Unravelling Phytochemical and Bioactive Potential of Three Hypericum Species from Romanian Spontaneous Flora: H. alpigenum, H. perforatum and H. rochelii

Hypericum perforatum L., also known as St. John’s Wort, is recognized worldwide as a valuable medicinal herb; however, other Hypericum species were intensively studied for their bioactive potential. To fill the research gap that exists in the scientific literature, a comparative evaluation between H. alpigenum Kit., H. perforatum L. and H. rochelii Griseb. & Schenk was conducted in the present study. Two types of herbal preparations obtained from the aerial parts of these species were analyzed: extracts obtained through maceration and extracts obtained through magnetic-stirring-assisted extraction. LC-DAD-ESI-MSn analysis revealed the presence of phenolic acids, flavan-3-ols and flavonoid derivatives as the main constituents of the above-mentioned species. Moreover, all extracts were tested for their antioxidant, enzyme-inhibitory and antimicrobial potential. Our work emphasizes for the first time a detailed description of H. rochelii phenolic fractions, including their phytochemical and bioactive characterization. In comparison with the other two studied species, H. rochelii was found as a rich source of phenolic acids and myricetin derivatives, showing important antioxidant, anticholinesterase and antibacterial activity. The study offers new perspectives regarding the chemical and bioactive profile of the less-studied species H. alpigenum and H. rochelii.


Introduction
The Hypericum genus comprises more than 500 species, which are commonly found in temperate regions both as spontaneous or cultivated plants [1]. Among them, Hypericum perforatum L. (St. John's Wort) is the most well-known one, especially for its applications in human medicine, which are related to various phytochemicals contained in the aerial parts of this species. Naphtodianthrones (i.e., hypericins), the main interest constituents of this herb, were intensively studied for their pharmacological applications, their effectiveness in treatment of mild to moderate depressive episodes being proven by multiple clinical trials
Compound 9 (λ max at 257 nm) showed a molecular ion [M−H] − at m/z 421, which provided, after MS 2 fragmentation, two major ions, namely m/z 331 and m/z 301; the fragmentation pattern was close to the one previously described for dibenzo-γ-pyrone-C-glycosides [28,29]. Hence, based on all these features, the compound was tentatively identified as mangiferin, the only xanthone derivative found in the analyzed species.

Total Phenolic (TPC) and Total Flavonoid Content (TFC)
The results obtained after the assessment of TPC and TFC for the herbal preparations of H. perforatum, H. alpigenum and H. rochelii are presented in Table 2. As can be observed, the extracts obtained through MSA extraction showed higher phenolic contents compared to those obtained through maceration. Conversely, TFC varied independent of extraction method, the highest value being observed for HPM extract (100.17 ± 1.27 mg RE/g). A comparison between species based on these parameters highlights the highest values of total phenolic and flavonoid contents for H. alpigenum.

Antioxidant Potential of the Extracts
The antioxidant potential of the extracts obtained from H. perforatum, H. alpigenum and H. rochelii was evaluated through five in vitro complementary methods, the results being presented in Table 2. The HPM and HRA extracts exerted the strongest free radical scavenger activity (TEAC assay), a similar trend being described for the same extracts regarding ferric ion reducing power (FRAP) and oxidative hemolysis inhibition (OxHLIA assay). Nevertheless, a clear interdependence between phenolic content and antioxidant activity could not be established, which indicates that other phytoconstituents contained in the samples made an important contribution to their antioxidant potential.
The high to moderate antioxidant potency of the analyzed extracts is supported also by the results obtained in TBARS and OxHLIA assays, which were proven as being more sensitive than the conventional antioxidant methods [43]. Usually, in the OxHLIA kinetic assay, IC 50 values (µg/mL) are calculated for a given period of time by correlating the extract concentrations to the ∆t values (min) (calculated based on half hemolysis time obtained from the hemolytic curves of each extract concentration minus the Ht 50 value of the negative control). For a 60 min ∆t, all extracts presented lower IC 50 values than the positive control (Trollox), the same trend being described for all samples in the TBARS assay.

