Optimization and Implementation of Fed-Batch Strategy to Produce Ligninolytic Enzyme from the White-Rot Basidiomycete Pycnoporus sanguineus in Bubble Column Reactor

Abstract

The current work evaluates the production of ligninolytic enzyme optimization via response surface methodology using different inducers: acid cellulignin (CA); MnSO₄ (Mn²⁺); CuSO₄·5H₂O (Cu²⁺); veratryl (3, 4-dimethoxybenzyl); alcohol (VA); Tween 80% (T80); and the carbon-to-nitrogen ratio (C/N). A further goal was implementing a fed-batch strategy to produce ligninolytic enzyme extracts from P. sanguineus 2512 using a bubble column reactor (BCR). The best optimized experimental condition in the shake flasks was a 7.5 C/N ratio, 0.025 g/L Cu²⁺, 1.5 mM Mn²⁺, 3.0 mM VA and 0.025 mM T80, resulting in 64,580, 9.10 and 80.72 U/L for Laccase (Lac), Manganese (MnP) and Lignin peroxidase (LiP) activities, respectively. In the BCR, three feedings were performed at 24 h intervals on the 6th, 7th and 8th days with a significant increase in Lac (99,600 U/L) and MnP (47.53 U/L) activities on the 8th day and a reduction on the 9th day of cultivation. The LiP activity peak was achieved on the 5th day (416 U/L) of cultivation, decreasing thereafter. Enzyme cocktails concentrated in hollow fiber in the third cultivation batch showed contents of 4 × 10⁵ U/L, 220 U/L and 2.5 g/L for Lac, MnP and total proteins, respectively. The enzymatic cocktail with the highest LiP activity (1200 U/L) was obtained in the first batch. The results showed that the optimization of the biosynthesis of the ligninolytic enzymes provided satisfactory improvement in terms of Lac and MnP production per run.

1. Introduction

White-rot fungi are the most effective natural producers of oxidative enzymes, such as laccase (Lac, E.C. 1.10.3.2), lignin peroxidase (LiP, E.C. 1.11.1.14) and manganese peroxidase (MnP, E.C. 1.11.1.13), which have very low specificity and are ecologically sustainable and versatile catalysts [1,2,3]. These enzymes oxidize a wide array of organic and inorganic substrates, being of great interest for many industrial applications [3]. The global market for enzymes with industrial applications is estimated to be worth USD 7.48 billion by 2025 [3]. The current industrial pretreatment of lignocellulosic material produces compounds that can be pollutants or can result in process inefficiencies. Pretreatment with lignin-degrading enzymes overcomes several of these undesirable effects [2]. Ligninolytic enzymes are of great interest to industry and have a wide variety of biotechnological applications in the pulp and paper, biofuel, textile, food, pharmaceutical and cosmetic industries, as well as in waste treatment. The highly oxidative capacity of these ligninolytic enzymes enables them to have a key role in various industrial fields, particularly in the biodegradation of industrial wastes rich in phenolic and non-phenolic compounds, and can be an effective solution to the threat to the ecological balance and human life presented by recalcitrant environmental pollutants.

The main challenge for the mass production of enzymes from natural organisms is that current methods do not fulfill market demands because of their low yields and associated high costs [4]. A way to overcome these difficulties is suggested by Couto and Toca-Herrera [1], who stated that controlling the substrate feeding rate in a culture of these fungi in a bioreactor is a crucial factor in improving enzyme expression in these conditions. Performing a fed-batch strategy in which fresh portions of medium are introduced into the culture during the process operation can increase production capacity and the level of reactivity of the enzymes in the same run by avoiding catabolite repression and sugar-overflow metabolism. The present work has the goals of evaluating the production of ligninolytic enzyme optimization and the implementation of a fed-batch strategy to produce ligninolytic enzyme extracts from Pycnoporus sanguineus 2512.

2. Materials and Methods

2.1. Basidiomycete and Growth Medium

P. sanguineus 2512 was provided by the São Paulo Institute of Botany (CCIBt/SP) and maintained at 4 °C in Petri plates containing PDA and 0.2% peptone. Five mycelial discs were removed from the plates and inoculated in a 250 mL Erlenmeyer flask containing 100 mL of 5% glucose, 1% yeast extract and 0.5% peptone. Cultivation was carried out on a rotary shaker (200 rpm) at 30 °C. After 48 h of growth, 10% (v/v) of the pre-inoculum was transferred into Erlenmeyer flasks and an instrumented bioreactor. The culture medium was prepared according to Mandels and Reese [5], and supplemented with different concentrations of glucose (5.0–30.0 g/L); casein (1.0–3.0 g/L); CuSO₄·5H₂O (Cu²⁺) (0–0.05 g/L); acid celulignin (CA) (0–2.0 g/L); MnSO₄ (Mn²⁺) (0–3.0 mM); veratryl (3,4-dimethoxybenzyl); alcohol (VA) (0–3.0 mM); and Tween 80% (polyoxyethylene sorbitan mono-oleate) (T80) (0–0.05% v/v).

