A new source of bacterial myrosinase isolated from endophytic Bacillus sp. NGB-B10, and its relevance in biological control activity

Plant metabolism interacts strongly with the plant microbiome. Glucosinolates, secondary metabolites synthesized by Brassica plants, are hydrolyzed by myrosinase into bioactive compounds of great importance in human health and plant protection. Compared with myrosinase from plant sources, myrosinase enzymes of microbial origin have not been extensively investigated. Therefore, seven endophytic strains corresponding to Bacillus sp. were isolated from Eruca vesicaria ssp. sativa plants that could hydrolyse glucosinolates (sinigrin) in the culture medium and showed myrosinase activity (0.08–19.92 U mL−1). The bglA myrosinase-related gene encoding the 6-phospho-β-glucosidase (GH 1) from Bacillus sp. NGB-B10, the most active myrosinase-producing bacterium, was successfully identified. Response surface methodology (RSM) was applied to statistically optimize culture conditions for myrosinase production from Bacillus sp. strain NGB-B10. The Plackett–Burman design indicated that nitrogen concentration, incubation period, and agitation speed were the significant parameters in myrosinase production. The application of the Box–Behnken design of RSM resulted in a 10.03-fold increase in enzyme activity as compared to the non-optimized culture conditions. The myrosinase was partially purified by 40% fractionation followed by SDS-PAGE analysis which yielded two subunits that had a molecular weight of 38.6 and 35.0 KDa. The purified enzyme was stable under a broad range of pH (5.5–10) and temperatures (10–65 °C). The hydrolysis products released by bacterial myrosinase from some glucosinolate extracts had higher and/or equivalent in vitro antagonistic activity against several phytopathogenic fungi compared to the nystatin (a broad-spectrum antifungal agent). This study provides original information about a new source of bacterial myrosinase and affords an optimized method to enhance myrosinase production. Supplementary Information The online version contains supplementary material available at 10.1007/s11274-022-03385-3.


Introduction
Glucosinolates (GLs) are an important class of sulfur-containing secondary metabolites that are present exclusively in plants of the Brassicales, mainly in the family Brassicaceae (Cruciferae) such as cabbage, broccoli, mustard, arugula, and cauliflower (Bhat and Vyas 2019). Generally, GLs are stored in the plant vacuole or specific S-cells, and upon tissue disruption, they are hydrolyzed by the myrosinase enzyme (β-thioglucoside glucohydrolase EC 3.2.3.1), which is stored in so-called myrosin cells separately from the GLs, into D-glucose and unstable intermediates (Martinez-Ballesta and Carvajal 2015; Wittstock et al. 2016). These intermediates spontaneously rearrange into isothiocyanates, thiocyanates, nitriles, epithionitriles, or oxazolidine-2-thiones ( Fig. 1) depending on the substrate, reaction conditions (such as pH values, temperature), and presence of certain proteins (Hanschen et al. 2017;Sikorska-Zimny and Beneduce 2021).
The GLs-myrosinase system not only plays a significant role in the plant's defense against biotic stressors but also contributes to beneficial effects for human health and plant protection. GLs derivatives particularly isothiocyanates (ITCs) are effective candidates against several herbivores, 1 3 215 Page 2 of 17 pests, and pathogens as these compounds have antifungal, antimicrobial as well as insecticidal activities (Wittstock et al. 2016;Wassermann et al. 2017;Bhat and Vyas 2019). Due to their high antimicrobial properties and natural plant origin, ITCs also have been previously studied in the food industry and preservation (Dufour et al. 2015;Yaqoob et al. 2020). For humans, the biological activities of GLs and their breakdown products have been well documented in the literature for their disease-preventive and therapeutic effects (Romeo et al. 2018). Several studies have described the protective effects of ITCs against various types of cancer (Rudolf et al. 2014;Lawson et al. 2015) in addition to their positive effect on cardiovascular and neurological diseases (Palliyaguru et al. 2018).
For the efficient production of bacterial enzymes, optimizing media components and culture conditions is very important which affects the economics and feasibility of the process (Ojha et al. 2020). Recently, response surface methodology (RSM) has been widely used as a mathematical modeling and a statistical tool for optimizing conditions in multivariable systems, while saving time and labor by minimizing the number of required experiments (Mefteh et al. 2019;Abdel-Rahman et al. 2020;Zhao et al. 2020). Box-Behnken Design (BBD) is one of the main experimental popular designs used in the RSM for optimizing the production of bacterial enzymes (Tabssum et al. 2018;Mefteh et al. 2019).
The plant-endophytic microbiome is considered a major source of secondary metabolites and bioactive natural products. Endophytic bacteria are ubiquitous and it is now well established that the individual plant hosts one or more endophytes (Gouda et al. 2016). Recently, several studies reported that endophytes and their hosts contribute partially to the metabolism of each other (Naik et al. 2019;Hagaggi and Mohamed 2020) and therefore some metabolites are coproducts of the endophytes and the plants (Ludwig-Müller 2015). For example, a recent matching of bioactive secondary metabolites (total phenolic content, flavonoids, saponins, and tannins) production was identified between Calotropis procera and its associated bacterial endophytes (Hagaggi and Mohamed 2020).
Therefore, we hypothesized that due to their high GLs content, E. vesicaria ssp. sativa (Mill.) Thell. (arugula) plants belonging to the family Brassicaceae may harbor  (Rask et al. 2000) a specific microbiome containing potential myrosinaseproducing bacteria. We also raised the following questions: Do endophytes have a potential myrosinase activity better than non-endophytic "rhizospheric" bacteria? What are the optimum culture conditions for myrosinase production? To which extent do GLs-myrosinase hydrolysis products have antagonistic activities against plant pathogenic fungi? To answer the above questions, rhizospheric and endophytic bacteria associated with E. vesicaria ssp. sativa plants were isolated and screened for GLs-degradation activities. GLsdegrading bacteria were identified using 16S rRNA sequence analysis and the sequence of the myrosinase-related gene (bglA) was determined. The culture conditions for myrosinase production by the highest producing bacterium (Bacillus sp. NGB-B10) were optimized using combined statistical approaches including Plackett-Burman experiment, and the BBD of RSM. The partially purified enzyme was also characterized to study enzyme biochemical characteristics and address its kinetic properties. Moreover, the antagonistic effect of several sources of GLs-hydrolysis products against phytopathogenic fungi was investigated.

