Influence of Gallic Acid and Thai Culinary Essential Oils on Antibacterial Activity of Nisin against Streptococcus mutans

Streptococcus mutans is a well-known oral pathogen commonly associated with a normal dental problem and life-threatening infection. A bacteriocin nisin and the plant-derived compounds including gallic acid (GA) and Thai culinary essential oils (EOs) have been reported to have activity against oral pathogens. However, their synergistic interaction against S. mutans has not been explored. The purposes of this study were primarily to investigate anti-S. mutans properties and the antibiofilm formation of nisin, GA, and five EOs by using the broth microdilution method. Besides, the morphological change, killing rate, and antibacterial synergism were determined by scanning electron microscopy (SEM), time-kill assay, and checkerboard method, respectively. The results demonstrated that kaffir lime leaf (KLL) oil, lemongrass (LG) oil, and GA showed a potent anti-S. mutans activity and inhibited biofilm formation with the possible mechanism targeted on the cell membrane. Additionally, KLL oil revealed anti-S. mutans synergism with GA, LG oil, and chlorhexidine with the fractional inhibitory concentration (FIC) indexes ≤ 0.5. Interestingly, GA displayed a high potential to enhance anti-S. mutans activity of nisin by lowering the minimum inhibitory concentrations (MICs) to at least 8-fold in a bacteriostatic manner. These results suggest that GA and KLL oil may be potentially used as an adjunctive therapy along with nisin and chlorhexidine to control S. mutans infection.


Introduction
A Gram-positive streptococcal bacterium, S. mutans, is an important oral pathogen that can cause common dental caries in humans and life-threatening infectious diseases such as infective endocarditis after entering the blood circulation [1]. To prevent the dental problems and the resulting complication, dental hygiene is necessary throughout the human lifespan [2,3]. e virulence determinant produced by S. mutans is the formation of cariogenic biofilms or dental plaques, which protect the sessile bacteria from antibacterial compounds. erefore, potential strategies to combat S. mutans are the inhibition of biofilm development and eradication planktonic cells [4]. Although mechanical cleaning is an effective approach to remove the cariogenic species and biofilms, the chemical antibacterial agents have been beneficially added to many oral healthcare products for similar purposes [5]. However, long-term use of chemical agents can cause teeth discoloration and disturbance of physiological microbiota [6,7]. Moreover, the clinically essential antiseptic or antibiotic agents are not worth for treatment of common oral diseases due to an increase in multidrug-resistant bacteria, which is currently one of global health problems [8]. erefore, natural-derived compounds including polypeptide bacteriocins, phenolic compounds, and herbal EOs became the preference of bacterial controlling agents because of their safety perception and historical usage.
In recent years, a polypeptide bacteriocin, namely, nisin, is attractive as a new generation antibiotic [9]. Nisin produced by safe lactic acid bacteria species Lactococcus lactis effectively inhibits food-borne pathogens in both vegetatively growing cells and spores of Clostridium botulinum and Bacillus cereus [10,11]. Additionally, this bacteriocin does not change the organoleptic properties of foods [12]. Nisin is therefore approved by the food and drug administration in over 48 countries and becomes widely used as a natural preservative in the various food industries [10]. Despite the fact that nisin is not considered as a common anticaries agent, reports of utilizing nisin to inhibit oral pathogens are still limited [11]. For examples, nisin has demonstrated a potential effect against various cariogenic bacteria including mutans and non-mutans streptococci, Lactobacillus sp., and Actinomyces sp. [13][14][15]. Sangcharoen et al. [16] reported that the antibacterial activity of nisin could be enhanced by weak organic acids such as ascorbic acid and citric acid.
GA is a phenolic acid commonly found in several plants [17,18]. It possesses varieties of health-promoting benefits such as antioxidant, anti-inflammatory, and anticancer, as well as antimicrobial effects [17,18]. Previous reports showed that GA had strong antimicrobial activities with potentially replacing the synthetic antimicrobial agents for food and biomedical products. Besides, it was documented to have antibacterial potency against S. mutans [19,20]. erefore, GA becomes an interesting agent used to investigate the enhancement activity of nisin in this study. Apart from using a weak organic acid, the antibacterial efficacy of nisin could be intensified by the addition of plant EOs [21], which are recognized as antimicrobial agents in traditional medicine [22]. e well-known ai culinary essential oils from finger root (Boesenbergia pandurate (Roxb.) Schltr.), kaffir lime (Citrus hystrix DC.), holy basil (Ocimum tenuiflorum L.), and lemongrass (Cymbopogon citratus (DC.) Stapf.) have been reported to contain antimicrobial properties in particular against medically important microorganisms [23][24][25][26]. However, research regarding the antibacterial synergism among these EOs and nisin against S. mutans is scarce. erefore, the purpose of this study was primarily to determine the anti-S. mutans and antibiofilm activities of nisin, GA, and five ai culinary EOs. e killing rate and the bacterial cellular morphology were also conducted to investigate the antibacterial characteristics and the possible mechanism, respectively. Moreover, the synergistic interactions among these compounds against S. mutans were studies based on a checkerboard microdilution assay to establish the FIC index.  (1,030,000 IU/g), and saturated alkane standard (C7-C40) were purchased from Sigma-Aldrich (MO, USA). Other chemicals and solvents were of analytical grade and obtained from local distributors.

