Optimization of Large-Scale Culture Conditions for the Production of Cordycepin with Cordyceps militaris by Liquid Static Culture

Cordycepin is one of the most important bioactive compounds produced by species of Cordyceps sensu lato, but it is hard to produce large amounts of this substance in industrial production. In this work, single factor design, Plackett-Burman design, and central composite design were employed to establish the key factors and identify optimal culture conditions which improved cordycepin production. Using these culture conditions, a maximum production of cordycepin was 2008.48 mg/L for 700 mL working volume in the 1000 mL glass jars and total content of cordycepin reached 1405.94 mg/bottle. This method provides an effective way for increasing the cordycepin production at a large scale. The strategies used in this study could have a wide application in other fermentation processes.


Introduction
Cordyceps militaris is an entomopathogenic fungus belonging to Ascomycota, Sordariomycetidae, Hypocreales, and Cordycipitaceae [1] and is one of the most important traditional Chinese medicinal mushrooms. Cordyceps militaris is the type species of Cordyceps, which internally parasitizes larva or pupa of lepidopteran insects and forms fruiting bodies on their insect hosts. Cordyceps militaris has long been recognized as a desirable alternative for natural Ophiocordyceps sinensis [2] as it has been given Chinese Licence number Z20030034/35. This is because the gathering of Ophiocordyceps sinensis is causing substantial reductions in populations [3]. Cordyceps militaris produces many bioactive compounds, including polysaccharides, cordycepin, adenosine, amino acid, organic selenium, ergosterol, sterols, cordycepic acid, superoxide dismutase (SOD), and multivitamins [4,5].
In previous work, cordycepin has been synthesized by chemical [14,15] and microbial fermentation using C. militaris [6] or Aspergillus nidulans [16,17]. Solid-state fermentation [18,19], submerged culture [4,[20][21][22][23][24], and surface liquid culture [25][26][27] have been used in microbial fermentation of cordycepin. Cordycepin obtained through chemistry 2 The Scientific World Journal pathways is hard to purify, and the cost is much higher than through biology fermentation. Thus a major need is to improve the biology methodology [28]. Fermentation time is too long and is difficult to achieve large scale production via solid-state fermentation [18,19]. Productivity is generally low, the costs are high, and fermentation processes are easily contaminated in submerged culture in large fermenters [4,20,21,29]. Productivity in surface culture techniques is higher as compared to other methods [29,30] and the cost is lower [23]. New technologies, such as space mutation treatment and high-energy ion beam irradiation, have been used to obtain better Cordycepin producing, novel mutants of C. militaris. The resulting mutants were higher cordycepin produces, than the wild strain [30,31]. Bu et al. [20] reported that the cordycepin in C. militaris was substantially increased by the elicitor of Phytophthora sp. Research result showed that glucose and yeast extract were effective media components for improved cordycepin production by C. militaris [32,33]. There have been other studies using different culture conditions [21,24,25,32], culture medium, and additives [4, 22-24, 26, 27] for the production of cordycepin via liquid culture. However, as far as we know, these reports studied cordycepin production in 250 mL or 500 mL Erlenmeyer flasks, and there have been no reports to improved cordycepin production using static liquid culture in 1000 mL glass jars. The latter process is a good way to scale up large scale cordycepin production from the laboratory to industry.
In this study, the effects of working volume, carbon sources, nitrogen sources, inorganic salts, growth factor, nucleoside analogue, and amino acid additions were studied in order to improve the cordycepin production by static liquid culture of C. militaris (strain CGMCC2459) in 1000 mL glass jars. The results suggested that the optimization medium conditions were helpful for improved large scale cordycepin production.

Microorganism and Seed
Culture. The isolate of C. militaris (strain CGMCC2459) used in the present study was collected from Mt. Qingcheng in Sichuan Province, China. The microorganism was maintained on potato dextrose agar (PDA) slants. Slants were incubated at 25 ∘ C for 7 days and then stored at 4 ∘ C. The seed culture was grown in a 250 mL flask containing 70 mL of basal medium (sucrose 20 g/L; peptone 20 g/L; KH 2 PO 4 1 g/L; and MgSO 4 ⋅7H 2 O 0.5 g/L) at 25 ∘ C on a rotary shaker incubator at 150 rev/min for 5 days [24].