Bacterial isolation
Ten plant samples of E. vesicaria ssp. sativa (Mill.) Thell. (arugula) were collected from the Agricultural Experimental and Research Station (Latitude 30.026455, Longitude 31.192168), Faculty of Agriculture, Cairo University. Rhizospheric bacteria was isolated as described in (Youseif 2018). Soil adhering with plant roots was collected and mixed. Then, 10 gm was taken and a serial dilution was made and plated onto nutrient agar (NA) plates (0.5% peptone, 0.15% yeast extract, 0.15% beef extract, 0.5% sodium chloride, and 1.5% agar). Endophytic bacteria were obtained from surface-sterilized plant tissues as described in (Hallmann et al. 2006). Root and leave samples were separated and surface-sterilized by washing for 2 min in 70% ethanol, immersed in 1.5% sodium hypochlorite for 3 min, and finally washed 3 times using sterile distilled water under aseptic conditions. Surface sterilization efficiency was tested by imprinting the surface-sterilized plant tissue onto NA plates and spreading aliquots of the last rinse wash on the same type of medium as described in (Youseif et al. 2021a). Individual leaves and roots were cut into small pieces and aseptically ground in sterile distilled water, streaked on the surface of NA plates supplemented with 25 µg mL −1 nystatin, and finally were incubated at 30 °C for 5-7 days. The bacterial colonies were initially screened and grouped by colony color and morphological characteristics.

Screening for myrosinase-active bacteria (GLs-degrading organisms)
The activity of bacterial isolates to produce myrosinase and decompose GLs was assessed as described in (Tani et al. 1974) using a mustard extract broth medium (0.01% sinigrin, 6% mustard extract, 0.1% KH 2 PO 4 , 0.1% NH 4 Cl, 0.1% NaCI, and 0.1% MgSO 4 ·7H 2 O), all listed chemicals were purchased from Sigma-Aldrich, Merck KGaA, Germany. Bacterial isolates were inoculated in a mustard extract medium and incubated for 48 h. Then, cultures were centrifuged at 6000 rpm for 10 min at 4 °C and the supernatants were screened for GLs breakdown based on the barium sulfate test as described by Sakorn et al. (2002). Myrosinase activity was determined according to Palmieri et al. (1982). One unit of myrosinase activity was defined as the amount of enzyme capable of hydrolyzing 1 μmol of sinigrin per minute per mL under the experimental conditions.

Molecular identification and phylogenetic analysis
Total genomic DNA of bacterial cells showing myrosinase activity was isolated and purified using GeneJet™ Genomic DNA purification Kit (Thermo Scientific®, Waltham, USA) according to the manufacturer's protocol. The 16S rRNA gene was amplified using 27f and 1492r universal oligonucleotides (Weisburg et al. 1991). PCR was performed using the standard reaction mixture as described in (Youseif 2018). PCR reaction conditions were: initial denaturation at 95 °C for 3 min, 35 cycles of denaturation at 94 °C for 1 min, annealing at 58 °C for 45 s, and extension at 72 °C for 2 min. PCR fragments were purified using QIAquick PCR Purification Kit (Qiagen, Hilden, Germany) and sent to a sequencing service (Macrogen, Seoul, South Korea). Sequence reads were edited and assembled using DNA Baser assembly software (www. DnaBa ser. com). Isolates were identified by comparing their 16S rRNA sequences with reference sequences available at Ezbiocloud (https:// www. ezbio cloud. net/) and GenBank (www. ncbi. nlm. nih. gov) databases as previously described in (Youseif et al. 2021b). The sequences were aligned using ClustalW matrix-based algorithm and were subjected to phylogenetic analyses using MEGA X software version 10 (Kumar et al. 2018). 16S rRNA phylogenetic tree was generated using the Maximum likelihood (ML) algorithm and p-distance model.