Enzyme-Inhibitory Activity of the Extracts
All the extracts were tested for their anti-glucosidase, anti-tyrosinase and anti-cholinesterase potentials, the results being summarized in Table 3; additionally, inhibition curves for each active extract are presented in Figure 1. Overall, the analyzed species showed weak inhibitory activity on tyrosinase (HPA was the only one that exerted anti-tyrosinase activity), promising anti-glucosidase activity (all the extracts reached IC 50 values at lower concentrations than the positive control) and moderate anti-cholinesterase potential. Several differences could be observed between species in terms of inhibitory potential, H. perfora-tum samples acting more as acetylcholinesterase and α-glucosidase inhibitors, while, for H. alpigenum, the most prominent α-glucosidase inhibition was observed (an IC 50 value of 17.35 ± 4.29 µg/mL extract for HMA).

Antimicrobial Properties of the Extracts
In addition to well-established pharmacological properties (i.e., antidepressant, antioxidant, etc.), some Hypericum species were cited as possessing antibacterial and/or antifungal activity, proven by in vitro methods [22,[44][45][46][47]. In this regard, we aimed to conduct a comparative evaluation of the antimicrobial potential of the extracts obtained from H. perforatum, H. alpigenum and H. rochelii, highlighting the main differences between this species in terms of antibacterial and antifungal effectiveness.
As can be observed in Table 4, all bacterial strains were sensible to the tested extracts, excluding HAM, which showed MIC and MBC values higher than 8 mg/mL. The most prominent antibacterial activity was exerted by H. perforatum, the lowest MIC and MBC being obtained for Gram-negative bacteria; conversely, H. rochelii was found as being more active on Gram-positive strains. In terms of the effectiveness of antibacterial potential, the extracts obtained through MSA extraction were found as being more active than those obtained by maceration. A comparative overview between the sensibility of the tested strains upon action of Hypericum extracts reveals Staphylococcus aureus as the most sensible among them.  The parameters that describe the antifungal activity of Hypericum extracts are summarized in Table 5. Overall, the tested strains showed moderate (A. fumigatus, A. niger, A. versicolor, T. viride) or weak (P. funiculosum, P. verrucosum var. cyclopium) sensibility. A similar trend for antibacterial potential was observed, H. perforatum extracts showing the lowest MIC and MFC values, especially for the extract obtained through MSA extraction (HPA), which exerted good antifungal activity on A. niger and A. versicolor (MIC-1 mg/mL, MFC-2 mg/mL for both strains).