Analytical Conditions for EOs.
e constituents of five EOs were analyzed using a gas chromatography/mass spectrometry (GC/MS-6890n, Agilent, USA) equipped with HP-5 capillary column (30 m, 0.25 mm; J&W Scientific, Folsom, CA). Helium was used as carrier gas at a constant flow rate of 1 mL/min. e oven temperature was initially 50°C for 3 min and then was increased to 200°C at a rate of 10°C/min for 3 min. Finally, the oven temperature was increased to 260°C at a rate of 15°C/min for 20 min. e injector temperature was 250°C. e sample was injected using a split ratio of 1:100. e retention indexes (RI) of constituents were determined with reference to a saturated alkane (C7-C40). Additionally, identification of component was evaluated by computer matching the fragmentation pattern with Wiley 7N spectral library.

Bacterial Strain and Culture
Condition. S. mutans ATCC 25175 was purchased from the American Type Culture Collection.
ree clinical S. mutans TLJ1-1, TLJ1-2, and TLJ1-3, which lacked collagen-binding adhesin encoded by the cnm gene [27], were kindly supported by Assoc. Prof. Dr. Jinthana Lapirattanakul, Department of Oral Microbiology, Faculty of Dentistry, Mahidol University, Bangkok, ailand. is project was approved by the Ethics Committee of Mahidol University (MU-DT/PY-IRB2021/PY022). Bacterial strains were maintained in a mixture of brain heart infusion (BHI; Difco, USA) broth and 20% w/v glycerol at -80°C until use. For experiments, S. mutans was grown separately on BHI agar at 37°C for 48 h. e isolated bacterial colonies of actively growing cultures from agar plates were transferred to a test tube with BHI broth and incubated at 37°C for 24 h. e turbidity of inoculum was adjusted spectrophotometrically at 600 nm to obtain an optical density (OD) of 0.2 (approximately 10 6 -10 7 CFU/mL) before use in the experiments.

Antibacterial Susceptibility Test.
e MICs of the test compounds were determined by a broth microdilution method [28]. e test compounds were prepared by dissolving in the BHI medium with Tween 80 (0.04% v/v) and absolute ethanol (0.03% v/v). en, the serial twofold dilutions of test compounds were mixed with BHI broth at a 1: 1 ratio (v/v) in 96-well sterile microtiter plates to obtain final concentrations of 0.006-0.80% v/v for EOs, 0.06-8.00 mg/ mL for GA, 31.25-4,000 IU/mL for nisin, and 0.008-1.00 mg/mL for chlorhexidine. 20 μL of the prepared inoculum was added to BHI broth supplemented with the test compounds to obtain a 100 μL final volume in each well. e microtiter plates were then incubated at 37°C for 24 h under aerobic conditions. e negative and positive controls were set in each test. A negative control included the test sample but not the organism, and a positive control included the organism but not the test sample. Chlorhexidine was used as a reference control. e mixture of 0.04% v/v Tween 80 and 0.03% v/v absolute ethanol in BHI medium was also tested to control the effect of solvent. e MIC was defined as the lowest concentration at which no bacterial growth was determined by the unaided eye. e growth endpoint in the wells containing test samples was observed by comparing with the growth in the control wells. 2 Advances in Pharmacological and Pharmaceutical Sciences To establish the minimum bactericidal concentration (MBC), 20 μL of each culture medium was removed from wells with no visible growth and placed into 80 μL of sterile BHI broth in 96-well plates. After incubation at 37°C for 24 h, the MBC was determined as the lowest concentration that produced no bacterial growth observed by the unaided eye. Each sample was tested in triplicate in separate experiments.
e enhancing effects of EOs, GA, and chlorhexidine on the antibacterial activity of nisin were also evaluated by the broth microdilution method. e MIC and MBC values of nisin were determined in combination with 0.5×MIC of the test compounds (or 0.4% v/v of EOs when the MIC value of EOs was higher than 0.8% v/v).

