Antimicrobial Diterpenes from Rough Goldenrod (Solidago rugosa Mill.)

Solidago rugosa is one of the goldenrod species native to North America but has sporadically naturalized as an alien plant in Europe. The investigation of the root and leaf ethanol extracts of the plant using a bioassay-guided process with an anti-Bacillus assay resulted in the isolation of two antimicrobial components. Structure elucidation was performed based on high-resolution tandem mass spectrometric and one- and two-dimensional NMR spectroscopic analyses that revealed (–)-hardwickiic acid (Compound 1) and (–)-abietic acid (Compound 2). The isolates were evaluated for their antimicrobial properties against several plant pathogenic bacterial and fungal strains. Both compounds demonstrated an antibacterial effect, especially against Gram-positive bacterial strains (Bacillus spizizenii, Clavibacter michiganensis subsp. michiganensis, and Curtobacterium flaccumfaciens pv. flaccumfaciens) with half maximal inhibitory concentration (IC50) between 1 and 5.1 µg/mL (5–20 times higher than that of the positive control gentamicin). In the used concentrations, minimal bactericidal concentration (MBC) was reached only against the non-pathogen B. spizizenii. Besides their activity against Fusarium avenaceum, the highest antifungal activity was observed for Compound 1 against Bipolaris sorokiniana with an IC50 of 3.8 µg/mL.


Introduction
Goldenrod plants belong to the genus Solidago which includes over 120 members, with the majority being native to North America. These herbaceous perennial plants with yellow flowers often grow up to 2 m in height. Solidago virgaurea L. is the only native goldenrod in Europe, but several North American goldenrods were introduced to Europe as ornamentals, and at least four of these species have become naturalized. S. gigantea Aiton and S. canadensis L. became particularly successful invaders, while S. graminifolia (L.) Elliot and S. rugosa Mill. occur only sporadically [1][2][3]. S. rugosa (rough or wrinkle-leaved goldenrod) is considered a naturalized alien already in Switzerland, Portugal, Norway, and Great Britain, probably present in France and Belgium [3][4][5] and its geographic expansion is still in progress. The plant has hairy stems lined with thick, firm, and rough-textured leaves, and at the tips, yellow flowerheads cascade ( Figure 1). Only limited information is available regarding the bioactivity of the rough goldenrod's chemical constituents. It was observed that, compared to other Solidago species, S. rugosa was more resistant against the rust fungus Coleosporium asterum, with a lower incidence of bright orange rust pustules on the leaves [6]. Furthermore, extracts of rough goldenrod tissues inhibited Mycobacterium tuberculosis in Despite the occasional occurrence of hypersensitivity reactions or gastrointestinal disorders, the aerial parts of several other Solidago species are traditionally used for the treatment of, e.g., minor urinary complaints and diabetes [11,12]. Goldenrod extracts showed a wide range of pharmacological effects, including antimicrobial, insecticidal, anti-obesity, antimutagenic, anti-inflammatory, and cholinesterase inhibitory activity [11,[13][14][15][16] that can be attributed to the secondary metabolites of the plants, such as phenolic acids, flavonoids, essential oils, polyacetylenes, diterpenes, triterpene saponins, and tannins [11,13,[17][18][19][20].
Phytopathogenic fungi and bacteria have shown a tendency to develop resistance to pesticides due to adaptive mutations as a consequence of the extensive use of plant protection agents [21,22]. The loss of efficacy of these agrochemicals poses a significant threat to crop production, resulting in poor-quality products, lower yields, and an increased risk of plant diseases [23]. Furthermore, it has a negative impact on human nutrition and also raises environmental and ecological concerns. Several modes of action of antimicrobial molecules are known, although microbes are increasingly evading these mechanisms [22]. Therefore, to find sustainable solutions for managing pesticide resistance is critical for maintaining crop productivity. There is an urgent demand for extending the chemical space by discovering new, effective, and environmentally friendly substances. Plants are considered an inexhaustible source of secondary metabolites possessing valuable, diverse structures and biological activity [24,25]. Natural products isolated from plants can serve as starting points for chemical modifications to improve their properties, acting as lead compounds during the development of small-molecule, potent antimicrobial biopesticides [26].
