Five soil cores were selected from each plot, and then a composite sample was  Table 5. Specific fertilizer application amounts for different treatments and years.