{Optimization of phenol biodegradation by immobilized bacillus subtilis isolated from hydrocarbons contaminated water using the factorial design methodology} J. Serb. Chem. Soc. 84 (7) 679–688 (2019) UDC 547.562+504.4.054:582.261: JSCS–5218 542.92.000.57 Original scientific paper 679 Optimization of phenol biodegradation by immobilized Bacillus subtilis isolated from hydrocarbons-contaminated water using the factorial design methodology HAMIDA HAMDI* and AMINA HELLAL Ecole Nationale Polytechnique, Laboratoire des Sciences et Techniques de l’Environnement, 10 Avenue Hacen Badi, BP182 El Harrach, 16200 Algiers, Algeria (Received 4 December 2018, revised 8 March, accepted 11 March 2019) Abstract: The ability of newly isolated bacteria, identified as Bacillus subtilis immobilized on alginate hydrogel beads, to degrade phenol was investigated under different parameters, such as phenol concentration, bead diameter and inoculums size, and was optimized using full factorial design methodology. A mathematical model that governs the degradation of phenol by the immobilized system was obtained and it fitted the experimental data very well. The model indicated that within the range of variables employed, all the parameters and their interactions influenced the biodegradation process, whereby the phenol concentration was the most significant factor. B. subtilis revealed a very high degradation activity and could be grown using phenol as the sole source of carbon. Phenol was degraded by the new bacteria in 8 h under the optimum conditions obtained by the desirability function: 100 mg L-1 phenol concentration, 3 mm beads diameter and 244.5 mg of cell dry per liter biomass size, with a desirability value of 91.25 %. Keywords: alginate beads; biodegradation; immobilized system; phenol. INTRODUCTION Environmental pollution is one of the major problems and the most important in the world. The development in agriculture, energy sources, and chemical ind- ustries is necessary in order to fulfill the needs and demands of the ever-growing human population. Almost all processes employed by man for the production of goods and services lead to the production of environmental pollutants.1,2 Several physico–chemical methods for removing organic pollutants have been proposed, but these methods are not cost effective for large-scale applications. Nowadays, biological methods are widely used as low-cost alternative treatment methods, which offer the possibility of complete mineralization of organic compounds.3 * Corresponding author. E-mail: hamida.hamdi@g.enp.edu.dz https://doi.org/10.2298/JSC181204022H 680 HAMDI and HELLAL Some of these pollutants are biologically recalcitrant and inhibitory organics, which greatly reduces the ability of microorganism to biodegrade the compounds during treatment or in nature.4,5 For this reason, the use of degradation of the pollutants by free system is not preferable. Immobilized systems in which cells are entrapped in polymeric matrices6 have received increasing attention in the field of contaminant treatment and may be an effective and economical technique. Immobilization provides higher cell concentrations, ease of use in continuous reactors, shorter start-up period for the bioreactor, and greater stability by protecting cells from direct exposure to toxic compounds compared with freely suspended cells and hence they can be stored for long periods without losing their degradative abilities.6–8 The rate and efficiency of the biodegradation of pollutants by immobilized systems depend on several factors. Among these parameters are initial concen- tration of pollutant, bead diameter and biomass size.9,10 The factorial experimental design has been used for studying, developing and optimizing a wide range of engineering systems.11,12 Factorial designs are widely used to investigate the effects and influence of experimental factors and the interactions between them.12 The advantages of factorial experiments include the relatively low cost, a reduced