Identification of Antibacterial Components and Modes in the Methanol-Phase Extract from a Herbal Plant Potentilla kleiniana Wight et Arn

The increase in bacterial resistance and the decline in the effectiveness of antimicrobial agents are challenging issues for the control of infectious diseases. Traditional Chinese herbal plants are potential sources of new or alternative medicine. Here, we identified antimicrobial components and action modes of the methanol-phase extract from an edible herb Potentilla kleiniana Wight et Arn, which had a 68.18% inhibition rate against 22 species of common pathogenic bacteria. The extract was purified using preparative high-performance liquid chromatography (Prep-HPLC), and three separated fragments (Fragments 1–3) were obtained. Fragment 1 significantly elevated cell surface hydrophobicity and membrane permeability but reduced membrane fluidity, disrupting the cell integrity of the Gram-negative and Gram-positive pathogens tested (p < 0.05). Sixty-six compounds in Fragment 1 were identified using Ultra-HPLC and mass spectrometry (UHPLC-MS). The identified oxymorphone (6.29%) and rutin (6.29%) were predominant in Fragment 1. Multiple cellular metabolic pathways were altered by Fragment 1, such as the repressed ABC transporters, protein translation, and energy supply in two representative Gram-negative and Gram-positive strains (p < 0.05). Overall, this study demonstrates that Fragment 1 from P. kleiniana Wight et Arn is a promising candidate for antibacterial medicine and food preservatives.


Introduction
Infectious diseases caused by pathogenic bacteria continue to be a global concern for public health, causing millions of deaths worldwide per year [1]. Since the introduction of sulfonamides in 1933, a large number of antibiotics have been applied in clinics [2]. Nevertheless, in recent decades, the overuse and/or misuse of antibiotics have accelerated the spread of antibiotic-resistant bacteria, leading to ineffective drug treatment [3]. It was estimated that at least 700,000 people worldwide die each year due to antimicrobial resistance [4].
Pharmacophagous plants are recognized as a rich source of phytochemicals with antimicrobial potential [5]. Phytocompounds extracted from such plants are long known for their therapeutic uses, and characterized by safety and low toxicity [6]. The application of herbal products may be a better choice for the extensive and imprudent use of synthetic antibiotics [7]. For example, In China, approximately 34,984 native higher plant species have been recorded [8]. Of these, the herbal plant Potentilla kleiniana Wight et Arn was first recorded in the earliest pharmaceutical book "Divine Farmer's Classic of Materia Medica" during the Warring States period (475-221 B.C.) in China. It belongs to the phylum of Angiospermae, the class of Dicotyledoneae, the order of Rosales Bercht. and J. Presl, and the family of Rosaceae Juss. P. kleiniana Wight et Arn is widely distributed in China, and many Asian countries such as Japan, India, Malaysia, Indonesia, and North Korea.
In this study, the methanol and chloroform extract method exhibited a broader antibacterial spectrum, consistent with our previous reports [15,16]. Previous studies also reported effective extraction of bioactive compounds from P. kleiniana Wight et Arn. For example, Tao et al. [13] extracted TFP in P. kleiniana Wight et Arn using an ethanol-water solution, and the obtained extract was further partitioned using petroleum ethers, chloroform and ethyl acetate. The extracted TFP inhibited survival and virulence of P. aeruginosa, and MRSA. Song et al. [14] extracted bioactive compounds from P. kleiniana Wight et Arn using ethanol and ethyl acetate, and the obtained extract showed antibacterial activity against P. aeruginosa, S. aureus, C. albicans, and E. coli. The difference in bioactive compounds ex-Foods 2023, 12, 1640 3 of 23 tracted from P. kleiniana Wight et Arn using the different methods may explain the distinct antibacterial profiles between this study and the previous reports [13,14]. the growth of 15 bacterial species, including one species of Gram-positive S. aureus, and 14 species of Gram-negative bacteria, P. aeruginosa ATCC9027, S. typhimurium ATCC15611, S. dysenteriae CMCC51252, S. flexneri CMCC51572, S. flexneri CMCC51574, S. sonnei ATCC25931, V. alginolyticus ATCC17749, V. cholerae Q10-54, V. fluvialis ATCC33809, V. mimicus bio-56759, V. parahemolyticus ATCC17802, and V. vulnificus ATCC27562, which showed a 68.18% inhibition rate (Table 1, Figure 1). In this study, the methanol and chloroform extract method exhibited a broader antibacterial spectrum, consistent with our previous reports [15,16]. Previous studies also reported effective extraction of bioactive compounds from P. kleiniana Wight et Arn. For example, Tao et al. [13] extracted TFP in P. kleiniana Wight et Arn using an ethanol-water solution, and the obtained extract was further partitioned using petroleum ethers, chloroform and ethyl acetate. The extracted TFP inhibited survival and virulence of P. aeruginosa, and MRSA. Song et al. [14] extracted bioactive compounds from P. kleiniana Wight et Arn using ethanol and ethyl acetate, and the obtained extract showed antibacterial activity against P. aeruginosa, S. aureus, C. albicans, and E. coli. The difference in bioactive compounds extracted from P. kleiniana Wight et Arn using the different methods may explain the distinct antibacterial profiles between this study and the previous reports [13,14].
We further determined minimum inhibitory concentrations (MICs) of the crude extracts from P. kleiniana Wight et Arn, and the results are shown in Table 1. The MICs of the chloroform-phase extract ranged from 12.5 mg/mL to 50 mg/mL against the eleven species of the bacteria. Notably, for the methanol-phase extract, the MICs were between We further determined minimum inhibitory concentrations (MICs) of the crude extracts from P. kleiniana Wight et Arn, and the results are shown in Table 1. The MICs of the chloroform-phase extract ranged from 12.5 mg/mL to 50 mg/mL against the eleven species of the bacteria. Notably, for the methanol-phase extract, the MICs were between 1.56 mg/mL and 50 mg/mL against the fifteen bacterial species. Of these, the growth of B. cereus A2-2 and V. parahemolyticus ATCC17802 was the most strongly repressed by the methanol-phase extract with the MICs of 1.56 mg/mL, followed by V. alginolyticus ATCC17749, V. mimicus bio-56759, V. parahemolyticus B3-13, V. parahemolyticus B5-29, V. parahemolyticus B9-35, and V. parahemolyticus A1-1 with MICs of 3.13 mg/mL. In addition, the growth of B. cereus A1-1, P. aeruginosa ATCC9027, S. typhimurium ATCC15611, S. flexneri CMCC51572, S. aureus ATCC8095, and V. parahemolyticus B4-10 was also inhibited by the methanol-phase extract with lower MICs (6.25 mg/mL). Of these pathogens, for example, V. alginolyticus is a foodborne marine Vibrio that can cause gastroenteritis, otitis media, otitis externa, and septicemia in humans [17]. V. mimicus can also cause gastroenteritis in humans due to contaminated fish consumption and seafood [18]. P. aeruginosa is an opportunistic pathogen and can cause serious infections, especially in patients with compromised immune systems [19].
Recently, Song et al. [14] reported that the ethyl acetate extract of P. kleiniana Wight et Arn inhibited E. coli, P. aeruginosa, and C. albicans, with MICs of 5 mg/mL, 2.5 mg/mL, and 5 mg/mL, respectively. Tao et al. reported the MIC value of the TFP against MRSA was 20 µg/mL [9].
These results indicated that the methanol-phase crude extract had a higher inhibition rate (68.18%), showing a more broad inhibitory profile with much lower MICs (1.56-50 mg/mL) against the pathogens tested, as compared to the chloroform-phase crude extract (50.00%; 12.5-50 mg/mL). Thus, the methanol-phase crude extract was chosen for further analysis in this study.

