Optimization of Fermentation Conditions of Artemisia capillaris for Enhanced Acetylcholinesterase and Butyrylcholinesterase

In this study, the fermentation of Artemisia capillaris by probiotic Leuconostoc mesenteroides MKJW (MKJW) was optimized to increase the acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) inhibitory and antioxidant activities using the response surface method (RSM). The independent variables were the contents of A. capillaris, Gryllus bimaculatus, and yeast extract, while the dependent variables were AChE inhibitory activity, BuChE inhibitory activity, and antioxidant activities such as FRAP, reducing power, and DPPH radical scavenging ability. Seventeen experimental runs were designed with RSM and analyzed after fermentation with MKJW. Quadratic models were used to analyze the inhibition of AChE and BuChE, and a linear model was used to analyze the FRAP. The three models were significantly appropriate (p < 0.0001). The highest optimal condition of the AChE inhibitory activity was derived by a multiple regression equation. When the optimum fermentation conditions were A. capillaris 6.75%, G. bimaculatus 0.18%, and yeast extract 1.27%, 91.1% was reached for AChE inhibitory, 74.0% for BuChE inhibitory, and 34.1 mM FeSO4 for FRAP. The predicted dependent variables were not significantly different from the experimental values (p > 0.05). In conclusion, the A. capillaris fermented by MKJW might be used as a natural antidementia improving agent with AChE inhibitory, BuChE inhibitory, and antioxidant activities.


Introduction
Neurological disorders are the leading causes of disability-adjusted life years (DALYs; the sum of years of life lost) and death, including chronic degenerative diseases, cancer, diabetes, and cardiovascular diseases [1,2]. The neurotransmitter acetylcholine is reduced in the brains of dementia patients compared to that in normal brains due to a decrease in acetylcholine synthesis and the breakdown of acetylcholine by cholinergic enzymes [3]. Symptoms of dementia include a decrease in memory and cognitive ability, among which Alzheimer's disease shows decreased memory, depression, decreased judgment, and confusion. The main biological function of acetylcholinesterase (AChE) is to rapidly break down the neurotransmitter acetylcholine into acetic acid and choline [4]. Butyrylcholinesterase (BuChE) is also involved in neurological disorders as a surrogate hydrolase for acetylcholine [5]. Choline deficiency due to cholinergic enzymes is associated with the onset of Alzheimer's disease, and the current drug approach to Alzheimer's disease is based upon suppressing cholinergic enzymes [6]. However, common adverse effects of drugs include loss of appetite, gastrointestinal symptoms such as nausea, vomiting, diarrhea, and a lot of weight loss that appears when used for a long period [7,8]. Furthermore, no single-target drug has been successful in preventing dementia. Consequently, interest in natural plants or fermented natural products has emerged.
Artemisia sp., including A. princeps, A. argyi, A. iwayomogi, A. princeps, A. capillaris, etc. is a perennial herb with a wide distribution in Europe and Southeast Asia, including Korea [9]. Artemisia sp. has various physiological activities and is widely used for food or Hot-water-extracted A. capillaris (40 • brix) was obtained from DHbio (Anseong, Korea). Briefly, A. capillaris was extracted with ten-fold water for 6 h at 120 • C. Then, six-fold water was added, and boiled for 3 h at 105 • C. G. bimaculatus, obtained from Cricket Farm Co. (Hwaseong, Korea), was hot-air-dried at 70 • C for 10 h at the farm, and was ground with a mixer and sieved with an 18 mm mesh sieve. MKJW (KCTC14459BP) was used from our laboratory stock.

