Three New Xanthones from Hypericum scabrum and Their Quorum Sensing Inhibitory Activities against Chromobacterium violaceum

Quorum sensing (QS) plays an important role in the production of virulence factors and pathogenicity in pathogenic bacteria and is, therefore, a hopeful target to fight against bacterial infections. During our search for natural QS inhibitors, two new xanthonolignoids (1 and 2), each existing as a racemic mixture, one new simple oxygenated xanthone (7), and eight known analogs (3–6, 8–11) were isolated from Hypericum scabrum Linn. Chiral separation of 1 yielded a pair of enantiomers 1a and 1b. The structures of these compounds were elucidated by spectroscopic analysis and ECD (electrostatic circular dichroism) calculations. All isolates were evaluated for their QS inhibitory activity against Chromobacterium violaceum. Both 9 and 10 exhibited the most potent QS inhibitory effects with minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) values of 31.25 and 62.5 μM, respectively. Crystal violet staining was used to further evaluate the biofilm inhibition potential of compounds 7, 9 and 10, and the formation of biofilms increased with decreasing drug concentration in a classic dose-dependent manner. The results of a cytotoxicity assay revealed that compounds 7, 9 and 10 exhibited no cytotoxic activity on PC-12 cells at the tested concentration.


Introduction
Microbial infection is still a major public health concern worldwide. Interfering with quorum sensing (QS) systems that regulate the expression of many genes associated with the virulence of pathogens was a promising alternative strategy for controlling bacterial infection, especially for those caused by multi-drug resistant strains [1,2]. Chromobacterium violaceum ATCC12472 is one of the most commonly used bacterial species for QS research. Quorum sensing inhibitors (QSIs) are different from conventional antibiotics in that they focus on reducing the virulence of bacteria without killing bacteria or inhibiting bacterial growth. Recently, there has been increasing evidence that phytochemicals can be an important source of QSIs [3,4].
Xanthones are the main secondary metabolites of plants belonging to the genus Hypericum, which is widely distributed in temperate zones of the world and many have been used as folk medicines. Xanthonolignoids are a relatively rare class of natural compounds, most of which have a phenylpropane moiety attached to the xanthone skeleton through a dioxane ring [5,6]. Hypericum scabrum is a perennial herb growing mainly in arid rocky hillsides or gravel-sloped lands of Altai in Xinjiang, China and Central Asia, where it is commonly used to treat a variety of disorders of the heart, liver, gallbladder, intestines and bladder [7,8]. In our ongoing research for natural QSIs, eleven xanthones, including two new xanthonolignoids and one new simple oxygenated xanthone (Figure 1), were mainly in arid rocky hillsides or gravel-sloped lands of Altai in Xinjiang, China and Central Asia, where it is commonly used to treat a variety of disorders of the heart, liver, gallbladder, intestines and bladder [7,8]. In our ongoing research for natural QSIs, eleven xanthones, including two new xanthonolignoids and one new simple oxygenated xanthone (Figure 1), were obtained from the whole plants of H. scabrum. Herein, we describe the isolation, structure elucidation, and QS inhibitory activities of these compounds against C. violaceum.

General Experimental Procedure
UV spectra were measured on a Shimadzu UV-1800 spectrophotometer. A Thermo Scientific Nicolet iN10 Microscope was used to obtain IR spectra. ECD spectra were measured on a JASCO J1500 CD spectrometer. NMR spectra were acquired in DMSO-d6 at 303 K on a Bruker AVANCE III-500 ( 1 H NMR, 500 MHz; 13 C NMR, 125 MHz) with TMS as the internal standard. HRESIMS data were acquired on a Thermo Q-Exactive orbitrap mass spectrometer. Preparative HPLC was performed on a Shimadzu HPLC system consisting of an LC-6AD pump and a Shimadzu SPD-20A detector, coupled with a Shim-park RP-C18 column (

