Curry Leaf Triggers Cell Death of P. gingivalis with Membrane Blebbing

Periodontal disease has become a serious public health problem, as indicated by accumulating evidence that periodontal disease is not only a major cause of tooth loss but is also associated with various systemic diseases. The present study assessed the anti-bacterial activities of three herbal products (curry leaf, clove, and cinnamon) against Porphyomonas gingivalis, a keystone pathogen for periodontal diseases. The curry leaf extract (CLE) showed the strongest growth inhibitory activity among them, and the activity was maintained even after extensive heat treatment. Of note, while clove and cinnamon extracts at sub-minimum inhibitory concentrations (sub-MICs) significantly enhanced the biofilm formation of P. gingivalis, CLE at sub-MIC did not have any effect on the biofilm formation. The MIC of CLE against P. gingivalis was higher than those against a wide range of other oral bacterial species. P. gingivalis cells were completely killed within 30 min after treatment with CLE. Spatiotemporal analysis using high-speed atomic force microscopy revealed that CLE immediately triggered aberrant membrane vesicle formation on the bacterial surface. Bacterial membrane potential assay revealed that CLE induced depolarization of the bacterial membrane. Taken together, these findings suggest the mechanism behind early bactericidal activity of CLE and its therapeutic applicability in patients with periodontal diseases.


Introduction
Periodontal disease is not only a major cause of tooth loss but is also associated with a range of systemic diseases [1]. Porphyromonas gingivalis is a major periodontopathic bacterium that possesses a diverse repertoire of virulence factors [2] and causes disruption of host-microbial homeostasis, leading to inflammatory bone loss [3,4]. P. gingivalis has thus been regarded as a primary target for prevention and treatment of periodontal diseases. However, sufficient progress has not been made to develop novel drugs for periodontal diseases, such as P. gingivalis-selective antimicrobials, for several decades. In addition, antimicrobial resistance (AMR) is an increasingly serious threat to global public health. In response to the serious situation, the world health organization (WHO) described the importance of natural sources of biodiversity and biorepositories as sources for the development of new antibiotics to combat AMR [5].
Bioactive natural product-based medicine for the prevention or cure of periodontitis has garnered much attention from both the public and scientific communities. In addition, the therapy using natural products is generally less expensive as compared with highly pure chemicals. Moreover, it is likely that a range of natural product-based therapies could become an alternative treatment option for periodontitis [6,7]. Curry leaf (Murraya koenigii) 2 of 12 is an herbal product frequently used especially in southern India for flavoring curries and chutneys. It has been reported that curry leaf has antibacterial activity against a range of pathogens, such as mycobacteria [8], oral streptococci, Pseudomonas aeruginosa, Candida albicans [9], and Plasmodium [10]. However, the anti-bacterial effect of curry leaf on P. gingivalis has not been investigated.
In the present study, we aimed to better understand the anti-bacterial activity of curry leaf extract (CLE) by focusing on electrophysiological/spatiotemporal features of the bacterial membrane. We found that CLE rapidly triggers cell death of P. gingivalis with membrane blebbing. CLE also showed a strong ionophoric activity against bacterial membrane. Based on these findings, we discuss the clinical applicability of potential new CLE-based treatments against periodontitis.

Curry Leaf Extract (CLE) Shows Anti-Bacterial Activity Together with Thermo-Stability
For screening of an herbal product that shows a potential inhibitory activity against P. gingivalis, we chose a high-throughput assay system using 96-well plates, by which both growth and biofilm formation can be assessed at the same time. Among 12 different ethanol-extracted herbs tested, we found that curry leaf, clove, and cinnamon showed P. gingivalis growth inhibitory activity. The MICs of extracts of curry leaf (CLE), clove, and cinnamon was determined as 16, 64, and 16 µg/mL, respectively (Figure 1), and each extract inhibited the growth in a dose-dependent manner (data not shown). Both clove and cinnamon extracts treated at 100 • C for 15 min partially lost their growth inhibitory activities; however, CLE after heat treatment completely inhibit P. gingivalis growth without any loss of activity (Figure 1), indicating that CLE contains thermostable antimicrobials. In the biofilm formation assay, CLE at a concentration of 16 µg/mL significantly inhibited biofilm formation ( Figure 1). Notably, biofilm formation significantly increased by the addition of extracts of clove or cinnamon at the sub-minimum inhibitory concentrations (sub-MICs). On the basis of these screening results, we decided to focus on the anti-bacterial effect of CLE in the following experiments. As a part of the authentication of CLE used in this study, the major components have been confirmed by HPLC analysis (Supplemental Figure S1) in reference to a published work characterizing curry leaf compounds [11].