Preliminary Experiments for Fermentation Conditions
In order to determine the range of A. capillaris content to be input into the RSM, fermentation patterns according to A. capillaris at 4%, 5%, 6%, 8%, and 10% were investigated. In addition, three types of the combination of 5% A. capillaris and protein supplements (first, 2% (w/v) G. bimaculatus [Cricket Farm, Hwaseong, Korea]; second, 2% (w/v) yeast extract [Difco, Becton Dickinson and Co., Sparks, MD, USA]; and third, 1% (w/v) G. bimaculatus and 1% (w/v) yeast extract) were confirmed. After adding 2.5% (w/v) glucose to each preliminary test group, the mixture was sterilized for 20 min at 121 • C. After cooling, MKJW cultured for 18 h was inoculated to 5% (v/v) of the total medium volume. Then, shaking incubation (130 rpm, 30 • C) was performed for 24 h. After fermentation, the solution was pasteurized at 60 • C for 30 min and centrifuged to separate the supernatant and the precipitate.

Box-Behnken Experiment Design
The Box-Behnken design (BBD) can identify the linear influence, interactions between factors, and quadratic effects, and is used for RSM purposes. Response surface analysis was performed using the Stat-Ease program (Design-Expert ® software, version 12, Stat-Ease, Inc., Minneapolis, MN, USA) to determine the optimal Artemisia fermentation conditions. The amount for A. capillaris was between 3% and 7% (X 1 ), for G. bimaculatus was between 0% and 2% (X 2 ), and for yeast extract, was between 0% and 2% (X 3 ), which were set as independent variables (Table 1). As dependent variables, acetylcholinesterase and butyrylcholinesterase inhibition activities and ferric reducing antioxidant activity (FRAP) were selected.

Fermentation
The BBD design planned 17 experiments with five replications of the center points ( Table 2). The corresponding fermentation was prepared. Briefly, different contents of A. capillaris (3, 5, and 7%), G. bimaculatus powder (0, 1, and 2%), and yeast extract (0, 1, and 2%) provided from the BBD design were added into each tube, autoclaved at 121 • C for 20 min, and cooled. All fermentation broth contained sterilized 2.5% glucose. Three-timessubcultured MKJW on MRS broth (Difco, Becton Dickinson and Co., Sparks, MD, USA) was inoculated to 5% (v/v) of the total broth volume (8 Log CFU·g −1 ). The fermentation was carried out with shaking incubation at 130 rpm and 30 • C for 12 h.
The fermented samples were serially diluted with 0.1% peptone water and spread onto the MRS agar. MRS agar was incubated for 24 h at 30 • C to determine total viable bacterial counts. For other analysis, the sample tubes were heated at 60 • C for 30 min, followed by centrifugation to collect the supernatant for use in the experiment. The pH of the samples was measured with a pH meter (Fisher Science Education, Waltham, MA USA).

Acetylcholinesterase (AChE) and Butyrylcholinesterase (BuChE) Inhibitory Activity
AChE and BuChE inhibitory assays were carried out in a 96-well plate using a modified method, as described by Sharififar et al. [25]. The reagents (Sigma Aldrich, St. Louis, MO, USA) were injected into the well in the following procedure: 5 µL enzyme solution (AChE or BuChE, 2 U/mL), 40 µL 5,5-dithio-bis-(2-nitrobenzoic acid) (DTNB, 1 mM), and 125 µL diluted sample solution. The mixture was pre-incubated for 5 min at 25 • C. The reaction was initiated by adding 7.5 µL substrate (acetylcholine iodide or butyrylthiocholine iodide, 7.5 mM), and then the absorbance was measured at 412 nm for 15 min at 25 • C. The control consisted of adding the phosphate buffer (200 mM, pH 7.7) instead of the sample. All solutions were dissolved in 200 mM phosphate buffer (pH 7.7). Absorbance was plotted against time and enzyme activity was calculated from the slope of the line, obtained and expressed as a percentage compared with an assay using a buffer without inhibitor. Galantamine was used as the positive control.  The activity to scavenge the free radical DPPH was measured according to the method of Park et al. [26], with modifications. Briefly, a volume of 300 µL of the sample solution was mixed with 200 mM 2,2-diphenyl-1-picrylhydrazyl (DPPH) solution dissolved in methanol, which was left to react for 30 min at room temperature in the dark. The amount of scavenged DPPH radical was determined by measuring the absorbance at 517 nm. Methanol was used as the control instead of the sample. Ascorbic acid was used as the positive control. The DPPH radical scavenging activity was calculated as follows: A control : Absorbance of the control A sample : Absorbance of the sample