Plant Material
The whole plants of Hypericum

General Experimental Procedure
UV spectra were measured on a Shimadzu UV-1800 spectrophotometer. A Thermo Scientific Nicolet iN10 Microscope was used to obtain IR spectra. ECD spectra were measured on a JASCO J1500 CD spectrometer. NMR spectra were acquired in DMSO-d 6 at 303 K on a Bruker AVANCE III-500 ( 1 H NMR, 500 MHz; 13 C NMR, 125 MHz) with TMS as the internal standard. HRESIMS data were acquired on a Thermo Q-Exactive orbitrap mass spectrometer. Preparative HPLC was performed on a Shimadzu HPLC system consisting of an LC-6AD pump and a Shimadzu SPD-20A detector, coupled with a Shim-park RP-C18 column (

Anti-QS Assay
C. violaceum ATCC12472 provided by Ocean University of China was used as the target bacterium. The antibacterial activities of isolated compounds (1-11) were screened according to Clinical and Laboratory Standards Institute (CLSI) guidelines. For the QS assay, 100 µL of different concentrations of each compound (0-1 mmol/L) and C. violaceum were incubated into a 96-well plate at 30 • C for 24 h without shaking,. Then, the OD 590 values were measured using a microplate reader (Thermo Fisher), and the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) values were determined.
To quantify violacein production in C. violaceum after treatment with tested compounds, microorganisms in each well were collected and centrifuged for 10 min at 13,000 rpm. After removing the supernatant, an equal volume DMSO was added to precipitates to dissolve the violacein. The solution was then vortexed at maximum velocity for 1 min; the extracted violacein was acquired after another centrifugation (13,000 rpm for 10 min). The absorbance of violacein solution was recorded with a spectrophotometer at 585 nm. The inhibition percentage was calculated based on the formula reported in literature [9].
To further study their antibacterial activity, inhibition of biofilm formation experiments of compounds 7, 9 and 10 were conducted using a crystal violet assay. Briefly, tested compounds and bacterial suspension were mixed and added into each well of the 96-well microplate. After incubation for 16 h at 37 • C, the medium was removed and washed with sterile water three times. The plates were dried at 65 • C, and 200 µL of 1% (m/v) aqueous solution of crystal violet (CV) was added to stain for 20 min, then thoroughly washed with PBS three times. Subsequently, 100 µL of 95% ethanol was added to each well to dissolve CV at 37 • C for 30 min. The absorbance at OD 470 nm was then measured to determine the biofilm biomass on a microplate reader [2].

ECD Calculations
Conformational analyses were carried out using the MMFF94S force field with an energy cutoff of 2.5 kcal/mol. Subsequently, the low-energy conformers were re-optimized using DFT at the B3LYP/6-311+G(d,p) level in MeOH using the polarizable conductor calculation model (CPCM). The energies, oscillator strengths, and rotational strengths (velocity) of the first 30 electronic excitations were calculated using the TDDFT methodology at the B3LYP/6-311+G(d,p) level in MeOH. The ECD spectra were simulated by the overlapping Gaussian function (half the bandwidth at 1/e peak height, sigma = 0.20 for all) [10]. To get the final spectra, the simulated spectra of the conformers were averaged according to the Boltzmann distribution theory and their relative Gibbs free energy (∆G).

