Taguchi approach significantly increases bioremediation process efficiency: a case study with Hg (II) removal by Pseudomonas aeruginosa

Letters in Applied Microbiology ISSN 0266-8254

ORIGINAL ARTICLE

Taguchi approach significantly increases bioremediation process efficiency: a case study with Hg (II) removal by Pseudomonas aeruginosa S.G. Tupe, J.M. Rajwade and K.M. Paknikar Agharkar Research Institute, Pune, India

Keywords bioremediation, mercury, parameters optimization, Pseudomonas aeruginosa, Taguchi approach. Correspondence K.M. Paknikar, Agharkar Research Institute, G.G. Agarkar Road, Pune, India 411004. E-mail: [email protected]

2006 ⁄ 1286: received 13 September 2006, revised 2 February 2007 and accepted 19 February 2007 doi:10.1111/j.1472-765X.2007.02152.x

Abstract Aim: Optimization of process parameters for mercury removal by an Hg (II)reducing Pseudomonas aeruginosa strain. Methods and Results: A strain of Ps. aeruginosa was found to reduce 10 mg l)1 Hg (II) to Hg0 with 70% efficiency in 24 h. To optimize process performance, a statistical tool – Taguchi design of experiments (DOE) – was used to carry out 18 well-defined experiments (L18 Orthogonal array) with eight variable parameters (viz. agitation, temperature, pH, carbon source, medium volume: flask volume ratio and concentrations of Hg (II), ammonium sulfate and yeast extract). When data obtained were analyzed using specialized software for Taguchi design, Qualitek-4 (Nutek Inc., MI, USA), Hg (II) reduction efficiency was predicted to be 95% in 24 h under the optimized process parameters (also suggested by the software). In the validation experiment, Hg (II) removal of 99Æ29% in 24 h was indeed obtained. Conclusions: Using Taguchi DOE, Hg (II) reduction (and hence its removal) using Ps. aeruginosa could be improved by 29Æ3%. Significance and Impact of the Study: Taguchi approach could be employed as an efficient and time-saving strategy for parameter optimization in bioremediation processes.

Introduction Mercury is a major pollutant and the problem of mercury pollution is rising at an alarming rate in many developing countries such as India. Effluents arising from various industries (mainly chlor–alkali plants) are sources of mercury in the environment and hence, urgent steps need to be taken to combat the problem (Anon 2003). Microbial reduction of Hg (II) to Hg0 has emerged as an efficient and economical alternative to the physical and chemical methods for the removal of mercury from wastewaters (Chang et al. 1998; Canstein et al. 1999). Being a microbial process, Hg (II) reduction efficiency is influenced by factors such as concentration of mercury, inoculum size, medium volume, temperature, pH, redox 36

potential, and the presence of inorganic and organic complexing agents, nutrient concentration, etc. (Chang et al. 1998; Ullrich et al. 2001). Therefore, to achieve maximal process efficiency various parameters need to be optimized. Because the final aim of the present study is to develop a continuous-mode bioreactor for treatment of mercury containing wastewater, parameters such as agitation, temperature, pH, nutrients, Hg (II) concentration, and medium volume are the most important to be optimized. Inoculum size, which is a significant parameter in batch studies, is not taken into consideration as the biomass in a continuous bioreactor is expected to attain many folds higher density, than that obtainable in a flask experiment or a bioreactor in batch mode (Barreto et al. 1991).

ª 2007 The Authors Journal compilation ª 2007 The Society for Applied Microbiology, Letters in Applied Microbiology 45 (2007) 36–41

S.G. Tupe et al.