P. gingivalis Is Highly Susceptible to Treatment with CLE
Next, we determined the MICs of three strains of P. gingivalis, as well as another 15 oral bacterial strains including two strains of Fusobacterium nucleatum, two Prevotella species, two strains of Aggregatibacter actinomycetemcomitans, and nine different species of oral streptococci. The MICs of CLE against these species are listed in Table 1. The MICs of P. gingivalis were 16 or 32 µg/mL, which were the lowest concentrations among the tested bacteria. Although A. actinomycetmcomitans ATCC 29522, S. gordonii ATCC 10558, S. oralis No. 10, and S. sobrinus ATCC 6715 were more resistant to CLE (64 µg/mL) than P. gingivalis, all tested 18 oral strains except the three species were even more resistant to CLE (more than 64 µg/mL) than P. gingivalis, demonstrating selective anti-bacterial activity of CLE against P. gingivalis. We also examined whether CLE showed toxicity on human oral epithelial cells. Treatment of the cells with CLE at the MIC against P. gingivalis (16 or 32 µg/mL) induced neither morphological changes nor LDH release even after 6 h (data not shown). Together, CLE at a concentration of 16 or 32 µg/mL shows selective anti-P. gingivalis activity without any in vitro toxicity to human cells. Shown are growth curves of P. gingivalis cultured in the presence of three herbal extracts at the MICs. The MICs of curry leaf, clove, and cinnamon were 16, 64, and 16 µg/mL, respectively (data not shown). The turbidity (OD600) of the bacterial culture was monitored for two days at different time points. The thermo-stability of each herbal product at MICs was also assessed (denoted as points/lines of red). Data shown are representative of three independent experiments performed in triplicate assays. Similar results were obtained in three independent experiments. (Right panels) The effects of herbal extracts on biofilm formation of P. gingivalis were examined in 96-well microplate assay. The Y-axes indicate the biofilm mass after 48-h culture (OD492). Data shown are the mean with SD from three independent experiments performed in triplicate assays. White bars indicate the vehicle control without herbal products. Black bars indicate the herbal products at different concentrations. NS: no statistical significance. ** p ≤ 0.01, **** p ≤ 0.0001 (vs. vehicle control without herbal product).

P. gingivalis Is Highly Susceptible to Treatment with CLE
Next, we determined the MICs of three strains of P. gingivalis, as well as another 15 oral bacterial strains including two strains of Fusobacterium nucleatum, two Prevotella species, two strains of Aggregatibacter actinomycetemcomitans, and nine different species of oral streptococci. The MICs of CLE against these species are listed in Table 1. The MICs of P. gingivalis were 16 or 32 µg/mL, which were the lowest concentrations among the tested bacteria. Although A. actinomycetmcomitans ATCC 29522, S. gordonii ATCC 10558, S. oralis No. 10, and S. sobrinus ATCC 6715 were more resistant to CLE (64 µg/mL) than P. gingivalis, all tested 18 oral strains except the three species were even more resistant to CLE (more than 64 µg/mL) than P. gingivalis, demonstrating selective anti-bacterial activity of CLE against P. gingivalis. We also examined whether CLE showed toxicity on human oral epithelial cells. Treatment of the cells with CLE at the MIC against P. gingivalis (16 or 32 µg/mL) induced neither morphological changes nor LDH release even after 6 h (data not shown). Together, CLE at a concentration of 16 or 32 µg/mL shows selective anti-P. gingivalis activity without any in vitro toxicity to human cells. Shown are growth curves of P. gingivalis cultured in the presence of three herbal extracts at the MICs. The MICs of curry leaf, clove, and cinnamon were 16, 64, and 16 µg/mL, respectively (data not shown). The turbidity (OD 600 ) of the bacterial culture was monitored for two days at different time points. The thermo-stability of each herbal product at MICs was also assessed (denoted as points/lines of red). Data shown are representative of three independent experiments performed in triplicate assays. Similar results were obtained in three independent experiments. (Right panels) The effects of herbal extracts on biofilm formation of P. gingivalis were examined in 96-well microplate assay. The Y-axes indicate the biofilm mass after 48-h culture (OD 492 ). Data shown are the mean with SD from three independent experiments performed in triplicate assays. White bars indicate the vehicle control without herbal products. Black bars indicate the herbal products at different concentrations. NS: no statistical significance. ** p ≤ 0.01, **** p ≤ 0.0001 (vs. vehicle control without herbal product).

