Sesquiterpene Lactones from Cotula cinerea with Antibiotic Activity against Clinical Isolates of Enterococcus faecalis

Cotula cinerea, belonging to the tribe Anthemideae, is a plant widespread in the Southern hemisphere. It is frequently used in folk medicine in North African countries for several of its medical properties, shown by its extracts and essential oils. The dichloromethane extract obtained from its aerial parts demonstrated antibiotic activity against Enterococcus faecalis and was fractionated by bioguided purification procedures affording five main sesquiterpene lactones. They were identified by spectroscopic methods (NMR and ESIMS data) as guaiantrienolides, i.e., 6-acetoxy-1β-,6-acetoxy-1α-, and 6-acetoxy-10-β-hydroxyguaiantrienolide (1–3), and germacrenolides, i.e., haagenolide and 1,10-epoxyhaagenolide (4 and 5). The absolute configuration was assigned by applying the advanced Mosher’s method to haagenolide and by X-ray diffraction analysis to 1,10-epoxyhaagenolide. The specific antibiotic and antibiofilm activities were tested toward the clinical isolates of Enterococcus faecalis. The results showed that compounds 3–5 have antibacterial activity against all the strains of E. faecalis, while compound 2 exhibited activity only toward some strains. Compound 1 did not show this activity but had only antibiofilm properties. Thus, these metabolites have potential as new antibiotics and antibiofilm against drug resistant opportunistic pathogens.


Introduction
There has been remarkable interest in the last twenty years in the discovery of natural substances as a potential reservoir of innovative therapeutic solutions for human health, with the prospect of integrating and sometimes replacing conventional drugs. In particular, there is a growing interest in the study of natural substances extracted from plants to counter the galloping phenomenon of antibiotic resistance, i.e., the development and spread of microorganisms resistant to the drugs used in anti-infective therapies [1][2][3][4]. The misuse and abuse of antimicrobials in humans and animals have accelerated the growing phenomenon of antimicrobial resistance worldwide. Over the last few decades, antibiotic resistance (AR) has become a global threat to health systems and public health around the world. If it cannot be effectively counteracted, it is estimated that by 2050 the spread of multidrug-resistant strains will hamper the control of many infectious diseases, undermining the achievements of modern medicine [5].
Today, two main problems affect the effectiveness of antibiotics. The first is that following the introduction of a new antibiotic, resistance to it will arise sooner or later; activity against Fusarium graminearum and Fusarium sporotrichioides [28]. Furthermore, the antioxidant, anti-inflammatory, cytotoxic, and antimicrobial activities of its hydroethanolic and infusion extracts were also described [25]. Finally, it was also used as a bioherbicide for the biological control of the noxious weed Melilotus indicus L. [29].
However, up to now any report that exists has described biological activities of the metabolites isolated from this plant, and some of them are only partially chemical characterized. In addition, the organic plant extracts showed variability in the quality and quantity of their components. The differences in chemical profile do not surprise, and as for other plants they could be assigned to several factors, such as the geographical origin, the inter-and intraspecific genetic variability, and the plant vegetative stage.
In our study, organic extracts from Cotula cinerea isolated in Algerian Sahara are tested against both Gram-positive and Gram-negative bacteria. The preliminary antibiotic activity of the CH 2 Cl 2 plant extract prompts a complete chemical and biological characterization of the secondary metabolites. Thus, this manuscript reports the complete spectroscopic properties, the absolute stereochemistry of five sesquiterpene lactones belonging to guaianolides and germacrenolides subgroups, and their specific antibiotic and antibiofilm activity against the clinical and reference isolates of Enterococcus faecalis.