Results and Discussion
Compound 1 was isolated as an amorphous white powder. Its molecular formula was assigned as C 24 H 20 O 9 on the basis of the HRESIMS ( Figure S3 were observed. These observations suggested that 1 was a xanthonolignoid. Overall, the 1D NMR data of 1 were highly similar to those of known compound kielcorin (4) [11,12]. The only difference was that 1 possesses an additional hydroxyl group, which was deduced to be at C-3 based on the HMBC correlations as shown in Figure 2. The HMBC spectrum also showed correlations from H-2 (δ H 7.05) to C-7 , from H-6 (δ H 6.90) to C-7 and from H-7 (δ H 5.05) to C-1 , confirming that the trisubstituted benzene ring attached at C-7 . Furthermore, HMBC cross-peaks from H-7 (δ H 5.05) to C-8 , from H-8 (δ H 4.37) to C-7 and from H-9 (δ H 3.70 and 3.42) to C-8 , suggesting that a hydroxymethyl unit linked at C-8 . Since compound 1 showed no Cotton effect in the ECD spectrum, it was very likely that it was a racemic mixture, which was consistent with the optical rotation value of +0.7. Subsequent chiral resolution of 1 yielded a pair of enantiomers 1a and 1b in an approximate ratio of 1:1 (Figure 3), which exhibited mirror image-like ECD curves ( Figure  4). The absolute configurations of 1a and 1b were defined as (7R, 8R) and (7S, 8S), respectively, by means of ECD calculations. Since compound 1 showed no Cotton effect in the ECD spectrum, it was very likely that it was a racemic mixture, which was consistent with the optical rotation value of +0.7. Subsequent chiral resolution of 1 yielded a pair of enantiomers 1a and 1b in an approximate ratio of 1:1 (Figure 3), which exhibited mirror image-like ECD curves (Figure 4). The absolute configurations of 1a and 1b were defined as (7R, 8R) and (7S, 8S), respectively, by means of ECD calculations.
Compound 2 had a molecular formula of C 25  . A comparison of the NMR data of 2 with those of 1 revealed that 2 was the 5 -methoxy derivative of 1, which was supported by the HMBC data ( Figure 2). Accordingly, the structure of 2 was elucidated as 6-hydroxy-5 -methoxy kielcorin. Based on the optical rotation value of +1.1, compound 2 was also considered as a racemic mixture. However, due to the limited amount available, chiral resolution of this compound was not performed. Since compound 1 showed no Cotton effect in the ECD spectrum, it was very likely that it was a racemic mixture, which was consistent with the optical rotation value of +0.7. Subsequent chiral resolution of 1 yielded a pair of enantiomers 1a and 1b in an approximate ratio of 1:1 (Figure 3), which exhibited mirror image-like ECD curves ( Figure  4). The absolute configurations of 1a and 1b were defined as (7R, 8R) and (7S, 8S), respectively, by means of ECD calculations.   Since compound 1 showed no Cotton effect in the ECD spectrum, it was very likely that it was a racemic mixture, which was consistent with the optical rotation value of +0.7. Subsequent chiral resolution of 1 yielded a pair of enantiomers 1a and 1b in an approximate ratio of 1:1 (Figure 3), which exhibited mirror image-like ECD curves ( Figure  4). The absolute configurations of 1a and 1b were defined as (7R, 8R) and (7S, 8S), respectively, by means of ECD calculations.  The UV spectrum showed absorptions at 236, 293, 332 and 383 nm, indicating that 7 was a xanthone [12][13][14]. The 1D NMR data of 7 were comparable to those of 1,2,6,7tetrahydroxyxanthone (9) [15]. The only difference was that the hydroxyl group at C-6 in 9 was placed at C-5 in 7, which was confirmed by analysis of HMBC data ( Figure 2). Therefore, the structure of 7 was concluded to be 1,2,5,7-tetrahydroxyxanthone.
The genus Hypericum is rich in xanthones that can be classified into xanthonolignoids, prenylated and simple (hydroxy and/or methoxy) xanthones [16], which were reported to possess various bioactivities [18] such as antibacterial activity [19][20][21], cytotoxicity, etc. Since H. scabrum has been proven to have antimicrobial activity and xanthones are the main chemical constituents of this plant [22,23], the antibacterial activities of all isolated compounds (1-11) against C. violaceum ATCC12472 were screened. As a result (Table 2), compounds 7, 9, 10 and 11 manifested antibacterial activity with MIC ranging from 31.25 to 500 µM. Notably, all the tested compounds did not inhibit the growth of C. violaceum at subinhibitory concentrations of MIC. Based on the results of this study, it appeared that the antibacterial activity of simple xanthones was stronger than that of xanthonolignoids. The presence of the phenylpropane moiety may decrease the antibacterial activity. In simple xanthones, 6-OH is essential for the maintenance of antimicrobial effects.  Purple pigment formation is one of the main forms of cell-cell communication used by C. violaceum to coordinate their group behavior in response to population density. Compounds 7, 9, 10 and 11 were further assessed for if they can inhibit the purple pigment formation of C. violaceum. As shown in Figure 5, at the concentration of 16.53 µM, these compounds could reduce the formation of purple pigment with an inhibition rate of 21.27%, 75.36%, 76.21% and 53.11%, respectively.

