Production of Verbascoside, Isoverbascoside and Phenolic Acids in Callus, Suspension, and Bioreactor Cultures of Verbena officinalis and Biological Properties of Biomass Extracts

Callus, suspension and bioreactor cultures of Verbena officinalis were established, and optimized for biomass growth and production of phenylpropanoid glycosides, phenolic acids, flavonoids and iridoids. All types of cultures were maintained on/in the Murashige and Skoog (MS) media with 1 mg/L BAP and 1 mg/L NAA. The inoculum sizes were optimized in callus and suspension cultures. Moreover, the growth of the culture in two different types of bioreactors—a balloon bioreactor (BB) and a stirred-tank bioreactor (STB) was tested. In methanolic extracts from biomass of all types of in vitro cultures the presence of the same metabolites—verbascoside, isoverbascoside, and six phenolic acids: protocatechuic, chlorogenic, vanillic, caffeic, ferulic and rosmarinic acids was confirmed and quantified by the HPLC-DAD method. In the extracts from lyophilized culture media, no metabolites were found. The main metabolites in biomass extracts were verbascoside and isoverbascoside. Their maximum amounts in g/100 g DW (dry weight) in the tested types of cultures were as follow: 7.25 and 0.61 (callus), 7.06 and 0.48 (suspension), 7.69 and 0.31 (BB), 9.18 and 0.34 (STB). The amounts of phenolic acids were many times lower, max. total content reached of 26.90, 50.72, 19.88, and 36.78 mg/100 g DW, respectively. The highest content of verbascoside and also a high content of isoverbascoside obtained in STB (stirred-tank bioreactor) were 5.3 and 7.8 times higher than in extracts from overground parts of the parent plant. In the extracts from parent plant two iridoids—verbenalin and hastatoside, were also abundant. All investigated biomass extracts and the extracts from parent plant showed the antiproliferative, antioxidant and antibacterial activities. The strongest activities were documented for the cultures maintained in STB. We propose extracts from in vitro cultured biomass of vervain, especially from STB, as a rich source of bioactive metabolites with antiproliferative, antioxidant and antibacterial properties.

Macroscopic analysis of the appearance of callus cultures (embryogenic callus) showed the presence of globular embryos. During the experiment, they were observed to divide and form larger aggregates. The central part of the embryos had intensive green colour whereas the outer cells were parenchymatic and connected to neighbouring embryos through the central plaques. However, single embryos were not found. The tissue did not form cell clusters, which would favour further cytodifferentiation. In the outer parts, the biomass was dying-was characterized by a dark brown colour ( Figure 1A). The biomass increments recorded during 2-week growth cycles varied depending on the inoculum to medium ratio. The values ranged from 6.66 to 16.77 times (fresh biomass) and from 6.34 to 18.06 times (dry biomass) ( Figure 2). With the increase in inoculum size, a gradual decrease in biomass increments was observed, from 16.77-through 10.5-to 6.66-fold (fresh), and from 18.06through 9.88-to 6.34-fold (dry biomass) ( Figure 2). The biomass increments recorded during 2-week growth cycles varied depending on the inoculum to medium ratio. The values ranged from 6.66 to 16.77 times (fresh biomass) and from 6.34 to 18.06 times (dry biomass) ( Figure 2). With the increase in inoculum size, a gradual decrease in biomass increments was observed, from 16.77-through 10.5-to 6.66-fold (fresh), and from 18.06-through 9.88-to 6.34-fold (dry biomass) ( Figure 2).
The estimated compounds were not found in the analyzed extracts from lyophilized samples of the experimental media. Neither iridoids nor flavonoids were found in the extracts of biomass from callus and in the culture media. Table 1. Phenylpropanoid glycosides accumulation (mg/100 g DW ± SD) and their specific productivity* (mg/L/day) in studied callus, suspension and bioreactor V. officinalis in vitro cultures. Values are expressed as the mean ± SD (n = 3). Vo-C-V. officinalis callus culture extract; Vo-S-V. officinalis suspension culture extract; Vo-BB-V. officinalis balloon bioreactor culture extract; Vo-STB-V. officinalis stirred-tank bioreactor culture extract. a-g Different letters within the same column indicate significant differences between mean values (P < 0.05).  100 g DW × number o f days a f ter which the maximum concentration o f the certain molecule was reached Values are expressed as the mean ± SD (n = 3). Vo-C-V. officinalis callus culture extract; Vo-S-V. officinalis suspension culture extract; Vo-BB-V. officinalis balloon bioreactor culture extract; Vo-STB-V. officinalis stirred-tank bioreactor culture extract. a-g Different letters within the same column indicate significant differences between mean values (P < 0.05).

Verbascoside
The estimated compounds were not found in the analyzed extracts from lyophilized samples of the experimental media. Neither iridoids nor flavonoids were found in the extracts of biomass from callus and in the culture media.
In biomass extracts, the amounts of verbascoside varied moderately from 7154.93 (with 0.3 g of inoculum) to 7248.66 mg/100 g DW (with 0.9 g of inoculum). With 0.6 g of inoculum, the verbascoside content was almost identical to that obtained with 0.3 g of inoculum (7176.14 mg/100 g DW). The amounts of isoverbascoside were more varied, ranging from 442.81 (with 0.3 g of inoculum) through 528.17 (0.6 g of inoculum) to the maximum amount of 609.26 mg/100 g DW (0.9 g of inoculum) ( Table 1). For the production of the two estimated phenylpropanoid glycosides, the determined optimal inoculum size was 0.9 g. With this amount of inocular biomass, the total amount of the two compounds was the highest (7857.92 mg/100 g DW). Values are expressed as the mean ± SD (n = 3). Vo-C-V. officinalis callus culture extract; Vo-S-V. officinalis suspension culture extract; Vo-BB-V. officinalis balloon bioreactor culture extract; Vo-STB-V. officinalis stirred-tank bioreactor culture extract. a-g Different letters within the same column indicate significant differences between mean values (P < 0.05).
Molecules 2020, 25, 5609 The total amount of the estimated phenolic acids also varied depending on the amount of inoculum, from 19.42 mg/100 g DW (with the smallest inoculum of 0.3 g) through 26.90 mg/100 g DW (0.6 g) to 25.00 mg/100 g DW (0.9 g) ( Table 2). The main phenolic acids were protocatechuic acid and ferulic acid. The amounts of these two compounds ranged from 6.91 to 9.58 mg/100 g DW, and from 4.91 to 7.35 mg/100 g DW, respectively. The maximum amounts of the other four phenolic acids, i.e., chlorogenic, vanillic, caffeic and rosmarinic acids, did not exceed 1.8, 1.3, 3.6 and 3.8 mg/100 g DW, respectively.
The sizes of inoculum that provide the best accumulation of the two groups of metabolites-phenylpropanoid glycosides and phenolic acids are 0.6 g and 0.9 g. It was with these amounts of inoculum that the high concentrations of both groups of compounds were obtained. However, the optimum amount of inocular biomass is 0.6 g, which provides both high production of secondary metabolites (7704.31 and 26.90 mg/100 g DW) and high dry biomass increases (9.88 times) (Tables 1 and 2).