Basal Medium and Static
Culture of Glass Jars. The basal medium composition for the fermentation was as follows: sucrose 20 g/L; peptone 20 g/L; KH 2 PO 4 1 g/L; and MgSO 4 ⋅7H 2 O 0.5 g/L. The pH was not adjusted, followed by autoclaving for 30 min on the 121 ∘ C. The static culture experiments were performed in 1000 mL glass jars (inner diameter 110 mm, height 150 mm) containing basal medium after inoculating with 10% (v/v; the biomass dry weight of seed culture is 54 mg/mL) of the seed culture. The culture was incubated at 25 ∘ C without moving for 35 days, and samples were collected at the end of the fermentation from the glass jars for analyzing biomass dry weight and cordycepin production.

Static Culture
Conditions. The effects of factors affecting cell growth and the production of cordycepin by C. militaris were studied using a one-factor-at-a-time method for static culture. The effects of carbon sources on cordycepinproduction were studied by substituting carbon sources such as sucrose, lactose, soluble starch, and dextrin for glucose at 25 ∘ C for 35 days. Effects of nitrogen sources (yeast extract, beef extract, NH 4 NO 3 , NaNO 3 , NH 4 Cl, casein, and carbamide) and inorganic salts ( were also studied using static culture. Growth factors (Vitamin B 1 (VB 1 ), Vitamin B 6 (VB 6 ), Vitamin B 7 (VB 7 ), Vitamin B 11 (VB 11 ),naphthylacetic acid (NAA), 3-Indoleacetlc acid (IAA), and 2,4-dichlorophenoxyacetic acid (2,4-D)) were supplemented for 10 mg/L in basal media. Nucleoside analogues (1 g/L) and amino acids (8 g/L) established in our previous study [23] as an initial concentration were separately added to the optimal concentration of carbon and nitrogen source, inorganic salts, and growth factors and cultivated at 25 ∘ C for 35 days. All experiments were carried out at triplicate, and mean of results is presented.

Analytical Methods.
Samples collected at 35 days from the glass jars were centrifuged at 2810 ×g for 20 min. The mycelium at the bottom of tubes was washed sufficiently with a large amount of distilled water and dried to a constant dry weight at 55 ∘ C.
For analysis of extracellular cordycepin, the resulting culture filtrate was obtained by centrifugation at 2810 ×g for 20 min. The supernatant was filtered through a 0.45 m membrane and the filtrate was analyzed by HPLC (1100 series, Agilent Technology, USA). Accurate quantities of cordycepin (Sigma, USA) were dissolved in distilled water, to give various concentrations for calibration. The mobile phase was 10 mmol/L KH 2 PO 4 , which was dissolved in methanol/distilled water (6 : 94). Elution was performed at a flow rate of 1.0 mL/min with column temperature at 45 ∘ C and UV wavelength of 259 nm. Mean values were computed from triplicate samples.

Plackett-Burman
Design. The Plackett-Burman design, an effective technique for medium-component optimization [35,36], was used to select factors that significantly influenced hydrogen production. Sucrose ( 1 ), peptone ( 2 ), K 2 HPO 4 ⋅3H 2 O ( 4 ), MgSO 4 ⋅7H 2 O ( 5 ), and VB 1 ( 6 ) were investigated as key ingredients affecting cordycepin production. Based on the Plackett-Burman design, a 15-run was applied to evaluate eleven factors (including two virtual variables). Each factor was prepared in two levels: −1 for low level and +1 for high level. Table 1 illustrates the variables and their corresponding levels used in the experimental design. The values of two levels were set according to our preliminary experimental results. The Plackett-Burman design and the response value of cordycepin production are shown in Table 2.  [35,36]. In this study, the basic nutrient (carbon sources, nitrogen sources, inorganic salts, and growth factors) and additives (amino acid, nucleoside analogue) were studied for cordycepin production using static liquid culture. In the first test, a three-factor, five-level central composite design with 20 runs was employed. Tested variables (sucrose, K 2 HPO 4 ⋅3H 2 O, and MgSO 4 ⋅7H 2 O) were denoted as 1 , 4 , and 6 , respectively, and each of them was assessed at five different levels, combining factorial points (−1, −1), axial points (−1.6818, +1.6818), and central point (0), as shown in Table 3. Based on the above results, another test, a threefactor, five-level central composite design with 20 runs was employed. Tested variables (amino acid, nucleoside analogue, and culture time) were denoted as , , and , respectively, and each of them was assessed at five different levels, combining factorial points (−1, +1), axial points (−1.6818, +1.6818), and central point (0), as shown in Table 4.

