MATERIALS AND METHODS

Microorganism
The bacterium used throughout this study is a new Egyptian soil isolate. Soil samples were suspended in sterile water in a ratio of 1:1. Three sequential enrichment steps were carried out by inoculating 1 mL of the soil suspension to yeast peptone medium supplemented with 1% carboxymethyl cellulose (CMC) under shaking conditions at 55°C. An inoculum from each enrichment step was streaked onto agar plates of the same medium for further purification. After 24 and 48 h of incubation at 50°C, clear zones were developed by spreading iodine solution on the surface of each culture. A single isolated colony was selected and the bacterium was identified by 16S rRNA gene sequencing, as previously described (El-Helow, 2001) and blasting the results in the GenBank.

Enzyme Production Conditions
Cultures were allowed to grow in 250 mL Erlenmeyer flasks containing 25 mL aliquots of fermentation basal medium of the following composition (g L^-1): Avicel, 20; yeast extract, 0.5; (NH₄)₂SO₄, 2; MgCl₂.6H₂O, 0.2; CaCl₂, 0.1; K₂HPO₄, 2 and KH₂PO₄, 0.6 (Malfait et al., 1984). Cultures were incubated at 55°C under shaking (250 rpm) or stagnant. Standard overnight seed cultures, prepared in nutrient broth, were used to inoculate fermentation cultures as 1% inoculum. Bacterial growth was monitored turbidimetrically. Turbidity readings were converted to dry weights using a standard curve. Samples were collected after indicated incubation times and cells were isolated by centrifugation at 10,000 rpm for 10 min and the supernatants were separated and stored at -20°C for few hours until the necessary analyses were carried out.

Enzyme Assay
A solution of 0.125% avicel (from Cetraria Islandica-Sigma, USA) dissolved in phosphate buffer of pH 5 was used as a substrate. The reaction was performed by adding 0.5 mL substrate to 0.5 mL diluted enzyme and incubation in a water bath at 55°C for half an hour. The reducing sugars produced were measured by the di-nitrosalisylic acid method (Miller, 1959). An enzyme activity unit is defined as the amount of enzyme producing 1 mg mL^-1 reducing sugars (glucose equivalent) 1 h^-1.

Table 1:	Plackett-Burman experimental design for elucidation of factors affecting avicelase production
Natural levels of variables (g L-1) are presented between brackets

Statistical Designs
Plackett-Burman Design
As a preliminary optimization experiment, various medium components and environmental factors have been evaluated. Based on the Plackett-Burman factorial design, each factor was examined in two levels: -1 for a low level and 1 for a high level (Plackett and Burman, 1946). Table 1 shows the factors under investigation as well as levels of each factor used in the experimental design.

The Plackett-Burman experimental design is based on the first order model:

Where Y is the response (avicelase activity), β₀ is the model intercept, β_I is the linear coefficient and x_i is the level of the independent variable. In the present study, seven assigned variables were screened in eight experimental designs. All experiments were carried out in duplicates and the averages of avicelase activity results were taken as the response. Variables whose confidence levels were 90% or higher were treated as significant variables.

Box-Behnken Design
In order to describe the nature of the response surface in the experimental region, a Box-Behnken design (Box and Behnken, 1960) was applied. Table 3 represents the design matrix of a 13 trials experiment. Using this design, factors of highest confidence levels were prescribed into three levels, coded –1, 0 and +1 for low, middle and high concentrations (or values), respectively. For predicting the optimal point, a second order polynomial function was fitted to correlate relationship between independent variables and response (avicelase activity). For three factors the equation is:

Where Y is the predicted response, β₀ is the model constant; X₁, X₂ and X₃ are independent variables; β₁, β₂ and β₃ are linear coefficients; β₁₂, β₁₃ and β₂₃ are cross product coefficients and β₁₁, β₂₂ and β₃₃ are the quadratic coefficients.

Experiments were performed in triplicates and mean values are given. Microsoft Excel was used for the regression analysis of the experimental data obtained. The fitting quality of the polynomial model equation was expressed by the coefficient of determination R². The optimal value of enzyme activity was estimated using the solver function of Microsoft Excel tools.