Media optimization by one factor at a time (OFAT) method
Optimal media components for myrosinase production by Bacillus sp. strain NGB-B10 were screened via the OFAT method. The experiments were conducted in 500 mL Erlenmeyer flasks containing 200 mL using a mustard extract broth medium. After autoclaving, the flasks were cooled and inoculated with 1 mL freshly-grown culture containing 10 6 CFU mL −1 . All experiments were conducted in three sets and data were presented as mean values with ± SD.
For carbon sources, the leaf extract of Eruca vesicaria ssp. sativa (Mill.) Thell. (arugula), Brassica oleracea L. (red cabbage and white cabbage), Raphanus sativus L. (red radish and white radish); root extract of red radish and white radish; and seed extract of Sinapis alba L. (white mustard) and Brassica nigra (L.) W.D.J. Koch (black mustard seeds) were selected to evaluate the effect of carbon sources as inducers of myrosinase activity. The taxonomy and nomenclature were done according to The Integrated Taxonomic Information System (ITIS, https:// www. itis. gov/). Plant samples were purchased from markets, stored at − 80 °C for 24 h, and then freeze-dried. The natural GLs extracts were prepared from cruciferous plants according to the hot methanol extraction method (ISO 9167-1) as previously described in (Ishida et al. 2011). A 20 g of freeze-dried plant sample was grinding and cooked in hot water for 25 min. Then, the cooked samples were homogenized with methanol in a 3:2 solvent to fresh weight ratio and subsequently boiled at 80 °C under reflux for 30 min. The extract was filtered then methanol was evaporated in a rotary evaporator. The residual extract of the GLs extract was dissolved in 50 mL sterile 0.05 M phosphate buffer (pH 7). Each carbon source was investigated at a concentration of 2% and 10% (w/v) in the production medium.
For nitrogen sources, a concentration of 200 mmol L −1 of sodium nitrate, potassium nitrate, ammonium nitrate, ammonium chloride, and ammonium sulfate was selected to study the effect of different inorganic nitrogen sources on myrosinase activity.

Plackett-Burman statistical design for effective constituents
Seven independent variables including essential media components and culture conditions were selected for Plackett-Burman design (PBD) analysis as follows; agitation, C-source concentration, N-source concentration, temperature, pH, incubation period, and inoculum size. These factors were tested in 12 experimental runs and each factor was investigated at 2 levels; low level (−) and high level (+) as illustrated in Table S1. One millimeter freshly-grown culture of Bacillus sp. strain NGB-B10 containing 10 6 CFU mL −1 was inoculated into the production media and incubated at 30 °C for 48 h. PBD is based on the first-order model: where Y is the response (myrosinase production), β0 is the model intercept and βi is the linear coefficient, and Xi is the level of the independent variable. The statistical discovery JMP software version 14.1 (www. jmp. com) was used for the Plackett-Burman experimental design and subsequent regression analysis. The experiments were carried out in 3 replicates and the response was calculated as the average value of the results. Regression analysis was performed and significant variables were further optimized by using the RSM design.

The Box Behnken experimental design
According to PBD results, the BBD of RSM was employed to optimize the most significant variables (nitrogen concentration, incubation period, and agitation speed) for enhancing myrosinase production. Each variable was analyzed at 3 levels, low (− 1), medium (0), and high (+ 1) levels (Table S2). The design of the matrix includes 15 trials, all the trials were performed in triplicates and the final response value was calculated as the mean of three replicates (± SD). Myrosinase production was fitted using a second-order polynomial equation for predicting the optimal point. The general form of the second-order polynomial equation is: where Y is the predicted response, β0 is the intercept term, βi is the linear coefficient, βij is the quadratic coefficient, βii is the interaction coefficient, and XiXj represents the independent variables. The fitness of the polynomial model

