Rice Physiological Response with Bacillus subtilis and Saccharomyces cerevisiae Inoculation into Soil under Reclaimed Water–Fresh Water Combined Irrigation

The increasing soil salinity levels under reclaimed water irrigation have a negative effect on plant growth. Greenhouse experiments were conducted in 2018 and 2019 under reclaimed water–fresh water combined irrigation. After transplanting (Day 1), rice was irrigated with clean water (tap water) for 10 days to facilitate rice root colonisation. Subsequently, rice was irrigated with reclaimed water for 50 days (Day 11 to 60), and then irrigated with clean water. B. subtilis and S. cerevisiae were mixed with clean water (tap water) and irrigated into soil at Day 61. B. subtilis (20 billion colony-forming units/g) and S. cerevisiae (20 billion colony-forming units/g) were mixed at the following proportions: 5 g and 0 (J1), 3.75 g and 1.25 g (J2), 2.5 g and 2.5 g (J3), 1.25 g and 3.75 g (J4), and 0 and 5 g (J5), respectively; rice treated with reclaimed water (CK) and clean water (J0) with no microorganisms applied were also used. We measured NO3--N and NH4+-N concentrations and electrical conductivity (EC) in the soil at 0–5, 5–15, and 15–25 cm layers; root activity; and malondialdehyde (MDA), soluble sugar, superoxide dismutase (SOD), peroxidase (POD), catalase (CAT), and glutamine synthetase (GS) activity in leaves at Day 71. B. subtilis and S. cerevisiae combination could promote rice physiological indices, and B. subtilis had a greater effect than S. cerevisiae. There are obvious differences in the physiological performance and soil N between 2018 and 2019 due to the EC of reclaimed water. Redundancy analysis revealed that soil NO3−-N and the mass of B. subtilis applied were major factors influencing leaf physiological indices. Five grams of B. subtilis is recommended to facilitate rice growth after irrigation with reclaimed water. Our research provides a new agronomic measure for the safe utilisation of reclaimed water.