number of experiments, and increased possi- bilities to evaluate interactions among the variables.12,13 In this study, the biodegradation of phenol by immobilized system using a newly isolated bacteria strain (isolated and identified in the authors’ laboratory), was opti- mized by application of a factorial design methodology. This mathematical method was used to determine the main and the interactive effects of different variables on the immobilized system and to predict a model that evaluates the process. EXPERIMENTAL Culture media The growth medium (GM) of the strain consisted of peptone, 10.0 g L-1; beef extract, 3.0 g L-1 and NaCl, 5.0 g L-1, at a pH of 7.0, autoclaved at 121 °C for 15 min. The mineral culture medium (MM) that contained the necessary nutrients for the growth of microorganisms was composed of: KH2PO4, 1.5 g L-1; K2HPO4, 0.5 g L-1; NaCl, 0.5 g L-1; MgSO4·7H2O, 0.5 g L-1; NH4NO3, 3.0 g L-1; FeSO4·7H2O, 0.02 g L-1 and CaCl2·2H2O, 0.02 g L-1.8 The MM was used for the isolation and the biodegradation tests with phenol as the sole carbon and energy source. Solidified with agar, this medium (MMA) was used for preservation of the strains. The nutrient agar (NA) contained 28 g of powdered nutrient agar dissolved in 1 L of dis- tilled water and autoclaved at 121 °C for 15 min. Isolation, selection and purification of new bacterial strains Samples of water contaminated by hydrocarbons were collected from the petroleum ref- inery situated at Hassi Massoud, south of Algeria. The contaminated water was inoculated on agar mineral culture medium (MMA) supple- mented with 100 mg L-1 of phenol and incubated for 7 days. The development of colonies was OPTIMIZATION OF THE PHENOL DEGRADATION BY IMMOBILIZED SYSTEM 681 monitored every day. Well-isolated colonies were subcultured on GM. This procedure was repeated several times. Phenol tolerant bacterial strains were preserved on MMA at 4 °C. The purity of the cultures was regularly checked during preservation. Identification and characterization of the isolated strain The genera and/or species of the most active selected strain were identified by perform- ing Gram staining, morphological characteristics and biochemical tests (using API system tests). Dry-cell-weight determination Inoculums size was measured using a spectrophotometer (Shimadzu UV–Vis 1240) at 600 nm and the concentration, expressed in cell dry weight, were determined from previously established curves. Phenol dosage Phenol was quantified by a colorimetric method based on the condensation of 4-amino- antipyrine with phenol in the presence of an oxidizing agent, potassium ferricyanide, in an alkaline medium to give a red complex.14,15 The absorbance was measured at 510 nm using a Shimadzu UV–Vis 1240) spectrophotometer. Immobilization of bacteria on calcium alginate beads First, the pure strain was cultivated into 100 mL of GM at 37 °C. Bacterial cells were then harvested at the stationary phase by centrifugation (6000 rpm at 4 °C for 10 min), washed three times with the phosphate buffer solution and suspended in sterile MM. The bio-beads of calcium alginate were obtained by inclusion of microorganisms in sodium alginate and then application of the extrusion technique. This method involves the preparation of an aqueous solution of sodium alginate (3 mass%) in which microorganisms were incorporated. The suspension containing the cells with the sodium alginate solution was extruded as drops into 100 mL of a 0.1 M calcium chloride solution. The formed beads were transferred into a solution of 0.2 M CaCl2 and incubated at 37 °C for 2 h to allow for complete replace- ment of the sodium ions by calcium to ensure a better stability. After washing with sterile distilled water, the beads were stored at 4 °C for preservation.3 Biodegradation tests Biodegradation tests by the immobilized system were performed in a batch bioreactor at 37 °C under aerobic conditions. A control test without bacteria allowed for verification of whether the phenomenon of