Statistical Analysis.
Dry weight and cordycepin production are expressed as means ± SD. An analysis of variance (ANOVA) followed by Tukey's test was applied for multiple comparisons of significant analyses at < 0.05. Statistical data analyses were performed in SPSS version 17.0 software packet. Design-Expert Version 8.0.5b software package (Stat-Ease Inc., Minneapolis, USA) was used for designing experiments as well as for regression and graphical analysis of the experimental data obtained.

Effects of Working Volume on the Biomass and Cordycepin
Production. Dissolved oxygen concentration is the key factor in the medium for cell growth and metabolite biosynthesis [21]. Dissolved oxygen does not only have an important function in the respiratory chain, but also in metabolite composition [37,38]. A previous study showed that the highest cordycepin production and productivity were obtained at lower dissolved oxygen levels [21]. Masuda et al. [25] also reported that a lower medium depth was most efficient for cordycepin production in C. militaris by surface culture.  Figure 1: Effects of working volume on the production of cordycepin, total production of cordycepin, and biomass dry weight (total content ofcordycepin (mg) = cordycepin production (mg/L) × working volume (mL)).
In this study, we tried to establish the most efficient working volume of medium for improved cordycepin production. Cultures of C. militaris were prepared at the working volumes of 100 to 900 mL (corresponding to a medium depth of 1.26 to 11.31 cm). As shown in Figure 1, cordycepin production reduced gradually with increasing working volume of the medium, from 100 to 700 mL. However, there was no significant difference in cordycepin production in different working volumes. Obviously, the highest working volume of 900 mL did not help in cordycepin production. The result indicates that there is an upper dissolved oxygen limit in the medium for cordycepin production [21]. Lower working volumes result in higher cordycepin productivity, with the highest peak (463.33 ± 56.72 mg of cordycepin) produced at using 700 mL of media. Changes in biomass values were small (between 300 and 700 mL) because of the restricted area and thickness of the mycelial mat. In order to obtain higher cordycepin production, the most effective medium amount was 700 mL (corresponding to an 8.8 cm medium depth) and used as the media volume for next experiment.

Effects of Carbon and Nitrogen Sources on Cordycepin
Production. To find a suitable carbon source for C. militaris cordycepin production we added various carbon sources at a concentration of 20 g/L to the sugar-free basal medium. Glucose was previously found to be an excellent precursor of cordycepin production [39]. However, as shown in Figure 2(a), sucrose and lactose proved to be better carbon sources for cordycepin production than glucose in this study. Cordycepin production reached 843.63 ± 66.70 mg/L of sucrose and 823.72 ± 85.64 mg/L of lactose, respectively. Therefore, sucrose was selected as the main carbon source in the remaining experiment. The Scientific World Journal  In previous work, nitrogen showed a regulating role important in cordycepin production and had two effects [40]. One effect was negative since, in excess, N promoted a faster mycelial growth and consequently diverted the source of carbon toward energy and biomass production. The other effect was positive because a moderate input contributed to the maintenance of citric acid productive biomass [40]. To investigate the effect of nitrogen sources on cordycepin production in C. militaris, various compounds containing nitrogen (inorganic and organic nitrogen) were added individually to nitrogen free basal medium at a concentration of 20 g/L. Among the 8 nitrogen sources tested, peptone, yeast extract, beef extract, casein, and NH 4 NO 3 were favorable to the cordycepin production ( Figure 2(b)). Organic nitrogen was advantageous to both growth and biosynthesis of metabolites. The result is consistent with the experimental data reported [18] and showed that maximum cordycepin production resulted when the peptone was used as a nitrogen source.