High-performance thin-layer chromatography coupled with effect-directed analysis (HPTLC-EDA) is a quick, straightforward, cost-efficient, and powerful hyphenated method for the non-targeted, high-throughput screening of plant extracts for bioactive compounds, avoiding the commonly used, suboptimal trial and error approach in the labor-intensive, expensive isolation procedure [13,27]. After the determination of the bioprofile of a sample by an HPTLC-direct bioautographic (DB) method [28], the detected inhibition zones can be characterized by various techniques, such as HPTLC-mass spectrometry (MS) [29]. Therefore, HPTLC hyphenations combined with preparative column chromatographic fractionations can contribute to the detection and subsequent separation, purification, and isolation of bioactive compounds from challenging matrices [30]. Despite the occasional occurrence of hypersensitivity reactions or gastrointestinal disorders, the aerial parts of several other Solidago species are traditionally used for the treatment of, e.g., minor urinary complaints and diabetes [11,12]. Goldenrod extracts showed a wide range of pharmacological effects, including antimicrobial, insecticidal, anti-obesity, antimutagenic, anti-inflammatory, and cholinesterase inhibitory activity [11,[13][14][15][16] that can be attributed to the secondary metabolites of the plants, such as phenolic acids, flavonoids, essential oils, polyacetylenes, diterpenes, triterpene saponins, and tannins [11,13,[17][18][19][20].
Phytopathogenic fungi and bacteria have shown a tendency to develop resistance to pesticides due to adaptive mutations as a consequence of the extensive use of plant protection agents [21,22]. The loss of efficacy of these agrochemicals poses a significant threat to crop production, resulting in poor-quality products, lower yields, and an increased risk of plant diseases [23]. Furthermore, it has a negative impact on human nutrition and also raises environmental and ecological concerns. Several modes of action of antimicrobial molecules are known, although microbes are increasingly evading these mechanisms [22]. Therefore, to find sustainable solutions for managing pesticide resistance is critical for maintaining crop productivity. There is an urgent demand for extending the chemical space by discovering new, effective, and environmentally friendly substances. Plants are considered an inexhaustible source of secondary metabolites possessing valuable, diverse structures and biological activity [24,25]. Natural products isolated from plants can serve as starting points for chemical modifications to improve their properties, acting as lead compounds during the development of small-molecule, potent antimicrobial biopesticides [26].
High-performance thin-layer chromatography coupled with effect-directed analysis (HPTLC-EDA) is a quick, straightforward, cost-efficient, and powerful hyphenated method for the non-targeted, high-throughput screening of plant extracts for bioactive compounds, avoiding the commonly used, suboptimal trial and error approach in the labor-intensive, expensive isolation procedure [13,27]. After the determination of the bioprofile of a sample by an HPTLC-direct bioautographic (DB) method [28], the detected inhibition zones can be characterized by various techniques, such as HPTLC-mass spectrometry (MS) [29]. Therefore, HPTLC hyphenations combined with preparative column chromatographic fractionations can contribute to the detection and subsequent separation, purification, and isolation of bioactive compounds from challenging matrices [30].
This study aimed to detect, isolate, and identify the antimicrobial root and leaf components of S. rugosa by the combination of a non-targeted, effect-directed screening and a highly targeted, bioassay-guided isolation involving high-performance thin-layer chromatography (HPTLC)-Bacillus subtilis assay, preparative flash chromatography, HPLChigh-resolution tandem mass spectrometry (HRMS/MS), and nuclear magnetic resonance (NMR) spectroscopy. The bioactivity of the isolated compounds was characterized by the determination of the minimal bactericidal concentration (MBC), the minimal inhibitory concentration (MIC), and the half maximal inhibitory concentration (IC 50 ) values mainly against plant pathogens.