Results and Discussion
The organic extracts of C. cinerea obtained using different solvents (in succession n-hexane, CH 2 Cl 2 and EtOAc) as detailed in Section 3 were assayed at a single high concentration on Staphylococcus aureus ATCC 29231, Enterococcus faecalis ATCC 29212, Pseudomonas aeruginosa ATCC 27853, Acinetobacter baumannii ATCC 747 by standard broth microdilution assay to test the bacterial growth inhibition. Among them, only the CH 2 Cl 2 extract showed an antibacterial activity (Table 1), inhibiting in particular only the growth of Gram-positive test strains with 49% ± 2.7 of inhibition for S. aureus and 90% ± 1.5 for E. faecalis. Table 1. Antibacterial activity of CH 2 Cl 2 extract (500 µg/mL), expressed as percentage of growth inhibition, with ± standard deviation (SD), against Gram-positive and Gram-negative test strains (n.d. = inhibition not detected).  Markouk et al. (1999) [19] showed that the n-butanol extract of C. cinerea was active against phytopathogenic strains of Pseudomonas and Bacillus. The C. cinerea n-butanol extract was found to be active also against clinical isolates of Klebsiella pneumoniae, Pseudomonas aeruginosa, and S. aureus [30]. Ghouti et al. (2018) [20] tested a hydroethanolic extract and infusion of C. cinerea against both Gram-negative and Gram-positive bacteria, and the findings showed a higher activity toward Gram-positive ones.