Reducing Power
The method developed by Lin and Yen [27] was used with modifications to evaluate the reducing activity of fermented samples. Then, 250 µL sample was mixed with an equal volume of 200 mM sodium phosphate buffer (pH 6.6) and 1% (w/v) potassium ferricyanide solution. The reaction mixture was incubated at 50 • C for 20 min, then 250 µL of 10% (w/v) trichloroacetic acid was added to stop the reaction. Next, 500 µL the upper phase was obtained by centrifugation at 10,000 rpm for 5 min. Then, 500 µL of distilled water and 100 µL of 0.1% (w/v) FeCl 3 was added. The absorbance was read at 700 nm. Ascorbic acid was used to the positive control. The reducing power was expressed as the absorbance value at Abs 700 nm.

Ferric Reducing Antioxidant Power (FRAP)
The FRAP value of the fermented samples was determined using the method described by Cermeno et al. [28], with some modifications. The FRAP reagent was prepared by mixing 300 mM acetate buffer, 10 mM 2,4,6-tripyridyl-s-triazine (TPTZ), and 20 mM FeCl 3 ·6H 2 O in a 10:1:1 (v/v/v) ratio and heating at 37 • C for 15 min. Then, 1 mL of FRAP reagent was added to 50 µL of the sample solution. Then, the absorbance was read at 590 nm after 30 min incubation at 37 • C. The FRAP value was expressed as the millimole (mM) of the FeSO 4 equivalents.

Total Polyphenol Contents
The total polyphenol content (TPC) was determined by the Folin-Ciocalteu method, according to a method adapted from Anesini et al. [29]. 0.2 mL of the sample solution was mixed with 1.0 mL of Folin-Ciocalteu reagent, which was diluted 10 times in water. Then, 0.8 mL of 7.5% (w/v) Na 2 CO 3 was added and the mixture was left to stand at room temperature for 1 hr. The absorbance was measured at 765 nm. The TPC was expressed as the mg of gallic acid equivalents per ml of sample (mg GAE/mL).

Statistical Analysis
The data were expressed as the means ± standard deviation of triplicates. Analysis of variance and Tukey's means comparison tests were performed to determine the significant differences (p < 0.05) in the results using Minitab 16.0 software (Minitab Inc., State College, PA, USA). The influence of the independent variable on the dependent variable was quantified by deriving a quadratic regression equation, and the optimal condition was predicted by plotting a three-dimensional response surface diagram using the Stat-Ease program (Design-Expert ® software, version 12, Stat-Ease, Inc., Minneapolis, MN, USA). Figure 1 shows the number of viable bacteria when different concentrations of A. capillaris (4-10%) were fermented by MKJW for 24 h. The fermentation was successfully performed when 4% and 5% A. capillaris were used. However, more than 5% A. capillaris did not proliferate. Artemisia sp. has been reported to inhibit the growth of even lactic acid bacteria such as Lactobacillus plantarum and Leuconostoc mesenteroides, which are commonly used as fermentation starters [30,31], as well as pathogenic bacteria such as Bacillus subtilis [32,33], Escherichia coli [32,33], Listeria monocytogenes [34], Pseudomonas aeruginosa [32], Staphylococcus aureus [32][33][34], and Saccharomyces cerevisiae [35].  Figure 2 shows the number of viable bacteria and pH changes that occurred during 24 h with the addition of each G. bimaculatus and yeast extract to 5% A. capillaris. The addition of supplemental materials such as G. bimaculatus and yeast extract to A. capillaris enhanced the growth of MKJW to approximately 9 log CFU/mL compared to the 5% A. capillaris only (7.58 log CFU/mL) during 24 h fermentation. The pH of 5% A. capillaris decreased from 5.11 to 4.73 for 24 h; 5% A. capillaris supplemented with 2% G. bimaculatus decreased from 5.45 to 3.66 after fermentation; 5% A. capillaris supplemented with 2% yeast extract decreased from 5.32 to 4.00; and 5% A. capillaris supplemented with 1% G. bimaculatus and 1% yeast extract decreased from 5.36 to 3.83.  Figure 2 shows the number of viable bacteria and pH changes that occurred during 24 h with the addition of each G. bimaculatus and yeast extract to 5% A. capillaris. The addition of supplemental materials such as G. bimaculatus and yeast extract to A. capillaris enhanced the growth of MKJW to approximately 9 log CFU/mL compared to the 5% A. capillaris only (7.58 log CFU/mL) during 24 h fermentation. The pH of 5% A. capillaris decreased from 5.11 to 4.73 for 24 h; 5% A. capillaris supplemented with 2% G. bimaculatus decreased from 5.45 to 3.66 after fermentation; 5% A. capillaris supplemented with 2% yeast extract decreased from 5.32 to 4.00; and 5% A. capillaris supplemented with 1% G. bimaculatus and 1% yeast extract decreased from 5.36 to 3.83.