Results and Discussion
HPTLC separation using n-hexane-isopropyl acetate 4:1 v/v mobile phase and detection with vanillin-sulfuric acid reagent and B. subtilis bioassay revealed active zones in both root and leaf at hR F 38 and 43, Compounds 1 and 2, respectively ( Figure 2). Both compounds were present in both tissues, but the root was richer in Compound 1, while Compound 2 was more prevalent in the leaves. Their isolation was achieved by preparative flash chromatographic fractionation and purification applying consecutive, orthogonal separation steps, switching from normal-phase silica gel to C 18 reversed-phase stationary phase and back to silica gel columns, yielding 25.5 mg of Compound 1 (white powder) and 9.4 mg of Compound 2 (pale yellow powder). Their structural characterization was performed based on mass spectrometric and NMR spectroscopic data. and a highly targeted, bioassay-guided isolation involving high-performance thin-laye chromatography (HPTLC)-Bacillus subtilis assay, preparative flash chromatography HPLC-high-resolution tandem mass spectrometry (HRMS/MS), and nuclear magneti resonance (NMR) spectroscopy. The bioactivity of the isolated compounds wa characterized by the determination of the minimal bactericidal concentration (MBC), th minimal inhibitory concentration (MIC), and the half maximal inhibitory concentratio (IC50) values mainly against plant pathogens.

Results and Discussion
HPTLC separation using n-hexane-isopropyl acetate 4:1 v/v mobile phase an detection with vanillin-sulfuric acid reagent and B. subtilis bioassay revealed active zone in both root and leaf at hRF 38 and 43, Compounds 1 and 2, respectively ( Figure 2). Bot compounds were present in both tissues, but the root was richer in Compound 1, whil Compound 2 was more prevalent in the leaves. Their isolation was achieved b preparative flash chromatographic fractionation and purification applying consecutive orthogonal separation steps, switching from normal-phase silica gel to C18 reversed-phas stationary phase and back to silica gel columns, yielding 25.5 mg of Compound 1 (whit powder) and 9.4 mg of Compound 2 (pale yellow powder). Their structura characterization was performed based on mass spectrometric and NMR spectroscopi data. In order to verify the assignments, the fragmentation patter of the protonated molecules was obtained at 20 eV collision energy, resulting in th subsequent losses of H2O and HCOOH moieties and the cleavage of the diterpen skeleton. The MS spectrometric data were confirmed by comparing them to the predicte spectrum [31] and to that reported earlier in the literature [32]. In order to verify the assignments, the fragmentation pattern of the protonated molecules was obtained at 20 eV collision energy, resulting in the subsequent losses of H 2 O and HCOOH moieties and the cleavage of the diterpene skeleton. The MS spectrometric data were confirmed by comparing them to the predicted spectrum [31] and to that reported earlier in the literature [32].   18) ppm. Based on the spectral data along with the 2D homo-and heteronuclear correlations, the structure of Compound 1 was elucidated as hardwickiic acid, a transclerodane diterpenoid carboxylic acid bearing β-substituted furan moiety. (-)-Hardwickiic acid has been isolated from S. rugosa by Lu et al. [10,33], which allowed the deduction of its absolute configuration. The levorotatory enantiomer was supported by the observed negative sign of its specific rotation value ( = −104.4 (c 0.4775, EtOH)), which is simi- Based on the spectral data along with the 2D homo-and heteronuclear correlations, the structure of Compound 1 was elucidated as hardwickiic acid, a trans-clerodane diterpenoid carboxylic acid bearing β-substituted furan moiety. (-)-Hardwickiic acid has been isolated from S. rugosa by Lu et al. [10,33], which allowed the deduction of its absolute configuration. The levorotatory enantiomer was supported by the observed negative sign of its specific rotation value ([α] 25 D = −104.4 (c 0.4775, EtOH)), which is similar to the previously reported data ([α] 25 D = −116.5 (c 1.09, EtOH), [34]). Furthermore, due to the lack of the reported NMR spectra recorded in the methanol-d 4 solvent, the published NMR spectroscopic data of its antipode, ent-(+)-hardwickiic acid [35], was used for confirmation as enantiomers have identical NMR spectra. The two spectra were in excellent agreement, which corroborated the proposed structure.