Purification of the myrosinase enzyme
The crude extract of myrosinase was prepared using a modified M9 medium (10% leaf extract of red cabbage, 0.2% NaNO 3 , 0.3% KH 2 PO 4 , 0.05% MgSO 4 ·7H 2 O, and 0.002% CaCl 2 ·6H 2 O) as revealed by the statistical media optimization. One-millimeter fresh culture (OD 600 of 1.0) of Bacillus sp. strain NGB-B10 was inoculated into a 2 L flask containing 1 L of autoclaved medium and incubated in a shaking incubator at 250 rpm for 12 h at 30 °C. Bacterial culture was centrifuged at 6,000 rpm at 4 °C for 15 min and myrosinase activity was detected in the supernatant as described in (Palmieri et al. 1982). The total protein amount was determined by measuring optical density at 280 nm (Layne 1957;Stoscheck 1990). The myrosinase enzyme was purified and precipitated using ammonium sulfate fractionation. The crude extract was subjected to 20, 40, 60, 80, and 100% ammonium sulfate saturation by adding the salt slowly under continuous magnetic stirring in an ice water bath and was kept overnight at 4 °C. The number of grams of ammonium sulfate for each fraction was calculated using the online calculator (https:// www. encor bio. com/ proto cols/ AM-SO4. htm). After the precipitation of each fraction, the precipitated protein was separated by centrifugation at 15,000 rpm at 4 °C for 15 min (Kapilan 2016). The obtained pellet was dissolved in 15 mL of 0.1 M phosphate buffer pH 7.0 and dialyzed against the same buffer overnight at 4 °C using Spectrum™ dialysis membrane (Thermo Fisher Scientific, Waltham, USA). The dialysis process was repeated till ammonium sulfate was completely desalted. After dialysis, myrosinase activity and protein content in each fraction was determined as abovementioned.
The subunits of purified myrosinase enzyme in the 40% fraction with the highest activity were determined by using a sodium dodecyl sulfate-polyacrylamide gel electrophoresis SDS-PAGE as described in (Mohammed et al. 2008). SDS-PAGE analysis was performed using 12% acrylamide separating gel and 4% acrylamide stacking gel on a Mini-PRO-TEIN II Cell electrophoresis unit (Bio-Rad, Hercules, CA). Protein bands were visualized by staining with coomassie blue R-250. GangNam˗Stain™ prestained protein ladder was used as a standard (iNtRON Biotechnology, South Korea).

Characterization of the partially purified myrosinase enzyme
The purified extract of 40% fraction with the highest enzymatic activity was used to determine the main characteristics of the enzyme. All the assays were performed in triplicate.
One hundred microliters of 2 mM of sinigrin dissolved in 0.05 M phosphate buffer pH 7.0 were incubated with 20 µl of pure enzyme extract. The enzyme activity was determined by incubating the reaction mixture at different temperatures in the range from 10 to 90 °C. The optimum temperature for myrosinase activity was calculated as previously described in (Palmieri et al. 1982) by plotting enzyme activity against temperatures. The effect of pH on myrosinase activity was measured in buffers at different pH values ranging from 3.0 to 7.0 (using 0.05 mM sodium citrate buffer), 7.5 to 8.0 (using 0.05 mM sodium phosphate buffer), and 8.6 to 10.6 (using 0.05 mM glycine-NaOH buffer). The velocity of sinigrin catalysis by myrosinase was determined by using different concentrations of substrate (1, 2, 5, 10, 15, and 20 mM at 0.05 M in phosphate buffer pH 7.0). Enzyme kinetic constants (K m and V max ) were determined according to the double-reciprocal Lineweaver-Burk plot methods.

Screening the antifungal activity of GLs-myrosinase hydrolysis products
The antifungal activity of the GLs-hydrolysis products was evaluated against 9 phytopathogenic fungal strains corresponding to the following genera: Alternaria, Aspergillus, Curvularia, Fusarium, Penicillium, Sclerotium, and Stemphylium by using the agar well diffusion method. The plant pathogenic taxa were provided from the Suez Canal University Fungarium (SCUF), Egypt (https:// ccinfo. wdcm. org/ colle ction/ by_ id/ 1180). GLs natural extracts from 7 Brassica plants were hydrolyzed by adding myrosinase purified from Bacillus sp. NGB-B10. Cruciferous plant samples were purchased from markets and GLs extracts were prepared using the hot methanol extraction method (ISO 9167-1) as abovementioned. A 10 U of the purified myrosinase was added to them and incubated at 37 °C for 60 min to release hydrolytic products. Five hundred microliters of freshly-grown fungal cultures (10 8 spores mL −1 ) were spread on Czapek Dox agar plates (0.2% NaNO 3 , 0.05% KCl, 0.1% K 2 HPO 4 , 0.05% MgSO 4 , 0.001% FeSO 4 , 3% sucrose, and 1.5% agar). On each plate, wells were punched in the medium using a 6 mm diameter sterilized borer. Then, 0.5 mL of GLs-myrosinase complex was added to the wells, and subsequently, the plates were kept for 3˗4 days at 30 °C. A well containing 0.5 mL nystatin (25 µg mL −1 ) was used as positive control while 0.5 mL of each extract and 10 U of myrosinase were served as individual negative controls. The experiments were carried out in triplicates and antifungal activities were expressed as the mean of the inhibition zone diameters (mm).

Statistical analyses
The data obtained were expressed as mean ± standard deviation (SD). The statistical significance of the model was estimated using a one-way analysis of variance (ANOVA) at a significance level of 5% (p > 0.05) using the statistical analysis software SPSS version 16.0 (www. ibm. com). The mathematical modeling was performed using the statistical discovery JMP software version 14.1 (www. jmp. com).