Introduction
In China, the exploitation of reclaimed water in agriculture is in its infancy, and extensive studies on its safe utilisation in agricultural activities, in addition to technological developments, are required. There have been breakthroughs in the fields of drip irrigation [1] and vegetable production [2] and flower growth [3] using reclaimed water, which promotes the growth of microbes involved in soil carbon (C) and nitrogen (N) transformation, and effects numerous changes in soil microbial communities [2], in addition to improving soil microbial biomass [4].
Reclaimed water contains sodium, calcium and hydrogen carbonate ions, in addition to N, P and potassium nutrients, among others, although it also poses soil salinisation risks [5] and could deteriorate soil physical and chemical properties [6]. However, reclaimed water is not particularly safe, considering potential risks to agriculture, the environment and human health [7]. Furthermore, water reclaimed from municipal treatment plants could have higher levels of electrical conductivity (EC) and higher sodium adsorption ratio when compared to other freshwater resources [8].
Soil microbes improve soil physical and chemical processes and influence soil nutrient concentrations, soil aeration and soil structure. Plant growth-promoting bacteria facilitate plant development by enhancing plant growth or influencing plant metabolism directly through their metabolic activities [9]. Bacillus spp. are capable of forming long-lived, stresstolerant spores and secreting metabolites that stimulate plant growth and prevent pathogen infection. Bacillus spp. also secrete exopolysaccharides and siderophores that inhibit the movement of toxic ions, facilitate ionic balance maintenance and water translocation in plant tissues, and inhibit pathogenic microbe growth [10]. Bacillus spp. are considered safe bacteria that produce substances that are beneficial for crops and that could be used in the production of industrial compounds [11]. B. subtilis GB03 regulates cell expansion and auxin synthesis in Arabidopsis [12], and B. subtilis FTC01 has been demonstrated to be a rich source of xylanolytic enzymes [13]. Yeast has been reported to be a rich source of phytohormones, vitamins, enzymes, amino acids and minerals [14]. Rhizosphere yeast promotes P nutrition in arbuscular-mycorrhizal maize, which is associated with enhanced specific root length and arbuscular mycorrhizal fungi (AMF) [15]. Soil yeast can promote rice growth [16]. Mohamed (2005) [17] observed that active foliar application of dry yeast improved plant growth and yield, and increased N concentrations in grain.
The positive effect of single microorganism has attracted much attention, but whether the combined application of two microorganisms can play the positive role of both has not been determined. Yeast (Candida sake, Cryptococcus aerius and Williopsis californica) and AMF (Glomus intraradices BEG140 and Glomus mosseae BEG95) resulted in increased shoot biomass [18] and improved crop resistance to environmental stress while decreasing malondialdehyde (MDA) [19]. A mixture of AM fungus and B. subtilis was previously found to increase the yield of geranium [20]. Inoculation of Chloris guyana with the S. cerevisiae alone did not affect plant growth [21]. It can be inferred that the mixed inoculation of the two microorganisms has the best ratio or might inhibit plant growth.
Rice cultivation requires a lot of water; consequently, the use of reclaimed water for irrigation could minimises shortages in fresh water resources, while reclaimed water contains high salt concentrations, which would increase soil salinity [22] and, in turn, salt stress and physiological disorders in plants [23]. Rice crops exhibit considerable genotypic variations in response to salt stress, with higher sensitivity during early seedling stages and reproductive stages [24]. The use of reclaimed water in rice irrigation would require investigation of appropriate agronomic strategies of alleviating the potential salt stress in the soil. B. subtilis facilitates the maintenance of ionic balance and secretes metabolites that stimulate plant growth, while S. cerevisiae improves soil nutrient supply. The combination of the two species does not enhance root activity and enzyme activity in the leaves, and there is a certain proportion of the two microorganisms that can improve the yield of rice.
Ammonium N (NH 4 + -N) and nitrate N (NO 3 − -N) are the two major types of N fertilisers that are absorbed and utilised by crops. Under salt and drought stress, free radical accumulation in plant cells is too high; superoxide dismutase (SOD), peroxidase (POD) and catalase (CAT) activities are high; and higher MDA concentrations are produced, which impair plant growth and development. Most of the soluble proteins in plants are enzymes involved in various metabolic processes, and the processes are closely linked to the plant enzymes. Soluble protein concentrations are critical indices that facilitate the understanding of the total metabolism activities of plants. Chlorophyll is the key pigment involved in plant photosynthesis, and chlorophyll concentrations can reflect N supply in the soil. Plants generate a greater concentration of soluble sugar to facilitate salt stress tolerance. In addition, root activity influences soil nutrient concentrations and water absorption by plants, which further influences plant growth. Soluble sugars provide energy and metabolic intermediates for plant growth.
Considering that the difference in EC in reclaimed water could change the effect of microorganisms on rice, pot experiments using rice as test material were conducted in 2018 and 2019, and the EC of reclaimed water is 1411 and 1723 µS·cm −1 , respectively. After 50 days of irrigation with reclaimed water, different proportions of B. subtilis and S. cerevisiae were inoculated into the soil. The soil N and EC, root activity, enzyme activity, chlorophyll and malondialdehyde (MDA) of rice leaves were investigated. The objectives were (i) to reveal the positive physiological effects of B. subtilis and S. cerevisiae on rice and (ii) determine the role of soil EC and N in rice growth under reclaimed water-fresh water combined irrigation.

Site Descriptions
The experiments were carried out in the greenhouse at the Agricultural Soil and Water Environment Field Scientific Observation and Experiment Station of the Chinese Academy of Agricultural Sciences. The test site is located in Xinxiang City, Henan Province, where the annual average temperature is 14.1 • C, the frost-free period is 210 days, the number total sunshine hours is 2398.8 h, the annual average precipitation is 589 mm, and the annual average evaporation is 2000 mm.

Experimental Materials
The experimental rice variety was "Wugeng 519". The plastic buckets used for the tests had bottom diameters of 20.5 cm, upper diameters of 25 cm and depths of 28.5 cm. The test soil was a sandy loam, which was obtained from a wheat field near the test station. The saturated moisture content (by mass) of the soil was 38.92%, and the other soil physical properties are listed in Table 1. B. subtilis and S. cerevisiae were cultured by Shandong Sukehan Bioengineering Co., Ltd. (Weifang, China) at a 20 billion CFU/g concentrations. The reclaimed water, which comes from the sewage treatment plant, was stored in the refrigerator at 4 • C, and the water quality properties are listed in Table 2.