Conventional process optimization procedures involve altering of one parameter at a time keeping all other parameters constant, which enables to assess the impact of those particular parameters on the process performance. The approach assumes that process parameters do not interact, and that the outcome is a direct function of a single variable (Beg et al. 2003). In reality, however, variables are interactive. Therefore, optimization of parameters by the conventional approach becomes erroneous. Statistical optimization methods can take into account the interaction of variables in generating the process response (Haaland 1989). One such tool is the Taguchi design of experiments (DOE) developed by Genichi Taguchi to improve the quality of manufactured goods (Taguchi 1986). The Taguchi method involves the study of a system by a set of independent variables (factors) over a specific region of interest (levels) (Roy 1990). Taguchi’s parameter design concept is related to finding the appropriate design factor levels to make the system insensitive to variations in noise (uncontrollable factors). The approach also facilitates the identification of the influence of individual factors, determines the relationship between variables and operational conditions and finally establishes the performance at the optimum levels obtained with a few well-defined experimental sets. Analysis of the experimental data using the analysis of variance (anova) and factor effects gives the output that is statistically significant in finding the optimum levels. The Taguchi methodology has been successfully applied to banking, electronics, Government policy making and communication and information networks (Ross 1988; Phadke 1989; Taguchi 1995). It has also been used in experimental optimization, such as various biochemical techniques, microbial fermentations (Caetano-Anolles 1998; Jeney et al. 1999; Khoudoli et al. 2004; Lim et al. 2004; Rao et al. 2004; Nagarjun et al. 2005; Prasad et al. 2005) and chemical synthesis (Barrado et al. 1996). Recently, the Taguchi approach was used for anaerobic treatment of complex chemical wastewater in a sequencing batch biofilm reactor (Mohan et al. 2005). In this paper, we show that the Taguchi approach could be applied to optimize metal bioremediation processes, using a case study involving Hg (II) removal by Pseudomonas aeruginosa. Materials and methods Culture The culture used in the study was a sewage isolate having 180 mg l)1 Hg (II) tolerance and capable of reducing it

Mercury bioremediation

to Hg0 under aerobic conditions. The isolate was identified as Ps. aeruginosa by 16S rDNA sequencing. Preliminary Hg (II) reduction experiment A medium (pH 8) containing (in g%) yeast extract 0Æ1, KNO3 0Æ1, (NH4)2SO4 0Æ1, KH2PO4 0Æ05, K2HPO4 0Æ05, glycerol 0Æ1 and supplemented with 10 mg l)1 Hg (II) as mercuric chloride was used. Pseudomonas aeruginosa (OD6001) was inoculated in the medium so as to achieve initial cell density of 105 CFU ml)1 and the flasks were incubated for 24 h at room temperature (25 ± 3C). After 24 h incubation, 2 ml aliquot was removed, centrifuged (8000 rev min)1, 10 min, 4C), and the supernatant was used for the estimation of residual Hg (II) by cold vapour atomic absorption spectrometry (2380; Perkin Elmer, Waltham, MA, USA). The experiment was performed in duplicate with uninoculated control run simultaneously. Percent mercury removal efficiency was calculated using the formula: 2 4



3

InitialHg0 conc:intest ResidualHg0 conc:intest 0  InitialHg0 conc:intest 5 InitialHg conc:incontrol ResidualHg0 conc:incontrol 0 InitialHg conc:incontrol

 100

To confirm Hg (II) reduction as volatile Hg0, Whatman No. 1 paper strips impregnated with cuprous iodide were positioned above the medium in all the flasks using cotton plugs. Taguchi experiment, data analysis and validation In the Taguchi DOE, eight factors were identified to evaluate their role in optimization of Hg (II) removal, viz. agitation, temperature, pH, carbon source, Hg (II), ammonium sulfate, yeast extract concentration and medium : flask volume ratio. Based on the number of control factors (agitation at two levels viz. shaking and no shaking, and three levels each for the remaining seven factors), the L18 (21 · 37) orthogonal array was used in the design. The different factors and their levels considered are shown in Table 1 while Table 2 shows the layout of the L18 orthogonal array. A late exponential phase culture of Ps. aeruginosa, grown in the medium (described above) for 24 h on rotary shaker (120 rev min)1) at 30C was used as inoculum. Experiments were performed in 250 ml Erlenmeyer flasks containing basal medium with (g%) potassium nitrate 0Æ1, K2HPO4 0Æ05 and KH2PO4 0Æ05. Other ingredients were added as mentioned in rows of Table 2 for respective experiment. Pseudomonas aeruginosa (OD6001) was inoculated so as to achieve initial cell density of 105 CFU ml)1 and the flasks were incubated

ª 2007 The Authors Journal compilation ª 2007 The Society for Applied Microbiology, Letters in Applied Microbiology 45 (2007) 36–41

37

Mercury bioremediation

S.G. Tupe et al.