The molecular formula of Compound 2 represents six double bond equivalents.  [37]), the proposed structure was confirmed. The complete 1 H and 13 C NMR resonance assignments of Compounds 1 and 2 are listed in Table 1. The measured 1D and 2D NMR spectra of Compounds 1 and 2 are presented in Figures S1-S30.  to the lack of the reported NMR spectra recorded in the methanol-d4 solvent, the published NMR spectroscopic data of its antipode, ent-(+)-hardwickiic acid [35], was used for confirmation as enantiomers have identical NMR spectra. The two spectra were in excellent agreement, which corroborated the proposed structure.  EtOH)), the structure of Compound 2 was identified as (-)-abietic acid, a tricyclic abietane diterpenoid carboxylic acid. Comparing the measured spectral data in the methanol-d4 solvent [36] and the observed specific rotation with the previously published value ( = −101.7 (c 0.6, EtOH), [37]), the proposed structure was confirmed. The complete 1 H and 13 C NMR resonance assignments of Compounds 1 and 2 are listed in Table 1. The measured 1D and 2D NMR spectra of Compounds 1 and 2 are presented in Figures S1-S30. to the lack of the reported NMR spectra recorded in the methanol-d4 solvent, the published NMR spectroscopic data of its antipode, ent-(+)-hardwickiic acid [35], was used for confirmation as enantiomers have identical NMR spectra. The two spectra were in excellent agreement, which corroborated the proposed structure.  EtOH)), the structure of Compound 2 was identified as (-)-abietic acid, a tricyclic abietane diterpenoid carboxylic acid. Comparing the measured spectral data in the methanol-d4 solvent [36] and the observed specific rotation with the previously published value ( = −101.7 (c 0.6, EtOH), [37]), the proposed structure was confirmed. The complete 1 H and 13 C NMR resonance assignments of Compounds 1 and 2 are listed in Table 1. The measured 1D and 2D NMR spectra of Compounds 1 and 2 are presented in Figures S1-S30.  (-)-Hardwickiic acid is a common constituent throughout the Solidago genus. Apart from S. rugosa, it was isolated from S. arguta [38], S. juncea [39], and S. serotina [40]. Moreover, its presence was also described in various families of plants, e.g., Hardwickia pinnata [41], Croton aromaticus [42], Grangea maderaspatana [43], and Echinodorus grandiflorus [44]. (-)-Abietic acid is a widespread resin acid in nature initially isolated from rosin and occurring mainly in trees, e.g., in the resin of Pinus species (Resina Pini) [45], the leaves of Pimenta racemosa var. grisea [46], and the cones of Abies nordmanniana ssp. equi-trojani [47]. However, to the best of our knowledge, we report here for the first time that (-)-abietic acid is also present and abundant in a plant (S. rugosa) belonging to the Asteraceae family.
The antibacterial experiments revealed that both compounds exhibited similar efficiency against all studied Gram-positive bacterial strains, with approximately 5-20 times higher IC 50 values (between 1 and 5.1 µg/mL) than that of the positive control gentamicin ( Table 2). In the used concentrations, the MBC was reached only in the B. spizizenii assay with 10.4 and 5.2 µg/mL for Compounds 1 and 2, respectively. Weak inhibition was noticed against the Gram-negative X. arboricola pv. pruni. Both Compounds 1 and 2 exhibited weak activity in the F. avenaceum antifungal assay with an IC 50 value 15 and 33 times higher than that of the reference fungicide benomyl. Compound 1 was also more effective against B. sorokiniana with an IC 50 of 3.8 µg/mL, which means a 20 times stronger antifungal potency when compared to that of benomyl.  (-)-Hardwickiic acid exerted mild cytotoxic activity towards HuCCA-1 (human cholangiocarcinoma cancer), KB (human epidermoid carcinoma of the mouth), HeLa (cervical adenocarcinoma), MDA-MB231 (human breast cells), and T47D (human mammary adenocarcinoma) cell lines with IC 50 values ranging from 28.0 to 45.0 µg/mL [36]. (-)-Abietic acid possessed moderate cytotoxicity against SK-BR-3 (breast) human cancer cell line with an IC 50 value of 37.5 µg/mL, although it proved to be inactive against HL60 (leukemia), A549 (lung), and AZ521 (stomach) human cancer lines [48]. (-)-Hardwickiic acid was found antibacterial against several Gram-positive and Gram-negative bacterial strains, including human pathogens (among others, Enterococcus faecalis, Klebsiella aerogenes, Pseudomonas aeruginosa, Salmonella typhi, Shigella flexneri, Staphylococcus aureus, etc.), with IC 50 values varying from 1.22 to 78.12 µg/mL [49], being comparable or less potent to the Grampositive bacteria investigated in our study. However, a higher efficiency was observed against Gram-negative bacteria when compared to our findings [49]. The antibacterial activity of (-)-abietic acid was tested on numerous Gram-positive and Gram-negative bacterial strains, including human pathogens (Acinetobacter baumannii, Cutibacterium acnes, K. pneumoniae, S. aureus, etc.). In the case of Gram-positive organisms, the antibacterial effects were comparable or weaker than in our experiments [50][51][52][53]. However, it inhibited the growth of Gram-negative bacteria to a greater extent when compared to our results [50][51][52][53]. The antileishmanial effect of (-)-hardwickiic acid against promastigotes of Leishmania major and L. donovani was evaluated, and the assays provided the IC 50 values of 62.82 and 31.57 µM, respectively [54]. (-)-Abietic acid displayed a neuroprotective effect against in vitro cerebral ischemia induced by iodoacetic acid treatment using the mouse HT22 hippocampal nerve cell line [55]. The in vitro anti-inflammatory activity of (-)-abietic acid was also demonstrated, as it inhibited the IL-1β-induced production of TNF-α, NO, and PGE2 and suppressed the COX-2 expression in human osteoarthritis chondrocytes [56]. (-)-Abietic acid inhibited the activity of the soybean 5-lipoxygenase enzyme with an IC 50 value of 29.5 µM, suggesting that it may be used as a therapeutic agent in the treatment of various human diseases, including allergy, asthma, arthritis, and psoriasis [47,57]. The observed antifungal potency of (-)-hardwickiic acid against F. avenaceum was similar to its previously reported inhibition against Candida albicans and C. glabrata, albeit it displayed a stronger activity against F. avenaceum than against C. krusei [49]. A prominent antifungal inhibitory activity of (-)-abietic acid was reported against Rhodotorula mucilaginosa and Cladosporium cladosporioides with the MIC 90 values of 31 and 63 µg/mL, respectively [50], in both cases indicating a stronger effect compared to that against the fungal strains in our work. However, these compounds have not yet been tested against plant pathogens to assess their potential as a basis for a new plant protection agent.
The preparation of the B. subtilis cell suspension and the performance of the direct bioautographic assay were described previously [59]. The test bacterium was grown in lysogeny broth (LB: 10 g/L tryptone (Reanal), 5 g/L yeast extract (Scharlau, Barcelona, Spain), and 10 g/L sodium chloride (Reanal)) at 37 • C on an orbital shaker (120 rpm) to reach late exponential phase (OD 600 = 1.2). The steps of the direct bioautography were: 1. immersion of the developed and dried HPTLC plates into the bacterial cell suspension; Molecules 2023, 28, 3790 9 of 13 2. incubation of the bioautogram for 2 h, at 37 • C and 100% humidity; 3. dipping of the bioautogram into a vital dye solution (MTT, 1 mg/mL in water); 4. further incubation for 30 min; 5. documentation of the bioautogram under visible light (bright zones against a purple background indicate the antibacterial compounds).