Discussion
After an overview of the LC-DAD-ESI/MS n results, it could be suggested that magneticstirring-assisted (MSA) extraction induced higher recovery yields for the phenolic compounds contained in the analyzed samples, this trend being more visible in the case of H. alpigenum and H. rochelii extracts. Moreover, our findings showed that each extraction method needs to be customized for each type of main compound that will be extracted; in fact, MSA increased the recovery of phenolic acids, xanthonoids and flavonoids, while maceration was more effective in terms of flavan-3-ols extraction yields. In addition, our phytochemical assessment provides the first detailed report about the qualitative and quantitative distribution of phenolic compounds from H. rochelii. As could be observed, our study emphasized the presence of myricetin derivatives, p-coumaroyl and caffeoylquinic acids as the main constituents of the above-mentioned species.
It must be noted that TPC and TFC assays offer a general overview about the amounts of phenolic and flavonoid compounds contained in different samples and present some limitations regarding the interferences with other phytochemicals found in the analyzed matrices [48,49]. Nonetheless, they are still useful complementary tools in the chemical evaluation of plant extracts, the results obtained through these methods revealing correlations between antioxidant capacity of the samples or their chemical profile assessed by more sensible methods (i.e., liquid chromatography). Referring to the present study, the results obtained for TPC both in chromatographic and classic spectrophotometric evaluation indicate higher extraction yields for total phenolic compounds using MSA extraction. As well, the trend described for quantitative distribution of flavonoidic compounds by using TFC and LC-MS was the same for HAA, HPA, HPM and HRM extracts.
Initially, a clear interdependence between phenolic content and antioxidant activity could not be established, which indicated that other phytoconstituents of the samples made an important contribution to their antioxidant potential; hence, based on Pearson's correlation coefficients, a correlogram was built in order to decipher the individual influence of each chemical constituent against the measured bioactivities, including the antioxidant one (Figure 2). A strong positive correlation was found between TPC and oxidative hemolysis inhibition, while TFC was positively correlated with the ferric reducing power of the extracts. Regarding the influence of the individual phenolic constituents identified after LC-DAD-ESI/MS n assessment, the most important contribution to the antioxidant activity of Hypericum samples could be attributed to the presence of 4-O-caffeoylquinic and 5-O-caffeoylquinic acids, as well as to the presence of (+)-catechin and procyanidin B5; the correlation coefficients obtained for the other identified constituents support the hypothesis that the antioxidant potential of the extracts could be influenced by some unidentified compounds, probably belonging to other chemical classes.
This aspect was previously highlighted for other Hypericum species; Radulović et al. showed that the antioxidant capacity of H. perforatum samples collected from the Balkans varied not exclusively depending on their phenolic content, revealing the contribution of several tannins to the total capacity of the extracts [45], while Gîtea et al. reported inconsistent variations in the antioxidant potential of several species rich in phenolic compounds collected from Romanian spontaneous flora (H. perforatum L., H. maculatum Cr var. typicum Frohlich., H. hirsutum L., H. tetrapterum Fr.) [50]. Moreover, the antioxidant potential of H. rochelii, was not reported yet by other studies. The extracts of this species strongly acted as free radical scavengers and lipid peroxidation inhibitors, showing also moderate capacity to act as reducing agents and oxidative hemolysis inhibitors. As could be observed, in comparison with H. perforatum and H. alpigenum, H. rochelii exhibited medium antioxidant capacity, the best results being described for the extracts obtained through MSA extraction.
Several studies were previously focused on evaluation of enzyme-inhibitory properties of Hypericum species [51][52][53][54]. Ethyl acetate, methanolic and aqueous extracts of H. perforatum L. were tested for their anti-cholinesterase and anti-tyrosinase activity by Altun et al., their findings indicating the highest acetylcholinesterase and low tyrosinase inhibition for the methanolic extract [51]. A moderate enzyme-inhibitory effect against acetylcholinesterase was also described for H. olympicum, H. pruinatum and H. scabrum collected from Turkey, while the same species showed important inhibition against α-glucosidase, correlated with the significant amounts of flavonoid derivatives quantified in their methanolic extracts [52]. To the best of our knowledge, there are no available data regarding the enzyme-inhibitory potential of H. rochelii, our study revealing for the first time the ability of the herbal preparations obtained from this species to act as α-glucosidase and acetylcholinesterase inhibitors. As can be observed, the anti-glucosidase effect was slightly enhanced in the case of the extract obtained through maceration, and the MSA extract showed better interaction with the acetylcholinesterase in terms of inhibitory activity. The correlation analysis revealed a strong interdependence between the anti-glucosidade activity of the extracts and their mangiferin and quercetin-O-deoxyhexoside content (positive correlation, r = 0.75 and r = 0.55, respectively), as can be observed in Figure 2. Even though the tyrosinase inhibition was weak, it could be positively correlated with the presence of quercetin-O-acetyl-hexoside in the HPA extract, while the anti-cholinesterase effects seem to be linked both with the highest total phenolic content of the extracts and several individual phenolic metabolites (especially caffeoylquinic acids and flavan-3-ols derivatives).
In terms of antimicrobial potence, H. alpigenum and H. rochelii were found as having comparable antifungal activity, little differences being observed only for A. niger and T. viridae, which were more sensible to the action of the extracts obtained from the second species. Ðordević et al. previously evaluated the antibacterial effect of H. rochelii against five bacterial (Bacillus subtilis ATCC 6633, Staphylococcus aureus ATCC 6538, Escherichia coli ATCC 8739, Pseudomonas aeruginosa ATCC 9027, Salmonella abony NCTC 6017) and two fungal strains (Aspergillus niger ATCC 16404 and Candida albicans ATCC 10231), revealing moderate activity of the volatile oil isolated from the aerial parts of this species against B. subtilis, S. aureus and C. albicans; no other studies indicated a antimicrobial effect of H. rochelii [15]. Hence, our findings provide new perspectives regarding potential use of the phenolic fraction obtained from the above-mentioned species as a mild antibacterial agent, encouraging further in-depth evaluation for additional mechanisms that could be involved in this activity (i.e., inhibition of biofilm or pyocyanin formation) [55,56].