Morphology, Inoculum Selection and Biomass Increases
As in the case of agar callus cultures, various inoculum sizes were tested during initiation of suspension cultures (1.5, 3.0 and 4.5 g of fresh biomass in 30 mL of medium). In the current work, V. officinalis suspension cultures were established for the first time ( Figure 1B).
The studied embryogenic suspension cultures were characterized by fast growth. However, a tendency to necrosis has also been observed. Noticeable browning of the experimental media and tissue, observed already in the early days of the experiment, indicated either an advantage in the population of lysing cells or the secretion of phenolic compounds into the media.
During 2-week growth cycles, very different increments in both fresh and dry biomasses were recorded depending on the inoculum to volume-of-medium ratio. These increases ranged from 4.92 through 2.62 to 1.81 times (fresh biomass), and from 5.02 through 2.48 to 1.87 times (dry biomass) ( Figure 2). As was the case with the agar cultures, with an increase in the size of inoculum there was a gradual decrease in biomass increments.

Phytochemical Analysis of Biomass Extracts
The qualitative metabolite profile found in the analyzed extracts from the biomass of the suspension cultures was identical to that obtained for the agar cultures (Tables 1 and 2). The presence of verbascoside and isoverbascoside, together with the same 6 phenolic acids was confirmed-these were protocatechuic, chlorogenic, vanillic, caffeic, ferulic and rosmarinic acids. As in the case of callus cultures, no iridoids and flavonoids were found in the analyzed samples, and also none of the metabolites was found in the extracts from the lyophilized media samples.
The analyses provided evidence of significant differences in the verbascoside content, dependent on the amount of inoculum. The concentrations ranged from 6288.95 (4.5 g of inoculum) to 7059.37 mg/100 g DW (3.0 g of inoculum). With the smallest inoculum size used (1.5 g of inoculum), the concentration of verbascoside was 6845.25 mg/100 g DW (Table 1). In the case of isoverbascoside, the analyses also showed large differences in its concentration. The amounts of this metabolite ranged from 264.36 mg/100 g DW (1.5 g of inoculum) through 387.19 mg/100 g DW (3.0 g of inoculum) to the maximum of 477.25 mg/100 g DW (4.5 g of inoculum) ( Table 1). The total concentration of both phenylpropanoid glycosides ranged from 6766.21 mg/100 g DW (4.5 g of inoculum) through 7109.61 mg/100 g DW (1.5 g of inoculum), to the maximum of 7446.56 mg/100 g DW (3.0 g of inoculum). The amount of inoculum found to be optimal for the production of these two compounds was 3.0 g ( Table 1).
The total phenolic acids content increased almost two-fold, depending on the inoculum size, from 28.09 mg/100 g DW (1.5 g of inoculum) through 49.46 mg/100 g DW (3.0 g) to a maximum of 50.72 mg/100 g DW (4.5 g) ( Table 2). Quantitatively dominant metabolites were three acids-protocatechuic acid, ferulic acid and additionally rosmarinic acid. The concentrations of these compounds ranged from 2.09 to 10.64 mg/100 g DW, from 6.78 to 10.44 mg/100 g DW, and from 13.39 to 26.34 mg/100 g DW, respectively. The concentrations of the remaining three phenolic acids, i.e., chlorogenic, vanillic and caffeic acids, were lower, with maximum values of 1.25, 1.43 and 4.80 mg/100 g DW, respectively ( Table 2).
The size of inoculum that should be considered as providing the best conditions for biomass growth and the accumulation of both groups of metabolites-phenylpropanoid glycosides and phenolic acids-is 1.5 g. With this inoculum size, the total concentration of verbascoside and isoverbascoside is high (7109.61 mg/100 g DW), and biomass increases are the highest-5.02-fold (dry biomass). However, the concentration of phenolic acids under these conditions is low (28.09 mg/100 g DW) (Tables 1 and 2).

Morphology, Inoculum Selection and Biomass Increases
The suspension cultures were cultured for the first time in two types of bioreactors: a balloon bioreactor (BB) and a stirred-tank bioreactor (STB). In both cases, the working volume of the system was 1000 mL and the experiment duration was 2 weeks.
Macroscopic and microscopic observations revealed that biomass grown in BB was more viable than cells cultured in STB ( Figure 1C,D). In the case of STB, during active growth, a faster browning of the growth medium and the presence of a suspension of broken cells accumulating in the vicinity of the bubbler was found. In the case of biomass incubated in BB, the medium did not turn brown during a 2-weeks growth cycle only showing slightly pearly opalescence ( Figure 1C,D).
In both types of bioreactor, during the 2-week culture cycles, similar dry biomass increases of 4.55-fold (BB) and 5.09-fold (STB) were recorded ( Figure 2).

Phytochemical Analysis of Biomass Extracts
In the bioreactor cultures, the qualitative composition of secondary metabolites was found to be the same as in the agar and suspension cultures: the presence of verbascoside, isoverbascoside, and phenolic acids (protocatechuic, chlorogenic, vanillic, caffeic, ferulic and rosmarinic acids) was confirmed. None of the analyzed iridoids or flavonoids were found in the biomass. The estimated compounds were not found in the analyzed extracts from media lyophilizates.
High concentrations of verbascoside were found in biomass extracts-7685.00 mg/100 g DW (BB) and 9176.87 mg/100 g DW (STB) ( Table 1). The amount obtained in the STB system was the highest obtained in the course of the study. The amounts of isoverbascoside obtained in BB and STB cultures were equal to 306.40 and 339.91 mg/100 g DW, respectively, and were markedly lower than those obtained in the agar callus cultures (max. 609.26 mg/100 g DW), and also lower than the maximum amounts found in the suspension cultures (477.25 mg/100 g DW) ( Table 1).
The total amounts of verbascoside and isoverbascoside obtained in the STB were significantly higher than those in the BB (9516.79 and 7991.40 mg/100 g DW, respectively) ( Table 1).
The confirmed total amounts of phenolic acids in the cultures grown in the BB were very low (19.88 mg/100 g DW) ( Table 2). On the other hand, they were almost than twice as high (36.78 mg/100 g DW) in the cultures grown in the STB. In both types of bioreactors, the quantitatively dominant phenolic acids were the same as in the agar and/or suspension cultures, i.e., rosmarinic acid (10.22 and 14.87 mg/100 g DW), protocatechuic acid (5.07 and 9.68 mg/100 g DW), and ferulic acid (2.12 and 7.10 mg/100 g DW in BB and STB, respectively). The amount of total phenolic acids obtained in the STB bioreactor was higher than those of the agar callus cultures (19.42-26.90 mg/100 g DW) but lower than their maximum amount (50.72 mg/100 g DW) in the suspension cultures (Table 2).