NMR Spectroscopy
The NMR samples were prepared by dissolving the isolated Compounds 1 and 2 in 0.6 mL of methanol-d 4 and were transferred to a standard 5 mm NMR tube for analyses. NMR spectra were recorded on a Varian DDR 600 ( 1 H: 599.9 MHz, 13 C: 150.9 MHz; 14.1 T) spectrometer equipped with a dual 5 mm inverse-detection pulsed-field gradient (IDPFG) probehead at 298 K. VnmrJ 3.2C software was utilized for instrument operation and instrument control as well as data acquisition and data processing. All applied pulse sequences were part of the Chempack 5.1 standard pulse program library of the spectrometer. 1 H and 13 C chemical shifts (δ) are given on the δ-scale, reported in ppm, and referenced to the applied NMR solvent (CHD 2 OD residual peak at δ( 1 H) 3.31 ppm and CD 3 OD at δ( 13 C) 49.0 ppm), whereas spin-spin coupling constants (J) are provided in Hz. The complete resonance assignments were established from comprehensive one-( 1 H and 13 C) and two-dimensional homonuclear ( 1 H-1 H COSY, 1 H-1 H TOCSY, and 1 H-1 H NOESY) and heteronuclear ( 1 H-13 C edHSQC ( 1 J C-H = 140 Hz) and 1 H-13 C HMBC ( n J C-H = 8 Hz), both of them gradient-enhanced with adiabatic pulses) NMR experiments. To achieve the appropriate resolution in the 2D heteronuclear measurements for Compound 2, band-selective HSQC (bsHSQC) and HMBC (bsHMBC) spectra were also collected.

Determination of Optical Rotations
Optical rotations of the isolated compounds were measured at 25 • C with a Perkin Elmer 341 LC polarimeter (Waltham, MA, USA) in ethanol (1-0.4775 g/100 mL; and 2-0.1225 g/100 mL) at 589.3 nm (D-line of sodium) with an optical path length of 100 mm and an integration time of 2 s.

Microdilution Assays
B. spizizenii and X. arboricola pv. pruni were grown by shaking at 120 rpm in LB at 37 • C and 28 • C, respectively. C. flaccumfaciens pv. flaccumfaciens and C. michiganensis ssp. michiganensis were grown in Nutrient Broth (16 g/L Nutrient Broth (Biolab, Budapest, Hungary)) by shaking at 120 rpm at 28 • C. Liquid cultures of both fungi, B. sorokiniana and F. avenaceum, were shaken in LB at 120 rpm, 21 • C, in the dark, for three days. Mycelia were collected, washed in fresh LB, then fragmented using a homogenizer (FastPrep ® -24 Classic, MP Biomedicals, Beograd, Serbia) by putting the mycelia in 1 mL of the same medium into 2 mL Eppendorf tubes containing 7 × 2 mm glass beads. The homogenizer's acceleration was 4.5 m/s, and the duration 20 s.
We determined the MBC, MIC, and IC 50 of the isolated Compounds 1 (10 mg/mL in ethanol) and 2 (5 mg/mL in ethanol) using 96-well microplates. Gentamicin (0.1 mg/mL in ethanol) or benomyl (25 mg/mL in ethanol) served as the positive and ethanol as the negative control. A two-fold dilution series starting from 5 µL of each isolate was prepared in ethanol in the wells. When the ethanol evaporated under a sterile laminar airflow, 150 µL of a bacterial suspension (10 5 CFU/mL in LB) or 120 µL of a mycelium suspension (OD 600 = 0.1 in LB) was added to each well. Starting values of the absorbance at 600 nm were measured by a microplate spectrophotometer (Labsystems Multiscan MS 4.0, Thermo Fisher Scientific, Budapest, Hungary). The bacterial cultures were shaken at 900 rpm and an appropriate temperature with a microplate shaker (Grant PHMP) for 24 h, while fungal cultures were kept at 21 • C for 72 h. Then, the OD 600 values were again observed. The experiment was repeated twice with three parallels, and the results were averaged. The MBC was determined by plotting 5 µL of each well onto appropriate agar layers, and after incubation, by checking the presence or absence of bacterial colonies.

Conclusions
Two antimicrobial diterpenes were detected and isolated from the roots and leaves of S. rugosa utilizing a bioassay-guided process. Both compounds were present in both tissues, but Compound 1 was dominant in the roots, and Compound 2 in the leaves. These compounds, identified as (-)-hardwickiic acid (1) and (-)-abietic acid (2), exhibited notable antimicrobial activity against Gram-positive bacterial and fungal strains, including plant pathogens. The results imply that suitable structural modification and formulation could makethese compounds promising agrochemical agents.