MIC (µM) MBC (µM) MBC/MIC Ratio
Recent reports suggest that the formation of rigid biofilms on natural and artificial surfaces drastically promotes the resistance capabilities of bacteria. The combination of antibiotics with bacterial biofilm-preventing or disturbing agents seems to be a promising strategy to prevent bacterial infection [24]. The formation of bacterial biofilms is regulated by the QS system. In this study, the biofilm inhibition potential of compounds 7, 9 and 10 was evaluated at different concentrations using crystal violet staining. As shown in Figure 6, the formation of biofilms increased with decreasing drug concentration in a classic dose-dependent manner. At a concentration of 15.63 µM, the biofilm inhibition rates of compounds 7, 9 and 10 were 55.17%, 86.23% and 74.25%, respectively. These results suggest that compounds 7, 9 and 10 have potential abilities to inhibit the formation of biofilms, which is considered a potential drug target for controlling drug-resistant chronic infections [25].
In light of several studies revealing the cytotoxic activity of xanthones, the cytotoxic effects of compounds 7, 9 and 10 toward PC-12 cells were evaluated using an MTT assay. As summarized in Figure 7, the three compounds exhibited no cytotoxic activity on PC-12 cells at the concentration of 0.97-15.63 µM. It's worth noting that compounds 7, 9 and 10 showed more than 50% bacterial biofilm reduction at the concentration of 15.63 µM but were still safe for PC-12 cells. They showed no obvious damage to PC-12 cells even at the concentrations of MIC and MBC.
Purple pigment formation is one of the main forms of cell-cell communication used by C. violaceum to coordinate their group behavior in response to population density. Compounds 7, 9, 10 and 11 were further assessed for if they can inhibit the purple pigment formation of C. violaceum. As shown in Figure 5, at the concentration of 16.53 μM, these compounds could reduce the formation of purple pigment with an inhibition rate of 21.27%, 75.36%, 76.21% and 53.11%, respectively. Recent reports suggest that the formation of rigid biofilms on natural and artificial surfaces drastically promotes the resistance capabilities of bacteria. The combination of antibiotics with bacterial biofilm-preventing or disturbing agents seems to be a promising strategy to prevent bacterial infection [24]. The formation of bacterial biofilms is regulated by the QS system. In this study, the biofilm inhibition potential of compounds 7, 9 and 10 was evaluated at different concentrations using crystal violet staining. As shown in Figure 6, the formation of biofilms increased with decreasing drug concentration in a classic dose-dependent manner. At a concentration of 15.63 μM, the biofilm inhibition rates of compounds 7, 9 and 10 were 55.17%, 86.23% and 74.25%, respectively. These results suggest that compounds 7, 9 and 10 have potential abilities to inhibit the formation of biofilms, which is considered a potential drug target for controlling drug-resistant chronic infections [25]. In light of several studies revealing the cytotoxic activity of xanthones, the cytotoxic effects of compounds 7, 9 and 10 toward PC-12 cells were evaluated using an MTT assay. As summarized in Figure 7, the three compounds exhibited no cytotoxic activity on PC-12 cells at the concentration of 0.97-15.63 μM. It's worth noting that compounds 7, 9 and 10 showed more than 50% bacterial biofilm reduction at the concentration of 15.63 μM but were still safe for PC-12 cells. They showed no obvious damage to PC-12 cells even at the concentrations of MIC and MBC.