adsorption or other interactions between the alginate beads and the phenol occurred. The reduced phenol was less than 2 %, which is the result of the chem- ical method used for the determination of phenol. Factorial design methodology Factorial design methodology was used to determine the main and interactive effects of different variables on the immobilized system and to predict a model to evaluate the process. The principle steps of factorial designed experiments are determination of the response that reflects the aim of the study, factors and their levels and choice of the experimental design and statistical analysis of the data. A full 23 factorial design was performed to evaluate the import- ance of three factors: the initial phenol concentration (X1), the diameter beads (X2), and the inoculums size expressed in dry cell weight (X3) on the response (Y), which is the time, exp- ressed in h, required for total phenol removal. The number of experiments conducted is con- sidered as 23. 682 HAMDI and HELLAL The high and low levels defined for the 23 factorial design are listed in Table I. The low and high levels for the factors were selected according to some preliminary experiments. TABLE I. Factors and levels used in the factorial design Independent variable Coded variables level Low (–1) Center (0) High (+1) X1 concentration of phenol, mg L -1 100 500 900 X2 diameter beads, mm 3 4 5 X3 dry cell weight, mg L -1 244.5 317.7 390.9 The results were analyzed with JMP® 8 software, and the main effects and interactions between factors were determined. RESULTS AND DISCUSSION Isolation and selection of phenol degrading strains Ninety bacterial strains were isolated and purified from contaminated waters. Among these strains, 44 were able to degrade phenol. Of these 44 strains, only 7 started growth from the second day. The rest of the strains started development from the third day. The period of appearance of the colonies constitutes a lag phase necessary for the adaptation of these strains to the new carbon source. The faster the growth of the strain, the better its adaptation to the pollutant. For this reason, one strain taken of the seven most active strains was chosen for the study. Identification and characterization of phenol degrading strain The identification and characterization of the isolated bacterial was per- formed using preliminary analysis: morphological and cultural characteristics (Table S-I of the Supplementary material to this paper) and biochemical tests (Table S-II of the Supplementary material). The identification of the strain by the preliminary analysis confirmed that the strain was B. subtilis species. Biodegradation of phenol by the immobilized bacteria Factorial design methodology. The design matrix of the coded values for factors and the response Y, measured in each factorial experiment, is shown in Table II. The response represents the time for total degradation. The data listed in Table II indicate a wide variation in the response Y, from 8 to 103, in the 10 trials. Application of factorial design methodology as an optimization technique requires the selections of a model at the beginning. The model employed for 23 factorial designs with interaction is represented by Eq. (1):16–18 Y = a0 + a1X1 + a2X2 + a3X3 + a12X1X2 + a13X1X3 + a23X2X3 + + a123X1X2X3 (1) OPTIMIZATION OF THE PHENOL DEGRADATION BY IMMOBILIZED SYSTEM 683 TABLE II. Experimental values for the total time of degradation Experiment number X1 X2 X3 Y (Response) 1 –1 –1 –1 08 2 +1 –1 –1 78 3 –1 +1 –1 09 4 +1 +1 –1 75 5 –1 –1 +1 10.5 6 +1 –1 +1 77 7 –1 +1 +1 13 8 +1 +1 +1 103 9 0 0 0 46 10 0 0 0 46 In Eq. (1), Y is the estimated response that represents the time of total degradation, a0 is the independent coefficient (a constant term), ai (i = 1, 2, 3) are the linear coefficients for the variables: phenol concentration, bead diameter and inoculums size, respectively, aij represents the coefficients of the interaction parameters Xi and Xj with i < j. The coefficients