Bioremediation of oil contaminated soil using agricultural wastes via microbial consortium

Agricultural wastes, such as wheat bran and swine wastewater, were used for bioremediation of oil-contaminated soil. Two optimised strains that could degrade oil efficiently were selected. The result showed that the best ratio of strain A to strain B was 7:3. Swine wastewater could be a replacement for nitrogen source and process water for bioremediation. Next, the Box-Behnken design was used to optimise the culture medium, and the optimal medium was as follows: microbial dosage of 97 mL/kg, wheat bran of 158 g/kg and swine wastewater of 232 mL/kg. Under the optimal medium, the oil degradation rate reached 68.27 ± 0.71% after 40 d. The urease, catalase, and dehydrogenase activities in oil-contaminated soil all increased, and the microbe quantity increased significantly with manual composting. These investigations might lay a foundation for reducing the pollution of agricultural wastes, exploring a late model for bioremediation of oil-contaminated soil.


Results
comparison of oil degradation rates among different strains and indigenous microorganisms. The main component of petroleum pollutants in soil is polycyclic aromatic hydrocarbons (PAHs), yet the solubility of PAHs in soil aqueous solution is very low, and PAHs can be adsorbed on soil particles, resulting in poor bioavailability. This retards the biodegradation rate of PAHs at the solid-liquid interface of soil 11 . Surfactant is a kind of macromolecule with both hydrophilic and hydrophobic groups. It can increase the contact area between PAHs and soil microorganisms, improving the availability and degradation rate by curling, solubilisation, and elution 12 . However, exogenous surfactants increase the cost of soil remediation. Therefore, a strain producing sophorolipid(strain B) was selected as the soil remediation strain in this study. At the same time, sophorolipid is degradable and does not easily cause secondary pollution.
In order to select better oil degrading bacteria, four strains were selected for comparison. The comparison of oil degradation rates among different strains and indigenous microorganisms is shown in Fig. 1. The results shows that A and B have high oil degradation rate. The oil degradation rate of C and D is relatively low, so no further study is required. Because the effects of A and B are high, the synergistic effect of A and B is also investigated. The sequence of oil degradation rate of strains is A + B > A (30.08 ± 1.08%)> B (21.19 ± 1.30%), which are higher than indigenous microorganisms. After 40 d culture of strain A + B, the oil degradation rate was 41.12 ± 1.50%, while the oil degradation rate of indigenous microorganisms after 40 d was only 5.02 ± 0.15%. The main reason for the treatment effect of the mixed bacteria being better than that of the single bacterium was that strain B was a sophorolipid-producing bacteria, which could degrade PAHs well. Moreover, there was no competition between strain B and strain A, which could be mixed cultured. Strain A degraded alkane, and strain B degraded PAHs. They played a complementary role, so the degradation rate reached the optimal value. Construction of dominant flora and determination of the optimal ratio. The effect of mixed proportions (A:B) on oil degradation rate were investigated; the results are shown in Fig. 2.  Box-Behnken design and response surface analysis. The best level of the three factors (inoculation amount of mixed strains, the amount of wheat bran and the amount of swine wastewater) was determined by BBD. The factor levels used in the BBD are shown in Table 1, and results are in Table 2. Variance for the quadratic design is in Table 3.
Values of "Prob> F" less than 0.0500 indicate model terms are significant. In this case F, F 2 , and G 2 are significant model terms. The final coding factor formula is as follows: where Y is oil degradation rate, E is the inoculation amount, F is the amount of wheat bran, and G is the amount of swine wastewater.
F test was used to judge the significance of each variable in the regression equation to the response value. The smaller the probability of P, the higher the corresponding variable. The first order term F (P < 0.05) of model (1) was significant, while E and G (P > 0.05) had no significant effects. The effects of secondary order item F 2 , G 2 (P < 0.05) was significant, and E 2 (P > 0.05) had no significant effects. There was no significant difference between EF, EG, and FG (P > 0.05).
These results clearly showed that experimental values had a linear distribution and a good correlation (R 2 = 0.9912). Therefore, the model could be used to predict the degradation rate of crude oil in the variable range. The maximum oil degradation rate of 68.22% was obtained at 96.53 mL/kg inoculation, 157.85 g/kg wheat bran and 232.29 mL/kg swine wastewater. To validate this prediction, three independent experiments were carried out, and an oil degradation rate of 68.27 ± 0.71% was obtained via 97 mL/kg inoculation, 158 g/kg wheat bran and 232 mL/kg swine wastewater.
The effects of the inoculation amount of mixed strains, amount of wheat bran and amount of swine wastewater on oil degradation rate were analysed by RSM. As shown in Fig. 3, three-dimensional response surface plots and contour plots were generated. This allows for studying the interaction of any two variables on the response. From the observation of the response surface graph of the interaction between the two factors, there was a strong interaction among the factors, and it was found that the steeper the curve trend, the more significant the influence. The smoother the curve trend, the smaller the influence. The shape of the contour plot represents the strength of interaction. The contour plot was round, indicating that the interaction between the two factors was weak. Moreoeer, the contour plot was oval, indicating that the interaction between the two factors was strong. Comparing the highest points and contours of the response surface in Fig. 3, there were extreme values in the selected range. www.nature.com/scientificreports www.nature.com/scientificreports/ comparison of soil properties before and after bioremediation. Oil-contaminated soil, oil contaminated soil treated by indigenous microbial aeration composting, and oil-contaminated soil treated by microbial consortium (optimal ratio and optimal culture conditions) were recorded as M1, M2, M3, respectively. The changes of the soil's physical and chemical properties, enzyme activities, and microbial population before and after bioremediation were investigated. The results are shown in Table 4.
After comparing three kinds of soil samples, M3 had the best recovery effect, and the oil degradation rate reached 68.27 ± 0.71%. The oil degradation rate of M2 was only 5.02 ± 0.27%. Large amounts of nutrients were put into the sample in the process of M3 repair, so the content of organics and total nitrogen in the soil were greatly improved, while the changes in the corresponding content in M2 were not as obvious as that in M3.
The activities of dehydrogenase and catalase in M3 were the highest. This was due to the large amount of soil microorganism reproduction during the repair process, and the microorganism participated in the degradation process of oil hydrocarbon, which improved the activity of dehydrogenase and catalase. The urease activity in M3 was the highest, indicating that the soil had a strong nitrogen conversion ability, which was beneficial to the growth of microorganisms in the soil. It also provided sufficient nutritional conditions for the microorganism to participate in the degradation of oil hydrocarbon. However, the activities of various enzymes in M2 were not as obvious as those in M3.
Bacteria counts, actinomycetes counts, and fungi counts in M3 were significantly higher than that of untreated soil sample M1 and soil sample M2 after ordinary composting, which were close to the actual microbial number in the natural world.