Preliminary Experiments for Fermentation Conditions
enhanced the growth of MKJW to approximately 9 log CFU/mL compared to the 5% A. capillaris only (7.58 log CFU/mL) during 24 h fermentation. The pH of 5% A. capillaris decreased from 5.11 to 4.73 for 24 h; 5% A. capillaris supplemented with 2% G. bimaculatus decreased from 5.45 to 3.66 after fermentation; 5% A. capillaris supplemented with 2% yeast extract decreased from 5.32 to 4.00; and 5% A. capillaris supplemented with 1% G. bimaculatus and 1% yeast extract decreased from 5.36 to 3.83.  As a result, MKJW did not grow well when fermented with only A. capillaris, but the bacteria grew well as a result of fermentation with the addition of the protein-rich G. bimaculatus and yeast extract. A. capillaris is thought to have poor growth due to a lack of nitrogen source necessary for the synthesis of protein, enzymes, nucleic acid, and other microbial components among the nutrients that are necessary for the growth of lactic acid bacteria. Park et al. [36] fermented glasswort witha mixture of rice bran and soybeans to supplement sugar and protein. When Yang et al. [37] mixed and fermented amaranth and quinoa to compensate for the protein source lacking in white rice, the number of viable bacteria increased compared to when only white rice was fermented.
Finally, the A. capillaris content was set to 5%, which allowed the bacteria to grow. The fermentation time was determined to be 12 h when the bacteria grew well, and the antioxidant activity was high.

Microbiological and Physiological Analysis of Fermented A. capillaris
Based on the preliminary experiments, the amounts of A. capillaris, G. bimaculatus, and yeast extract were selected as independent variables. Fermentation of A. capillaris was carried out under 17 conditions according to the Box-Behnken design. Table 2 shows the changes in the viable cell counts, pH value, and dependent variables during fermentation for the 17 experimental runs. All samples increased in viable cells and decreased in pH during fermentation. This result shows that MKJW eats nutrients such as glucose and yeast extracts, and as it grows, it produces fermentation products such as organic acids [38].
The substrate combination of 3% A. capillaris, 2% G. bimaculatus, and 1% yeast extract, had a significant increase in viable cell counts (8.99 Log CFU/mL), and a significant decrease in pH values. The fermentation with supplementation of G. bimaculatus or/and yeast extract to A. capillaris boosted the fermentation with increasing cell growth and lowering the pH values.