CLE Showed Strong and Rapid Bactericidal Activity
To better understand the anti-bacterial activity of CLE, we performed a killing assay by counting CFUs of P. gingivalis after treatment with CLE at the MIC ( Figure 2). Two standard antibiotics, ampicillin and tetracycline, that show bactericidal and bacteriostatic effects, respectively, were also used as controls ( Figure 2). Treatment with tetracycline for 30 or 120 min did not significantly reduce the CFUs as compared to the vehicle control ( Figure 2). Treatment with ampicillin for 30 or 120 min dramatically reduced the CFUs to approximately 10% or 5% of that of the vehicle control, respectively ( Figure 2). Surprisingly, complete killing was achieved within 30 min after adding CLE ( Figure 2). This killing activity occurred in a dose-dependent manner ( Figure 2). These results demonstrated that CLE showed strong bactericidal effects on P. gingivalis, and the bactericidal action of CLE occurred more rapidly than that of ampicillin, a standard β-lactam with bactericidal activity.

CLE Showed Strong and Rapid Bactericidal Activity
To better understand the anti-bacterial activity of CLE, we performed a killing assay by counting CFUs of P. gingivalis after treatment with CLE at the MIC ( Figure 2). Two standard antibiotics, ampicillin and tetracycline, that show bactericidal and bacteriostatic effects, respectively, were also used as controls ( Figure 2). Treatment with tetracycline for 30 or 120 min did not significantly reduce the CFUs as compared to the vehicle control ( Figure 2). Treatment with ampicillin for 30 or 120 min dramatically reduced the CFUs to approximately 10% or 5% of that of the vehicle control, respectively ( Figure 2). Surprisingly, complete killing was achieved within 30 min after adding CLE ( Figure 2). This killing activity occurred in a dose-dependent manner ( Figure 2). These results demonstrated that CLE showed strong bactericidal effects on P. gingivalis, and the bactericidal action of CLE occurred more rapidly than that of ampicillin, a standard β-lactam with bactericidal activity.