Extraction Procedure
For ensuring uniformity of the plant material used for extraction, it was powdered using a laboratory mill (Grindomix ® GM 200, Retsch Gmbh., Germany) and manually sieved (1 mm standard sieve according to PhEur 10.6). In order to achieve extraction of bioactive compounds, two parallel extraction methods were implemented: magneticstirring-assisted (MSA) extraction and conventional maceration. For MSA extraction, 5 g of each powdered plant material was mixed with ethanol 70% (1:10 w/v ratio), and the resulted mixture was subsequently placed on the magnetic stirrer for 15 min, at a temperature of 40 • C. Likewise, 5 g of each powdered plant material was mixed with ethanol 70% in a ratio of 1:10 (w/v), and the resulted mixture was subsequently placed in a dark place, at room temperature, for 10 days, for maceration to occur. After the extraction procedure, the extracts were filtered under reduced pressure, concentrated until complete evaporation of the alcohol (using a rotary evaporator), freeze-dried and stored in a desiccator, protected from light and at room temperature. Hence, 6 herbal preparations were obtained, namely:

LC-DAD-ESI/MS n Characterization of Phenolic Profile
Ten milligrams of each dry extract were redissolved in 2 mL of ethanol/water (20:80, v/v) and filtered through 0.22-µm disposable LC filter disks before injection. An Dionex Ultimate 3000 UPLC (Thermo Scientific, San Jose, CA, USA) system equipped with a diode array detector coupled to an electrospray ionization mass detector (LC-DAD-ESI/MS n ) was employed for analysis of phenolic compounds using a method previously described [39]. Chromatographic separation was conducted on a Spherisorb S3 ODS-2C18 column (3 µm, 4.6 mm × 150 mm, Waters, Milford, MA, USA) using as solvents 0.1% aqueous formic acid (A) and acetonitrile (B) in gradient elution: isocratic 15% B (5 min), 15% B to 20% B (5 min), 20-25% B (10 min), 25-35% B (10 min), 35-50% B (10 min) and re-equilibration of the column using a flow rate of 0.5 mL/min. Online detection was achieved using a Diode Array Detector DAD (280, 330 and 370 nm as preferential wavelengths) coupled with an ESI mass spectrometer working in negative mode (Linear Ion Trap LTQ XL mass spectrometer, Thermo Finnigan, San Jose, CA, USA). Identification of phenolic compounds was made by comparing their retention times and UV-Vis and mass spectra with those obtained from standard compounds (when available); otherwise, compounds were tentatively identified comparing the obtained information with available data reported in the literature. For the quantitative evaluation, a calibration curve for each available phenolic standard (Extrasynthèse, Genay, France) was constructed based on the UV signal; for the identified phenolic compounds for which a commercial standard was not available, the quantification was performed through the calibration curve of the most similar available standard, and results were expressed as mg/g of extract [57].