2.2. Experimental Design and Optimization for Maximum Production of Ligninolytic Enzymes by P. sanguineus 2512

Before carrying out the response surface methodology (RSM) studies, a factorial design 2⁶⁻² with three center points was used to evaluate the relative importance of the inducers (CA, VA, Cu²⁺, Mn²⁺ and T80) and the C/N ratio for ligninolytic enzyme production by P. sanguineus 2512 using submerged fermentation (

Table 1. Factors and levels used in design of experiment for ligninolytic enzyme production by P. sanguineus 2512 in submerged cultivation.

Factors	Minimum	Central Point	Maximum
	−1	0	+1
Cu2+ (g/L)	0	0.025	0.05
Mn2+ (mM)	0	1.5	3.0
VA (mM)	0	1.5	3.0
CA (g/L)	0	1.0	2.0
T80 % (v/v)	0	0.025	0.05
C/N ratio	5	7.5	10.0

). The second step was focused on optimizing these medium components for the maximal production of ligninolytic enzymes using a central composite design (CCD) with 47 experiments and five factors (both in an Erlenmeyer flask). An analysis of variance (ANOVA) was performed using the STATISTICA^® 13 software and the significance level considered was 5%.

2.3. Fed-Batch Cultivation in a Bubble Column and Tangential-Flow Filtration

The best conditions obtained in the experimental design were chosen for fed-batch cultivation in a bubble column reactor (BCR) (Biostat B, B. Braun Biotech International, Allentown, USA) inoculated with a 10% v/v ratio. Based on the experimental design, assays were scaled up in a bioreactor as follows: 17.5 g/L glucose, 2.0 g/L casein, 0.025 g/L Cu²⁺, 1.5 mM Mn²⁺, 3.0 mM VA and 0.025 mM T80. The BCR operational approach used were 2 L, 0.667 vvm, 30 °C, and the initial pH was 5. The feeding of the BCR started after the carbon source had been consumed in a previous simple batch, which corresponded to enzyme with higher the first peak of Lac activity (the extracellular secretion) and happened on the 6th day of fungus cultivation. A set of repeated fed-batch cultivations was carried out, each one with the identical steps of removing 500 mL of culture medium and supplementing the BCR with 500 mL of fresh medium. In all cases, culture medium samples of 50 mL were collected once per day under aseptic conditions to measure the biomass, substrate and product concentrations. Lac, MnP and LiP activity were determined from the oxidation of 2,2-azino-bis-ethylbenzthiazoline (ABTS), phenol red and veratryl alcohol according to Wolfenden and Willson [6], Kuwahara et al. [7] and Tien and Kirk [8], respectively. For all activities, one activity unit (U) was the amount of enzyme that oxidized 1 μmol of substrate/min and was expressed in U/L. The total concentration of sugars and the total protein content (PTN) were determined according to Miller [9] using 3,5-dinitrosalicylic acid, and Bradford [10] using bovine serum albumin as standard, respectively. Tangential-flow filtration experiments were performed using polysulfone hollow fiber columns (QuixStand, benchtop system, GE Healthcare). First, the raw extract was subjected to microfiltration with a 0.22 µm-pore-size membrane at a maximum pressure of 5 psi at the entrance of the column, in order to eliminate cellular residues and particulate materials. Subsequently, the filtrate enzymatic extract was submitted to ultrafiltration using a membrane of 10 kDa (maximum pressure of 10 psi).