Experimental Designs
The experiments were carried out from May 2018 to October 2018 and from May 2019 to October 2019. The experimental design used reclaimed water and fresh water (tap water). The soil moisture was determined by weighing the pots using an electronic scale (20 kg) daily at 8:00 am. Each pot contained 11 kg dry soil in two years, and the urea, potassium sulphate and potassium dihydrogen phosphate concentrations were 2.5 g, 1.0 g and 3.0 g in each pot, respectively.
The day of transplanting was considered the first day of the rice growth period, marked as S1. In 2018, the seedbed was prepared and soaked on 3 May, seeded on 5 May, loaded on 9 June, soaked on 12 June, transplanted on 14 June (S1, rice growth stage day 1) and harvested on 18 October (S127, rice growth stage day 127). In 2019, the seedbed was prepared and soaked on 17 April, and seeded on 29 April. Thereafter, the soil was loaded on 26 May, the was fertiliser applied, water was added to soak the soil on 31 May and the seedlings were transplanted on 2 June (S1). The plants were harvested on 8 October (S129, rice growth stage day 129).
Five B. subtilis and S. cerevisiae combination treatments (wt/wt) were established (5 g and 0, 3.75 g and 1.25 g, 2.5 g and 2.5 g, 1.25 g and 3.75 g, and 0 and 5 g, and referred to as J1, J2, J3, J4 and J5, respectively) ( Table 3). Soaking the soil with tap water Soaking the soil with tap water 14 June (S1) Transplanting seedlings Transplanting seedlings S1-S10 Irrigation with tap water (FI) Irrigation with tap water (FI) S11-S60 Irrigation with reclaimed water (CI) Irrigation with reclaimed water (CI) 13 August (S61) No microorganisms Inoculating microorganisms S61-S127 Irrigation with reclaimed water (CI) Irrigation with tap water (CI) 18 October (S127) Harvest Harvest 2019 26 May Loading the bucket with dry soil Loading the bucket with dry soil 31 May Soaking the soil with tap water Soaking the soil with tap water 2 June (S1) Transplanting seedlings Transplanting seedlings S1-S10 Irrigation with tap water (FI) Irrigation with tap water (FI) S11-S60 Irrigation with reclaimed water (CI) Irrigation with reclaimed water (CI) 1 August (S61) No microorganisms Inoculating microorganisms S61-S129 Irrigation with reclaimed water (CI) Irrigation with tap water (CI) 8 October (S129) Harvest Harvest S1-S127 (S129), represent rice growth stages 1 to 127 in 2018 (129 in 2019) starting from the day of transplantation to harvest. At 60 days after transplanting, Bacillus subtilis (BS) and Saccharomyces cerevisiae (SC) were mixed in water proportionally and then irrigated into the soil. FI is flooded irrigation with 0-5 cm water depth; CI is controlled irrigation: no water layer will be established in other growth periods except for the 0-50 mm water layer after transplanting for 10 days; the upper limit of soil water control in the root layer is the saturated water content, and the lower limit is 60-80% of the saturated water content.
After irrigation with reclaimed water for 50 days (S11-S60), B. subtilis and S. cerevisiae were mixed in tap water (200 mL) and the water used to irrigate the soil. Tap water was used to irrigate J1, J2, J3, J4 and J5 treatments from S61 to harvest. Treatments with no B. subtilis and S. cerevisiae application in the case of the reclaimed water (control, CK) and tap water (J0) were also established. Each treatment had 3 replicates.