Isolation of myrosinase-producing bacteria
A total of 36 bacterial isolates (15 rhizospheric and 21 endophytic) were isolated in this study. The endophytic bacteria were isolated from surface-sterilized roots (9 isolates) and leaves (12 isolates) of E. vesicaria ssp. sativa plants. Out of the 36 bacterial isolates, 9 endophytic isolates (2 isolates from roots and 7 isolates from leaves) could produce myrosinase activity in the range of 0.08 ± 0.03 to 19.92 ± 1.23 U mL −1 (Table 1). However, all rhizospheric isolates failed to produce any myrosinase activity.

Bacterial identification
The 16S rRNA genes from the 9 endophytic bacteria with myrosinase activity were sequenced and deposited (accessions: LC589981-LC589989) in the Genbank database (www. ncbi. nlm. nih. gov). The sequences of the 16S rRNA gene were blasted to the EZBioCloud database (https:// www. ezbio cloud. net/) (Table 1). Isolates NGB-B11 and NGB-B14 shared 99.91 and 99.63% similarity, respectively, with the type strain Bacillus velezensis CR-502 T . While, the rest of the isolates had 99.04-99.93% 16S rRNA sequence similarity with the type strain Bacillus siamensis KCTC 13613 T , a plant growth-promoting bacterium with antimicrobial compounds against plant pathogens (Jeong et al. 2012). According to the ML-phylogenetic tree based on 16S rRNA sequences (Fig. 2), all bacterial isolates were classified into the genus Bacillus. All myrosinase-producing bacteria were clustered with five species of the genus Bacillus: B. amyloliquefaciens, B. nakamurai, B. siamensis, B. vallismortis, and B. velezensis. Due to the low phylogenetic power at the species level, the newly isolated bacteria in this study were assigned only to the genus level and were identified as Bacillus sp.

Classical optimization of media components to enhance myrosinase production
To achieve optimal enzyme production by the highest myrosinase-producing bacterium (Bacillus sp. strain NGB-B10), various studies on medium composition had been carried out to select the optimum carbon and nitrogen sources by using the OFAT. The effects of different carbon (2 and 10% w/v) and nitrogen sources (200 mmol L −1 ) on the production of myrosinase by Bacillus sp. strain NGB-B10 are shown in Fig. 3a, b. Out of the 9 carbon sources investigated, the leaf extract of red cabbage exhibited the highest myrosinase activity (22.79 ± 0.50 U mL −1 ), followed by the leaf extract of white cabbage (12.04 ± 1.58 U mL −1 ) at a concentration of 10% (w/v). Generally, increasing the concentration of cruciferous plant extracts resulted in higher myrosinase yield except for leaf extract from arugula, and seed extract from black and white mustard. For nitrogen source, sodium nitrate was found to be the most effective for myrosinase production (7.45 ± 0.62 U mL −1 ) which was followed by ammonium sulfate (5.38 ± 0.09 U mL −1 ) and ammonium nitrate (5.18 ± 0.09 U mL −1 ). While potassium nitrate drastically decreased enzyme production (1.99 ± 0.03 U mL −1 ).

Fig. 2
The phylogenetic relationships between myrosinase-producing bacterial isolates (bold) and reference strains based on 16S rRNA gene sequences. NCBI accession numbers are in parentheses.

Plackett-Burman experimental design
A total of 7 parameters were screened for their effects on myrosinase production by Bacillus sp. strain NGB-B10 in 12 trials using the Plackett-Burman experimental design ( Table 2). The maximum myrosinase production (21.93 ± 4.40 U mL −1 ) was attained in trial No. 2, while no myrosinase activity was detected in trials No. 3, 7, and 8 ( Table 2). The Statistical analysis using the PBD (Tables 3, S3) indicated that nitrogen concentration (sodium nitrate), incubation period, and agitation speed significantly affected myrosinase production, while the other components were found to be insignificant. According to the design (Table S3), nitrogen concentration, temperature, pH and agitation speed have positive effects while three variables including the inoculum volume, the incubation period, and carbon source concentrations (leaf extract of red cabbage) have negative effects on the myrosinase activity. The interactive effects between selected critical components for maximum myrosinase production were further analyzed using the BBD of RSM.