Strain
The high efficacy showed by the CH 2 Cl 2 plant extract in inhibiting the planktonic growth of E. faecalis appeared to be very interesting. E. faecalis, a commensal organism of human gastrointestinal tract, is recognized as an important nosocomial pathogen whose control is of paramount importance since this bacterium can cause bacteremia, urinary tract infections, endocarditis, intra-abdominal and pelvic infections, and burn and surgical site wound infections [31][32][33]. It has also become increasingly recognized as an important pathogen that causes infections of catheters and other implanted medical devices, including those that are orthopedic [34][35][36]. E. faecalis is in fact the most frequently identified species in enterococcal prosthetic joint infection (PJI) [37,38]. In dentistry, E. faecalis is particularly prevalent in root canals with a diagnosis of apical periodontitis and has been implicated as the main pathogen in persistent secondary endodontic infections [39,40]. The success of E. faecalis as a pathogen mostly linked to hospital-acquired infections (HAIs) is due to its intrinsic resistance to many antimicrobials and its capacity to acquire new genetic determinants of resistance to different classes of antibiotics. Furthermore, it is capable of forming a well-organized biofilm, which makes it more resistant to antibiotic killing and the action of human immune defenses, and it tolerates extreme environmental conditions (poor nutrient availability, low O 2 potential, alkaline pH) [41]. Furthermore, it can induce hydroxyapatite precipitation in a mature biofilm to form a calcified biofilm [42].
Thus, the active CH 2 Cl 2 extract was fractionated by combined column and TLC on direct and reverse phase to afford five pure metabolites. The first investigation of their 1 H NMR and ESI MS spectra showed that they are sesquiterpenoids belonging to different subgroups. They were identified as the guaianolides, i.e., 6-acetoxy-1β-, 6-acetoxy-1αand 6-acetoxy-10-β-hydroxyguaiantrienolide (1-3, Figure 1), and the germacrenolides, i.e., haagenolide and 1,10-epoxyhaagenolide (4 and 5, Figure 1), by comparing their 1 H NMR and MS data with those reported in literature, in particular for 1-3 with the data reported by Jakupovic et al. (1988) [43] when isolated from the same plant and for 4 and 5 with those reported when isolated, respectively, from the Inula species [44] and from Inula heterolepis [45].
gical site wound infections [31][32][33]. It has also become increasingly recognized as an important pathogen that causes infections of catheters and other implanted medical devices, including those that are orthopedic [34][35][36]. E. faecalis is in fact the most frequently identified species in enterococcal prosthetic joint infection (PJI) [37,38]. In dentistry, E. faecalis is particularly prevalent in root canals with a diagnosis of apical periodontitis and has been implicated as the main pathogen in persistent secondary endodontic infections [39,40]. The success of E. faecalis as a pathogen mostly linked to hospital-acquired infections (HAIs) is due to its intrinsic resistance to many antimicrobials and its capacity to acquire new genetic determinants of resistance to different classes of antibiotics. Furthermore, it is capable of forming a well-organized biofilm, which makes it more resistant to antibiotic killing and the action of human immune defenses, and it tolerates extreme environmental conditions (poor nutrient availability, low O2 potential, alkaline pH) [41]. Furthermore, it can induce hydroxyapatite precipitation in a mature biofilm to form a calcified biofilm [42].
Thus, the active CH2Cl2 extract was fractionated by combined column and TLC on direct and reverse phase to afford five pure metabolites. The first investigation of their 1 H NMR and ESI MS spectra showed that they are sesquiterpenoids belonging to different subgroups. They were identified as the guaianolides, i.e., 6-acetoxy-1β-, 6-acetoxy-1αand 6-acetoxy-10-β-hydroxyguaiantrienolide (1-3, Figure 1), and the germacrenolides, i.e., haagenolide and 1,10-epoxyhaagenolide (4 and 5, Figure 1), by comparing their 1 H NMR and MS data with those reported in literature, in particular for 1-3 with the data reported by Jakupovic et al. (1988) [43] when isolated from the same plant and for 4 and 5 with those reported when isolated, respectively, from the Inula species [44] and from Inula heterolepis [45].  However, the authors that were cited above only partially reported the 1 H NMR data of the five sesquiterpenoids. Thus, by extensive use of 1D and 2D 1 H and 13 C NMR However, the authors that were cited above only partially reported the 1 H NMR data of the five sesquiterpenoids. Thus, by extensive use of 1D and 2D 1 H and 13 C NMR spectra (COSY, HSQC and HMBC), the chemical shifts to all the carbons and protons were assigned for the first time and reported in Tables 2-6.     The relative configuration of 1-5 were supported by correlations observed in their NOESY spectra (see Figures S6, S12, S18, S24 and S30).
The 1 H NMR spectra of both 6 and 7 (Table 7) substantially differed from that of 4 for the downfield shift (∆δ 1.27 and 1.24, respectively) of H-9 and for the presence of the proton of benzene ring system and the singlet of the methoxy group at δ 7.508-7.384 and 7.532-7.367 and 3.519 and 3.557, respectively. Subtracting the chemical shifts of the protons (Table 6) of 9-O-S-MTPA of 4 (7) from those of 9-O-R-MTPA of 4 (6) esters, the ∆δ (7-6) values for all of the protons were determined as reported in Figure 2. The positive ∆δ values are located on the right-hand side and the negative values on the left-hand side of model A as reported previously [46]. This model allowed us to assign the S configuration to C-9 and consequently the S and R configuration to C-7 and C-6, respectively. Thus, 4 was formulated as (6R,7S,9S)-5-hydroxy-6,10-dimethyl-3-methylene-3a,4,5,8,9,11a-hexahydro- For compound 5, the relative configuration was confirmed by an X-ray analysis carried out on the single crystal obtained by the slow evaporation of CHCl 3 -i-PrOH (1/1, v/v) in a water-saturated atmosphere at 5 • C. The ORTEP view of the epoxyhaagenolide molecular structure is reported in Figure 3.   For compound 5, the relative configuration was confirmed by an X-ray analysis carried out on the single crystal obtained by the slow evaporation of CHCl3-i-PrOH (1/1, v/v) in a water-saturated atmosphere at 5 °C. The ORTEP view of the epoxyhaagenolide molecular structure is reported in Figure 3. The 1R, 6R, 7S, 9S, and 10R absolute configurations of 5 at the C1, C6, C7, C9, and C10 stereogenic centers were determined on the basis of the anomalous scattering in the X-ray diffraction data by performing X-ray diffraction data collection according to the   For compound 5, the relative configuration was confirmed by an X-ray analysis carried out on the single crystal obtained by the slow evaporation of CHCl3-i-PrOH (1/1, v/v) in a water-saturated atmosphere at 5 °C. The ORTEP view of the epoxyhaagenolide molecular structure is reported in Figure 3. The 1R, 6R, 7S, 9S, and 10R absolute configurations of 5 at the C1, C6, C7, C9, and C10 stereogenic centers were determined on the basis of the anomalous scattering in the X-ray diffraction data by performing X-ray diffraction data collection according to the The 1R, 6R, 7S, 9S, and 10R absolute configurations of 5 at the C1, C6, C7, C9, and C10 stereogenic centers were determined on the basis of the anomalous scattering in the X-ray diffraction data by performing X-ray diffraction data collection according to the literature methods reported to assigning the absolute configuration in light-atom structures when MoKα radiation is used (see Section 3 for details).
As shown in Figure 3, the molecule is a sesquiterpene lactone with a 10-membered ring transfused with a 5-membered lactone ring. The lactone ring is in the conformation twisted on C6-C7. The 10-membered ring adopts the chair-chair-chair conformation with methyl groups in axial positions and hydroxy and 1,10-epoxy groups in equatorial positions (see the Supplementary Materials for crystallographic data).
Sesquiterpenes 1-5 were independently assayed against E. faecalis reference strains and four clinical isolates typed by studying the susceptibility profile to antibiotics (Table 8). Compounds were tested by broth microdilution method in a concentration range between 300 and 9 µg/mL. At the highest tested concentrations of 150 and 300 µg/mL, compounds 3, 4, and 5 showed activity against all tested strains while compound 2 was active on four of the five tested strains. Growth inhibition rates ranged from 50 to 90% (Table 9). Compound 1, on the other hand, did not show inhibitory activity against any of the test strains. Table 9. Antibacterial activity, expressed as percentage of maximum growth inhibition (GI) (mean percentage ± SD) of compounds 1-5 at concentrations ranged from 300 to 9 µg/mL against E. faecalis test strains. Considering the activity in the subgroup of guaianolides (1-3), the substituent at C-10 seems not important for the activity, while an important structural feature appeared to be the configuration. In fact, the presence of a hydroxy group protruding from the α-side determined inactivity as observed in 1. In the subgroup of germacrenolides, the nature of the substituents at C-1 and C-10 did not seem to play a role to impart activity.