3. Results and Discussion

The Pareto chart from the factorial design revealed that the C/N ratio was statistically significant (p level ≤ 0.05) and was the most relevant factor that positively influenced all the evaluated responses. Both the carbon and nitrogen sources were critical for increasing enzyme secretion and enzyme concentration. Schneider et al. [11] conducted an optimization study with Marasmiellus palmivorus VE111 and reported that the best carbon and nitrogen sources were glucose and casein, respectively. The authors also reported that there was a significant increase in Lac secretion at low concentrations of glucose and medium concentrations of casein, which corroborates the results found herein. The most effective inducers of Lac production by fungi were 2,5-xylidine and copper (CuSO₄) [12]. Copper acts as a cofactor in the catalytic center of Lac, and minimal amounts of this metal are sufficient to increase the production of these enzymes [12,13]. For LiP and MnP, only a few studies report the inductive action of Cu²⁺. However, Vrsanska et al. [13] studied, in detail, the inductive effect of copper and CuSO₄ complexes on five different white-rot fungi and found improvement in the enzymatic secretion of Lac, LiP and MnP activities.

Once the CA factor was not significant for MnP and marginally significant for Lac (Figure 1), it was removed from the study. Additionally, a statistically significant curvature was found in all of the addressed cases, signaling that the experimental design could be optimized by using CCD, which was performed next.

The main effects, linear (L) factors, quadratic (Q) factors and their interactions, resulting from the ANOVA table for each of the Lac, MnP and LiP activity response variables, are summarized in

Table 2. Influence of linear (L) and quadratic (Q) factors for relative activities of Lac, LiP and MnP by p level from ANOVA.

Factors	Relative Lac (%) R² = 0.8788; R-Adj. = 0.81416	Relative LiP (%) R² = 0.68323; R-Adj. = 0.47959	Relative MnP (%) R² =0.77506; R-Adj. = 0.6432
C/N	(L) and (Q) p ≤ 0.05	(L) and (Q) p ≤ 0.05	(L) and (Q) p ≤ 0.05
Cu2+	(L) and (Q) p ≤ 0.05	(L) and (Q) p ≤ 0.05	(L) and (Q) p ≤ 0.05
Mn2+	(L) and (Q) p ≤ 0.05	(L) p ≤ 0.05	(L) p ≤ 0.05
VA	(L) and (Q) p ≤ 0.05	(L) and (Q) p > 0.05	(L) p > 0.05 and (Q) p ≤ 0.05
T80	(L) p > 0.05 and (Q) p ≤ 0.05	(L) and (Q) p ≤ 0.05	(L) p ≤ 0.05 and (Q) p > 0.05

* p > 0.05, not statistically significant; p ≤ 0.05, statistically significant.

. The linear and quadratic C/N ratio factors significantly influenced the enzymatic secretion of Lac, LiP and MnP, followed by Cu²⁺ concentration. The statistical significance of the model for the dependent variables studied was evaluated using the F test (Fisher’s test) by analysis of variance; when the p values are lower than 0.05, they indicate that the model terms are significant. The coefficient of determination R² values were approximately 0.88, 0.68 and 0.77, indicating that the model can explain 88, 68 and 77% of total variation around the mean for Lac, LiP and MnP, respectively (Table 2). These low values may be acceptable for biological experiments. Table 2 shows that the factor VA was not necessary for LiP; all the other factors were significant. The inductive effect of Mn²⁺ on MnP expression is well understood [14]. However, considering the (L) and (Q) effects and their interactions in this study (Table 2), the presence of this ion was shown to be more significant for Lac expression and relatively equal among the peroxidases. Furthermore, the inductive action of T80 was also investigated, and it was detected in the following order of influence as a function of the enzymes studied: LiP (L, Q and interactions effects) > MnP (L and interactions effects) > Lac (Q and interactions effects) (Table 2). These results agree with those of Bettin et al. [15], who reported low influence on the induction of laccases by the action of T80. Regarding the presence of VA, the order of influence was Lac > MnP > LiP, considering the L, Q and interaction effects.

Figure 2 illustrates the response surfaces for the dependent variables Lac, MnP and LiP as a function of Cu²⁺ and C/N, with Mn²⁺, VA and T80 fixed at the center point (1.5 mM, 3.0 mM and 0.025% v/v, respectively).

Figure 2a shows that the higher the Cu²⁺ concentration and median C/N ratios, the higher the Lac secretion. The optimum region for this enzyme lies between 6 and 7 for C/N and is equal to or greater than 0.05 g/L for Cu²⁺. For LiP and MnP, there was no strictly defined optimum region, as can be observed in Figure 2b,c. However, these enzymes most expressive secretion values are at high Cu²⁺ concentrations (≥0.05 g/L). Regarding the C/N ratio, the highest LiP activity values range from 5 to 9 (Figure 2b), while MnP is increased at lower C/N values (4–8) (Figure 2c). The maximum secretion of Lac activity was predominant over LiP and MnP activity, with the following values: 64,580, 9.10 and 80.72 U/L, respectively. The predicted values according to the experimental-condition-optimized region (7.5 C/N ratio, 0.025 g/L Cu²⁺, 1.5 mM Mn²⁺, VA 3.0 mM and T80 0.025% v/v) and the results for the Lac, MnP and LiP relative activities from the experiments applying the predicted optimal conditions are shown in

Table 3. Validation of the optimum conditions predicted for Lac, MnP and LiP relative activities.