CLE Attacks Bacterial Outer Membrane
SEM analysis revealed that treatment with CLE dramatically altered the appearance of outer membrane of P. gingivalis (Figure 3). The whole cell surfaces were covered by numerous, aberrant blebs, and the blebs were also released into the extracellular milieu ( Figure 3). We also examined the nanometer-scale dynamics of CLE-treated P. gingivalis cells using a high-speed atomic force microscopy (HS-AFM) system as an alternative to nanometerscale imaging. Spatiotemporal HS-AFM analysis showed that CLE triggered numerous membrane blebbing at least at 7 min and the number and size of the blebs continuously increased over time for 28 min (Figure 3, and Supplemental Movies S1 and S2). of outer membrane of P. gingivalis (Figure 3). The whole cell surfaces were covered by numerous, aberrant blebs, and the blebs were also released into the extracellular milieu ( Figure 3). We also examined the nanometer-scale dynamics of CLE-treated P. gingivalis cells using a high-speed atomic force microscopy (HS-AFM) system as an alternative to nanometer-scale imaging. Spatiotemporal HS-AFM analysis showed that CLE triggered numerous membrane blebbing at least at 7 min and the number and size of the blebs continuously increased over time for 28 min (Figure 3, and Supplemental Movies S1 and S2). Nanometer-scale dynamics were continuously monitored. Shown are images of the cells before (Pre) and 7, 16, and 28 min after treatment with curry leaf (t = 7, 16, and 28). See also Supplemental Movies S1 and S2.
To further characterize the influence of CLE on P. gingivalis, we examined how CLE affects the membrane physiology. Using P. gingivalis strain ATCC 33277 and a membrane impermeable dye TO-PRO-3, we firstly monitored membrane permeability to CLE ( Figure  4A). Surprisingly, sublethal and lethal doses of CLE significantly decreased the membrane permeability ( Figure 4A). We also examined the membrane potential of P. gingivalis using DiOC2(3)-based assay [12]. We found that the fluorescence intensity of this membrane potential-sensitive dye, significantly, was dramatically decreased by treatment of CLE in a dose-dependent manner ( Figure 4A), assuming that CLE affected the membrane potential. However, we could not assert that CLE affected the membrane potential of P. gingivalis  To further characterize the influence of CLE on P. gingivalis, we examined how CLE affects the membrane physiology. Using P. gingivalis strain ATCC 33277 and a membrane impermeable dye TO-PRO-3, we firstly monitored membrane permeability to CLE ( Figure 4A). Surprisingly, sublethal and lethal doses of CLE significantly decreased the membrane permeability ( Figure 4A). We also examined the membrane potential of P. gingivalis using DiOC 2 (3)-based assay [12]. We found that the fluorescence intensity of this membrane potential-sensitive dye, significantly, was dramatically decreased by treatment of CLE in a dose-dependent manner ( Figure 4A), assuming that CLE affected the membrane potential. However, we could not assert that CLE affected the membrane potential of P. gingivalis because no change in the signal intensity of DiOC 2 (3) was observed, even when P. gingivalis cells were treated with an ionophoric compound, CCCP ( Figure 4A). Thus, the DiOC 2 (3)-based membrane potential assay with DiOC2(3) was not functional with P. gingivalis. Nevertheless, these findings suggest that the killing activity of CLE against P. gingivalis may not be due to the loss of membrane integrity but rather to the abnormality of the bacterial membrane accompanied with disorder of the membrane permeability and membrane potential. To determine whether CLE certainly has ionophoric activity, we performed the same assays using E. coli strain BW25113 ( Figure 4B), which was highly resistant to CLE (MIC > 64 µg/mL, Table 1). The E. coli's natural resistant property to CLE could offer a methodological advantage because bacterial membrane response after treatment with CLE might be investigated in a broad concentration range without inducing cell death. The results of the assays revealed that CLE induced membrane depolarization in a dose-dependent manner without increasing membrane permeability ( Figure 4B). Notably, CLE at 10.24 µg/mL induced membrane depolarization nearly as strongly as a representative ionophore, CCCP. On the other hand, the ionophoric activity in ampicillin was not observed ( Figure 4B).