Phytochemical Analysis of Parent Plant Material
The HPLC-DAD analysis of V. officinalis herb showed the presence of phenylpropanoid glycosides, phenolic acids, iridoids and flavonoids ( Figure 3).

Artemia salina Leach Lethality Bioassay
The extract obtained from V. officinalis grown in vivo (Vo-in vivo) was found to be non-toxic for A. salina, with an LC50 > 1000 μg/mL; by contrast, the extracts obtained from in vitro biomasses resulted in toxic effects against brine shrimps. Based on Clarkson's toxicity criterion, V. officinalis agar callus culture extract (Vo-C) showed low toxicity (LC50 = 698.67 μg/mL), whereas V. officinalis suspension culture extract (Vo-S) and V. officinalis stirred-tank bioreactor culture extract (Vo-STB) were moderately toxic, with an LC50 value of 479.34 μg/mL and 187 μg/mL, respectively [28].

Effect on SH-SY5Y Human Neuroblastoma Cell Proliferation
Each extract was tested for its potential anti-proliferative activity on human neuroblastoma SH-SY5Y cell line, at concentrations ranging from 125 to 750 μg/mL, for 24-72 h. As shown in Figure 4, each extract was able to hinder SH-SY5Y cell growth, yet with different efficacy.
Considering the results in detail, both Vo-in vivo and Vo-STB significantly reduced cell viability by more than 70% at 500 μg/mL after 24 h of treatment, whereas the latter was the only one among the other tested extracts to reach an inhibition of more than 90% at 500 μg/mL after 72 h. Vo-C showed the weakest anti-proliferative activity among the others, reducing cell growth by more than 70% at 500 μg/mL and by 90% at 750 μg/mL after 72 h of treatment. Vo-S followed the same trend as Vo-C, reaching the 90% inhibition at the highest concentration after the longest exposure time; however, unlike Vo-C, it lowered growth rate by more than 70% at 500 μg/mL after 48 h of treatment.
These results were further confirmed by the extrapolation of IC50 values. In particular, Vo-STB showed the lowest IC50 value of 240.6 μg/mL at 72 h, followed by Vo-in vivo with an IC50 value of 275.1 μg/mL at 72 h. Vo-C and Vo-S showed IC50 values of 454.7 μg/mL and 419.9 μg/mL after 72 h of treatment, respectively. The analyzed herb extracts were found to contain a high amount of verbascoside (1728.97 mg/100 g DW) and a much lower concentration of isoverbascoside (78.34 mg/100 g DW). From the phenolic acids investigated, in extracts three compounds-protocatechuic acid (25.75 mg/100 g DW), ferulic acid (29.76 mg/100 g DW) and rosmarinic acid (2.53 mg/100 g DW) were estimated. Of iridoids, verbenalin and hastatoside were found in the analyzed herb extracts. The verbenalin content was very high-4210.88 mg/100 g DW. The concentration of hastatoside was lower-398.87 mg/100 g DW. Among flavonoids, scutellarin and apigenin 7-glucuronide were found in the analyzed herb extracts. The scutellarin content was 172.13 mg/100 g DW while apigenin 7-glucuronide content was about 4-fold lower-41.77 mg/100 g DW ( Figure 3).

Artemia salina Leach Lethality Bioassay
The extract obtained from V. officinalis grown in vivo (Vo-in vivo) was found to be non-toxic for A. salina, with an LC 50 > 1000 µg/mL; by contrast, the extracts obtained from in vitro biomasses resulted in toxic effects against brine shrimps. Based on Clarkson's toxicity criterion, V. officinalis agar callus culture extract (Vo-C) showed low toxicity (LC 50 = 698.67 µg/mL), whereas V. officinalis suspension culture extract (Vo-S) and V. officinalis stirred-tank bioreactor culture extract (Vo-STB) were moderately toxic, with an LC 50 value of 479.34 µg/mL and 187 µg/mL, respectively [28].

Effect on SH-SY5Y Human Neuroblastoma Cell Proliferation
Each extract was tested for its potential anti-proliferative activity on human neuroblastoma SH-SY5Y cell line, at concentrations ranging from 125 to 750 µg/mL, for 24-72 h. As shown in Figure 4, each extract was able to hinder SH-SY5Y cell growth, yet with different efficacy. officinalis stirred-tank bioreactor culture extract. Each value is the mean ± SEM of three different experiments performed in eight replicates. * P < 0.05 vs. ctrl, ** P < 0.01 vs. ctrl, *** P < 0.001 vs. ctrl and **** P < 0.0001 vs. ctrl.

Antioxidant Activity
The estimation of total phenolics contained in V. officinalis extracts was carried out using spectrophotometric analysis with Folin-Ciocalteu phenol reagent. As shown in Table 3, the total phenolic content varied from 126.55 ± 0.78 mg GAE/g extract (Vo-C) to 189.91 ± 2.93 mg GAE/g extract (Vo-S) and it decreased in the order: Considering the results in detail, both Vo-in vivo and Vo-STB significantly reduced cell viability by more than 70% at 500 µg/mL after 24 h of treatment, whereas the latter was the only one among the other tested extracts to reach an inhibition of more than 90% at 500 µg/mL after 72 h. Vo-C showed the weakest anti-proliferative activity among the others, reducing cell growth by more than 70% at 500 µg/mL and by 90% at 750 µg/mL after 72 h of treatment. Vo-S followed the same trend as Vo-C, reaching the 90% inhibition at the highest concentration after the longest exposure time; however, unlike Vo-C, it lowered growth rate by more than 70% at 500 µg/mL after 48 h of treatment.
These results were further confirmed by the extrapolation of IC 50 values. In particular, Vo-STB showed the lowest IC 50 value of 240.6 µg/mL at 72 h, followed by Vo-in vivo with an IC 50 value of 275.1 µg/mL at 72 h. Vo-C and Vo-S showed IC 50 values of 454.7 µg/mL and 419.9 µg/mL after 72 h of treatment, respectively.