Evaluation of Total Phenolic (TPC) and Total Flavonoid (TFC) Contents
To evaluate total phenolic content (TPC) of the assessed species, Folin-Ciocalteu (F-C) method was implemented. In a 96-well plate, 100 µL of 10% F-C solution were mixed with 20 µL of sample solution and pre-incubated for 3 min at room temperature in a dark place. Subsequently, the mixture was completed with 80 µL of 7.5% Na 2 CO 3 solution and the resulted mixture was incubated for another 30 min in the same conditions. Finally, the absorbance of the mixture was read at 760 nm, and the results were expressed as milligrams of gallic acid equivalents/g of lyophilized extract (mg gallic acid equivalents − GAE/g extract) [43].
Conversely, to evaluate total flavonoid content (TFC) of assessed plants, in a 96-well plate, 100 µL of 2% AlCl 3 solution were mixed with 100 µL of sample solution and incubated for 10 min at room temperature in a place free of light. The absorbance of the mixture was read at 420 nm, and the results were expressed as milligrams of rutin equivalents/g of lyophilized extract (mg rutin equivalents − RE/g extract) [43].

TEAC Assay
To generate a radical stock solution, 50 mL of ABTS + (2.13 mM) were mixed with 50 mL of K 2 S 2 O 8 (1.38 mM), both reagents being dissolved in ultrapure water, which, after an incubation time of~16 h in a dark place and at 20 to 25 • C, was subsequently diluted with distilled water to reach a final absorbance of the radical stock solution of 0.70 ± 0.02 at 734 nm. Afterwards, in a 96-well plate, 220 µL of reaction mixture, consisting of 200 µL of radical stock solution and 20 µL of sample solution, were incubated at room temperature in a dark place. After 6 min, the absorbance of the reaction mixture was read at 734 nm. The TEAC radical scavenging activity of the extracts was expressed as milligrams of Trolox equivalents/g of lyophilized extract (mg TE/g extract) [57].

FRAP Assay
A FRAP reagent was generated by mixing 50 mL of acetate buffer (0.3 M, pH 3.6), 5 mL of FeCl 3 solution (20 mM) and 5 mL TPTZ solution (10 mM) (both FeCl 3 and TPTZ were dissolved in 40 mM HCl). Afterwards, in a 96-well plate, 200 µL of reaction mixture, consisting of 175 µL of FRAP reagent and 25 µL of sample solution, was incubated for 30 min in a dark place at room temperature; the final absorbance of the mixture was read at 593 nm. The FRAP radical scavenging activity of the extracts was expressed as milligrams of Trolox equivalents/g of lyophilized extract (mg TE/g extract) [43].

DPPH Assay
A 0.004% DPPH radical solution was initially prepared by dissolving 5 mg of DPPH in 125 mL of methanol. Afterwards, in a 96-well plate, 270 µL of DPPH radical solution were mixed with 30 µL of sample solution. Following 30 min incubation period in a dark place and at room temperature, the absorbance of the reaction mixture was read at 517 nm, and DPPH radical scavenging activity of the extracts was expressed as milligrams of Trolox equivalents/g of lyophilized extract (mg TE/g extract) [57].

TBARS Assay
In a pre-incubation phase, 200 µL of sample solution (extracts of each sample were serially diluted) was mixed with 100 µL of FeSO 4 (10 µM) and 100 µL of ascorbic acid (0.1 mM) in an Eppendorf reaction tube (2 mL). The mixture was pre-incubated for 1 h at 37 • C. Afterwards, the reaction mixture was completed with 500 µL of trichloroacetic acid (28% w/v) and 380 µL of thiobarbituric acid (TBA, 2% w/v). The new resulted mixture was heated for 20 min at 80 • C. Finally, the reaction tubes were centrifuged at 3000× g for 10 min, and, in order to quantify malondialdehyde (MDA)-TBA complex, the absorbance of supernatant was read at 532 nm and results were expressed as EC 50 values (µg/mL) [43]

Enzyme-Inhibitory Activity
Evaluation of enzyme-inhibitory activity of H. alpigenum, H. perforatum and H. rochelii included screening of these plants against α-glucosidase, tyrosinase and acetylcholinesterase using in vitro protocols.