Antibiofilm Formation Assay.
e effect of test compounds on biofilm formation was determined as described by Wongsariya et al. [29] with modification. e prepared test compounds were mixed in BHI broth supplemented with 1% w/v sucrose by using twofold dilutions method to obtain the final concentrations of 0.006-0.80% v/v for each EO, 0.06-8.00 mg/mL for GA, 31.25-4,000 IU/mL for nisin, and 0.008-1.00 mg/mL for chlorhexidine. 20 μL of the prepared inoculum was added to each well. After incubation at 37°C for 24 h, the medium was aspirated; the biofilm was then washed twice with 100 μL of sterile saline (0.9% w/v NaCl). e adherent biofilm was fixed with absolute ethanol (100 μL) for 15 min and stained with 0.1% w/v crystal violet for 15 min. After washing the samples with 200 μL of distilled water three times, the dye bound to the biofilm was solubilized by adding 100 μL dimethyl sulfoxide. e extracted dye was measured with a microtiter plate reader (Varioskan LUX, ermo Fisher Scientific) at the absorbance of 590 nm.
e experiments were carried out in triplicate and the percentages of biofilm inhibition were calculated using where OD (mean,control) was defined as the average absorbance of untreated cells, and OD (mean,treatment) was defined as the average absorbance of treated cells. e biofilm inhibition curves were constructed by plotting the percentage of inhibition against concentrations.

Time-Kill
Assay. e bactericidal activities of the test compounds were determined according to the time-kill assay of Koo et al. [30] with modification. e bacterial suspension (240 μL, approximately 10 6 CFU/mL) was added to BHI broth (960 μL) containing the test sample at 1, 2, and 4 × MICs. After incubation at 37°C, sample (20 μL) was collected at different time intervals (0, 2, 4, 8, 12, and 24 h) and a tenfold serial dilution was prepared in sterile saline.
ereafter, 20 μL of each dilution was placed on a BHI agar plate and incubated at 37°C for 24-48 h. A bacterial viability count was performed and recorded as the number of CFU/ mL. In each assay, a bacterial growth control was included and consisted of 0.04% v/v Tween 80 and 0.03% v/v absolute ethanol without the addition of test samples. Chlorhexidine was also used as the reference antiseptic agent. All experiments were carried out in triplicate and the experimental results were expressed as mean ± standard deviation (SD). Time-kill curves were established by plotting log 10 CFU/mL against time. Bactericidal activity was defined as a ≥3 log 10fold decrease in the number of survivors at each time point compared with the initial number inoculum.

SEM.
e prepared inoculum of S. mutans ATCC 25175 was incubated with the test compounds at concentration of 4 × MIC. Bacterial growth controls were performed with the addition of 0.04% v/v Tween 80 and 0.03% v/v absolute ethanol without the test samples and the bacteria treated with chlorhexidine at a concentration of 4 × MIC were used as a reference compound. After incubation at 37°C for 12 h, bacterial cells were collected by centrifugation at 3,000 rpm for 10 min. en, samples were fixed in 2.5% w/v of glutaraldehyde in 0.1 M phosphate buffer solution (pH 7.2) for overnight and post-fixed in 1% w/v osmium tetroxide in 0.1 M phosphate buffer solution for 1-2 h. e cells were passed through a filter disc (pore size 1.2 micron) and dehydrated using serial concentrations of ethanol (30,50,70,95, and 100% v/v). After critical point drying and coating with a gold sputter, samples were examined using a scanning electron microscope (JSM-IT500HR InTouchScope ™ , JEOL, Tokyo, Japan).