Strains
As described above, enterococci can form biofilms on foreign biomaterials. Bacterial subpopulations within the biofilm characterized by low metabolism have a reduced absorption of antibiotics, especially for the active molecules on the cell wall as beta-lactams and glycopeptides, making the implant-device infections difficult to treat [47][48][49].
Our interest was to investigate the antibiofilm activity of the five natural metabolites on the test strains of E. faecalis, using crystal violet as a staining method for total biofilm biomass quantification. Serial dilutions of compounds 1-5 were tested in microplates starting from concentrations corresponding to 1/8 or 1/4 of the concentration inhibiting 80% and 50% of planktonic growth, respectively. The compounds, which did not show inhibitory activity against the planktonic form, were tested starting from the concentration of 300 µg/mL (Table 10). Table 10. In vitro biofilm formation inhibition (BI) of E. faecalis test strains, presented as mean percentage ± SD, following overnight treatment with compounds 1-5 at serial dilutions of sub-planktonic growth inhibitory concentrations. Compound 1 inhibited biofilm formation by all test strains at a concentration of 150 µg/mL. No significant differences were found between biofilm inhibition percentage at 150 and 300 µg/mL (p > 0.005). Compounds 2-5 exhibited a strain-dependent activity at the maximum concentration tested, showing a biofilm inhibition ranging from 50 to 70% for most of the test strains.

Strains
The E. faecalis biofilm is highly resistant to the action of ampicillin, vancomycin, and linezolid, despite prolonged times of treatment [50,51]. Some authors have tested the in vitro efficacy of antibiotic combinations against enterococci in biofilms, and they showed how rifampicin-containing antibiotic regimes, rifampicin-tigecycline combinations, and fosfomycin-based combinations were able to reduce the number of bacteria in biofilms formed within 24 h [52][53][54]. However, to date, management strategies in enterococcal PJI are controversial and nonstandardized as well as those in enterococcal endodontic infections, and the disinfection of root canals with commonly used antimicrobial irrigations is very challenging due to cytotoxicity [55][56][57][58].
In light of these considerations, these preliminary data encourage further studies aimed to reach improved antibiofilm activity by combining these natural metabolites with conventional antibiotics to fight PJI as well as with endodontic irrigating agents for more effective intracanal medicaments against this microorganism.