Suitability of the Regression Model
The results were analyzed using multiple regression, and the coefficients of each model were evaluated for their significance by regression analysis ( Table 3). The model of the AChE inhibitory activity was adopted as a quadratic model in which independent variables interact. The F-value of 32.83 implied that the model was significant. p-values less than 0.05 indicated that model terms (factors X 1 , X 2 , X 3 , X 1 X 2 , X 2 X 3 , X 1 2 ) were statistically significant. In addition, F-value (1.94) of the lack of fit was not significant (p = 0.26) relative to the pure error. A nonsignificant lack of fit is a good indication that the model fits the actual relationships of the response parameters within the chosen ranges. Likewise, the model of the BuChE inhibitory activity was adopted as a quadratic model. The F-value of 99.08 and p-values less than 0.05 showed that the model was significant. The factors, X 1 , X 2 , X 1 X 2 , X 1 X 3 , X 2 X 3 , X 1 2 , X 2 2 , and X 3 2 , were highly significant. The lack of fit was also not significant (F = 1.21, p = 0.4129) indicating fit goodness of the proposed model [34]. The values of the F-(77.12) and p-(<0.0001) was to indicate that FRAP (Y 3 ) is suitable for the linear model. Furthermore, the p-value for all primary terms was less than 0.05 as a result of the analysis of variance. This means that each independent variable of all linear terms affects the dependent variable. F-value of the lack of fit was 1.77 and the p-value was 0.31, which was more than 0.1, thus indicating that the model is appropriate. The lack of fit of the reducing power (Y 4 ) and DPPH radical scavenging ability (Y 5 ), which are for the remaining antioxidant experiments, were 0.0983 and 0.0742, respectively. For each, the lack of fit was lower than 0.1, so the two experiments were determined to be inappropriate models.  Table 4 shows the actual and predicted values of each dependent variable (AChE inhibitory activity, BuChE inhibitory activity, and FRAP) based on the experimental design. The observed and predicted values matched reasonably well.