Discussion
P. gingivalis is regarded as a late colonizer of biofilms in subgingival pockets, as well as a pathobiont that leads to dysbiosis in oral cavity [13]. In the present study, we found that biofilm formation of P. gingivalis was strongly enhanced by sub-MICs of ethanol extracts of clove and cinnamon (Figure 1). Some antibiotics or disinfectants at sub-MIC have been reported to induce biofilm formation of some bacterial pathogens, such as Staphylococcus aureus [14], Pseudomonas aeruginosa [15], Staphylococcus epidermidis [16], Staphylococcus saprophyticus, and uropathogenic Escherichia coli [17]. To the best of our knowledge, this is the first report of P. gingivalis biofilm formation enhanced by low doses of natural products with anti-bacterial activities. Therefore, we would like to call attention to the possibility that biofilm formation is enhanced by low doses of anti-P. gingivalis products such as clove and cinnamon, such that the increased biofilm formation might deteriorate the subgingival environment and develop periodontal diseases.
The basis for every living organism depends on the ability of the cell to maintain ion gradients across biological membranes, i.e., membrane potential. Ionophoric drugs are attractive in terms of its application to a wide range of pathogens. In general, ionophoric antibiotics induce cell death through creating an imbalance in the cytoplasmic membrane's ion leak-pump relationship. For example, valinomycin is a representative ionophoric antibiotic with K + -selective ionophoric activity, resulting in an increase in K + permeability of the membrane. Gramicidin, a representative peptide ionophore, forms a continuous channel that spans the membrane. Very recently, Hards et al. reported a novel ionophoric antibiotic candidate that can inhibit ATP synthesis in E. coli by functioning as a H + /K + ionophore, causing transmembrane pH and potassium gradients to be equilibrated [18]. In terms of the difference in sensitivity of between Gram-positive and Gram-negative bacteria to ionophores, Tempelaars et al. reported that both cereulide and valinomycin, two antibiotics that share highly similar cyclic dodecadepsipeptides with K + -selective ionophoric activity, showed anti-bacterial activity against Gram-positive bacteria, but not against Gram-negative bacteria [19]. The difference in sensitivity between Gram-negative and Gram-positive bacteria may be due to the inability of these ionophores to cross the outer membrane. On the other hand, in the present study, FACS analysis using an E. coli membrane model revealed that CLE showed striking ionophoric activity against the cytoplasmic membrane of even Gram-negative bacteria. The activity of CLE at 10-fold MIC (against P. gingivalis) was nearly as strong as that of CCCP at the final concentrations of 10 µM, which is known to induce cytoplasmic membrane depolarization of Gram-positive bacteria as well as Gram-negative bacteria (Figure 4). In addition, notably, P. gingivalis, a Gram-negative anaerobe, was the most sensitive species among the oral bacterial species examined in this study, while Gram-positive oral commensals were relatively insensitive. These findings suggest the applicability of CLE as a narrow-spectrum therapeutic to treatment of periodontitis, if CLE would selectively eliminate P. gingivalis in the periodontal pockets while maintaining the homeostatic benefit provided by oral commensals.
In a previous report regarding chemical composition analysis, curry leaf extract contains plentiful and varied alkaloids and flavonoids [11], which is in agreement with the present study showing the presence of a range of alkaloids in CLE by HPLC analysis with the leading compound mahanine (Supplemental Figure S1). Mahanine, a novel carbazole alkaloid derived from curry leaf, inhibits the growth of prostate cancer cells via blocking androgen receptor signaling [20] and also induces cell death in pancreatic adenocarcinoma cells by inducing reactive oxygen species production [21], thereby implicating a therapeutic role for mahanine in cancer treatment. On the other hand, no mechanistic insight into the antibacterial effect of curry leaf in infectious disease research has been shown. Regarding the antimicrobial compounds, further biochemical and functional investigations should proceed on the basis of the chromatographical analysis of CLE (Supplemental Figure S1). Nevertheless, we suggest that curry leaf contains compound(s) with strong ionophoric activity that are responsible for its anti-bacterial activity to both Gram-positive and Gramnegative bacteria.
It has been also reported that curry leaf showed anti-inflammatory effect in in vitro [22] and clinical studies [23]. Adebajo et al. reported that Methanol-extracted curry leaf shows antioxidant and anti-inflammatory properties and proposed applicability for the treatment of gastro-intestinal inflammation, bronchitis, and hepatitis [24]. Therefore, CLE may also be useful for prevention of periodontal diseases, in an alternative view of attenuation of gingival inflammation. Further studies to clarify the possible therapeutic effect on inflamed gingival tissues in patients are needed.
Antimicrobial resistance (AMR) of bacterial pathogen is a serious risk to global health. The WHO's "global action plan on AMR" [5] was followed by worldwide support from governments, health ministries, and health agencies. However, despite such extensive efforts, researchers are still in the process of establishing scientific and rational strategies to combat globally emerging and re-emerging infectious diseases. In particular, regarding development of antibiotics, no new class of antibiotics with activity against Gram-negative bacteria has been approved in over fifty years. In the context of natural products that combat pathogens, we have provided here a feasible strategy to selectively kill P. gingivalis. Future studies to identify the responsible compound in curry leaf and to better understand the mechanisms behind its activity would further propose the therapeutical applicability of curry leaf for treatment of other infectious diseases.