Plant Material
Whole aerial parts of Cotula cinerea plants were collected in winter 2019 from the Oued Souf area located in Southern Algeria, which is characterized by dry weather, Saharan climate (33 • 21 21 N 6 • 51 47 E; mean annual temperature: 21 • C; mean annual precipitation: 75 mm; elevation: 80 m). Plant samples were dried naturally on laboratory benches at room temperature (24 • C) for 1 month. The dried plant materials were powdered by using a blender and then stored away from humidity in bags of conservation. A specimen of the plant was deposited in the collection of the Nationale Supérieure Agronomique, Département de Zoologie Agricole et Forestière, El-Harrach, Alger, Algeria.
3.4. Crystallographic Data of 1,10-Epoxyhaagenolide (5) Single crystals of 5, which were suitable for X-ray structure analysis, were obtained by slow evaporation of CHCl 3 -i-PrOH (1/1, v/v) in a water-saturated atmosphere at 5 • C. One selected crystal was mounted in flowing N 2 at 173 K on a Bruker-Nonius Kappa CCD diffractometer equipped with an Oxford Cryostream apparatus (graphite monochromated MoKα radiation λ = 0.71073 Å; CCD rotation images, thick slices; ϕ and ω scans to fill the asymmetric unit). A semiempirical absorption correction (multiscan, SADABS) was applied. The structure was solved by direct methods using SIR97 program [60] and anisotropically refined by the full matrix least-squares method on F 2 against all independent measured reflections using the SHELXL-2018/3 [61] and WinGX software [62]. Crystallization solvent water molecules were found in the crystal structure. All the hydroxy and water H atoms were individuated in difference Fourier maps and freely refined with U iso (H) equal to 1.2 U eq of the carrier atom. All the other H atoms were placed in calculated positions and refined accordingly to a riding model with C-H distances in the range 0.95-1.00 Å and U iso (H) equal to 1.2 U eq of the carrier atom (1.5 U eq for C methyl ). The 1R, 6R, 7S, 9S, and 10R absolute configurations of 5 at C1, C6, C7, C9, and C10 stereogenic centers were determined by performing X-ray diffraction data collection according to the literature methods reported in order to assign the absolute configuration in light-atom structures when MoKα radiation is used. The calculated absolute parameters and the Bayesian statistics p values indicate a very high probability that the absolute configuration is correctly assigned [63][64][65]. Bijvoetpair analysis and Bayesian statistics were performed using the program PLATON [66]. The figures were generated using ORTEP-3 [62] and Mercury CSD 4.0 [67].  (14), Bijvoet pairs 3724, Friedel coverage 97%. Absolute structure parameters were calculated using the programs SHELXL-2018/3 and PLATON-v30118.

Test Bacterial Strains and Culture Conditions
Bacterial strains used in this study were reference and clinical strains: Pseudomonas aeruginosa ATCC 27853, Acinetobacter baumannii ATCC 747, Staphylococcus aureus ATCC 29213, Enterococcus faecalis ATCC 29212, and four clinical strains of E. faecalis (EF-91823 and EF-91804 isolated from PJI; EF-165 and EF-91705 isolated from bloodstream infections). The strains were obtained from a collection of anonymous clinical isolates previously established at the Department of Molecular Medicine and Medical Biotechnologies (University of Naples Federico II). No ethical approval was required for the study because there was no access to patients' data. All strains were stored as 15% (v/v) glycerol stocks at −80 • C. Before each experiment, cells were subcultured from the stocks onto TSA plates at 37 • C for 24 h. The strains were identified by MS MALDI-TOF (Bruker Daltonics, Bremen, Germany) and were characterized by studying the antibiotic susceptibility profile performed on Vitek 2 (bioMérieux, Marcy-l'Étoile, France).