Antioxidant Activity
The estimation of total phenolics contained in V. officinalis extracts was carried out using spectrophotometric analysis with Folin-Ciocalteu phenol reagent. As shown in Table 3, the total phenolic content varied from 126.55 ± 0.78 mg GAE/g extract (Vo-C) to 189.91 ± 2.93 mg GAE/g extract (Vo-S) and it decreased in the order: Vo-S > Vo-STB > Vo-in vivo > Vo-C. Values are expressed as the mean ± SD (n = 3). Vo-in vivo-V. officinalis parent plant grown in vivo-herb extract; Vo-C-V. officinalis callus culture extract; Vo-S-V. officinalis suspension culture extract; Vo-STB-V. officinalis stirred-tank bioreactor culture extract; ASE-ascorbic acid equivalent; GAE-gallic acid equivalent; BHT-butylated hydroxytoluene; EDTA-Ethylenediaminetetraacetic acid. a-e Different letters within the same column indicate significant differences between mean values (P < 0.05).
The primary antioxidant efficacy of V. officinalis extracts was examined using the DPPH test and the reducing power assay; the secondary antioxidant properties were determined by measuring the chelating activity of the extracts. Furthermore, linear regression analysis was used to explore the relationship between the antioxidant activity of V. officinalis extracts and their phenolic content [29]. The results of the DPPH assay showed that all V. officinalis extracts exhibited strong radical scavenging ability ( Figure 5). Among the extracts, Vo-S and Vo-STB were the most effective, as confirmed also by the IC 50 values (0.110 mg/mL and 0.137 mg/mL); indeed, at the dose of 0.25 mg/mL both the extracts showed an activity superimposable to that of the standard BHT (88%), and much higher than that of Vo-in vivo and Vo-C extracts. At the concentration of 0.5 mg/mL, all the tested extracts reached their maximum effect, which was comparable to that of BHT (96%). Based on the comparison of IC 50 values, the scavenging activity of extracts and standard decreased in the order: BHT > Vo-S > Vo-STB > Vo-C = Vo-in vivo (Table 3). A strong positive correlation was found between total phenolic content and DPPH radical scavenging activity (R 2 = 0.9839).    Table 3 show the results of the reducing power assay. All the extracts displayed good reducing power, that increased with raising concentrations; in this test, Vo-S and Vo-STB also exhibited the strongest activity, as confirmed by the calculated ASE/mL values (ASE/mL = 1.762 and 1.779, respectively). At the concentration of 1 and 2 mg/mL, the reducing power of both the extracts was found to be slightly higher than that of BHT. Comparing the ASE/mL values, the reducing power of extracts and standard decreased in the following order: BHT > Vo-S = Vo-STB > Vo-C = Vo-in vivo. Finally, a strong positive correlation was found between total phenolic content and reducing power (R 2 = 0.9629).   The results of the Fe 2+ chelating activity assay showed that Vo-S, Vo-STB and Vo-C extracts interfered with the formation of the ferrous-ferrozine complex starting from the concentration of 0.5 mg/mL, displaying at the maximum tested dose (2 mg/mL) an activity slightly lower than that of the standard EDTA, whereas Vo-in vivo exhibited weak chelating effects only at the concentration of 1 and 2 mg/mL ( Figure 7). According to the calculated IC 50 values, Vo-S was found to be the most effective, followed by Vo-C (IC 50 = 0.767 mg/mL and 0.908 mg/mL).

Antibacterial Activity of Extracts
The MIC (Minimal Inhibitory Concentration) and MBC (Minimal Bactericidal Concentration) values of extracts obtained from V. officinalis were determined to evaluate their activity against selected pathogenic bacteria. Vo-in vivo, Vo-C, Vo-S and Vo-STB showed antibacterial activity against tested bacterial strains ( Table 4). The V. officinalis extracts showed stronger bacteriostatic activity against Gram-positive bacteria (MIC 0.6−4.5 mg/mL), and lower in relation to Gram-negative bacteria (MIC 0.6-9.0 mg/mL). The extracts had similar bactericidal activity, and MBC values ranged from 0.6 to 18.0 mg/mL. The test strains had different sensitivity to the V. officinalis extracts. The most sensitive bacteria were Y. enterocolitica, K. pneumoniae (MIC/MBC 0.6-2.2 mg/mL) and S. epidermidis (MIC 0.6-1.1 mg/mL, MBC 0.6-9.0 mg/mL). Differences in the bacteriostatic and bactericidal activity of the extracts were also observed.
Vo-C showed a weaker antibacterial effect than the other tested extracts. The growth of most Gram-positive strains was inhibited at the MIC of 9.0 mg/mL, and a bactericidal effect was observed at an MBC of 18.0 mg ml. L. monocytogenes, E. aerogenes and E. coli were more resistant to Vo-C.
Vo-S had weaker antibacterial activity compared to Vo-in vivo, but stronger than Vo-C especially against Gram-positive bacteria. Vo-S strongly inhibited the growth of S. epidermidis (MIC 0.6 mg/mL), but weakly E. coli (MIC 9.0 mg/mL).  The metal chelating effect of the V. officinalis extracts and EDTA decreased in the order: EDTA > Vo-S > Vo-C > Vo-STB > Vo-in vivo (Table 3). Moreover, no correlation was found between metal chelating activity and total phenolic content (R 2 = 0.2391).

Antibacterial Activity of Extracts
The MIC (Minimal Inhibitory Concentration) and MBC (Minimal Bactericidal Concentration) values of extracts obtained from V. officinalis were determined to evaluate their activity against selected pathogenic bacteria. Vo-in vivo, Vo-C, Vo-S and Vo-STB showed antibacterial activity against tested bacterial strains ( Table 4). The V. officinalis extracts showed stronger bacteriostatic activity against Gram-positive bacteria (MIC 0.6−4.5 mg/mL), and lower in relation to Gram-negative bacteria (MIC 0.6-9.0 mg/mL). The extracts had similar bactericidal activity, and MBC values ranged from 0.6 to 18.0 mg/mL. The test strains had different sensitivity to the V. officinalis extracts. The most sensitive bacteria were Y. enterocolitica, K. pneumoniae (MIC/MBC 0.6-2.2 mg/mL) and S. epidermidis (MIC 0.6-1.1 mg/mL, MBC 0.6-9.0 mg/mL). Differences in the bacteriostatic and bactericidal activity of the extracts were also observed.
Vo-C showed a weaker antibacterial effect than the other tested extracts. The growth of most Gram-positive strains was inhibited at the MIC of 9.0 mg/mL, and a bactericidal effect was observed at an MBC of 18.0 mg ml. L. monocytogenes, E. aerogenes and E. coli were more resistant to Vo-C. Vo-S had weaker antibacterial activity compared to Vo-in vivo, but stronger than Vo-C especially against Gram-positive bacteria. Vo-S strongly inhibited the growth of S. epidermidis (MIC 0.6 mg/mL), but weakly E. coli (MIC 9.0 mg/mL).
Vo-STB had the best antibacterial effect among the investigated V. officinalis extracts. Vo-STB inhibited the growth of Gram-positive bacteria at MICs in the range of 0.6−2.2 mg /mL and Gram-negative at MICs in the range 0.6−4.5 mg/mL. In particular, strains of S. epidermidis, Y. enterocolitica and K. pneumoniae were more sensitive (MIC/MBC 0.6 mg/mL). Vo-STB also showed lower MBC values in relation to most strains compared Vo-in vivo, Vo-C and Vo-S.
Comparing the MIC/MBC values the antibacterial activity of extracts decreased in the following order: Vo-STB > Vo-in vivo > Vo-S > Vo-C.