Model Verification
The final estimated response model equation for the AChE inhibitory activity was as follows: The coefficients of A. capillaris (X 1 ) and the yeast extract (X 3 ) were positive; G. bimaculatus (X 2 ) was negative of the dependent variable Y 1 . This means that when the contents of A. capillaris (X 1 ) and yeast extract (X 3 ) increase and the content of G. bimaculatus (X 2 ) decreases, the AChE inhibitory activity (Y 1 ) increases. The response increased as the content of yeast extract (X 3 ) increased, but the influence on the response was less than that of A. capillaris (X 1 ). The adjusted coefficient of determination for the multiple regression equation was 0.9471 and the p-value was less than 0.0001.
Equation (3) was visualized as a 3D response surface and is shown in Figure 3. Figure 3a shows that the AChE inhibitory activity was rapidly increasing in a curved shape when the A. capillaris (X 1 ) content increased, which then decreased in response to the increase in the G. bimaculatus (X 2 ) content. Figure 3b shows the effect of the A. capillaris (X 1 ) content and yeast extract (X 3 ) content response, and as in 3a, it increased as the A. capillaris (X 1 ) content increased, but with little effect on the increase of yeast extract (X 3 ) content. Figure 3c shows that the lower the content of G. bimaculatus (X 2 ), the lower the change in the AChE inhibitory activity as the yeast extract (X 3 ) increased. Conversely, as the content of G. bimaculatus (X 2 ) increased, the effect of the yeast extract (X 3 ) on the AChE inhibitory activity increased. The coefficients of A. capillaris (X1) and the yeast extract (X3) were positive; G. bimaculatus (X2) was negative of the dependent variable Y1. This means that when the contents of A. capillaris (X1) and yeast extract (X3) increase and the content of G. bimaculatus (X2) decreases, the AChE inhibitory activity (Y1) increases. The response increased as the content of yeast extract (X3) increased, but the influence on the response was less than that of A. capillaris (X1). The adjusted coefficient of determination for the multiple regression equation was 0.9471 and the p-value was less than 0.0001.
Equation (3) was visualized as a 3D response surface and is shown in Figure 3. Figure  3a shows that the AChE inhibitory activity was rapidly increasing in a curved shape when the A. capillaris (X1) content increased, which then decreased in response to the increase in the G. bimaculatus (X2) content. Figure 3b shows the effect of the A. capillaris (X1) content and yeast extract (X3) content response, and as in 3a, it increased as the A. capillaris (X1) content increased, but with little effect on the increase of yeast extract (X3) content. Figure  3c shows that the lower the content of G. bimaculatus (X2), the lower the change in the AChE inhibitory activity as the yeast extract (X3) increased. Conversely, as the content of G. bimaculatus (X2) increased, the effect of the yeast extract (X3) on the AChE inhibitory activity increased. The coefficient for BuChE inhibitory activity (Y2) was positive for A. capillaris (X1) and yeast extract (X3), and negative for G. bimaculatus (X2). This result was the same as Y1. As the contents of A. capillaris (X1) and yeast extract (X3) increased and the content of G. bimaculatus (X2) decreased, the BuChE inhibitory activity (Y2) value increased. The adjusted coefficient of determination was 0.9822 and the p-value was less than 0.0001. The final estimated response model equation for BuChE inhibitory activity is as follows:  The coefficient for BuChE inhibitory activity (Y 2 ) was positive for A. capillaris (X 1 ) and yeast extract (X 3 ), and negative for G. bimaculatus (X 2 ). This result was the same as Y 1 . As the contents of A. capillaris (X 1 ) and yeast extract (X 3 ) increased and the content of G. bimaculatus (X 2 ) decreased, the BuChE inhibitory activity (Y 2 ) value increased. The adjusted coefficient of determination was 0.9822 and the p-value was less than 0.0001. The final estimated response model equation for BuChE inhibitory activity is as follows: Y 2 = 50.10 + 16.13 X 1 − 2.96 X 2 + 0.3375 X 3 − 6.52 X 1 X 2 − 6.08 X 1 X 3 + 3.65 X 2 X 3 + 2.87 X 1 2 + 4.15 X 2 2 − 4.80 X 3 2 (4) Equation (4) was consistent with the three-dimensional graph that shows the interaction between factors, which is presented in Figure 4. Figure 4a,b show that the BuChE inhibitory activity increased rapidly as the A. capillaris (X 1 ) increased. In addition, the response was decreased when the content of G. bimaculatus (X 2 ) or yeast extract increased. Figure 4c shows that the response decreased as the yeast extract (X 3 ) and G. bimaculatus (X 2 ) increased.
Equation (4) was consistent with the three-dimensional graph that shows the interaction between factors, which is presented in Figure 4. Figure 4a,b show that the BuChE inhibitory activity increased rapidly as the A. capillaris (X1) increased. In addition, the response was decreased when the content of G. bimaculatus (X2) or yeast extract increased. Figure 4c shows that the response decreased as the yeast extract (X3) and G. bimaculatus (X2) increased. The coefficient for FRAP (Y3) was positive for A. capillaris (X1) and yeast extract (X3), whereas G. bimaculatus (X2) was negative. G. bimaculatus (X2) has less influence on the dependent variable than do X1 and X3. Its R 2 was 0.9468. The final estimated response model equation for the FRAP value is as follows: (5) Figure 5 shows the changes in the independent variables A. capillaris (X1) and G. bimaculatus (X2) when the independent variable yeast extract (X3) was fixed. FRAP was only significant for linear terms (p < 0.05). As with the regression equation, as A. capillaris (X1) increased, the FRAP value increased linearly.  The coefficient for FRAP (Y 3 ) was positive for A. capillaris (X 1 ) and yeast extract (X 3 ), whereas G. bimaculatus (X 2 ) was negative. G. bimaculatus (X 2 ) has less influence on the dependent variable than do X 1 and X 3 . Its R 2 was 0.9468. The final estimated response model equation for the FRAP value is as follows: Y 3 = 24.33 + 7.84 X 1 − 1.68 X 2 + 3.09 X 3 (5) Figure 5 shows the changes in the independent variables A. capillaris (X 1 ) and G. bimaculatus (X 2 ) when the independent variable yeast extract (X 3 ) was fixed. FRAP was only significant for linear terms (p < 0.05). As with the regression equation, as A. capillaris (X 1 ) increased, the FRAP value increased linearly. Equation (4) was consistent with the three-dimensional graph that shows the interaction between factors, which is presented in Figure 4. Figure 4a,b show that the BuChE inhibitory activity increased rapidly as the A. capillaris (X1) increased. In addition, the response was decreased when the content of G. bimaculatus (X2) or yeast extract increased. Figure 4c shows that the response decreased as the yeast extract (X3) and G. bimaculatus (X2) increased. The coefficient for FRAP (Y3) was positive for A. capillaris (X1) and yeast extract (X3), whereas G. bimaculatus (X2) was negative. G. bimaculatus (X2) has less influence on the dependent variable than do X1 and X3. Its R 2 was 0.9468. The final estimated response model equation for the FRAP value is as follows: Y3 = 24.33 + 7.84 X1 − 1.68 X2 + 3.09 X3 (5) Figure 5 shows the changes in the independent variables A. capillaris (X1) and G. bimaculatus (X2) when the independent variable yeast extract (X3) was fixed. FRAP was only significant for linear terms (p < 0.05). As with the regression equation, as A. capillaris (X1) increased, the FRAP value increased linearly.  Based on our results, the models for Y 1 , Y 2 , and Y 3 can be useful tools to optimize the fermentation conditions because these are statistically significant and provide solid solutions as a function of independent variables.