Antimicrobial Assays
Standard broth microdilution method in 96-well polystyrene plates was used to determine the minimum inhibitory concentration (MIC) of plant extracts and secondary metabolites using Mueller-Hinton Broth 2 (MHB2) as culture medium. For each test strain, starting from bacterial suspensions with a turbidity of 0.5 McFarland (corresponding to 1-5 × 10 8 cells/mL) and subsequently adjusted to approximately 5 × 10 6 CFU/mL −1 , 100 µL aliquots of the suspension were dispensed in wells in triplicate. Then 100 µL of 500 µg/mL of plant extract solution were added to the wells, and plates were incubated at 37 • C for 19 h under shaking (300 rpm) and under aerobic conditions, except for E. faecalis strains, which were incubated under anaerobic conditions. Instead, to test the activity of the five metabolites, the wells were added with equal volumes (100 µL) of the suspensions and twofold serial dilutions starting from 300 µg/mL for each metabolite. Wells with no compounds were used as positive controls (100% growth). Conventional antibiotics as ampicillin and amikacin, selected depending on antibiotic-susceptibility profiles of the test strains, were included as control for Gram positive and Gram negative, respectively. The medium turbidity was measured by a spectrophotometer at 595 nm (Bio-Rad Laboratories S.r.l., Hercules, CA, USA). Antimicrobial activity was expressed as a percentage of microbial growth inhibition. Each assay was performed twice. To be sure that the 2% of DMSO present in the 2× stock solutions of the plant extracts and that metabolites did not act on bacterial growth, the effect of serial dilutions of DMSO starting from 1% on the growth of test strains was separately tested.

Biofilm Formation Inhibition Assay
Biofilm biomass formed in vitro in the presence of the five secondary metabolites was measured using the crystal violet (CV) staining method in flat-bottomed 96-well microplates as described by Stepanović et al. with some modifications [68]. For each strain, a cell suspension in BHI supplemented with 1% (w/v) glucose was prepared and diluted to obtain a suspension of 1.5 × 10 6 CFU/mL. One hundred microliters of this suspension were incubated with 100 µL of serial dilutions of sub-MIC compounds concentrations, showing no influence on planktonic growth of the test strains. The positive controls were compound-free wells. The microtiter was incubated at 37 • C for 24 h. The nonadherent cells were then removed with gentle aspiration and gentle washings with PBS, and the biofilm was dried at 60 • C for 30 min and subsequently stained with 0.1% (w/v) crystal violet solution for 30 min. After washing with PBS and solubilization with absolute ethanol to release the dye from the biofilm, absorbance measurement values at 570 nm were obtained using the spectrophotometric reading. The absorbance recorded was correlated to the amount of biofilm produced. The percentage of biofilm mass reduction was calculated using the formula [(Ac-At)/Ac] × 100, where Ac is the OD570 for control wells, and At is the OD570 in the presence of the tested compound.

Statistical Analysis
Data were represented as the mean ± standard deviation and analyzed for statistical significance using ordinary one-way analysis of variance (ANOVA) and Tukey's multiple comparisons test. For all tests, p < 0.005 was considered to indicate a statistically significant difference.

Conclusions
This is the first report on the antimicrobial and antibiofilm activities of three guaiantrienolides and two germacrenolides isolated from Cotula cinerea against clinical isolates of Enterococcus faecalis, a common drug-resistant opportunistic pathogen responsible for important biofilm-related infections. Our results are preliminary data, but we can hypothesize that these natural substances could be a potential alternative for new antibiofilm formulations in strategies for the prevention of persistent infection by Enterococcus faecalis. The introduction of modifications in their structure on the basis of structure-function stud-ies could lead to the synthesis of derivatives with higher antibiofilm activity. In addition, the absolute configuration of two of the three germacranolides was also assigned.