Box-Behnken experimental design
Based on the PBD analysis, the BBD of RSM was applied to determine the optimal levels of the three selected variables (sodium nitrate, incubation period, and agitation speed). A total of 15 experiments with different combinations of the selected parameters at 3 levels were performed. The design matrix and the corresponding responses are presented in Table 4. The experimental results were analyzed by standard ANOVA. The BBD was fitted with the second-order polynomial equation: where Y is the predicted myrosinase production, X 1 , X 2 and X 3 correspond to incubation period, agitation speed, and sodium nitrate, respectively. The statistical analysis of obtained data revealed that the model F value was 120.68 with a low probability value (p < 0.0001), indicating that the proposed model is highly significant ( Table 5). The R 2 value of the model was 0.976368 and the adjusted R 2 value of the model was 0.994116 demonstrating the fitting of the tested model. Three-dimensional response surface curves were plotted to explain the interaction effect of the variables and to determine the optimum level of each variable for maximum response (Fig. 4a-c). Each figure demonstrated the effect of two factors while the third factor was fixed at the zero level (intermediate level). The optimum conditions for myrosinase production by Bacillus sp. strain NGB-B10 were determined using the response surface analysis and regression equation.
The predicted values of the regression equation were generally consistent with the experimental values obtained by three repeated experiments. At the optimum experimental conditions, the maximum myrosinase activity was found to be 135.68 ± 1.44 which was very close to the response predicted by the regression model verifying the validity of the model.
Based on the surface response graphs and the regression analysis, we found that the optimal conditions to achieve the maximum production of myrosinase enzyme were at 2.25 g L −1 sodium nitrate, 9.4 h of the incubation period, and 250 rpm. At these optimized conditions, the predicted enzyme activity was found to be 199.848 U mL −1 with a 10.03-fold enhancement as compared to the initial production medium (19.92 U mL −1 ).

Purification of myrosinase
Myrosinase was purified from the culture media of Bacillus sp. strain NGB − B10 by ammonium sulfate fractionation (20-100%). The results showed that the 40% fraction was the most adequate amount of salt to precipitate the myrosinase, as 477.6 U mg −1 protein of the enzyme activity was recovered in the precipitate (Fig. 5). However, the rest of the fractions had a lower efficiency for myrosinase precipitation (0.0-79.38 U mg −1 protein). The specific activity of crude enzyme extract increased from 2.1 to 477.7 (227.5-fold) with an 89.8% recovery of myrosinase activity by the ammonium sulfate precipitation (Table 6).
Considering the higher enzymatic recovery of the 40% fraction, the purified enzyme in this fraction was analyzed using SDS-PAGE, and two distinct bands were yielded on the gel (Fig. 6). In addition, Log M W vs. Relative Mobility (R f ) graph was plotted which showed that the enzyme subunits had an approximate molecular weight of 38.6 and 35.0 KDa (Fig. S1).

Characterization of partially purified myrosinase
The influence of temperature on the activity of purified myrosinase is shown in Fig. 7a. The results indicated that there was no activity below 10 °C and it increased as the incubation temperature increased until it reached the highest enzyme activity at 40 °C. The activity then declined rapidly at temperatures higher than 40 °C until completely disappear at 70 °C. The effect of pH on the myrosinase activity was evaluated using different buffer solutions (Fig. 7b). The enzymatic activity gradually increased as the pH increased from pH 5.0 until it reached its maximum at pH 8.0. The activity then steadily decreased until it was completely inactivated at pH 10.5. The K m and V max parameters were determined to understand the reaction kinetics between purified myrosinase and different concentrations of the substrate (sinigrin). According to the resulting Lineweaver-Burk plot, the purified myrosinase demonstrated a K m value of 2.5 mM and a V max value of 5 mM sinigrin min −1 mg −1 (Fig. S2).

Antifungal activity of GLs-myrosinase system
The growth inhibition of different phytopathogenic fungi growing on Czapek Dox agar medium and exposed to GLshydrolysis products was estimated using the agar well diffusion method, and the results had been expressed in mm as diameters of inhibition zones ( Fig. 8 and Table S4). The results showed that GLs-hydrolysis products from various plant sources varied in their antifungal activities against tested fungal pathogens except for C. tuberculate which was resistant to all GLs extracts. It is worth mentioning that, hydrolytic products in GLs extracts from some Brassicaceae species had higher or equivalent inhibitory effects to nystatin (a broad-spectrum antifungal agent). For example, the antagonistic effects of GLs-hydrolysis products in seed extract of white mustard against A. flavus and F. oxysporium were similar to those obtained by nystatin. Also, the antifungal activities of GLs-hydrolysis products released from the root extract of both red radish and white radish towards P. expansum were greater than those recorded by nystatin. Out of all tested phytopathogenic fungi, A. flavus was the most sensitive which was inhibited by most (7/9) of investigated extracts.