Test Indexes and Analysis Methods
The chlorophyll and MDA concentrations as well as superoxide dismutase (SOD), peroxidase (POD), catalase (CAT), glutamine synthetase (GS) activity, soluble sugar, soluble protein and root activity were measured at S71 in 2018 and 2019. For chlorophyll, 0.2 g of weighted fresh samples of leaves was soaked in 20 mL 95% ethanol until the colour of leaves disappeared completely, the extracting solution was measured at 470 nm, 649 nm and 665 nm. MDA, SOD, POD, CAT and GS activities were determined using a kit manufactured by Nanjing Jiancheng Bioengineering Institute (http://www.njjcbio.com/). Phosphate buffer (PBS, pH = 7.4) was used as homogenising medium, and the reaction liquid was measured with spectrophotometer. For MDA, 0.5 g of weighted fresh samples of leaves was milled with the help of a motor, 2 mL 10% TCA and a small amount of quartz sand were added and ground to homogenate, 3 mL TCA was added and then further ground. The homogenised sample was centrifuged at 12,000 rpm for 10 min. To 2 mL supernatant, 0.67% TBA was added, and then the solution was mixed and boiled for 15 min in a 100 • C water bath. The sample was cooled at room temperature and centrifuged again. The absorption values of samples were measured at 532 nm, 600 nm and 450 nm, respectively [25]. For SOD, POD, CAT and GS, 0.5 g of fresh samples of leaves was milled with the help of a motor and pestle and standardised in 0.05 mol/l PBS under chilled condition; the standardised mixture was centrifuged at 12,000 rpm for 10 min at 4 • C after sieving through four layers of muslin cloth.
Soluble sugar concentrations were measured using anthrone colorimetry following the method by Jing et al. [26]. Weighted fresh leaves (0.2 g) were cut into pieces and put into a triangular flask, then 20 mL distilled water was added and the triangular flask was sealed with plastic film and extracted in boiling water for 30 min. The samples were cooled, filtered and then the volume was fixed to 100 mL for testing. We then transferred 0.5 mL of the solution to be tested into a 20 mL test tube, added 1.5 mL of distilled water, 0.5 mL of anthrone ethyl acetate 0.5 mL and 5 mL of concentrated sulphuric acid, fully vibrated and immediately put the test tube into boiling water, kept it warm for 1 min, took it out and cooled it to room temperature and then the samples were measured at 630 nm.
Soluble protein concentrations were measured using the Coomassie Brilliant Blue method based on the method by Bradford [27] using the standardised mixture for determination of enzyme activity in leaves. Root activity (Dehydrogenase activity) was measured using 2,3,5-triphenyltetrazolium chloride (TTC) based on the method by Bon, et al. [28]. Weighed fresh root tip samples (0.2 g) were put it into a 25 mL beaker, we then added 5 mL of 0.4% TTC solution and 5 mL of phosphate buffer (pH = 7.0), kept it in the dark at 37 • C for 1-3 h and then added 2 mL of 1 mol/L sulphuric acid. The sample was extracted with ethyl acetate and measured at 485 nm.
Three soil layers (0-5, 5-15 and 15-25 cm) were obtained at S71, the soil electricity conductivity (EC) of dry soil was measured using Lei-ci DDB-303A (INESA Scientific Instrument Co., Ltd., Shanghai, China) and the NO 3 − -N and NH 4 + -N concentrations in fresh soil were determined using an Auto Analyzer 3 (Brown rupee company, Hamburg, Germany) [29]. We took the average value of three layers for correlation analysis. At harvest, the plant height was measured by ruler (the length from the soil to the top of panicle). The number of tillers included invalid tillers. All ears in each pot were stored in paper bags and dried at 80 • C for 24 h.