Biomass Increments
As demonstrated in numerous studies on cell cultures of higher plants, the selection of optimum inoculum size is crucial for obtaining high biomass yield [30]. In the V. officinalis agar callus cultures cultivated now on the best growth and production medium selected in our previous studies (MS, BAP-1 mg/L, IBA-1 mg/L), high, but strongly varied, increases in dry biomass were obtained, from 6.34 to 18.06 times, during 2-week culture cycles, depending on the ratio of the amount of inoculum to the volume of the medium. An increase in the inoculum size from 0.3 to 0.9 g resulted in biomass increments becoming gradually lower.
The biomass increases obtained in the current work were higher than the maximum increase obtained in our previous experiments with vervain (a 5.4-fold increase over 4-week growth cycles).
In the suspension cultures of V. officinalis, successfully established for the first time in the present work, also grown on the previously selected MS medium (containing 1 mg/L BAP and 1 mg/L IBA), the increases in biomass varied from 1.87 to 5.02 times (Figure 2), depending on the inoculum to volume-of-medium ratio. Satisfactory 5.02-fold increases were obtained during 2-week growth cycles with the smallest tested amount of inoculum (1.5 g per 30 mL of medium). Such biomass increments are satisfactory enough for suspension cultures.
After the suspension cultures of V. officinalis had been established, attempts were made to introduce them to bioreactor cultivation which is a crucial step in scale-up operations. In the current work, two bioreactors were employed: BB (balloon bioreactor) and STB (stirred-tank bioreactor). Both of these bioreactor types have been commonly used for plant cell cultivation, in the laboratory and on an industrial scale [31]. In the bioreactor cultures established by us, cultivated in BB and STB, the obtained increases in biomass during 2-week cycles were comparable (4.55 and 5.09 times) and satisfactory enough (Figure 2).

Phytochemical Analysis
Phytochemical analysis of herb was carried out to make a reliable assessment of the biotechnological research results.
The phytochemical analyses performed by us have demonstrated that biosynthetic profile of cells from in vitro cultures is different as compared to those of parent plant grown in vivo. Iridoids (verbenalin and hastatoside) dominate quantitatively in the herb of the plant. Quite large amounts of phenylpropanoid glycosides (verbascoside and isoverbascoside) and moderate amounts of phenolic acids and flavonoids were also found.
The secondary metabolism pathways that are dominant in the cells of in vitro cultured biomass are shikimate and cinnamate pathways leading to the formation of verbascoside and, to a lesser extent, isoverbascoside. Cell culture extracts were also found to have a phenolic acid profile comparable to that of the parent plant grown in vivo. The maximum amounts of phenylpropanoid glycosides obtained in cell cultures exceeded the values recorded for parent plant material. These results suggest that tissue differentiation in Verbena does not favour the accumulation of verbascoside and isoverbascoside. On the other hand, iridoids were absent in embryogenic calli and suspension cultures of vervain. These observations are in agreement with other studies. Sesterhenn et al. [32] demonstrated that callus cultures of Scrophularia nodosa were devoid of iridoids which are otherwise abundant in organized biomass of the species. Another example is the study by Grąbkowska et al. [33] which showed that tissue dedifferentiation in Harpagophytum procumbens in vitro cultures is positively correlated with verbascoside content and negatively correlated with the production of iridoids. The ability of non-organized cultures to accumulate verbascoside was also documented in the study on callus and suspension cultures of Scrophularia striata [34].
As revealed by HPLC analyses of media samples, all of the synthesized metabolites in V. officinalis suspension and bioreactor cultures were stored intracellularly which is a common attribute of plant in vitro cultures [31].
In the present work, high amounts of verbascoside were obtained in the agar callus cultures (max. 7248.66 mg/100 g DW), suspension cultures (max. 7059.37 mg/100 g DW), BB bioreactor (max. 7685.00 mg/100 g DW), and extremely high in the STB bioreactor (max. 9176.87 mg/100 g DW). Cultures were maintained in 14 days growth cycles which gives us the specific productivity 120.49, 74.36, 104.39 and 115.20 mg/L/day, respectively ( Table 1). The amounts of this metabolite were, respectively, 3.2, 3.1, 3,4 and 4.1 times higher than its maximum concentration in the soil-grown parent plant (2263.85 mg/100 g DW). The amounts of isoverbascoside were also significant: max. 609.26 mg/100 g DW (specific productivity 10.13 mg/L/day) in the agar callus cultures, max. 477.25 mg/100 g DW (4.85 mg/L/day) in the suspension cultures, max. 306.40 mg/100 g DW (4.16 mg/L/day) in the BB bioreactor, and max. 339.91 mg/100 g DW (4.27 mg/L/day) in the STB bioreactor. The total amounts of verbascoside and isoverbascoside obtained in both suspension (7446.56 mg/100 g DW) and bioreactor (BB-7991.4 mg/100 g DW, STB-9516.79 mg/100 g DW) cultures are very high (Table 1).
High concentrations of both constituents were obtained within a short culture period (2 weeks), the established system can thus be considered as economically viable. The best results were obtained in the STB which belongs to the most commonly used bioreactor types [31]. The specific productivity in callus cultures is even higher than in STB cultures but obtaining biomass from bioreactor cultures is much easier and economically more profitable than obtaining it from agar cultures. The in vitro system of V. officinalis established by us can be proposed as a rich source of phenylpropanoid glycosides, in particular verbascoside. The concentrations of this glycoside obtained are of practical interest, and what is most important, the highest amount was produced in bioreactor cultures (STB bioreactor-9.17 g/100 g DW) ( Table 1).
Phenylpropanoid glycosides are often found in in vitro cultures of various plants. The highest concentration of verbascoside among plants grown in in vitro conditions were observed in Buddleja cordata cell suspension cultures and reached 11.6 g/100 g DW [35]. By comparison, the maximum contents of verbascoside as the main component in suspension cultures of Salvia vulgaris and Cistanche salsa have been 16 g/100 g DW and 689 mg/L, respectively [36,37]. It has to be noted, however, that those high amounts were obtained only after feeding with suitable precursors.
In other cell cultures investigated, the concentrations of verbascoside are much lower than those obtained in our research but concentrations of isoverbascoside were higher. For instance, Verbascum thapsus callus cultures contain a maximum of 4.7 g/100 g DW verbascoside and 5.0 g/100 g DW isoverbascoside. Suspension cultures of this species reached 3.9 g/100 g DW verbascoside and 2.9 g/100 g DW isoverbascoside [38]. Callus tissue of H. procumbens reached a maximum level of 3.1 g/100 g DW verbascoside, but the concentration of isoverbascoside-0.7 g/100 g DW, was higher than in our research [33]. Scutellaria alpina cell suspension cultures reached maximal content of verbascoside equal to 2.7 g/100 g DW [39].
From literature reports, up to now, the highest verbascoside specific productivity-165 mg/L/day, was reached in suspension cultures of Harpagophytum procumbens grown in pulse-aerated column bioreactor. In our research, the highest specific productivity was also high and reached in BB and STB 104 and 115 mg/L/day, respectively (Table 1) [40].
High amounts of phenylpropanoid glycosides are also observed in transgenic root cultures of various plants. The maximum verbascoside amounts obtained by us, above 9.2 g/100 g DW (Table 1), are comparable with those obtained in Paulownia tomentosa hairy root cultures, (max. 9.5 g/100 g DW DW), and higher than in transformed roots of Rehmannia glutinosa (max. 1.69 g/100 g DW). Moreover, the highest concentration of isoverbascoside obtained in our research (0.61 g/100 g DW) is higher than in R. glutinosa transformed roots (0.35 g/100 g DW) [24,41,42]. In adventitious root cultures of Castilleja tenuiflora, the maximum level of verbascoside was 1.5 g/100 g DW and the maximum concentration of isoverbascoside was 3.7 g /100 g DW [43].
The amounts of verbascoside obtained by us are among the highest recorded in in vitro cultures. From a practical point of view, the significant amounts of verbascoside that we were able to obtain in undifferentiated biomass of vervain should be considered an unquestionable success.
It should be emphasized that verbascoside and the accompanying isoverbascoside are compounds with very important medicinal values. They exhibit antioxidant, anti-inflammatory, anti-proliferative, antiviral and antibacterial properties, among others [24][25][26].
To sum up, the best biomass growth increments in V. officinalis in vitro cultures we reached in agar callus cultures. Although, the biomass increments of suspension and bioreactor cultures were high, too.
In the all tested types of in vitro cultures, identified metabolites were the same. The cells cultured in vitro produced primarily verbascoside and isoverbascoside, and lower amounts of phenolic acids. The amounts of these metabolites varied between the types of tested cultures. Noteworthy is the extremely high production of verbascoside in STB. Moreover, in this system, isoverbasoside production was also high, and the accumulation of phenolic acids was sufficient. The biomass increments in this system were about 5 times in two weeks, which is very good result from a biotechnological point of view.