Discussion
The remediation of crude oil contaminated soil is a worldwide problem. Although there exist many remediation methods, this technology is still a promising remediation method. The comparison of different method was shown in Table 5. The order of oil degradation rate of different methods saw this study as better than Besalatpour et al. which was better than Wu et al. In this study, the remediation time was the shortest, and the problem of piggery wastewater pollution was solved. Moreover, it had the following advantages: (1) The treatment cost was low. In the process of treatment, all materials except strains were wastes. (2) The process could offset the disposal costs of agricultural wastes. (3) No exogenous chemical surfactants were needed during the remediation process, because strain B produced sophorolipids. (4) Swine wastewater could be a replacement for process water in bioremediation.
comparison of oil degradation rates between two strains and indigenous microorganisms. Two strains were actived in seed medium. The concentration of microbial cells was 1.5×10 8 CFU/mL.
The oil contaminated soil was packed into 48 conical bottles (50 g each). The conical flasks were divided into three groups, with 16 flasks in each group. Then, 50 mL/kg of strain A and strain B suspension were added to two groups, and put in the incubator (28 °C). They were watered and ploughed every day to keep the water content around 20%. The remaining group (control group) did not receive strain suspension for investigating the effect of indigenous microorganisms. Samples were taken at intervals, and the oil degradation rate was calculatedby gravimetric method to checkthe mass fraction of oil in soil samples 14 .
Construction of dominant flora and determination of the optimal ratio. Combinations of two strains were carried out. The mixed bacteria of different combinations were added into the oil-contaminated soil according to a certain amount of inoculation, and the total inoculation amount (50 mL/kg) of the mixed bacteria suspension in each group was equal. They were watered and ploughed every day to keep the water content around 20%. After 40 days in the incubator (28 °C), the mass fraction of oil in the soil samples was determined, and the oil degradation rate was calculated.
Box-Behnken design and response surface analysis. Three factors affecting the oil degradation rate (inoculation amount of mixed strains, the amount of wheat bran and the amount of swine wastewater) were selected. Box-Behnken design (BBD) was used to determine optimal concentrations of factors using Design-Expert software (Version 8.0.6, Stat-Ease, Inc, USA), and to understand the relationship among various factors. The three factors were studied at three levels (Table 1), and 17 sets of experiments were carried out ( Table 2). All experiments were carried out in triplicate. The optimal values of factors were obtained by analysing 3D plots. The statistical analysis of the model was represented as an analysis of variance (ANOVA).
Analytical methods. The pH of soil samples was measured by potentiometric titration, and the volume ratio of water to soil was 2.5. The content of organic matter in soil samples was determined by potassium dichromate volumetric method. The total nitrogen content of the soil sample was determined by the semi-micro kjeldahl method. The content of total phosphorus in a soil sample was determined by the Mo-Sb antispetrophotography method 15 . The total petroleum hydrocarbons (TPH) content in soil was determined by the gravimetric method 14 . The calculation formula for the oil degradation rate was as follows: (2020) 10 where ODR was the oil degradation rate (%), W 1 was the content of petroleum hydrocarbon in the soil before degradation, and W 2 was the content of petroleum hydrocarbon in the soil after degradation. The amount of polycyclic aromatic hydrocarbon and alkane in soil were measured by the gravimetric method 14 .
Total nitrogen, total phosphorus, suspended solid, NH 3 -N, BOD 5 , and COD Cr in the swine wastewater were measured according to the procedure described in standard methods for the examination of water and wastewater 15 .
The activity of dehydrogenase was determined by using triphenyltetrazolium chloride as hydrogen acceptor, which was reduced to form red formazan. It was determined by colorimetry. The amount of formazan produced in 1 g of dry soil for 6 h was an active unit of dehydrogenase. The urease activity in the soil sample was determined by Nessler's reagent colorimetry. The amount of NH 3 -N produced in 24 h in 1 g of dry soil was an activity unit of urease. Polyphenol oxidase activity in a soil sample was determined by pyrogallol colorimetry. The amount of gallic acid produced in 3 h in 1 g of dry soil was taken as an activity unit of polyphenol oxidase. Catalase activity in a soil sample was determined by Potassium Permanganate titration. The enzyme activity was expressed by the volume number of potassium permanganate consumed in 1 g dry soil in 1h 16 . The amount of microbial population in the soil was determined by colony forming unit 17 .