Soil EC, NO 3
− -N and NH 4 + -N Ten days after the restoration of fresh water irrigation (Table 4), the soil EC (mean value of three layers of soil) of J0-J5 treatment decreased at S71, and J2 had the largest decrease rate, and the difference between J0-J5 and CK was significant in 2019 (p < 0.05). In 2018, soil EC of J2 and J4 treatments was also lower than J0, and in 2019, soil EC of J2-J5 was lower than J0, indicating that the application of microbial agents can further reduce soil EC. The reclaimed water irrigation with different EC significantly affected the soil EC of 0-5 cm (F = 6.690, p < 0.05), 15-25 cm (F = 22.099, p < 0.01) and the average EC of three layers (F = 27.848, p < 0.01).
In 2018, the fresh water irrigation treatment (J0) and the B. subtilis and S. cerevisiae treatments (J1-J5) increased the NO 3 − -N ( Figure 1a) and NH 4 + -N (Figure 1d) concentrations in different soil layers, excluding NH 4 + -N in the 15-25 cm layer in the J5 treatment). Specifically, in the 0-5 cm soil layer, compared with CK, the NO 3 − -N concentrations in the J0, J2 and J3 treatments decreased, while those in the J1, J4 and J5 treatments increased. In the 5-15 cm soil layer, compared with CK, the NO 3 − -N concentrations in the J1 and J4 treatments increased by 72.50% and 118.04%, respectively, and the J4 treatment had significant differences with the CK, J0, J2, J3 and J5 treatments (p < 0.05). In the 15-25 cm layer, compared with CK, the NO 3 − -N concentrations in the J4 and J5 treatments increased, and the concentrations in the J4 treatment were significant when compared with the concentrations in the CK (p < 0.05).   NH 4 + -N increased following B. subtilis and S. cerevisiae application. In the 0-15 cm soil layer, compared with the CK, the NH 4 + -N concentrations in the J0, J1, J2, J3, J4 and J5 treatments increased. In the 15-25 cm soil layer, the NH 4 + -N concentrations in the J0, J1, J2, J3 and J4 treatments increased by 5.67-9.49%, while the concentrations in the J5 treatment decreased by 0.74%, and there were significant differences between the J4 treatment and the CK, J3 and J5 treatments (p < 0.05).
In 2019, the NO 3 − -N ( Figure 1c) and NH 4 + -N (Figure 1d) concentrations in the soil increased under J0 treatment, while the NO 3 --N concentrations in the J2, J3, J4 and J5 treatments in the soil decreased. In the 0-5 cm layer, the NO 3 − -N concentrations in the J0 and J1 treatments increased by 34.14% and 20.97%, respectively, which were also significantly higher than the concentrations in the J2, J3, J4 and J5 treatments (p < 0.05). In the 5-15 cm soil layer, the concentrations in the J1 treatment were significantly higher than that in the J5 treatment (p < 0.05). In the 15-25 cm soil layer, the NO 3 − -N concentrations in the J1 treatment increased by 47.61%, and J0, J2, J3, J4 and J5 decreased by 12.12%, 3.72%, 21.51%, 47.15% and 64.70%, respectively, compared to CK, and the J1 treatment was significantly higher than the J3, J4 and J5 treatments (p < 0.05).
Unlike in 2018, the NH 4 + -N concentrations in some of the treatments decreased in 2019. In the 0-5 cm layer, compared with CK, the NH 4 + -N concentrations decreased in the J0, J1 and J4 treatments, while the concentrations in the J2, J3 and J5 treatments increased, and the concentrations in the J5 treatment were the highest, and the differences between the J5 and the CK treatments were significant (p < 0.05). In the 5-15 cm soil layer, there were no significant differences between the J0-J5 treatments and CK; in the 15-25 cm layer, the NH 4 + -N concentrations increased by 2.52-37.11%, but the difference was not significant.

Plant Height, Tiller Number and Yield of Rice
As Table 5 shows, there are differences in plant height, tiller number and yield between 2018 and 2019. In 2018, the J0-J5 treatments increased the plant height of rice, in which the results of J0-J4 were significantly different to CK (p < 0.05), the J2 treatment increased by 15.61% and J0-J5 treatments significantly reduced the number of tillers by 19.44-32.78% (p < 0.05) and significantly increased the rice yield by 333.65-527.88% (p < 0.05), in which J2 treatment increased by 527.88%. In 2019, there was no significant difference in plant height between J0-J5 and CK, but significantly increased the number of tillers by 76.32-128.95% (p < 0.05), and the yield of J0 and J1 treatments increased by 251.85% and 311.11% (p < 0.05), respectively; there was no significant difference between other treatments and CK. On the whole, the J1 treatment can increase yield.

Chlorophyll Concentrations in Rice Leaves
Chlorophyll a (Chl a), chlorophyll b (Chl b), carotenoid and total chlorophyll (Chl a + Chl b) concentrations, in addition to the ratio of chlorophyll a to chlorophyll b (Chl a/b) in rice leaves at S71, are illustrated in Figure 2. In 2018 (Figure 2a), J3-J5 increase chlorophyll concentrations, but the difference in J0-J5 treatments with CK on Chlorophyll concentrations was not significant. In 2019 (Figure 2b), there was significant difference between J1, J2 and J5 treatments with CK on Chl a +b (p < 0.05), which increased by 34.59%, 41.44% and 29.11%, respectively.

Root Activity and Soluble Sugars and MDA Concentrations
In 2018, the root activity (Dehydrogenase) (Figure 3a) in J0, J1, J2, J3, J4 and J5 treatments increased when compared with the root activity in the CK. The differences in root activity between the J4, J4 and J5 treatments and CK were significant (p < 0.05), and the root activity in the J4 and J5 treatments was significantly higher than in the J0 treatment (p < 0.05). In 2019, the root activity in the J0, J2, J3, J4 and J5 treatments decreased when compared to the activity in the CK, while that in the J1 treatment increased by 13.89% with no significant difference. In addition, root activity in the J1 treatment was significantly higher than in the J3 treatment (p < 0.05).
In 2018, compared with CK, the soluble sugar concentrations of J0-J5 treatments had no significant differences with CK (Figure 3b), but the J1 and J5 treatments had significant differences with the J0 treatments (p < 0.05). In 2019, the soluble sugar concentrations in the J1 and J2 treatments increased by 131.21% and 112.68% (p < 0.05), respectively, and J1 was significantly higher than J0 (p < 0.05). The above results showed that the inoculation of B. subtilis and S. cerevisiae increased the soluble sugar concentrations in leaves.
In 2018, the MDA concentrations in the J0, J2, J3 and J4 treatments decreased when compared with the concentrations in the CK (Figure 3c), and there were significant differences between J4 and CK (p < 0.05). In 2019, the MDA concentrations in the J0, J1, J2, J3 and J4 treatments decreased, and the concentrations in the J4 treatment were significantly different from the concentration in the CK, J1 and J5 treatments.