Table 1 Factors and their levels used for L18 orthogonal array design Levels Control factors

1

2

3

A. Agitation B. Temperature (C) C. pH D. Hg (II) concentration (mg l)1) E. Carbon source (01% w ⁄ v) F. Ammonium sulfate concentration (mg l)1) G. Yeast extract concentration (mg l)1) H. Medium : flask volume ratio

Shaking 35 6 10

No shaking 30 7 20

Glucose

Sucrose

Glycerol

500

1000

1500

500

1000

1500

1:5

2:5

3:5

40 8 30

Table 2 Design of experiment by L18 orthogonal array (OA) for mercury removal

Expt. no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Control factors assigned to the columns A

B

C

D

E

F

G

H

Experimental Hg removal (%)

1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2

1 1 1 2 2 2 3 3 3 1 1 1 2 2 2 3 3 3

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

1 2 3 1 2 3 2 3 1 3 1 2 2 3 1 3 1 2

1 2 3 2 3 1 1 2 3 3 1 2 3 1 2 2 3 1

1 2 3 2 3 1 3 1 2 2 3 1 1 2 3 3 1 2

1 2 3 3 1 2 2 3 1 2 3 1 3 1 2 1 2 3

1 2 3 3 1 2 3 1 2 1 2 3 2 3 1 2 3 1

45Æ90 88Æ18 91Æ80 63Æ50 26Æ10 90Æ40 45Æ60 80Æ01 71Æ00 10Æ91 81Æ36 76Æ38 25Æ10 73Æ91 71Æ70 46Æ14 69Æ30 82Æ10

The columns in the OA indicate the factor and its corresponding levels, and each row in the OA constitutes an experimental run which is performed at the given factor settings.

for 24 h at conditions mentioned in Table 2. Uninoculated controls (18, corresponding to 18 experiments) were maintained under similar conditions. After 24 h incubation, aliquots (2 ml) were removed, centrifuged (8000 rev min)1, 10 min, 4C), and the supernatant was used for the estimation of residual Hg (II) by cold vapour atomic absorption spectrometry and mercury removal efficiency was calculated as described above. All the experiments were performed in duplicates. 38

The experimental data were processed by using Qualitek-4 software (Nutek Inc., MI, USA) for DOE using Taguchi approach. Using the software, individual influence of the factors at the assigned levels, Severity Indices for different interactions between factors and anova were calculated. Optimized process conditions were automatically calculated by the software based on pooled anova. The validation experiment was conducted under optimized process conditions. During this experiment, mercury removal and biomass concentration (OD600) were monitored at an interval of 4 h for 24 h. Optimized mercury removal efficiency was calculated after 24 h. Results In preliminary studies, it was observed that the culture could remove 70% Hg (II) from 10 mg l)1 initial concentration aerobically within 24 h. Hg removal was due to reduction to volatile Hg0 as colourless cuprous iodide strips turned red (due to formation of cuprotetraiodomercuriate). The results obtained for the Taguchi experiment are summarized in Table 2 in terms of % Hg removal from the medium in comparison to controls. As seen in the table, Hg (II) removal efficiency of Ps. aeruginosa under the 18 different experiments ranged between 10% and 92%. The influence of individual factors at the assigned levels on Hg (II) reduction is shown in Table 3. The table also indicates the relative influence of each factor (L2–L1) on Hg (II) removal. Here, larger (numeric) difference indicates stronger influence of the factor. The sign of the Table 3 Individual performances of the factors at the assigned levels (%) and the resulting influence of each factor on Hg (II) removal Factor

Level 1

Level 2

Level 3

L2–L1

Agitation Temperature pH Hg (II) Carbon source Ammonium sulfate Yeast extract Medium : flask volume ratio

66Æ94 65Æ76 39Æ53 67Æ13 69Æ88

59Æ66 58Æ45 69Æ81 57Æ24 70Æ99

– 65Æ69 80Æ56 65Æ53 49Æ04

-7Æ29 -7Æ30 30Æ29 -9Æ88 1Æ11

64Æ52

64Æ93

60Æ45

0Æ42

56Æ57 52Æ79

62Æ68 67Æ03

70Æ64 70Æ08

6Æ11 14Æ22

Out of the 18 experimental trials, the agitation factor has nine level one and nine level two conditions, whereas other seven factors have six conditions for all the three levels. The average performance of a factor at a particular level was calculated by taking the average of the results of its nine (for agitation) and six (for other factors) conditions. Column denoting L2-L1 shows the difference between the average values of each factor at these levels.