Herbs and Preparation of Ethanol-Extracted Herb
Dried whole curry leaf, clove powder, and cinnamon powder were purchased from S&B Food Inc. (Tokyo, Japan), which is approved by international organization for standardization (ISO) 9001. The dried whole curry leaves were ground into a fine powder by using a hand mill CM-50GT (Kyocera Corp., Kyoto, Japan). Curry leaf (Murraya koenigii) used in this study was collected in the north western province of Sri Lanka, situated between 7 •  To prepare each herbal extract, 100 mg of the powder was stirred in 2 mL of 100% ethanol for 15 min at room temperature. After stirring, the extract in the supernatant after centrifugation was collected, and the concentration of each extract was determined as its protein content based on Bradford assay [25]. The extract was concentrated after evaporation by centrifugation under vacuum and was eventually redissolved with ethanol to standardize at 6.4 mg/mL. Thus, ethanol was used as solvent for all herbal extracts in our experiments. Regarding curry leaf extract used in this study, the chromatographic analysis has been also performed (Supplemental Figure S1).

Bacterial Strains and Growth Conditions
All strains used in this study are shown in Table 1. All Porphyromonas and Prevotella strains were grown in brain heart infusion (BHI) broth (Beckton Dickinson Co., Franklin Lakes, NJ, USA) supplemented with hemin (5 µg/mL) (Fujifilm Wako Chemicals Co., Tokyo, Japan) and menadione (1 µg/mL) (Fujifilm Wako Pure Chemicals, Osaka, Japan), or on BHI blood agar plates (BAP) containing hemin and menadione. Porphyromonas gingivalis strain ATCC 33277 was mainly used for this study to screen the herbal products for anti-bacterial activity as well as to further examine the activity in detail. BHI broth and BHI BAP (without hemin and menadione) were used to maintain strains of the other oral bacteria including streptococci and Aggregatibacter actinomycetmcomitans, Fusobacterium nucleatum. All oral bacteria were grown in an anaerobic chamber (miniMACS anaerobic workstation; Don Whitley Scientific Ltd., Shipley, UK) in 80% N 2 , 10% H 2 , and 10% CO 2 at 37 • C. A laboratory Escherichia coli strain BW25113, which was used for bacterial membrane potential assay, is maintained in LB (Becton Dickinson) broth and on LB agar under aerobic conditions.

Growth Assay and Biofilm Formation Assay
Strain ATCC 33277 was used for the growth and biofilm formation assay of P. gingivalis. Growth in the presence of the herbal extracts at different concentrations was monitored at different time points, using 96-well plates (3595; Corning, New York, NY, USA). Two-fold dilution series of each herbal extracts of curry leaf, clove, and cinnamon (2.5 µL) were added to the bacterial suspension (200 µL, 1 × 10 8 CFU/mL) in BHI-HM broth. The turbidity of the bacterial suspension (OD 620 ) during growth was measured at different time points using a micro-plate reader (Cytation 5, BioTek, Winooski, VT, USA). Bacterial strains listed in Table 1 were assayed for determination of MICs. The MIC for each strain was determined using microdilution methods according to the protocols of the Clinical and Laboratory Standards Institute [26,27], with some modifications. MIC was defined as the minimum concentration that restricted growth to a level of A 600 < 0.05 at after incubation for 48 h. The broth contained a 2-fold dilution series of extract of each herbal product at final concentrations ranging from 0.125 to 64 µg/mL. After monitoring the growth for two days, biofilm formation was also assessed using the same 96-well plates, as described previously [28]. In brief, 2-day-old biofilms grown in the 96-well plates under static conditions were stained with 0.25% safranin. All dye associated with the attached biofilms was dissolved with 80% ethanol, and then the optical density at 492 nm (OD 492 ) was measured using a micro-plate reader (Cytation 5, BioTek, Tokyo, Japan).

Killing Assay to Assess Bactericidal Activity
The bactericidal activity of the curry leaf extract against P. gingivalis was evaluated at 30 and 120 min using a killing assay, as described previously in our laboratory [29].

Scanning Electron Microscopy (SEM)
SEM analysis of P. gingivalis cells treated with the curry leaf extract for 30 min was performed using an S-5200, (HITACHI, Hitachi, Japan), as described previously in our laboratory [30].

Cytotoxicity toward Oral Epithelial Cells
Cytotoxicity of the samples toward the oral squamous cell carcinoma cell line Ca9.22 was evaluated by quantifying release of lactate dehydrogenase (LDH) with plasma membrane damage (Cytotoxicity Detection Kit Plus (LDH), Roche Diagnostics GmbH, Mannheim, Germany), according to the manufacturer's instructions with some modifications, as described previously in our laboratory [29].