Optimization of A. capillaris Fermented Product
The optimal fermentation conditions were determined by maximizing the desirability of the responses. The maximal desirability should be initially at the highest values for AChE and BuChE inhibitory activities, and FRAP, and then be at the higher concentration of A. capillaris. Finally, the optimal fermentation conditions for AChE and BuChE inhibitory activities are as follows: X 1 = 6.75%, X 2 = 0.18%, and X 3 = 1.27%. Inoculum volume was 6.82 increased to 8.14 with a decrease in pH values from 5.16 to 4.93. Using the above conditions, the predicted values for each model are Y 1 = 91.6%, Y 2 = 74.5%, and Y 3 = 33.3 mM FeSO 4 . After fermentation under optimal conditions, the predicted and observed values for the three dependent variables were analyzed ( Figure 6). The observed values of each dependent variable were the AChE inhibitory activity (Y 1 ) of 91.1 ± 4.75% (p = 0.864), BuChE inhibitory activity (Y 2 ) of 74.0 ± 1.96% (p = 0.681), and FRAP (Y 3 ) of 34.1 ± 0.43 mM FeSO 4 , (p = 0.060). There was no significant difference between the predicted value and the observed value. These results reveal that the model used in this study is valid. Based on our results, the models for Y1, Y2, and Y3 can be useful tools to optimize the fermentation conditions because these are statistically significant and provide solid solutions as a function of independent variables.