Biological Activities
The second part of our research was focused on the comparative investigation of the antiproliferative, antioxidant, and antibacterial activities of V. officinalis extracts obtained from in vitro grown biomass and from the parent plant. Nowadays, finding new candidate drugs from the plant kingdom is highly appealing since an ever-increasing number of people is turning to natural remedies to treat or at least to ameliorate the symptoms of many diseases. Among these, chronic pathologies, such as cancer or degenerative ones, represent the main target of interest since, during longer therapies, it is more likely to come across the unwanted side effects of synthetic drugs, that natural ones instead are less prone to have. Moreover, natural extracts have the advantage that they can be used as multitarget drugs that target different intracellular targets [44]. Several studies reported the importance of natural products for this purpose [45]. In this regard, our study emphasized the biological value of extracts rich in both verbascoside and isoverbascoside, indicating that there is a clear link between their high contents in in vitro grown biomass and their proven biological activities. Indeed, the V. officinalis extract with the highest levels of verbascoside, Vo-STB, among the others present in this study, showed the strongest anti-proliferative activity in SH-SY5Y cells, with IC 50 values from 1.5 to 1.9-fold lower than those of the values for other two extracts (Figure 4). This is in line with the scientific literature, where verbascoside is widely acknowledged as an anti-cancer agent that acts following several mechanisms to achieve this outcome, like activating the apoptotic machinery, influencing cell cycle or interacting with damaged DNA [24][25][26]46].
The results of A. salina lethality bioassay indicated Vo-STB as the most promising cytotoxic extract, which is in agreement with experiments conducted on human cell culture.
The antioxidant potential of V. officinalis extracts was investigated and compared through three in vitro tests based on different mechanisms of action. Several literature reports indicate a linear correlation between antioxidant activity and phenolic content of plant extracts [29,47]. Therefore, linear regression analysis was used to explore this relationship for V. officinalis extracts.
Strong primary antioxidant properties were highlighted for all the extracts and a positive correlation between polyphenol content and both DPPH and reducing power assays was found. Among the tested extracts, Vo-S and Vo-STB showed a higher primary antioxidant capacity than that of the parent plant grown in vivo. Compared to Vo-in vivo extract, Vo-S and Vo-STB contain higher quantities of phenylpropanoid glycosides and quite a similar amount of phenolic acids. Vo-C contains a concentration of verbascoside and isoverbascoside superimposable to Vo-S but about half the amount of phenolic acids, which could explain the lower antioxidant efficacy observed for the extract. These findings suggest that the antioxidant ability of the extracts may be attributed both to the phenylpropanoid glycosides and to the phenolic acids, whose antioxidant properties have been previously demonstrated [48][49][50][51].
As for the secondary antioxidant properties, all the biomass extracts displayed higher chelating activity than that of Vo-in vivo, which does not seem to be related to the phenolic content. In fact, Vo-S, exhibiting the highest total phenolic content, showed the best chelating properties, followed by Vo-C, containing the lowest total phenolic concentration. The lack of correlation between chelating activity and phenolic content is in agreement with previous researches published by some of the authors [52,53] and indicates that the observed activity might depend also on non-phenolic chelators present in the extracts. However, the partial involvement of phenylpropanoid glycosides and phenolic acids, whose chelating capacities have been reported, cannot be excluded [54,55]. The greater activity observed for Vo-C in comparison to Vo-STB could be explained, at least in part, by the presence of higher amounts of isoverbascoside whose Fe 2+ chelating activity was previously demonstrated to be 2-fold stronger than that of verbascoside [54], and of chlorogenic acid, a strong iron chelator [55].
The antibacterial potential of V. officinalis extracts was assessed by MIC/MBC determination (Table 4). All the extracts showed bacteriostatic and bactericidal activity, although the power of the effect was different. The strongest antimicrobial activity was found for Vo-STB and Vo-S, which is an agreement with the results of antioxidant assays and the highest content of total phenolics, verbascoside, isoverbascoside and phenolic acids in these extracts. The strong mode of action of verbascoside against S. aureus is based on the inhibition of protein synthesis [56]. The antibacterial activity of V. officinalis extracts has been the subject of several studies. Results of previous research are consistent with ours, indicating that the verbena herb extracts were more active against Gram-positive bacteria than Gram-negative bacteria. The Gram-negative bacteria E. coli and K. pneumoniae showed resistance toward V. officinalis extracts [57]. The high sensitivity of bacteria Y. enterocolitica to other plant extracts has already been confirmed by other authors. The MIC values of plant extracts for these bacteria were comparable to the values obtained for S. aureus [58,59]. The similar study performed by another team also showed differences in the antibacterial activity of ethanol extracts of V. officinalis prepared from various parts of the plant. Extracts from the stems showed, in general, higher antibacterial activity than leaf and root extracts. However, the extracts were shown to be active against bacterial strains that cause serious infections, including methicillin-resistant S. aureus strains and multidrug-resistant S. typhi [60].
The results of our biological activity studies suggest that verbascoside and isoverbascoside, present in high quantities in in vitro grown biomass of vervain, likely contribute to the observed effects. Antioxidant, antiproliferative and antibacterial activities of the examined extracts could be connected with the high production of verbascoside which is present in much higher amounts than other compounds. The strongest potential was highlighted for Vo-S and Vo-STB cultures which accumulated the highest amounts of this constituent [24].