Effects of Inorganic Salt and Growth
Factor on the Cordycepin Production. Inorganic ion was one of the most important nutrition components of medium for the mycelial growth [41]. In order to investigate the effects of inorganic salt for the cordycepin production in C. militaris, we tested nine types (at 1 g/L) of inorganic salts (Figure 3(a)). Media with only 20 g/L glucose and 20 g/L peptone were used as the control. The Growth factor is essential for growth response and metabolite production [42]. In order to find the optimal growth factor for cordycepin production, C. militaris was cultured in a basal medium with different vitamins and plant growth hormones in static liquid culture. Cordycepin production increased in media with added 10 mg/L of VB 1 , NAA, and VB 11 (Figure 3(b)). Maximum cordycepin production (1159.34 ± 109.01 mg/L) occurred when VB 1 was used as the growth factor.

Screening of Important Variables Using Plackett-Burman
Design. The data ( Table 2) indicated wide variation in cordycepin production in the 15 tests. The data suggested that process optimization is important for improving the efficiency of cordycepin production. Analysis of the regression coefficients and values of 7 factors (Table 5) showed that 1 , 2 , 4 , and 5 had positive effects on cordycepin production. 6 had negative effects. The variable affects with a confidence level above 95% are considered as significant factors. Based on these results, three factors ( 1 , sucrose; 4 , K 2 HPO 4 ⋅3H 2 O; The Scientific World Journal    The Scientific World Journal   Table 6 shows the detailed experimental design and results. Regression analysis was performed to fit the response function (cordycepin production) with the experimental data. From the variables obtained ( Results of -test analysis of variance (ANOVA) showed that the regression was statistically significant at 95% and 99% confidence levels ( Table 7). The " value" of the model was 9.21, and the value of "Prob > " < 0.01 indicated that the model was significant. In this case, linear terms of 1 and quadratic terms of 2 1 , 2 4 , 2 6 were significant of model terms for cordycepin production. The "Lack of Fit F value" of 0.0903 implied that the "Lack of Fit" was not significant relative to the pure error ( > 0.05). The Pred-2 of 0.3183 was not as close to the Adj-2 of 0.7954 as one might normally expect. The result suggested that some factors were not considered in the model. However, the "Adeq Precision" of 8.173 indicated that the model was adequate for prediction production of cordycepin.
The response surface plot obtained from (1) is shown in Figure 4. It is evident that cordycepin production reached its maximum at a combination of coded level ( 1 , sucrose, level 0.47; 4 , K 2 HPO 4 ⋅3H 2 O, level 0.21; 6 , MgSO 4 ⋅7H 2 O, level −0.20) when using canonical analysis of the Design-Expert Version 8.0.5b software package. The model predicted a maximum response of 1451.43 mg/L cordycepin production at levels of sucrose 24.7 g/L, K 2 HPO 4 ⋅3H 2 O 1.11 g/L, and MgSO 4 ⋅7H 2 O 0.90 g/L as optimized medium components. [43] reported that adenine and adenosine were precursors for cordycepin synthesis. Amino acids were regarded as the best substance for improved cordycepin production [4,23]. Based on these results, among 10 different kinds of nucleoside  analogue were supplemented for 1 g/L in this study. As shown in Figure 5(a), cordycepin production increased obviously in the medium with hypoxanthine, thymine, and thymidine additives. The highest production of cordycepin was achieved, when hypoxanthine was used as the nucleoside analogue. Hypoxanthine's molecular structure is similar to purine bases found in cordycepin. Substituent on purine bases structure is -OH on hypoxanthine rather than -NH 2 .

Effects of Nucleoside Analogue and Amino Acid on the Production of Cordycepin. Chassy and Suhadolnik
The -OH should be replaced in metabolic pathways. In addition, among 14 different amino acids were tested for 8 g/L. As shown in Figure 5(b), L-alanine can improve cordycepin production. Previous research showed that adenine, adenosine, and glycine were good additives for increased cordycepin production [4,23,26,27]. L-alanine may be an important nutritional element for C. militaris or component of cordycepin production. Hypoxanthine and L-alanine were the best additives to promote cordycepin production in this study.