ai, aij and aijk were determined using JMP® 8 software.17,19 Substitution of coefficient ai, aij and aijk in Eq. (1) by their values, the mathematical model selected is the following: Y = 46.55 + 36.5625X1 + 3.3125X2 +4.187 X3 + 2.4375X1X2 + + 2.5625X1X3 + 3.8125X2X3 +3.4375X1X2X3 (2) Eq. (2) showed how the experimental variables and their interactions influ- ence the time of total degradation.13,20 The phenol concentration (X1) had the greatest effect on Y, followed by the inoculum size, bead diameter–inoculum size interaction (X2X3), phenol concentration–bead diameter–inoculum size interact- ion (X1X2X3), bead diameter (X2) and phenol concentration–bead diameter inter- action (X1X2). According to the obtained model, all parameters had a positive effect on the response. The positive sign shows that there is a direct relation between the parameter and the dependent variable. Thus, an increase in the phenol concen- tration, bead diameter or inoculums size increases the time of total degradation. The predicted values vs. the experimental values of the time of total degrad- ation of phenol are shown in Fig. 1. The model presented a high square correl- ation coefficient (R2) of 99.993 % and an adjusted square correlation coefficient (R2 adj) of 99.9697 %, which confirmed that the chosen of 23 factorial design with the interaction model was found to be adequate for the prediction within the range of variables employed.18,21–23 The Student’s t-test The significance of the regression coefficients was determined by applying a Student’s t-test (Table III).13,16 684 HAMDI and HELLAL Fig. 1. Comparison of the experimental and the estimated responses at the levels of the variable. TABLE III. Estimated regression coefficients for the time of total degradation Parameter Estimate coefficients Standard error t-value P-valuea Constant 46.55 0.194454 239.39 ˂.0001* X1 36.5625 0.217407 168.18 ˂.0001* X2 3.3125 0.217407 15.24 0.0043* X3 4.1875 0.217407 19.26 0.0027* X1X2 2.4375 0.217407 11.21 0.0079* X1X3 2.5625 0.217407 11.79 0.0071* X2X3 3.8125 0.217407 17.54 0.0032* X1X2X3 3.4375 0.217407 15.81 0.0040* aP-value < 0.05 The Student’s t-test and P-value were used to determine the significance of the regression coefficients of the parameters. With the 95 % confidence level and two degrees of freedom, the value of t-critic is equal to 2.92. The coefficient of the regression is statistically significant if the corresponding t-value is higher than 2.92. It was observed from Table III that the linear effects of phenol concen- tration, bead diameter, and inoculum size and the interaction effect between these factors were significant. The values for each effect are shown in the Pareto chart by the horizontal columns (Fig. 3). The importance of the data can also be judged by their P-value, which is the probability value that is used to determine the statistically significant effects in the model, with values closer to zero denoting greater significance. For the 95 % confi- dence level, the P-value should be less than or equal to 0.05 for the effect to be considered statistically significant.24,25 Hence, the final model is given in Eq. (3): Y=46.55+36.5625 X1 + 3.3125 X2 +4.1875 X3 + 2.4375 X1X2 +2.5625X1X3 + 3.8125 X2X3 +3.4375X1X2X3 (3) OPTIMIZATION OF THE PHENOL DEGRADATION BY IMMOBILIZED SYSTEM 685 Interaction plots The plots of the interaction effects are shown in Fig. 2. An interaction is effective when the change in the response from low to high levels of a factor is dependent on the level of a second factor, i.e., when the lines do not run paral- lel.13 Thus, significant interactions between the parameters could be concluded. Fig. 2. Interaction plot for the time of total degradation (X1: phenol concen- tration, mg L-1; X2: diameter of beads, mm; X3: dry-cell weight, mg L-1). The standardized Pareto chart (Fig. 3) depicts the main effect of the inde- pendent variables and interactions on the time of total degradation. The length of each bar in the graph indicates the effect of these factors and the level of their effects on the response.26,27 Fig.3. Pareto chart of the standardized effects. This graphic (Fig. 3) shows both the magnitude and the importance of the effects (variables and interactions).26 According to Fig. 3, it could be inferred that the main factors (X1, X2 and X3) and their interactions (X1X2, X1X3, X2X3 and X1X2X3) were significant at the 0.05 level. The concentration of phenol represents the most significant effect on the time of total degradation. 