Factors	Confidence Limit		Predicted Value	Experimental Result
	−95%	+95%		Optimized Medium	Non-Optimized Medium
Relative Lac (%)	60.93	81.58	71.22	79.01	7.43
Relative MnP (%)	26.11	37.65	31.88	34.15	23.06
Relative LiP (%)	20.10	29.96	25.03	22.42	12.45

. According to these values, the experimental results were inside the confidence limits, which means that these findings were in agreement with the model. After optimization using the response surface methodology, the Lac relative activity increased 10.6 times (from 7.43 to 79.01), MnP relative activity increased 1.48 times (from 23.06 to 34.15), and LiP relative activity 1.8 times (from 23.06 to 34.15) when compared to the results from the non-optimized medium.

Therefore, this experimental condition was used for the fed-batch experiments in the BCR reactor. The results of P. sanguineus 2512 cultivation in the fed-batch culture are presented in Figure 3.

As noted, the kinetic profile in the BCR confirms the results previously obtained in the Erlenmeyer flasks, showing that the biosynthesis of ligninolytic enzymes is strictly dependent on the composition of the culture medium, mainly the C/N ratio [11,16]. High mycelial growth (6 g/L) and a marked pH (3.2) and OD (20%) reduction occurred on the 3^rd day of cultivation, as shown in Figure 3a. The increase in OD is probably associated with the change in metabolism and the beginning of the stationary phase of mycelial growth from the fourth day onwards. Only a tiny amount of Lac and MnP was produced during this cultivation phase, in contrast to the behavior of LiP, which exhibited a remarkable increase (Figure 3b). Among the three target enzymes, Lac exhibited the most prominent secretion, with significant production starting on day 4 of cultivation (Figure 3b). On day 6, batch feeding was initiated with the removal of 500 mL of secretome cells and the insertion of fresh medium with low C/N concentrations (500 mL). Three feedings were performed at 24 h intervals on days 6, 7 and 8, with a significant increase in Lac (99,600 U/L) and MnP (47.53 U/L) activities on the 8th day, and reduction thereafter until the culture stopped on the 9th day. The LiP secretion showed a kinetic profile with peak activity on the 5th day (416 U/L) and a decrease from the 6th day onwards.

Batch feeding with low and non-repressive amounts of glucose increased the production of Lac and MnP 1.25- and 3-fold, respectively, when compared to the culture without feeding, considering that, in a simple batch, the production of Lac finished on the 6th day of culture, as reported in our previous study [17]. The implementation of this strategy was not efficient in increasing LiP, since there was a 2.6-fold reduction after the start of feeding (Figure 3b). The reduction in LiP activity may be associated with the physiological modification characteristics of a shift from primary to secondary metabolism, which is activated in low concentrations of carbon sources. Antecka et al. [16] report that Lac activity has secondary metabolism, and the enzyme is secreted in high amounts when the carbon source is very low. The results obtained in the BCR are in agreement with the CCD results (Figure 2): once increasing, the C/N ratio allows greater LiP secretion. On the other hand, reduced glucose concentration and its maintenance at low levels (5 g/L) during batch feeding favors Lac secretion, but not LiP secretion.

Figure 3 shows that P. sanguineus 2512 has a remarkable capacity for Lac secretion since the cocktail isolated in this study is composed mainly of this enzyme, which exhibits expressly higher activity compared to LiP and MnP. Georris et al. [18] reported that this basidiomycete is a model organism for Lac production among many laccase-producing fungi. The results found herein are in accordance with those reported in the literature, especially when the carbon source comes from lignocellulosic residues, which also act as inducers [4,12,18]. There are no studies in the literature that report the production of LiP and MnP by the Picnoporus genus in a fed-batch culture. However, some studies corroborate the present work regarding Lac production by basidiomycetes. Galhaup et al. [19] found that when operating in a fed-batch culture, laccase production by Trametes pubescens increased 2-fold, achieving remarkably high Lac activity (740,000 U/L). Antecka et al. [16] implemented different cultivation strategies with Cerrena unicolor. These authors concluded that a fed-batch culture with the partial removal of spent medium and the addition of fresh medium substantially increased the fungus-secreting potential (up to 12,000 U/L) and final Lac recovery (118,650 U) compared to the simple batch operation, which resulted in 2200 U/L and 19,560 U, respectively.