Checkerboard Microdilution Assay.
e antibacterial synergism among test compounds, which could be determined the MIC values, was further studied by checkerboard microdilution assay as previously described by Botelho [31] with modification. 20 μL of the prepared inoculum of S. mutans was added to the mixed concentrations of two test compounds, which were in a range of 0.0625-4 × MIC. e experiments were performed in triplicate. e FIC index was defined as the lowest concentration of the combination of test compounds with no visible growth of the test organisms. FIC indexes for the double and triple combinations were calculated using formulas (2) and (3), respectively.
Advances in Pharmacological and Pharmaceutical Sciences 2.9. Statistical Analysis. In the time-kill assay, the statistical analyses were performed using SPSS (version 26.0, SPSS Inc., Chicago, IL, USA). An analysis of variance (ANOVA) was performed, and significant differences between means were determined using Tukey's honesty significant difference test or Dunnett's T3 test at a significance level of p < 0.05.

Chemical Constituents of EOs.
e top ten compositions of five EOs analyzed by GC/MS system are reported in Table 1. Terpenes and terpenoids were the major constituent found in FR (Δ-3-careen, 24.4%), KL (L-limonene, 25.1%), and KLL (citronella, 73.3%) oils, whereas lactone (c-dodecalactone, 33.1%) was mainly found in LG oils. A phenolic compound (3-allyl-6-methoxyphenol, 29.7%) was a major compound in HB oil. For the phenolic compound, GA showed similar MIC and MBC values of 4 mg/mL against both standard and clinical isolates. As shown in Table 2, nisin failed to inhibit the growth of S. mutans ATCC 25175 and two clinical isolated at a concentration lower than 4,000 IU/mL, but nisin was able to inhibit one clinical isolate with the MIC and MBC values of 2,000 and 8,000 IU /mL, respectively. is result was inconsistent with the previous report that nisin illustrated the antibacterial activity against S. mutans UA 159 strain with MIC value in the range of 625-1250 IU/mL [15]. It could be noted that the different antibacterial efficacy of natural compounds could depend on the type of chemical compounds, the mechanism of action, and the strain of the test microorganism [33]. erefore, the different bacterial strain could possibly explain this phenomenon.

Antibacterial Susceptibility.
For the reference agent, chlorhexidine demonstrated similar MIC and MBC values against all tested bacteria with a concentration lower than 0.0156 mg/mL. ese results indicated similar or higher efficacy of all test compounds against the standard strain than those against the clinical isolates. erefore, S. mutans ATCC 25175 was only used for further studies. e mixture of 0.04% v/v Tween 80 and 0.03% v/v absolute ethanol in BHI medium produced visible turbidity of bacterial growth similar to the positive control against all test strains. is result implied that Tween 80 and absolute ethanol at the tested concentrations had no inhibition effect on bacterial growth. In addition, this result was consistent with the negative control results of the time-kill assay and SEM study.

Antibiofilm Assay.
e antibiofilm formation properties of the test compounds are demonstrated in Figure 1. e results showed that LG oil had the highest antibiofilm activity with greater than 86.44% inhibition when treating at the concentration range of 0.1-0.8% v/v. In the case of KLL oil, the biofilm formation of S. mutans was inhibited in a concentration-dependent pattern and showed the maximum inhibition of >73.32% after treatment at 0.8% v/v. is finding was similar to the study of Wongsariya et al. [29] where the antibiofilm formation efficacy of KLL oil was greater than 90% inhibition after treatment at the MIC.
On the other hand, although HB, FR, and KL oils could not determine the MIC in the concentration range of 0.006-0.8% v/v, these oils could inhibit the biofilm formation less than 50% in the same concentration range (Figure 1(a)).
is result indicates that the biofilm suppression of EOs may not relate to their antibacterial activity. e composition of EOs was a complex mixture of compounds in different amounts; however, terpene represented the biggest composition along with other non-terpene compounds [34]. Similarly, the major terpene and terpenoid constituents of LG oils were neral dimethyl acetal, citral, and (Z)-citral about 51%, whereas KLL oil mainly consisted of citronella approximately 73% (Table 1). It has been documented that monoterpene-based oils were able to cause a loss of membrane integrity of biofilm cells; therefore the target sites for EOs seemed to be the cell membrane [35]. In addition, bacterial cells with damaged membranes often fail to attach and form biofilm structures [36]. erefore, KLL and LG oils containing terpenes as a major constituent may be diminished the biofilm formation by these mechanisms [37]. As shown in Figures 1(b) and 1(c), nisin at all test concentrations could not inhibit the formation of biofilm of S. mutans, whereas GA at the concentration of 8 mg/mL interfered the formation of biofilms with 88.98% inhibition. Besides, chlorhexidine at a concentration of ≥0.0156 mg/mL could prevent the biofilm formation greater than 85.75% (Figure 1(d)).