RESULTS AND DISCUSSION

Isolation and Characterization of a Novel Avicelase Producing Bacterium
Based on clear zone diameters on carboxy methyl cellulose (CMC) supplemented plates, the isolate WL1 was selected out of a group of purified bacterial colonies. Microscopic examinations revealed that the bacterium is a Gram-positive rod-shaped endospore former.

Table 2:	Statistical analyses of the Plackett-Burman experimental results

Table 3a:	Natural values of variables examined in the two sequential optimization trials

Table 3b:	The Box-Behnken factorial experimental design

For molecular characterization, genomic DNA of the WL1 isolate was prepared, purified and used for PCR amplification of the gene encoding 16S rRNA using universal primers. Double-strand nucleotide sequences of two fragments (536 and 746 base pairs) were determined and edited. Alignment of these sequences (accession numbers AY 738088 and AY 738089), using Biology WorkBench, showed 99.8% identity to the homologous fragments of G. stearothermophilus. To the best of our knowledge, it is the first time to investigate the celluolytic activity of this species.

Using the basal enzyme production medium, preliminary studies were carried out on WL1 with respect to culture initial pH, temperature and shaking. Enzyme assay results of these experiments revealed that the bacterium expresses a promising level of avicelase activity (0.425 U mL^-1) in a shake culture with an initial pH of 7 and an incubation period of 16 h at a temperature of 55°C. Under this culture condition, a maximum specific growth rate (0.94 h^-1) was recorded within 24 h of incubation. In a parallel study (data not shown), the partially purified enzyme showed high thermal stability, where it retained 80% of its activity when exposed to 60°C for three hours. The enzyme lost 50% of its activity at 80°C. Optimum enzyme temperature and pH were 60°C and 6.0°C, respectively.

Evaluating the Significance of Medium Constituents
To test the significance of seven culture variables at a time on avicelase formation by the isolate, The Plackett-Burman experimental design was applied. Table 2 represents the design matrix together with the response observations.

Fig. 1:	Effect of nutritional factors on avicelase activity produced by G. stearothermophilus in shaked culture at 55°C based on results of Plackett-Burman design

The linear polynomial model describing the correlation between examined variables and expressed avicelase activity could be presented as follows:

Based on the analysis of regression coefficients of the seven variables as well as the calculated main effect figures presented graphically in Fig. 1, avicel, yeast extract, magnesium chloride, calcium chloride and dipotassium hydrogen phosphate showed positive effects on the expression of avicelase. On the other hand, ammonium sulfate and potassium dihydrogen phosphate contributed negatively. As shown in Table 2, statistical analyses of the results suggest that yeast extract is the most significant factor affecting avicelase expression by WL1.

The presence of yeast extract as a nitrogen source in the medium stimulates the formation of readily metabolizable compounds that can be utilized for growth and avicelase formation (Mawadza et al., 1996). On the other hand, low ammonium sulfate concentrations and high avicel levels, within the tested constrains, also enhanced the expression of avicelase. In accordance, it has been demonstrated by Bok et al. (1998) that low ammonium sulfate and high avicel levels stimulate the activities of the promoters controlling cellulase genes in Thermotoga neapolitana.

According to the results of this experiments, a medium of the following composition (g L^-1) is predicted to be near optimum: avicel, 25; yeast extract, 1; (NH₄)₂SO₄, 1.5; MgCl₂.6H₂O, 0.3; CaCl₂, 0.15; K₂HPO₄, 1.8 and KH₂PO₄, 0.7. A verification experiment demonstrated that the enzyme activity expressed in this medium was increased by 1.18 fold (0.5 U mL^-1) when compared to the basal condition.

Fig. 2:	Avicelase activity (U mL-1) response surface from G. stearothermophilus as affected by culture conditions (first optimization set)

Fig. 3:	Avicelase activity (U mL-1) response surface from G. stearothermophilus as affected by culture conditions (second optimization set)

Optimization of Avicelase Production
In this phase of optimization, the most effective variables were further investigated according to the Box-Behnken factorial design (Box and Behnken, 1960). The variables applied were; avicel (X₁), yeast extract (X₂) and (NH₄)₂SO₄ (X₃). Levels of the other medium components were selected based on their main effect signs and treated as constant factors throughout the optimization experiment. Table 3 represents the design matrix, which comprises thirteen trials, accompanied with avicelase activity in two sequential optimization sets with different X₁, X₂ and X₃ levels.