Discussion
A hallmark trait of the Brassicaceae family is their ability to produce GLs that can be hydrolyzed into glucose and bioactive products via the myrosinase enzyme (Kissen et al. 2009). The GLs-myrosinase system has been well studied in plants (Bhat and Vyas 2019) but little research has been done on the metabolism of GLs by microbial myrosinase (Narbad and Rossiter 2018). Arugula (E. vesicaria ssp. sativa) like other cruciferous vegetables produces diverse secondary metabolites, particularly the GLs which can influence its associated microbiota. This study used a cultivation-dependent approach to isolate GLs-degrading bacteria associated with arugula plants. All GLs degraders were only endophytic bacteria and were identified as Bacillus sp. using the 16S rRNA sequence analysis. The genus Bacillus has been previously described as an endophyte of many Brassica vegetables (Card et al. 2015). However, few studies have reported the isolation of strains belonging to Bacilli that are capable of in-vitro metabolism of GLs, for example, the ability of four Bacillus isolates to hydrolyze both sinigrin and other natural glucosinolates from rapemeal has been previously documented (Brabban and Edwards 1994). Diverse species of the genus Lactobacillus were also identified as GLs-degrading microorganisms (Palop et al. 1995;Sikorska-Zimny and Beneduce 2021).
Although there are a large number of reports on bacteria with myrosinase-like activities that can metabolize GLs and produce bioactive compounds (Narbad and Rossiter 2018), identification of the myrosinase enzyme in these microorganisms is still limited to only a few studies. It is now documented that, myrosinase metabolic activity is not restricted to a single phylotype or a family of bacterial species Wassermann et al. 2017;Tie et al. 2021;Cebeci et al. 2022). In the present study, we isolated the bglA gene, one of the genes responsible for myrosinase activity, from Bacillus sp. strain NGB-B10. Protein analysis using InterPro (https:// www. ebi. ac. uk/ inter pro/) revealed that the bglA identified in this study belongs to the GH1 family which is similar to some myrosinases of bacterial origin (Cordeiro et al. 2015;Wassermann et al. 2017). The bglA gene was previously detected in E. cloacae strain KS50, an active myrosinase-producing bacterium isolated from the endosphere of turnip cabbage (Wassermann et al. 2017). Furthermore, the ability E. coli strain 0157:H7, a high active GLs-metabolizing bacterium, to metabolize sinigrin was significantly reduced when the bglA was disrupted (Cordeiro et al. 2015). However, the bglA of the current study differs from other bacterial myrosinases belonging to glycoside hydrolase family 3 (GH3) which have been identified from Citrobacter sp. strain Wye1 Cebeci et al. 2022) and L. adecarboxylata strain G1 (Tie et al. 2021). It has been reported that myrosinase enzymes corresponding to GH1 and GH3 differ in their properties and mechanisms of hydrolysis (Tie et al. 2021). Taken together, our data indicated that the GLs-metabolizing mechanism of Bacillus sp. NGB-B10 appears to be similar to that of E. coli strain 0157:H7, not Citrobacter strain WYE1 or L. adecarboxylata strain G1. It is noteworthy that the detected similarity between bglA from Bacillus sp. NGB-B10 and both plant myrosinase and cabbage aphid myrosinase could be largely due to both of them falling into the glycoside hydrolase family 1 (GH1) (Husebye et al. 2005;Lv et al. 2022).
Previous studies showed that K m values, using sinigrin as a substrate, varied according to the type of myrosinase (Hang 1994;Jones et al. 2001;Pontoppidan et al. 2001;Andersson et al. 2009). Compared to other myrosinases of microbial origin, the K m value of the myrosinase enzyme identified from Bacillus NGB-B10 has a lower affinity for sinigrin than the cymr myrosinase identified from Citrobacter sp. strain WYE1 Cebeci et al. 2022). Similarly, it has a lower affinity for sinigrin than the plant myrosinase (Li and Kushad 2005) and aphid myrosinase (Husebye et al. 2005). While it has a higher affinity for sinigrin compared to fungal myrosinase identified from A. niger AKU 3302 (Ohtsuru and Hata 1973). We do not have a clear explanation for this inconsistency that is reflected in the literature, but it could be largely attributed to the source of the myrosinase, the type of substrate, or the experimental conditions. The production of enzymes by microorganisms depends on the availability of carbon and nitrogen sources (Mefteh et al. 2019). In the current study, various sources of carbon and nitrogen were investigated to reach the maximum myrosinase production by Bacillus sp. strain NGB-B10. We found that the leaf extract of red cabbage and sodium nitrate were the best sources of carbon and nitrogen, respectively. These results are consistent with the previous studies indicating the use of natural extracts of Brassica plants as carbon sources to induce myrosinase production (Rakariyatham et al. 2005;Szucs et al. 2018). The high efficiency of leaf extract of red cabbage to increase myrosinase production may be attributed to its high content of several GLs (Wu et al. 2021). Stimulation of the activity of various bacterial enzymes by sodium nitrate as a nitrogen source has been also reported in previous studies (Rajoka 2004;Sereen et al. 2015).