α-Glucosidase Inhibition Assay
Using a protocol adapted for microplate reader, in a 96-well plate, a reaction mixture consisting of 50 µL of extract solution of different concentrations was mixed with 50 µL of α-glucosidase enzyme solution (0.75 U/mL), and 50 µL potassium phosphate buffer (100 mM, pH = 6.8) was pre-incubated for 10 min at 37 • C. Afterwards, 50 µL of p-NPG were added to the reaction mixture and plate was incubated for another 10 min at 37 • C. Finally, the absorbance was read at 405 nm and results were expressed as IC 50 value (µg/mL). Acarbose was used as a positive control [43].

Tyrosinase Inhibition Assay
Analogous to α-glucosidase inhibition assay, the protocol used for evaluating tyrosinase inhibition activity of the extracts was adapted for microplate reader. Therefore, 40 µL of different concentration of sample solutions were mixed with 80 µL of potassium phosphate buffer (50 mM, pH = 6.5) and 40 µL tyrosinase enzyme solution (125 U/mL) in a 96-well plate. The resulted mixture was pre-incubated for 5 min at 37 • C. Subsequently, 40 µL of L-DOPA (10 mM) were added and the new resulted mixture was incubated for another 15 min. Finally, the final absorbance of the reaction mixture was read at 492 nm, and results were expressed as IC 50 value (µg/mL). Kojic acid was used as a positive control [57].

Acetylcholinesterase Inhibition Assay
In a similar way as α-glucosidase and tyrosinase inhibition assays, a protocol adapted for microplate reader was used to evaluate acetylcholinesterase inhibition capacity of the extracts. Thus, 25 µL of different concentration of sample solutions were mixed with 50 µL of Tris-HCl buffer (50 mM, pH = 8), 125 µL of DTNB (0.9 mM) and 25 µL of acetylcholinesterase enzyme solution (0.078 U/mL). The reaction mixture was pre-incubated for 15 min at 37 • C. Subsequently, 25 µL of ATCI (4.5 mM) were added to the reaction mixture, and the plate was incubated for another 10 min at 37 • C. Finally, the absorbance was read at 405 nm and results were expressed as IC 50 value (µg/mL). Galantamine was used as a positive control at varying concentrations [57,58].

Statistical Analysis
For each species, all the assays were carried out in triplicate. Statistical analysis was performed using GraphPad Prism 9 program. Differences were significant at the level of α = 0.05 by using one-way analysis of variance (ANOVA) followed by Tukey's HSD. To analyze the relationship between different outcome variables, correlation analysis was performed using RStudio software (RStudio Desktop 2022.07.2+576) [43,60]. All the data were expressed as mean values with standard deviations (mean ± SD).

Conclusions
We evaluated the phenolic and bioactive profile of the extracts obtained through maceration and magnetic-stirring-assisted extraction from the aerial parts of H. perforatum, H. alpigenum and H. rochelii collected from Romanian spontaneous flora. Even though the correlation analysis proved interdependence between the bioactive profile of this species (antioxidant and enzyme-inhibitory properties) and their phenolic profile (both total and individual phenolic contents), their therapeutic potential could be linked with the presence of the other chemical constituents found in the extracts. The originality of the present study consists in the first report about the phenolic composition and bioactive profile of H. rochelii, a less studied species belonging to the Hypericum genus. Herbal preparations obtained from H. rochelii were found as containing high amounts of phenolic acids and myricetin derivatives, exerting at the same time promising antioxidant and antibacterial activity, followed by moderate inhibitory potential against a-glucosidase and acetylcholinesterase. The obtained results encourage future in-depth evaluations on the chemical constituents of this species and the mechanisms involved in its bioactivities demonstrated in the present study.