Experimental in vitro Cultures
All experimental biomasses, including callus, suspension and bioreactor cultures, were maintained in a phytotron at 23 ± 1 • C, with a light intensity of 88 ± 8 mol × m −2 × s −1 (Philips-Flora-TLD 35W/33 fluorescent lamps, Philips, France). The length of the day and night-photoperiod (cyclical lighting-17/24 h) was regulated using an electric system.
In the experiment, all agar callus cultures inocular biomasses of 0.3, 0.6 and 0.9 g were used. The biomasses were placed into in vitro culture jars (Sigma-Aldrich) containing 30 mL MS medium with 1 mg/L BAP and 1 mg/L IBA solidified with agar (7.2 g/L) and supplemented with 30 g/L sucrose. Callus growth and morphology were observed over 14-day growth cycles (3 series, 10 samples each).
For the establishment of suspension cultures, the biomass propagated on agar medium was transferred (1.5, 3 and 4.5 g of callus) to the liquid medium by gently dispersing the biomass with a steel tweezer, to minimize mechanical damage of tissue. Biomasses were placed into 100 mL Erlenmeyer flasks containing 30 mL of MS medium with 1 mg/L BAP, 1 mg/L IBA and 30 g/L sucrose, and closed with silicone foam stoppers. The cultures were incubated on an orbital shaker (Innova 2300 Platform Shaker, New Brunswick Scientific, USA) for 14 days at 120 rpm (3 series, 10 samples each).
The two types of bioreactors were applied for experimental suspension cultures: commercially available stirred-tank bioreactor (STB) (Bellco Glass, Vineland, NJ, USA), and custom-made balloon bioreactor (BB) ( Figure 1C,D). The details of the used STB were presented earlier [61]. In the current work, the system was operated at 33 rpm and continuous 0.5 vvm aeration. The design of the BB bioreactor used was similar to the BB system previously used by the authors [62]. In the present study, a larger vessel was employed (170 mm i.d., 300 mm height). Continuous aeration (0.5 vvm) was provided through silicone foam sparger.
Based on the results obtained at the stage of suspension cultures, both devices were operated at 50 g biomass/1000 mL medium ratio. The composition of the growth medium was the same as in the case of suspension cultures. The single growth cycle lasted 14 days and for each bioreactor type, the experiment was run in triplicate.
The biomasses from all tested types of in vitro cultures were separated from the growth medium (agitated and bioreactor cultures were vacuum filtrated using G4 glass funnel). The collected biomasses and media samples were lyophilized using GT 2 apparatus (Finn Aqua Santasolo-Sohlberg, Tuusula, Finland). Harvested herb was air-dried in room temperature.

Extraction
For analyses of phenylpropanoid glycosides, 0.30 g of lyophilized biomass/dried herb was extracted five times with 3 mL portions of methanol in an ultrasonic bath for 30 min and the extracts were centrifuged and after filtration through a syringe filter (Millex ® GP, Millipore, 0.22 µm, Filter Unit) analyzed using HPLC-DAD method.
For analyses of phenolic acids, flavonoids and iridoids, 10 mL of each extract were dried in crystallizers, dissolved in 2 mL of methanol and after filtration through syringe filters, they were subjected to HPLC-DAD analysis.
Lyophilized samples of 20 mL media were dissolved in 1 mL of methanol, filtered and analyzed by HPLC.
For the biological investigations, lyophilized biomasses and dried herb were extracted five times with methanol in an ultrasonic bath (Polsonic 3, Poland) for 30 min. Extracts were centrifuged, filtered, placed in the crystallizers and evaporated to dryness at room temperature to obtain the following dry extracts: V. officinalis callus culture extract (Vo-C), V. officinalis suspension culture extract (Vo-S), V. officinalis stirred-tank bioreactor culture extract (Vo-STB)and V. officinalis, grown in vivo, herb extract (Vo-in vivo).

Phenylpropanoid Glycosides and Iridoids
Chromatographic analysis of phenylpropanoid glycosides (verbascoside and isoverbascoside) and iridoids (verbenaline and hastatoside) was carried out using the previously developed HPLC-DAD method according to Schönbichler, et al. [63]. The used analytical column was Kinetex C-18 analytical column (150× 4.6 mm, 2.7 µm; Phenomenex, Torrance, California, USA). The mobile phase consisted of two solvents: 0,1% trifluoroacetic acid and acetonitrile (gradient program). Compounds were estimated using a DAD detector. The detection wavelength was set at 330 nm for phenylpropanoid glycosides and at 235 nm for iridoids. The detected compounds were identified by the method of addition and comparison with UV-DAD spectra and retention times (Rt) of commercial standards. The quantification was carried out based on calibration curves.

Phenolic Acids and Flavonoids
Chromatographic analysis of phenolic acids was carried out by the HPLC-DAD method according to Ellnain-Wojtaszek and Zgórka [64,65] using HPLC-DAD-system (Merck-Hitachi) and a Purospher RP-18e analytical column (4× 250 mm, 5 mL; Merck, Darmstadt, Germany). The mobile phase consisted of two solvents: methanol with 0.5% acetic acid (1:4 v/v) and methanol (gradient program). The detection wavelength was set at 254 nm. The identity of detected compounds was confirmed by the method of standard addition and the comparison of UV-DAD spectra and the Rt values.

Artemia salina Leach Lethality Bioassay
Determination of median lethal concentration (LC 50 ) was carried out according to the method previously reported [66]. Ten brine shrimp larvae, taken 48 h after initiation of hatching in artificial seawater, were transferred to Petri dishes, and artificial seawater was added to obtain a final volume of 5 mL. Subsequently, the V. officinalis extracts (Vo-C, Vo-S, Vo-STB, and Vo-in vivo), diluted in artificial seawater, were added at the final concentrations of 10, 100, 500 and 1000 µg/mL. After 24 h of incubation at 25-28 • C, the surviving larvae were counted. The assay was carried out in triplicate, and LC 50 values were determined by the Litchfield and Wilcoxon's method. The extract was considered active if the LC 50 was lower than 1000 µg/mL.

Effect on SH-SY5Y Human Neuroblastoma Cell Proliferation
The human neuroblastoma cell line SH-SY5Y was originally obtained from ATCC (Rockville, MD, USA), maintained and kept as described by Cirmi et al. [67]. Cell culture reagents were from Carlo Erba (Milan, Italy). For cell viability assays, stock solutions of the four extracts were prepared in DMSO at the highest concentration depending on each solubility. Aliquots were stored at −20 • C and defrosted prior to the dilution in culture media. For each extract, the highest DMSO concentration was used as a control for the corresponding extract concentration.
The effect of V. officinalis extracts (Vo-C, Vo-S, Vo-STB, and Vo-in vivo) on the growth rate of SHSY5Y was assessed by 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, as stated by Celano et al. [68]. Cells were seeded at a density of 5 × 10 3 cell/well in 96-well plates and incubated overnight at 37 • C. Then, the medium was removed and replaced with fresh one containing extracts dilutions ranging from 125 to 750 µg/mL, along with a control (untreated cells) and solvent control (DMSO). After incubation for 24, 48 and 72 h at 37 • C, the medium was replaced with a fresh one containing MTT at a concentration of 0.5 mg/mL and incubated for further 4 h. Afterwards, MTT solution was removed and the formazan crystals derived from the metabolic activity of viable cells were solubilized in 100 µL of HCl/isopropanol 0.1 N and regularly shaken. The optical density was spectrophotometrically measured at 570 nm with a reference filter at 690 nm using a microplate reader (iMark™, Bio-Rad Laboratories, Milan, Italy). All experiments were performed in eight replicates for three times. Results are calculated as a percentage of viable cells compared to untreated cells and expressed as mean ± SEM.