Conclusions
As part of our ongoing investigation on natural QSIs, two new xanthonolignoids (1 and 2) each existing as a racemic mixture, one new simple oxygenated xanthone (7) and eight known analogs (3-6, 8-11) were obtained from H. scabrum. Chiral separation of 1 yielded a pair of enantiomers 1a and 1b. The structures of these compounds were elucidated by spectroscopic analysis and ECD calculations. The evaluation for the QS inhibitory activity against C. violaceum indicated that both 9 and 10 exhibited the most potent QS inhibitory effects with the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) values of 31.25 and 62.5 μM, respectively. The antibacterial activity of simple xanthones was stronger than that of xanthonolignoids. The presence of the phenylpropane moiety may decrease its antibacterial activity, and 6-OH in simple xanthones is essential for the maintenance of antimicrobial effect, which can be deduced from the analysis of structure-activity relationship. Crystal violet staining was employed to determine the biofilm inhibition potential of compounds 7, 9 and 10. The results indicated that the formation of biofilms increased with decreasing drug concentration in a classic dose-dependent manner. The biofilm inhibition rates of these three compounds (15.63 μM) were 55.17%, 86.23% and 74.25%, respectively. The cytotoxic assay results suggested that the three compounds exhibited no cytotoxic activity on PC-12 cells at the concentration of 0.97-15.63 μM, MIC and MBC, which suggested that com- QS is a cell-density-dependent communication process to measure population density and trigger appropriate responses and conduct behavioral regulation [26]. QS inhibitors inhibit the QS system by reducing its virulence factor production and biofilm formation without affecting bacterial growth, thus making it difficult to cause drug resistance [27]. The above results signified that bacterial killing was not responsible for the reduction of biofilm formation, and compounds 7, 9 and 10 acted as quorum sensing inhibitors rather than antimicrobial agents. These compounds inhibit QS of C. violaceum at sub-inhibitory concentrations representing their unique mechanism for anti-quorum sensing other than growth inhibition or cell death, which was congruent with previous findings [9,28]. Purple pigment and biofilm inhibition potential demonstrated that they act by disturbing the bacterial communication system and attenuating microbial pathogenicity without killing the pathogens. Owing to the insufficient amounts, further anti-quorum sensing mechanism of these compounds was not investigated.

Conclusions
As part of our ongoing investigation on natural QSIs, two new xanthonolignoids (1 and 2) each existing as a racemic mixture, one new simple oxygenated xanthone (7) and eight known analogs (3-6, 8-11) were obtained from H. scabrum. Chiral separation of 1 yielded a pair of enantiomers 1a and 1b. The structures of these compounds were elucidated by spectroscopic analysis and ECD calculations. The evaluation for the QS inhibitory activity against C. violaceum indicated that both 9 and 10 exhibited the most potent QS inhibitory effects with the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) values of 31.25 and 62.5 µM, respectively. The antibacterial activity of simple xanthones was stronger than that of xanthonolignoids. The presence of the phenylpropane moiety may decrease its antibacterial activity, and 6-OH in simple xanthones is essential for the maintenance of antimicrobial effect, which can be deduced from the analysis of structure-activity relationship. Crystal violet staining was employed to determine the biofilm inhibition potential of compounds 7, 9 and 10. The results indicated that the formation of biofilms increased with decreasing drug concentration in a classic dose-dependent manner. The biofilm inhibition rates of these three compounds (15.63 µM) were 55.17%, 86.23% and 74.25%, respectively. The cytotoxic assay results suggested that the three compounds exhibited no cytotoxic activity on PC-12 cells at the concentration of 0.97-15.63 µM, MIC and MBC, which suggested that compounds 7, 9 and 10 acted as quorum sensing inhibitors rather than antimicrobial agents. They act by inhibiting the formation of purple pigment and biofilm so as to disturb bacterial communication systems and attenuate microbial pathogenicity without killing the pathogens. These findings highlight the importance of the genus Hepricum as a source of QS inhibitory compounds.