686 HAMDI and HELLAL Analysis of variance (ANOVA) The adequacy of the model was checked using analysis of variance (ANOVA), as shown in Table IV.18,23 TABLE IV. Analysis of variance (ANOVA) Source Sum of squares DF Mean square F-value t-F Model 11233.469 7 1604.78 4244.5 19.4 Residual (error) 0.756 2 0.38 Prob ˃ F Correlation total 11234.225 9 0.0002 The F-value of the model obtained (4244.5) was higher than the tabular value t-F (F0.05,5,4 = 19.4), suggesting that the model is highly significant. Estimation of optimal design conditions by the method of desirability function. The optimization by the desirability function improved the performance of the analytical process and discovering the conditions at which the best response is obtained.28 The optimum condition for the degradation of phenol by the immobilized process was obtained from the desirability plot (Fig. 4). X1 X2 X3 Desirability X D es ira bi lit y Y / h Fig. 4. Desirability functions for the optimization of the response (X1: phenol concentration, mg L-1); X2: bead diameter, mm; X3: dry-cell weight, mg L-1). The best combination of factor settings for achieving the optimum response (minimum time of total degradation) was found to be phenol concentration 100 mg L–1, bead diameter 3 mm and inoculums size 244.5 mg L–1 for the predicted response of 7.86 h and a desirability value of 91.25 % (Fig. 4). CONCLUSIONS In conclusion, a new bacterial strain capable of degrading phenol was iso- lated from hydrocarbons contaminated water, collected from the petroleum refin- OPTIMIZATION OF THE PHENOL DEGRADATION BY IMMOBILIZED SYSTEM 687 ery, situated in the south of Algeria. It was identified as B. subtilis species based on morphological and cultural characteristics and biochemical tests The B. subtilis species presented a very high activity for phenol biodegrad- ation. The best time of complete degradation by immobilized cells determined using factorial design methodology was 7.86 h, which corresponds to the opti- mum conditions: phenol concentration 100 mg L–1, bead diameter 3 mm and biomass size 244.5 mg L–1. The concentration of phenol represented the most significant effect on the total time of degradation. Full factorial design methodology can be an efficient method for testing the effect of operating conditions and optimize the degradation of phenol by the studied immobilized process. SUPPLEMENTARY MATERIAL Additional data are available electronically from http://www.shd.org.rs/JSCS/, or from the corresponding author on request. И З В О Д ОПТИМИЗАЦИЈА БИОДЕГРАДАЦИЈЕ ФЕНОЛА ИМОБИЛИЗОВАНОМ БАКТЕРИЈОМ Bacillus subtilis ИЗОЛОВАНОМ ИЗ ВОДЕ КОНТАМИНИРАНЕ УГЉОВОДОНИЦИМА ПРИМЕНОМ МЕТОДЕ ФАКТОРИЈАЛНОГ ДИЗАЈНА HAMIDA HAMDI и AMINA HELLAL Ecole Nationale Polytechnique, Laboratoire des Sciences et Techniques de l’Environnement, 16200 Algiers, Algeria Испитана је примена новоизоловане бактерије Bacillus subtilis, имобилизоване на честицама алгинатног хидрогела, за деградацију фенола, варирајући концентрацију фенола, пречник честица и величину инокулума, а за оптимизацију је коришћена метода потпуног факторијалног дизајна. Израђен је математички модел разлагања фенола имобилизованим системом и он се добро уклапао са експерименталним резултатима. Према моделу, све променљиве које су тестиране, као и њихов примењени опсег, су утицале на процес биодег- радације, а концентрација фенола је била најутицајнији фактор. Показано је да B. subtilis испољава јаку деградациону активност и да може расти користећи фенол као извор угљеника. Ова бактерија разлаже фенол у року од 8 h са 91,25 % ефикасности, под следећим оптималним условима: концентрација фенола 100 mg L-1, пречник честица 3 mm и 244,5 mg сувих ћелија по литру биомасе. (Примљено 4. децембра 2018, ревидирано 8. марта, прихваћено 11. марта 2019) REFERENCES 1. A. Hazrat, Water Air Soil Poll. 213 (2010) 251 (https://doi.org/10.1007/s11270-010-0382-4) 2. D. Suryaman, K. Hasegawa, J. Hazard. Mater. 