Optimization by Response Surface Methodology for Amino
Acid, Nucleoside Analogue, and Fermentation Time. Similarly, central composite design was also applied to study the significant factors (hypoxanthine, L-alanine, and culture time) and their optimal levels. Figure 6 shows the morphological characteristics of C. militaris in 1000 mL glass jars after fermentation by static liquid fermentation. Table 8 shows the detailed experimental design and results. Regression analysis was performed to fit the response function (cordycepin production) with the experimental data. From the variables obtained (Table 9) Results of -test analysis of variance (ANOVA) showed that the regression was statistically significant at 95% and 99% confidence level ( Table 9). The " value" of the model was 11.91, and the value of "Prob > " < 0.01 indicated that the model was significant. In this case, linear terms    Figure 5: Effects of nucleoside analogue and amino acid on the production of cordycepin: nucleoside analogue (a); amino acid (b); * 5% significance level (test group versus control group); * * 1% significance level (test group versus control group); # 5% significance level (control group versus test group); ## 1% significance level (control group versus test group).
of ; interactive terms of ; and quadratic terms of 2 , 2 , and 2 were significant in model terms for cordycepin production. The "Lack of Fit F value" of 0.1707 implied that the "Lack of Fit" was not significant relative to the pure error ( > 0.05). The Pred-2 of 0.4162 was not as close to the Adj-2 of 0.8378 as one might normally expect. The result suggested that some factors were not also considered in the model. However, the "Adeq Precision" of 11.222 indicated that the model was adequate for prediction production of cordycepin.
The response surface plot obtained from (2) is shown in Figure 7. It is evident that cordycepin production reached its maximum with a combination of coded level ( , hypoxanthine, level 0.11; , L-alanine, level 0.06; , time, level −0.23) by canonical analysis of the Design-Expert Version 8.0.5b software package. The model predicted a maximum response of 2008.48 mg/L cordycepin production at levels of hypoxanthine 5.45 g/L, L-alanine 12.23 g/L, and time 8.6 days (in the practical test 8 days) as optimized medium components.
In previous work, the orthogonal design method [44][45][46], Box-Behnken design [34,47], and central composite design [31] were used to optimize culture conditions for cordycepin production by Cordyceps sp. These experimental designs have been successfully used to optimize medium for the mycelial growth and microbial metabolite production in liquid culture processes. In this study, static liquid culture conditions are optimized for the cordycepin production using response surface methodology and are an effective way to enhance the productivity of cordycepin and biomass in C. militaris. The result confirmed the model suited the predictive of hyperproduction of cordycepin by C. militaris in static liquid culture. Batch culture was carried for cordycepin production under optimized culture conditions (Figure 8).

In Vitro Cordycepin Production Using Liquid Culture in
Other Studies. The highest report for cordycepin production was 14300 mg/L by Masuda et al. [29] (Table 10). In our experiment, cordycepin production at 2008.48 mg/L was lower. However, a maximum total content of cordycepin (1405.94 mg) was achieved in our study. This is a second higher report of cordycepin production in one single fermenter. The results showed that the culture conditions will provide an effective way for increasing cordycepin production.

Conclusion
In this work, single factor design, Plackett-Burman design, and central composite design were employed to establish the key factors and identify optimal culture conditions which improved cordycepin production by C. militaris CGMCC2459. Optimal media contained peptone 20 g/L; sucrose 24.7 g/L; K 2 HPO 4 ⋅3H 2 O 1.11 g/L; MgSO 4 ⋅7H 2 O 0.90 g/L; VB 1 10 mg/L; hypoxanthine 5.45 g/L; and L-alanine 12.23 g/L. Hypoxanthine and L-alanine were added to the optimal medium at 8.6 days. Optimal incubation conditions were 25 ∘ C at an unaltered pH of 35 days. Using these culture conditions, a maximum production of cordycepin was 2008.48 mg/L for 700 mL working volume in the 1000 mL glass jars, and total content of cordycepin reached 1405.94 mg/bottle (700 mL/1000 mL). This method provides