ª 2007 The Authors Journal compilation ª 2007 The Society for Applied Microbiology, Letters in Applied Microbiology 45 (2007) 36–41

S.G. Tupe et al.

Mercury bioremediation

difference (+ ⁄ )) indicates whether the change from one level to the other increased or decreased the result. It is evident that at individual level pH had most pronounced influence on Hg (II) removal followed by carbon source, medium : flask volume ratio and yeast extract concentration. Qualitek-4 software plots line graphs (not shown) for the average performance of two factors at their three levels and interprets the interaction graphs in terms of interaction severity index (SI%). The severity index is considered as 100% for perpendicular lines and 0% for parallel lines. Interaction severity indices for all the possible 28 interactions were thus calculated. Table 4 shows severity indices for ten most significant interactions between factors. These data indicate that factors such as agitation, ammonium sulfate concentration, temperature, Hg concentration (which have less influence on the process individually) showed higher interaction severity index in combination with other factors. Factors such as medium : flask volume ratio, yeast extract concentration which had significant impact individually also showed Table 4 Severity index for different interactions between factors* Interaction factor pairs No. (in order of severity) 1 2 3 4 5 6 7 8 9 10

Hg concentration · medium : flask volume ratio Carbon source · nitrogen source concentration Hg concentration · carbon source Carbon source · yeast extract concentration Agitation · nitrogen source concentration Agitation · Hg concentration Hg concentration · yeast extract concentration pH · nitrogen source concentration pH · carbon source pH · Hg concentration

Columns SI (%) Col§ Opt– 4·8

19Æ84

12

[3,3]

5·6

16Æ39

3

[2,1]

4·5

12Æ75

1

[2,2]

5·7

12Æ71

2

[1,3]

1·6

8Æ64

7

[1,2]

1·4 4·7

6Æ95 5Æ1

5 3

[1,3] [3,3]

3·6

4Æ89

5

[3,1]

3·5 3·4

1Æ91 0Æ92

6 7

[3,1] [3,3]

*Out of 28 different interactions, ten most important (with least severity indices) are showed. Columns – represent the column locations to which the interacting factors are assigned. SI – interaction severity index (100% for 90 angle between the lines, 0% for parallel lines). §Col. – shows column that should be reserved if this interaction effect were to be studied (2-L factors only). –Opt. – indicates the factor levels desirable for the optimum condition (based strictly on the first two levels). If an interaction is included in the study and found significant (in ANOVA), the indicated levels must replace the factor levels identified for the optimum condition without considerations of any interaction effects.

higher interaction severity index. Surprisingly, pH, which was shown to be the most important factor at individual level (Table 3), was found to have lower interaction severity index. anova with the percentage of contribution of each factor with interactions was calculated by Qualitek-4 based on the ratio of pure sum to the total sum of the squares. In this analysis, pH was found to be the most significant factor (50Æ76% contribution) in Hg (II) reduction, whereas temperature and ammonium sulfate concentration showed no contribution. The optimum conditions achieved based on anova (with individual percent contributions for performance enhancement) are as follows: shaking (3Æ64); 35C (2Æ46); pH 8 (17Æ26); Hg (II) concentration 10 mg l)1 (3Æ83); sucrose (7Æ69); ammonium sulphate 1000 mg l)1 (1Æ63); yeast extract 1500 mg l)1 (7Æ35) and medium : flask volume ratio 3 : 5 (6Æ78). The parameters were ranked based on contribution to the process performance enhancement. pH, carbon source, yeast extract concentration and medium : flask volume ratio were major contributors for the performance under optimum conditions. For predicting the performance under optimum conditions, factors with confidence level