Leaf Enzyme Activity and Protein Levels
The enzyme activity and protein quantity in rice leaves in 2018 are listed in Table 6. SOD activity in the J0, J1, J3, J2, J4 and J5 treatments increased when compared with the activity in the CK; the differences between the J2 and J3 treatments and CK were significant (p < 0.05); and the concentrations in the J3 treatment were significantly higher than those in the J0, J1, J4 and J5 treatments (p < 0.05). Compared with CK, the CAT activities of the J0, J1, J3, J2, J4 and J5 treatments increased, and the differences between the J2 and J3 treatments and the CK were significant (p < 0.05). In addition, the activity in the J3 treatment was significantly higher than the activity in the J0, J1, J2, J4 and J5 treatments. Compared with CK, the POD and GS activity in the J0-J5 treatments had no significant difference with CK. However, the POD activity in the J0 treatment was significantly higher than in the J1 and J4 treatments (p < 0.05). The protein concentrations in the J1, J2, J3 and J5 treatments were lower than the concentrations in the CK treatment, and J2 and J3 treatments significantly decreased by 29.23% and 47.69% (p < 0.05), respectively.
In 2019, SOD activity of J5 was significantly decreased by 23.07% compared to CK, which was also significantly different from J4. CAT activity in the J0 and J2 treatments increased by 21.31% and 7.66%, respectively, although the differences were not significant, whereas that in the J1, J3, J4 and J5 treatments decreased (p < 0.05). The POD activity in the J0-J5 treatments had no significant difference with CK. The GS activity in the J3, J4 and J5 treatments were significantly different from CK. The soluble protein concentrations in the J0, J1, J2, J3, J4 and J5 treatments increased, and the differences between J1, J3, J4 and J5 and CK were significant (p < 0.05). Fresh water irrigation facilitated rice physiological recovery, while continuous use of reclaimed water for irrigation caused damage. B. subtilis and S. cerevisiae improved protein levels in the leaves.   Table 7 showed the relationship between soil N, EC (the mean value of the three layers) and physiological indices.  Note: * and ** represent that there was a significant correlation at p < 0.05 and p < 0.01, respectively.

The Salt Content of Reclaimed Water and Microorganism Changed the Soil EC
As increasing the amount of reclaimed water will increase soil salinity [30], salts in irrigation water should be controlled to allow sustainable irrigation [31]. In this study, the EC of soil restored to fresh water irrigation was lower than that of reclaimed water irrigation, which is an important basis for the combination of saline and fresh water irrigation [32]; that is, to control the soil salinity within a certain range, because reclaimed water irrigation with different salt content significantly affected soil silt and clay content and pH value [33]. We found that the EC of experimental soil is similar in 2018 and 2019 (Table 1), but the EC of reclaimed water in 2019 is 312 µS/cm higher than that in 2018 ( Table 2). Although the irrigation amount in 2019 is lower than that in 2018 (Table S1), at S61, the EC of 5-15 cm soil layer in 2019 is 459 µS/cm higher than that in 2018, but the EC of 0-5 cm and 15-25 cm layers was close, which indicated that the soil salinity increased significantly under the condition of high EC reclaimed water irrigation (in 2019), which directly led to the effect of different years on soil EC exceeding that of microorganism treatments (Table 4). However, with the use of fresh water irrigation at S61-S71, the soil salt gradually moved to the bottom layer (15-25 cm) (Table 4). Although there were differences in the effect of different proportion of microorganisms on reducing soil EC between 2018 and 2019, J2 treatment could significantly reduce soil EC, which further confirmed the positive effect of mixed inoculation of the two microorganisms [34]. Nevertheless, water with an EC above 750 µS/cm might cause yield loss to crops [35,36]. Therefore, when using reclaimed water for irrigation in agricultural production, apart for controlling the irrigation level of reclaimed water [30] and irrigation cycle [37], it is also necessary to pay attention to the salt content of reclaimed water.