183 (2010) 490 (https://doi.org/10.1016/j.jhazmat.2010.07.050) 3. O. Ali, A. Namane, A. Hellal, J. Ind. Eng. Chem. 19 (2013) 1384 https://doi.org/10.1016/j.jiec.2012.12.045) 4. L. Zhang, C. Xu, Z. Chena, X. Li, P. Li, J. Hazard. Mater. 173 (2010) 168 (https://doi.org/10.1016/j.jhazmat.2009.08.059) 5. R. Daghrir, P. Drogui, D. Robert, J. Photochem. Photobiol. A: Chem. 238 (2012) 41 (http://dx.doi.org/10.1016/j.jphotochem.2012.04.009) 688 HAMDI and HELLAL 6. M. C. Dictor, N. Berne, O. Mathieu, A. Moussay, A. Saada, Oil Gas Sci. Technol. 58 (2003) 481( https://doi.org/10.2516/ogst:2003031) 7. D. Z. Chen , J. Y. Fang , Q. Shao, J. X. Ye, D. J. Ouyang, J. M .Chen, Bioresour. Technol. 139 (2013) 87 ( https://doi.org/10.1016/j.biortech.2013.04.037) 8. N. K. Patil, Y. Veeranagouda, M. H. Vijaykumar, S. A. Nayak, T. B. Karegoudar, Int. Biodeter. Biodegrad. 57 (2006) 82 (https://doi.org/10.1016/j.ibiod.2005.11.007) 9. A. S. Liffourrena, G. I. Lucchesi, Int. Biodeter. Biodegrad. 85 (2013)337 (https://doi.org/10.1016/j.ibiod.2014.02.008) 10. I. Ani, S. Wahidinn, Process Biochem. 41 (2006) 1117 (https://doi.org/10.1016/j.procbio.2005.12.002) 11. M. Mäkelä, Energ. Convers. Manage. 151 (2017) 630 (http://dx.doi.org/10.1016/j.enconman.2017.09.021) 12. A. Hannachi, S. Hammami, N. Raouafi, H. Maghraoui-Meherzi, J. Alloys Compd. 663 (2016) 507 (https://doi.org/10.1016/j.jallcom.2015.11.058) 13. D. Bingol, N. Tekin, M. Alkan, Appl. Clay Sci. 50 (2010) 315 (https://doi.org/10.1016/j.clay.2010.08.015) 14. Q. S. Liu, T. Zheng, P. Wang, J. Jiang, N. Li, Chem. Eng. J. 157 (2010) 348 (https://doi.org/10.1016/j.cej.2009.11.013) 15. R. D.Yang, A. E. Humphyrey, Biotechnol. Bioeng. 17 (1975) 1211 (https://doi.org/10.1002/bit.260170809) 16. J. Goupy, L. Creighton, Introduction aux plans d’expériences (Introduction to experimental design), 3rd ed., Dunod, 2006 17. W. Tinsson, Plans d’experience: constructions et analyses statistiques, Mathematiques et Applications (Plans of experience: constructions and statistical analyzes, Mathematics and Applications), 67. Springer-Verlag, Berlin, 2010, in French (https://doi.org/10.1007/978-3-642-11472-4) 18. O. B. Ayodele, J. K. Lim, B. H. Hameed, Chem. Eng. J. 197 (2012) 181 (https://doi.org/10.1016/j.cej.2012.04.053) 19. M. Proust, JMP® 8 Introductory Guide, 2nd ed.; Cary, NC: SAS Institute Inc., 2009 20. R. Y. Sheeja, T. Murugesan, J. Hazard. Mater. 89 (2002) 287 (https://doi.org/10.1016/S0304-3894(01)00319-3) 21. V. A. Sakkas, M. A. Islam, C. Stalikas, T. A. Albanis, J. Hazard. Mater. 175 (2010) 33 (https://doi.org/10.1016/j.jhazmat.2009.10.050) 22. D. Juretic, H. Kusic, N. Koprivanac, A. L. Bozic, Water Res. 46 (2012) 3074 (https://doi.org/10.1016/j.watres.2012.03.014) 23. J. Zhou, X. Yu, C. Ding, Z. Wang, Q. Zhou, H. Pao, W. Cai, J. Environ. Sci. 23 (2011) 22 (https://doi.org/10.1016/S1001-0742(10)60369-5) 24. S. Saadat, A. Karimi-Jashni, Chem. Eng. J. 173 (2011) 743 (https://doi.org/10.1016/j.cej.2011.08.042) 25. W. Lazli, D. Hank, S. Zeboudj, A. Namane, A. Hellal, Desalin. Water Treat. 57 (2016) 6044 (https://doi.org/10.1080/19443994.2015.1004112) 26. A. Salhi, A. Aarfane, S. Tahiri, L. Khamliche, M. Bensitel, F. Bentiss, M. El Krati, Mediterr. J. Chem. 4 (2015) 59 (http://dx.doi.org/10.13171/mjc.4.1.2015.16.01.20.30/salhi) 27. A. S. Zidan, O. A. Sammour, M. A. Hammad, N. A. Megrab, M. J. Habib, M. A. Khan, Int. J. Pharm. 96 (2007) 2409 (https://doi.org/10.1002/jps.20824) 28. L. V. Candioti, M. M. De Zan, M. S. Cámara, H. C. Goicoechea, Talanta 124 (2014) 123 (http://dx.doi.org/10.1016/j.talanta.2014.01.034).