As presented graphically in Fig. 2, the correlation presents a saddle where optimal value could not be found. However, this optimization step reflected an image how the response could be increased and as a result, other levels chosen in the direction of increasing the response were also examined. The second optimization set was run according to the design matrix shown in Table 3.

Again, the response observations presented graphically in Fig. 3, demonstrate a saddle formation. Further application of other levels chosen in the direction of increasing the response in a third experiment decreased the activities when compared to the responses obtained in the second optimization set. Accordingly, the experimental results of the second optimization set were handled and fitted by regression to a polynomial equation as follows:

The quality of the fit of this polynomial model equation was expressed by R, the multiple correlation coefficient and R², the coefficient of determination. The closer the value of R to 1, the better is the correlation between the observed and the predicted values. In the second optimization experiment, the value of R = 0.99. This value indicates high degree of correlation between the experimental and the predicted values. The value of determination coefficient R² = 0.98, being a measure of goodness of fit of the model, indicates that about 2% of the total variations are not explained by the activity models.

The optimal levels of the three examined independent variables as predicted from the model are (g L^-1): Avicel, 42; yeast extract, 3.6 and ammonium sulfate, 0.8 with a predicted avicelase activity of 0.79 (U mL^-1). Although, otherwise, Krishna (1999) reported that cellulose production in Bacillus subtilis CBTK 106 increases by increasing the concentration of ammonium sulfate, it is worthwhile to mention that the increment was not significant. On the other hand, Enari and Markenan (1977) showed that good production of cellulase was obtained with organic nitrogen sources like yeast extract and peptone.

In order to verify the optimization results, an experiment was carried out under the predicted optimal condition, where the dry weight and avicelase activity were monitored at different time intervals. The basal culture medium was used as a control. As shown in Fig. 4, the optimized medium recorded a relatively higher exponential growth rate. With respect to enzyme expression this formula showed an actual maximum activity of 0.8 U mL^-1, which is 101% of the predicted value with a 1.89 fold increase when compared to the pre-optimized medium control culture. The great similarity between predicted and observed results in this application and others (Bloor and England, 1991; Teruel et al., 1997; Kimmel et al., 1998) confirms the high accuracy and applicability of the Box-Behnken model in the optimization process.

It is also likely that the optimized culture condition accelerated the rate of avicelase expression in WL1 as it showed maximum enzyme activity within 12 h instead of 16 h in the case of the control culture. According to these results, a medium of the formula (g L^-1): avicel, 42; yeast extract, 3.6; (NH₄)₂SO₄, 0.8; MgCl₂.6H₂O, 0.3; CaCl₂, 0.15; K₂HPO₄, 1.8 and KH₂PO₄, 0.7 is suggested for efficient production of avicelase by G. stearothermophilus WL1.

Fig. 4:	Growth (Δ) and avicelase activity (Ο) of G. stearothermophilus on control culture and growth (▲) and avicelase activity (●) in the optimized culture at optimum concentrations as obtained from the validation experiment

Fig. 5:	Effect of different cellulosic substrates on the expression of avicelase by G. stearothermophilus

This optimized medium contains an avicel concentration of more than twice that in the basal medium suggesting that avicelase expression in this bacterium is possibly regulated by induction.

Effect of Different Cellulosic Substrates on Avicelase Production by G. stearothermophilus
An important fermentation consideration is the cost of the carbon source, which contributes significantly to the overall production costs (Li et al., 1998). Also, it has been reported that the type of cellulose has an important role for the production of active cellulases (Li et al., 1997). Accordingly, a group of low cost cellulosic materials were individually examined to replace avicel in the optimized medium. The tested substrates namely; raw cotton fibers, Whattman filter paper No.4, cane bagasse, rice straw, rice polish, barely grains, corn stems, bran, corn cobs and wheat straw, were added separately, at carbon equivalent weights, to the optimized medium formula. Cultures were incubated for 12 h at 55°C under shaking conditions. Clear culture filtrates recorded enzyme assay results that ranged between 0.40 up to 0.83 U mL^-1. The observed differences in enzyme production levels could be attributed to structural factors, which affect the expression of individual enzymes of the total cellulase enzyme system (Krishna, 1999). As shown in Fig. 5, the most promising substrates are raw cotton fibers and filter papers, which attained avicelase activities of 104 and 101.3% of the optimized pure avicel culture, respectively.