Figure 4 shows the Lac, LiP, MnP and PTN values after tangential filtration and concentration in a hollow fiber system. The filtration was efficient, retaining values above 90% for all the enzymatic species in all three batches.

As observed, after the tangential filtration steps (microfiltration and ultrafiltration), four cocktails highly rich in enzymes of the ligninolytic complex were obtained. The enzymatic activities of Lac and MnP were maximal for the third batch, while for LiP the highest enzymatic content was recorded in the first batch. The adoption of the fed-batch method proved to be efficient and allowed the obtaining of a cocktail with high enzyme contents, whose final recovery balance reached 6.8 × 10⁵, 1.4 × 10⁴ and 3.5 × 10³ U for Lac, MnP and LiP, respectively.

Table 4. Kinetic parameters by enzymatic production of P. sanguineus 2512 in 2L-BCR yield factors, productivity and specific activity.

Kinetic Parameters	1 SB	2 FB 1	FB 2	FB 3
Lac activity (× 104 U/ L)	8.02	9.01	9.96	6.67
MnP activity (U/L)	15.00	26.00	45.00	27.00
LiP activity (U/L)	358.00	254.00	136.00	81.00
Time (days)	6.00	7.00	8.00	9.00
Residual glucose (g/L)	3.80	1.80	1.80	1.80
Mycelial biomass X (g/L)	6.66	6.00	6.06	4.60
Lac volumetric productivity (× 103 U/L∙d)	13.37	12.87	12.45	7.41
MnP volumetric productivity (U/L∙d)	2.50	3.71	5.62	3.00
LiP volumetric productivity (U/L∙d)	59.67	36.28	17.00	9.00
Substrate yield on Lac YE/S (× 104 U/g)	2.11	5.00	5.53	3.70
Substrate yield on MnP YE/S (U/g)	3.95	14.44	25.00	15.00
Substrate yield on LiP YE/S (U/g)	94.21	141.11	75.55	45.00
Substrate yield on biomass YX/S (g/g)	0.57	0.30	0.30	0.39
Biomass volumetric productivity PX (g/L∙d)	1.11	0.85	0.75	0.51
Lac specific activity SALac (× 104 U/g of biomass)	1.20	1.50	1.64	1.45
MnP specific activity SAMnP (U/g of biomass)	2.25	4.33	7.42	5.86
LiP specific activity SALiP (U/g of biomass)	53.75	42.33	22.44	17.60

¹ SB: simple batch; ² FB: fed-batch

shows the kinetic parameter values of the fed-batch culture for the four batch cycles. The best volumetric productivity of Lac and LiP was recorded in the first batch, with values of 13,373.67 and 59.67 U/L∙d, respectively. On the other hand, the volumetric productivity of MnP increased 2.25 times in the third batch, reaching a value of 5.62 U/L∙d. Maintaining the glucose concentration at approximately 5 g/L may have provided a higher substrate yield (Y_E/S) and specific activities (SA) of Lac and MnP to the maximum mycelial biomass for each run, which are important factors in the industrial scale-up process.

4. Conclusions

The results obtained in this study show that the concentrations of carbon, nitrogen and inducers affect the secretion of ligninolytic enzymes produced by P. sanguineus 2512. The best condition for the three enzymes treated in this study, in shake flasks, was a C/N ratio of 7.5, Cu²⁺ 0.025 g/L, Mn²⁺ 1.5 mM, VA 3.0 mM and T80 0.025% (v/v), showing values of 64,580, 9.10 and 80.72 U/L for Lac, MnP and LiP activities, respectively. Cultivation in the BCR operating in a fed-batch culture with three feeding cycles resulted in maximum Lac and MnP activity of approximately 1 × 10⁵ and 45 U/L, respectively, on the 8th day of cultivation. For LiP, the maximum secretion was achieved on the 5th day of cultivation (416 U/L), decreasing from the 6th day onwards. The fed-batch implementation applied in this research to cultivate P. sanguineus 2512 resulted in a concentrated cocktail with high Lac, MnP and LiP activity recovered per run (6.8 × 10⁵, 1.4 × 10⁴ and 3.5 × 10³ U, respectively) and allowed the prolongation of stable mycelial activity.