Optimization of A. capillaris Fermented Product
The optimal fermentation conditions were determined by maximizing the desirability of the responses. The maximal desirability should be initially at the highest values for AChE and BuChE inhibitory activities, and FRAP, and then be at the higher concentration of A. capillaris. Finally, the optimal fermentation conditions for AChE and BuChE inhibitory activities are as follows: X1 = 6.75%, X2 = 0.18%, and X3 = 1.27%. Inoculum volume was 6.82 increased to 8.14 with a decrease in pH values from 5.16 to 4.93. Using the above conditions, the predicted values for each model are Y1 = 91.6%, Y2 = 74.5%, and Y3 = 33.3 mM FeSO4. After fermentation under optimal conditions, the predicted and observed values for the three dependent variables were analyzed ( Figure 6). The observed values of each dependent variable were the AChE inhibitory activity (Y1) of 91.1 ± 4.75% (p = 0.864), BuChE inhibitory activity (Y2) of 74.0 ± 1.96% (p = 0.681), and FRAP (Y3) of 34.1 ± 0.43 mM FeSO4, (p = 0.060). There was no significant difference between the predicted value and the observed value. These results reveal that the model used in this study is valid. The antioxidant activities and TPC of fermented A. capillaris under optimal conditions are shown in Table 5. The DPPH radical scavenging activity was 86.5 ± 0.89%. It was higher than Lactobacillus-fermented A. annua (80%) during 12 h, which was reported by Lee et al. [39]. The reducing power was 2.14 ± 0.02 at Abs 700 nm. The total polyphenol content of fermented A. capillaris was 1.52 ± 0.61 mg GAE/mL. It was a much higher content compared to other fermented Artemisia sp. The total polyphenol contents of A. annua L. fermented by Lactobacillus sp. for 12 h were 0.3 mg GAE/mL [39], and those of A. princeps Pamp. fermented by Monascus sp. were about 0.7 mg GAE/mL [40]. This is thought to be because the Artemisia sp. type and starter are different. A significant correlation between total polyphenol content and AChE inhibitory activity among the five Asteraceae sp. plants has been reported [41]. The aromatic phenyl groups (-C6H5) of polyphenols structurally The antioxidant activities and TPC of fermented A. capillaris under optimal conditions are shown in Table 5. The DPPH radical scavenging activity was 86.5 ± 0.89%. It was higher than Lactobacillus-fermented A. annua (80%) during 12 h, which was reported by Lee et al. [39]. The reducing power was 2.14 ± 0.02 at Abs 700 nm. The total polyphenol content of fermented A. capillaris was 1.52 ± 0.61 mg GAE/mL. It was a much higher content compared to other fermented Artemisia sp. The total polyphenol contents of A. annua L. fermented by Lactobacillus sp. for 12 h were 0.3 mg GAE/mL [39], and those of A. princeps Pamp. fermented by Monascus sp. were about 0.7 mg GAE/mL [40]. This is thought to be because the Artemisia sp. type and starter are different. A significant correlation between total polyphenol content and AChE inhibitory activity among the five Asteraceae sp. plants has been reported [41]. The aromatic phenyl groups (-C 6 H 5 ) of polyphenols structurally interact well with the active site of AChE, which forms as an aromatic residue to inhibit the decomposition of acetylcholine [42]. Aside from the onset of Alzheimer's disease by cholinesterase, another risk factor for Alzheimer's disease is that oxidative stress increases in aging patients due to the accumulation of reactive oxygen species [43]. Oxidative stress can be eliminated by antioxidant active substances such as polyphenols and exopolysaccharides. Papandreou et al. [44] reported that when polyphenol-rich blueberry extracts were administered in the mouse brain, both oxidative stress and AChE activity reduced. The antioxidant active substances can be produced or increased as a result of fermentation using lactic acid bacteria [45]. Therefore, it is believed that the antioxidant activity of the raw A. capillaris increased further due to the fermentation of MKJW as well as AChE inhibitory activity.

Conclusions
This study showed that optimized fermentation of A. capillaris by L. mesenteroides MKJW increased its cholinesterase inhibition activity and antioxidant activity. The maximal activity of acetylcholinesterase (91.6%) and butyrylcholinesterase (74.5%) inhibition and FRAP (33.3 mM FeSO 4 ) was achieved when 6.75% A. capillaris was fermented by MKJW in the presence of 0.18% G. bimaculatus and 1.27% yeast extract. Therefore, it is thought that it can be used as probiotic A. capillaris possessing antidementia and antioxidant properties, but further research should be undertaken.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.