e Enhancing Effects of EOs, GA, and Chlorhexidine on the Antibacterial Activity of Nisin.
e enhancing activities of test compounds on the antibacterial activity of nisin are shown in Table 3. It was found that the addition of either KLL oil at 0.4% v/v or GA at 2 mg/mL led to a dramatic decrease in the MIC of nisin. e addition of KLL oil was able to reduce the MIC of nisin from >4,000 IU/mL to ⩽31.25 IU/mL (at least 125-fold), whereas the addition of GA showed the capacity to decrease the MIC of nisin to 500 IU/ mL (at least 8-fold). Moreover, GA at the concentration of 2 mg/mL demonstrated the capacity to reduce the MBC of nisin by at least 4-fold. It is known that nisin is stable in an acidic medium [38], and GA is a weak acid with pKa values of 4.0 (carboxylic acid). erefore, this synergistic antimicrobial activity against S. mutans could be explained by the decrease of the pH in solution resulting in the increasing    Advances in Pharmacological and Pharmaceutical Sciences solubility of nisin. In addition, it could be similar to ascorbic acid, which enhanced the antibacterial activity of nisin by binding on nisin molecule to stabilize its structure [16,39].

Advances in Pharmacological and Pharmaceutical Sciences
However, the addition of other EOs and chlorhexidine at sub-MIC values of 0.4% v/v and 0.0078 mg/mL, respectively, demonstrated no enhancing effects on the antibacterial activity of nisin. ese results imply that the intensifying ability of EOs and chlorhexidine on the antibacterial activity of nisin against S. mutans may not be associated only with their antibacterial activity. Although KLL oil showed a higher ability to reduce the MIC value of nisin greater than GA, only GA displayed the capacity to decrease the MBC value of nisin. erefore, the combination of nisin and GA was selected to further determine the killing rate and to observe the morphological change.