Effect of Microorganisms on Soil NO 3 − -N and NH 4 + -N
The documented benefits of the inoculation of plants with beneficial microbes include reduced pathogen infection; improved fertiliser use efficiency; and improved resistance to abiotic stress factors such as drought, mineral deficiency, salinity and phosphate solubilisation [38,39]. Yeast in the soil can promote crop growth [40], mainly because of p solubilisation improved p uptake [41], and induced indole acetic acid production [42]. B. subtilis facilitates the maintenance of ion balance and water movement in plant tissues. N is essential for plant growth, particularly in the early and middle stages, and as long as the root system receives NO 3 − -N, aboveground crop growth will not be limited [43]. Currently, our understanding of the effect of mixed application of B. subtilis and S. cerevisiae on soil nitrogen remains poor.
In the present study, B. subtilis and S. cerevisiae could improve soil NH 4 + -N concentrations; however, their effects on soil NO 3 --N varied in the two years of study. The addition of B. subtilis and S. cerevisiae accelerated oxygen (O 2 ) consumption in the soil and created an increasingly anaerobic environment, which was conducive for denitrification. The soil was alkaline, which was also conducive for N denitrification. As Table S3 shows, air temperature and humidity in 2019 were lower than in 2018; temperature influences N mineralisation and B. subtilis and S. cerevisiae reproduction [44,45], which can alter soil nitrogen supply. While yeast stimulated the nitrification of added ammonium, the hydrolysis of urea, and the subsequent nitrification of the released ammonium [46], the reclaimed water improved soil salinity. High salinity concentrations could inhibit nutrient absorption by roots and affect N transformation [47,48]. In addition, the presence of NO 3 − -N could promote NH 4 + -N absorption [49]. Although NH 4 + -N in soil is converted easily into NO 3 − -N under irrigation, how B. subtilis and S. cerevisiae improve NH 4 + -N requires further investigation, particularly regarding the roles of soil nitrifying and denitrifying bacteria.
Most notably, the nitrate nitrogen content in the 15-25 cm soil layer was higher than 0-15 cm, which increased the risk of nitrate nitrogen loss [50] and caused non-point source pollution [51]. Moreover, the higher soil salt content in the 15-25 cm layer inhibited the absorption of nitrogen by roots. Therefore, it is necessary to reduce the nitrate nitrogen migration to the lower layer through cutting down the amount of irrigation or fertilisation, which is in favour of making full use of the nitrogen provided by reclaimed water [52].