Determination of Total Phenolic Content
The total phenolic content of V. officinalis extracts (Vo-C, Vo-S, Vo-STB, and Vo-in vivo) was determined by the Folin-Ciocalteu method, using a calibration curve of gallic acid as a reference [69]. Briefly, 100 µL of each sample solution were mixed with 0.2 mL Folin-Ciocalteu reagent, 2 mL of H 2 O, and 1 mL of 15% Na 2 CO 3 , and the absorbance was measured at 765 nm, after 2 h incubation at room temperature, with a model UV-1601 spectrophotometer (Shimadzu, Milan, Italy). The total phenolics were estimated as gallic acid equivalent (GAE) and expressed in mg GAE/g extract ± standard deviation (SD). The data were obtained from the average of three determinations.

Free Radical Scavenging Activity
The free radical scavenging activity of V. officinalis extracts (Vo-C, Vo-S, Vo-STB, and Vo-in vivo) was determined using the DPPH (1,1-diphenyl-2-picrylhydrazyl) method [69]. An aliquot (0.5 mL) of a methanol solution containing a different amount of each extract (0.0625-2 mg/mL) was added to 3 mL of daily prepared methanol DPPH solution (0.1 mM). The optical density change at 517 nm was measured, 20 min after the initial mixing, with a model UV-1601 spectrophotometer (Shimadzu). Butylated hydroxytoluene (BHT) was used as a reference.The results, obtained from the average of three independent experiments, are reported as mean radical scavenging activity percentage (%) ± SD and mean 50% inhibitory concentration (IC 50 ) ± SD.

Reducing Power Assay
The reducing power of V. officinalis extracts (Vo-C, Vo-S, Vo-STB, and Vo-in vivo) was evaluated by spectrophotometric detection of Fe 3+ →Fe 2+ transformation method [69]. Different amounts of samples (0.0625-2 mg/mL) in 1 mL solvent were mixed with 2.5 mL of phosphate buffer (0.2 M, pH 6.6) and 2.5 mL of 1% potassium ferricyanide [K 3 Fe(CN) 6 ]. The mixture was incubated at 50 • C for 20 min. The resulting solution was cooled rapidly, mixed with 2.5 mL of 10% trichloroacetic acid, and centrifuged at 3000 rpm for 10 min. The resulting supernatant (2.5 mL) was mixed with 2.5 mL of distilled water and 0.5 mL of 0.1% fresh ferric chloride (FeCl 3 ), and the absorbance was measured at 700 nm after 10 min. Ascorbic acid and BHT were used as a reference. The results, obtained from the average of three independent experiments, are expressed as mean absorbance values ± SD and ascorbic acid equivalent (ASE/mL) ± SD.

Ferrous Ions (Fe 2+ ) Chelating Activity
The Fe 2+ chelating activity of V. officinalis extracts (Vo-C, Vo-S, Vo-STB, and Vo-in vivo) was estimated by measuring the formation of the Fe 2+ -ferrozine complex, according to the method previously reported [69]. Briefly, different concentrations of each sample (0.0625-2 mg/mL) in 1 mL solvent were mixed with 0.5 mL of methanol and 0.05 mL of 2 mM FeCl 2 . The reaction was initiated by the addition of 0.1 mL of 5 mM ferrozine. The mixture was shaken vigorously and left standing at room temperature for 10 min, then the absorbance of the solutions was measured spectrophotometrically at 562 nm. Ethylenediaminetetraacetic acid (EDTA) was used as a reference. The results, obtained from the average of three independent experiments, are reported as mean inhibition of the Fe 2+ -ferrozine complex formation (%) ± SD and IC 50 ± SD.

MIC and MBC Determination of Extracts
MIC and MBC of extracts against the test strains were carried out by the method of serial micro dilutions [70,71]. Two series of dilutions of the extracts were prepared in Mueller-Hinton Broth (MHB) medium in the concentration range from 36.0 to 0.3 mg/mL. An inoculum was added to each well of the polystyrene 96-well plates, each of a volume of 250 µL so that the final number of cells was 5 × 10 5 cfu/mL. A well without extract was chosen as a positive control. Negative control was the medium (without the addition of extracts) together with the tested bacteria. The plates were incubated at 37 • C for 20 h. Afterwards, 25 µL of resazurin at 0.02% (m/v) (filtered through a 22 µm sterile filter) was then added to each well and the plates were incubated at 37 • C for 2 h. Bacterial growth was assessed based on the change in resazurin colour from violet to pink [72] and compared with the control sample. MIC examination of extracts was repeated three times. The MIC value was defined as the lowest concentration of extract, in which no growth of bacteria was observed and expressed in mg/mL.
To determine MBC, 100 µL bacterial culture from each well showing no bacterial growth were re-inoculated onto Mueller-Hinton agar (MHA) medium, which were incubated at 37 • C for 24 h. The plates were checked for growth of colonies. MBC was defined as the lowest concentration of extract, which resulted in complete inhibition of bacterial growth and expressed in mg/mL.

Statistical Analysis
Statistical comparison of data was carried out by using one-way analysis of variance (ANOVA) followed by Tukey-Kramer multiple comparisons test (GraphPAD Prism Software for Science). P-values lower than 0.05 were considered statistically significant.

Conclusions
Optimization of the best culture conditions for biomass growth and for the production of bioactive compounds in three types of vervain in vitro cultures (callus, suspension, bioreactors cultures) was performed. Very good results were obtained in suspension culture and bioreactors cultures (balloon bioreactor-BB and stirred-tank bioreactor-STB) both established for the first time.
The biological activities of the biomass extracts from different types of in vitro systems-antiproliferative, antioxidant and antibacterial properties, were also documented for the first time and they were greater compared to that of the parent plant grown in vivo.
Based on the results, the STB was selected as the best in vitro system for the increased production of phenylpropanoid glycosides. In addition, the extract obtained from biomass produced in STB displayed significant efficacy in all the biological activity tests carried out in this study. Our results suggest that cultures maintained in STB could be a valuable source of bioactive compounds alternative to the parent plant grown in vivo.
The obtained results have the applicable character. We propose V. officinalis bioreactor culture maintained in the STB as an efficient, rich potential source of verbascoside and isoverbascoside independent of environmental conditions. The obtained maximal quantities of verbascoside (9.18%) and additionally the high amount of isoverbascoside (0.34%) in STB are interesting from a practical point of view.