Time-Kill
Assay. e bacteriostatic and bactericidal effects of the test compounds against S. mutans are shown in Figure 2. e ability of KLL oil to kill S. mutans was in a time-dependent manner. Treatment with KLL oil at concentrations of 1, 2, and 4×MICs could reduce the number of survival bacteria greater than 3 log 10 CFU/mL within 12 h and induced complete bacterial cell death within 24 h (Figure 2(a)). e bactericidal activity of LG oil against S. mutans was in concentration-and time-dependent pattern. Treatment with LG oil at the concentrations of 2 and 4×MICs exerted the most bactericidal activity (≥3 log 10 -fold decreases) within 24 h, but LG at the concentration of MIC only suppressed the number of survival S. mutans constantly for 24 h (Figure 2(b)). Although GA showed the bacteriostatic effect at a concentration of 4×MIC by preventing bacterial growth for 24 h, it failed to inhibit the cell growth at either concentration of 1 or 2×MIC (Figure 2(c)). is result was in agreement with the study of Kang et al. [40] where GA was able to prevent the growth of periodontal bacteria in a bacteriostatic characteristic.
Because the samples at each time point were diluted with normal saline and transferred to a new BHI agar plate, the survival of adaptive isolates may occur from a dynamic manner of the time-kill assay [41,42]. erefore, the MIC and MBC values obtained from a static view of broth microdilution assay may not agree with the time-kill assay for a bacteriostatic agent [28]. In the case of chlorhexidine, a bactericidal effect with ≥3 log 10 -fold reductions was observed after treatment at 4×MIC for 24 h, whereas a bacteriostatic effect with <2 log 10 -fold reductions was detected after treatment at either 1 or 2×MIC (Figure 2(d)).
Despite the fact that the ability of plant-derived compounds to potentiate an antibacterial property of bacteriocin has been reported in particular against food-borne pathogens including Listeria monocytogenes and Salmonella sp. [43], the antibacterial synergism of GA and nisin against oral pathogen was studied using time-kill assay as shown in Figure 2(e). Treatment with the combination of GA and nisin at sub-MIC value was able to suppress the bacterial growth for 24 h, whereas treatment with either GA or nisin alone failed to inhibit bacterial growth. is result indicated that GA could intensify the antibacterial activity of nisin in a bacteriostatic manner.
3.6. SEM. SEM images of S. mutans ATCC 25175 are shown in Figure 3. e SEM images revealed that KLL and LG oils at a concentration of 4×MIC induced alteration in cell morphology. Control cells in the presence of 0.04% v/v Tween 80 and 0.03% v/v absolute ethanol showed an oval shape with a smooth cell surface (Figure 3(a)). In contrast, cells treated with either KLL oil or LG oil at 4×MIC displayed irregular oval shape with a concavity on the cell surface (Figures 3(b) and 3(c), arrows). In accordance with the previous study of Guimarães et al. [44], Escherichia coli treated with citronellol demonstrated irregular cell size with the rough surface. erefore, the anti-S. mutans activities of KLL and LG oils could be principally attributed by terpene constituents along with the contribution of other compounds.
Additionally, the treatment of GA at 4×MIC caused the unseparated spherical cells with coating materials (Figure 3(d), arrows). Apart from lowering environmental pH, GA could also act as a sequestering agent of divalent ions and consequently caused a disruptive effect on the cell membrane [45].
is explanation supports the morphological results of SEM. Similar results could be observed after treating S. mutans cells with Galla chinensis extracts, which are rich in GA content [46]. e bacterial cells also displayed irregular cell wall structure and showed fewer cells in the chain [46]. In addition, the antibacterial mechanism of GA was suggested to interfere biofilm composition and structure, inhibit glucosyltransferase activity, and directly suppress bacteria growth [18,47,48]. When S. mutans was treated with chlorhexidine, perforation on the cell surface was observed (Figure 3(e), arrows).
As shown in Figures 3(f ) and 3(g), cells treated with nisin (1,000 IU) or GA (0.5×MIC) alone showed similar shape and cell membrane to the control, which was consistent with the results from the time-kill assay. Although nisin was known to play an important role in pore-forming on bacterial membrane [49], the damaged cell membrane could not be observed after treatment with nisin at 1,000 IU. However, after treatment with the combination of nisin and GA, cells showed a small unseparated oval shape with the presence of cellular matrix Advances in Pharmacological and Pharmaceutical Sciences 7

Checkerboard Microdilution Assay.
e interaction among the test compounds was primarily studied using the checkerboard assay. In addition to lowering MIC of nisin, KLL oil together with LG oil, GA, and chlorhexidine demonstrated the synergistic effect against S. mutans with the FIC indexes lower than 0.37 as shown in Table 4. However, the additive interaction was found in the triple combination of KLL oil, LG oil, and chlorhexidine. is result was not surprising because LG combined with chlorhexidine showed an indifferent effect with the FIC index of 0.56. Although GA could decrease the MIC of nisin against S. mutans, the combination of GA with LG oil or chlorhexidine showed the indifferent effect with the FIC indexes greater than 2. However, no antagonistic effect was observed among these compounds. Because EOs are plantbased products containing a complex mixture of substances, the interaction within each constituent could lead to additive, synergistic, and antagonistic effects [22]. erefore, the mode of action associated with each constituent should be further investigated.

Conclusions
GA and two ai culinary EOs including KLL and LG oils exhibited potent inhibitory effects against S. mutans and biofilm formation. e damage to cell membranes resulting in an alteration in bacterial cell morphology was the possible mechanism of action. Despite a weak antibacterial activity of nisin against S. mutans, the addition of GA displayed markedly capacity to enhance the anti-S. mutans activity of nisin with the bacteriostatic character by 8-fold decrease in MIC values. Additionally, KLL oil combined with LG oil, GA, and chlorhexidine revealed the synergistic antibacterial interaction. erefore, these results may prove valuable information that GA, and KLL could be potentially utilized as an adjunctive therapy with nisin or chlorhexidine for overcoming S. mutans-associated infection.

Data Availability
e authors declare that all data supporting the findings in this study are provided in the results and discussion section within the article. e datasets used in the current study are available from the corresponding author on reasonable request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.