The Effect of Soil Salinity and Nitrate Nitrogen on Physiological Indexes
Wanas (2002) [53] reported that yeast enhanced the formation of chlorophyll and delayed its degradation and senescence in bean plants. In 2018 and 2019, chlorophyll (Chl a + b) and carotenoid concentrations were enhanced by application of B. subtilis and S. cerevisiae; one of the reasons for this is that B. subtilis and S. cerevisiae could increase soil nutrient supply, and mineral nutrition increases chlorophyll concentrations [54]. Upregulation of nitrate and ammonium uptake genes together with the N assimilation genes increased N content and relative nitrogen use efficiency, and then increased the chlorophyll content [55]. Furthermore, a combination of bryophytes and soil yeast could also increase chlorophyll concentrations significantly [56], and the decrease in soil EC (Table 4) is beneficial to the increase of chlorophyll content [57].
Yeast facilitates root colonisation of rice seedlings [16], and the change in soil nutrients after inoculation could promote root growth. In 2018, the application of B. subtilis and S. cerevisiae increased root activity, which was the opposite in 2019 excluding in the J2 treatment. A potential reason for this is that long-term salt stress decreased root activity [58,59]. The EC of reclaimed water in 2019 increased by 312 µS/cm compared to that in 2018 (Table  2), which meant the soil EC of all treatments in 2019 increased by 25.97% compared to that in 2018, and 5-15 cm soil EC at S61 in 2019 increased markedly when compared to that in 2018 (Table S2). The root activity is closely related to the utilisation of nitrogen in rice [60], the decrease in root activity reduced the absorption of nitrogen by plants, as nitrate uptake depends on internal factors related to N demand of the plant, rather than on nitrate availability in the soil volume [61], which explains the reason why soil NO 3 − -N of the treatments in 2019 is higher than that in 2018 ( Figure 1). Moreover, Khelil et al. [62] also found that greater N rates from RW did not improve total N plant uptake. Besides, the soluble sugar content of leaves in 2019 was higher than that in 2018, which was also the effect of soil salinity [63].
Rice planting in saline soil significantly increased MDA [24]. Our results show that inoculating microorganisms can reduce MDA in leaves. Inoculation of Aspergillus aculeatus reduced O 2 accumulation and MDA concentrations in rice roots, and increased antioxidant enzyme activity (SOD, POD and CAT) [64]. In another study, inoculation with yeast (Macrophomina phaseolina) decreased MDA concentrations in sunflower plants [19]. GS activity is an indicator of NH 4 + -N tolerance capacity of root systems [65]; we also observed that 15-25 cm NH 4 + -N was positively correlated with GS in 2018, and GS activity was correlated negatively with 0-5 cm NH 4 + -N in 2019. CAT and SOD can control the potential damage caused by excess reactive oxygen species under environmental stress [66] and maintain normal crop growth. Compared to CK, the mean value of SOD, CAT and protein concentrations of J1-J5 treatments decreased in 2018 and increased in 2019 (Table 6), which is mainly because soil salinity in 2019 is higher than that in 2018, and salt stress could increase protein level in the leaves [67] and cause physiological disorder of rice [68,69]. Despite the fact that yeast provides hormones, vitamins, enzymes, amino acids and minerals for plants [14], and B. subtilis could secrete plant hormones or metabolites with plant hormone activity [70], based on our results, the positive effect of microorganisms was weakened by higher soil salinity, as the plant height in 2018 was higher than that in 2019 ( Table 5). The decrease in air temperature and the increase in nitrogen application could increase soluble protein in leaves [71,72], which are also the reasons why the soluble protein of leaves in 2019 is lower than that in 2018, as the temperature in 2019 is lower than that in 2018 (Table S3), and the NH 4 + -N in 0-25 cm soil layer is lower than that in 2018 (Figure 1b,d). In addition, sowing time also had significant effect on rice growth [73], the date of sowing and transplanting in 2019 is earlier than that in 2018 (Table 3), which could decrease rice yield (Table 5). Specifically, the effect of J1 treatment on rice was relatively stable, redundancy analyses results also confirmed that B. subtilis had greater influence than S. cerevisiae ( Figure 4) and J1 treatment was beneficial to increase the yield (Table 5), mainly because B. subtilis has strong adaptability to extreme conditions and enhances the host's resistance to plant pathogens [74]. Further, the amount of Na + and Cl − in leaves increased with the increase of salt content in reclaimed water [50], and the photosynthesis of leaves was inhibited [75], which is not conducive to dry matter accumulation. Furthermore, considering the potential influence of soil temperature and moisture content, inoculated bacteria may not have a positive role, similar to the differences in NO 3 − -N in 2018 and 2019. Based on the characteristics of B. subtilis and S. cerevisiae, changes in soil p and soil enzyme activity should also be investigated to better understand the changes in plant physiology.

Conclusions
In the present study, we investigated NO 3 − -N and NH 4 + -N concentrations in the soil 10 days after the inoculation of B. subtilis and S. cerevisiae under reclaimed water-fresh water combined irrigation, and experiments were carried out in 2018 and 2019 using reclaimed water with different EC. The inoculation of B. subtilis and S. cerevisiae increased the amount of soil NH 4 + -N and leaf chlorophyll. Both single and mixed application of B. subtilis and S. cerevisiae increased soluble sugar concentrations in leaves, but mixed application could reduce MDA concentrations and GS activity in leaves. Reclaimed water with high EC (2019) increased soil EC compared to reclaimed water with low EC (2018), which led to more severe stress on rice and weakened the ability of microorganisms to increase the physiological activity of rice, resulting in a decrease in plant height and yield. NO 3 − -N influenced the physiological indices the most, especially rice root activity and leaf GS, while B. subtilis contributed more to the changes in physiological indices than S. cerevisiae. The combination of B. subtilis and S. cerevisiae has unstable influence on the growth of rice, which is easily affected by soil EC and air temperature under pot experiment, and the application of 5 g B. subtilis was recommended after irrigation with reclaimed water.