The use of statistical models such as RSM to optimize the production of microbial enzymes and important metabolites has been successfully applied in recent research, due to the applicability and suitability of this approach for a wide range of studies (Mefteh et al. 2019;Ahmad et al. 2020). To the best of our knowledge, this is the first report which gives a detailed study of the application of RSM for myrosinase production from bacteria. Based on the Plackett-Burman experimental design, sodium nitrate concentration, incubation period, and agitation speed were the significant variable affecting myrosinase production from Bacillus sp. strain NGB-B10. The Box-Behnken experimental design of RSM was then applied to address the optimum levels of each significant parameter and their interactive effects on myrosinase yield. The application of RSM resulted in a myrosinase activity with a 10.03 increase higher than that obtained before optimization. Various reports documented that nitrogen source has a major effect on regulating bacterial growth and enzyme synthesis (Mefteh et al. 2019;Sahu et al. 2020). The incubation period is also a key parameter influencing the production of bacterial enzymes in culture media or submerged fermentation which is correlated to the reduction of nutrients available to the bacterial cells (Sahu et al. 2020). Furthermore, agitation speed plays an important role in promoting cell growth and thus enzymatic production which is most likely due to the high oxygen supply and nutrient transfer rate in the culture medium (Sai-Ut et al. 2014).
The environmental conditions especially the temperature and pH have a significant role in the induction or inactivation of enzyme activity. Experimental results indicated that the myrosinase enzyme from Bacillus sp. strain NGB-B10 could work at a broad range of pH (5.5-10) and temperatures (10-65 °C), which increases the possibility of its application under various conditions. In contrast with other bacterial myrosinases whose maximal activities occur at 25 °C (Tie et al. 2021;Cebeci et al. 2022), the optimum temperature for myrosinase activity in this study was found to be 40 °C which is consistent with the activity observed for some myrosinases of plant origin (Andersson et al. 2009;Bhat et al. 2015). We do not have a clear explanation for the high thermal stability of myrosinase under this investigation which is higher than that reported so far for other microbial myrosinases, even those purified from fungi (Rakariyatham et al. 2005). However, it can be attributed to the nature of its source, as the genus Bacillus is known as a potential source of producing thermostable enzymes (Cotta et al. 2021;Aktayeva et al. 2022). Additionally, the myrosinase in this study appears to prefer alkaline conditions (with an optimal pH of 8.00) compared to myrosinases previously isolated from other bacteria that had optimal pH values in the range of 6.0-6.6 (Tie et al. 2021;Cebeci et al. 2022). Similar to our findings, higher optimal pH values ranging from 6.0 to 7.5 were reported for myrosinases purified from different fungal strains corresponding to Aspergillus sp. (Rakariyatham et al. 2005). Although there was no clear evidence, we suggested that it could be related to the alkaline nature of the Egyptian soils where the host plant of Bacillus sp. NGB-B10 was cultivated. In agreement with our hypothesis, (Cebeci et al. 2022) suggested that the pH stability profile of myrosinase isolated from Citrobacter sp. strain Wye1 might be correlated to the soil pH value where that bacterium was isolated.
The GLs-hydrolysis products have important impacts on plant protection (Wassermann et al. 2017;Arroyo et al. 2019). In this context, the experimental results demonstrated that GLs-hydrolysis products from various cruciferous plants released by the myrosinase from Bacillus sp. NGB-B10 had diverse antagonistic effects against several phytopathogenic fungi. These various antifungal activities may be related to the presence of different GLs contents in cruciferous plants (Bhandari et al. 2015;Rhee et al. 2020), which leads to the release of diverse hydrolytic bioactive compounds (Plaszkó et al. 2021). The high activities observed in the case of white and red radish root extracts against P. expansum could be largely due to their high GLs content of glucoraphasatin and glucoraphanin (Luo et al. 2022), the precursors of ITCs such as raphasatin and sulforaphane which have strong antimicrobial activities (Kim et al. 2015;Abukhabta et al. 2021). Whereas, the high antagonistic effects of white mustard seed extract against A. flavus and F. oxysporium are likely attributed to the hydrolysis products of sinalbin, the most common GLs in white mustard seeds and the precursor of benzyl-ITCs (Lietzow 2021) which are known as potent antifungal agents (Wang et al. 2020). Notably, in the current study, the GLs-hydrolysis products from all investigated cruciferous plants had no inhibitory effects against C. tuberculate fungus. This could be explained by the ability of some pathogenic microorganisms such as Sclerotinia sclerotiorum to detoxify GLs-hydrolysis products via enzymatic hydrolysis and therefore exhibit resistance to the GLs-myrosinase system (Chen et al. 2020).

Conclusion
The present study showed that bacterial endophytes of Brassica vegetables such as arugula are key participants in the GLs-myrosinase system and afford a new source of active microbial myrosinase. We identified the gene encoding this GH1 myrosinase and studied some characteristics of the purified enzyme. RSM was an effective tool for improving myrosinase production and resulted in a 10.03-fold enhancement in enzyme activity. The present work provides preliminary data that confirmed the presence of bioactive compounds in the GLs-hydrolysis products that have potent antifungal activities against several phytopathogenic fungi. These findings offer new sustainable approaches for plant protection by using Brassica endophytes. However, further studies are required to scale-up myrosinase production to control plant pathogens. Future perspectives of this study also include structural analysis of this GH1-myrosinase as