Purification of the Methanol-Phase Crude Extract from P. kleiniana Wight et Arn
Based on the obtained results, a large amount of the methanol-phase crude from P. kleiniana Wight et Arn was prepared and further purified using Prep-HPLC analysis. As shown in Figure S1, three separated fragments (designated Fragments 1-3) were observed via scanning at OD 211 for 12 min, including Fragment 1 (2.45 min), Fragment 2 (6.75 min), and Fragment 3 (9.83 min). The main peak of the methanol-phase crude was observed to occur at 2.45 min, wherein the absorption peak of Fragment 1 reached its maximum.
The three single fragments were subjected for antibacterial activity analysis. Fragment 1 had strong inhibitory effects on V. parahemolyticus ATCC17802, V. parahemolyticus B5-29, V. parahemolyticus B9-35, V. parahemolyticus B3-13, and V. parahemolyticus B4-10. In addition, the growth of the other six strains was also effectively repressed, including B. cereus A2-2, V. parahemolyticus A1-1, S. flexneri CMCC51572, S. aureus ATCC25923, S. aureus ATCC8095, and S. aureus ATCC6538 ( Table 2). Of these, V. parahaemolyticus is a Gram-negative halophilic bacterium that can cause diseases in marine animals, leading to huge economic losses to the aquaculture. V. parahaemolyticus can also cause gastrointestinal infections and other health complications in humans [20]. B. cereus is a Gram-positive foodborne pathogen that can cause diarrhea and emesis [21]. S. flexneri is a Gram-negative intracellular pathogen that invades colonic cells and causes bloody diarrhea in humans [22]. S. aureus is a Grampositive opportunistic pathogen leading to food poisoning as well as human and animal infectious diseases [23,24]. We also determined MICs of Fragment 1 against the four species of pathogenic bacteria ( Table 2). The synergistic effect may explain the observed MICs of Fragment 1 (6.25-50 mg/mL), as compared to the methanol-phase extract from P. kleiniana Wight et Arn. Among the Gram-negative pathogens, V. parahemolyticus ATCC17802 and V. parahemolyticus B5-29 were the most sensitive strains to Fragment 1, with MICs of 6.25 mg/mL. For the Gram-positive pathogen, the growth of S. aureus ATCC8095 and S. aureus ATCC25923 was also effectively repressed, with MICs of 6.25 mg/mL and 12.5 mg/mL, respectively.
Conversely, the other two peaks (Fragments 2 and 3) showed weak or no antibacterial activity. To further investigate possible antibacterial modes of Fragment 1, the two Gramnegative strains V. parahemolyticus ATCC17802 and V. parahemolyticus B5-29, and two Grampositive stains S. aureus ATCC8095 and S. aureus ATCC25923 were chosen for the further analysis in this study. Cell surface hydrophobicity is an important cellular biophysical parameter that affects cell surface interactions and cell-cell communication [25]. In this study, the hexadecane was used as a probe to assess cell surface hydrophobicity change. The difference between before and after the absorbance value of bacterial fluid can indicate the change of hydrophobicity, and the larger the difference, the more hydrophobicity of the surface [26]. The cell surface hydrophobicity of the four experimental groups (1× MIC of Fragment 1) was significantly increased, as compared to the control groups (p < 0.05) (Figure 2A). For instance, after being treated with Fragment 1 for 2 h, bacterial cell surface hydrophobicity was significantly increased, including V. parahaemolyticus B5-29 (8.62%, 1.42-fold), V. parahaemolyticus ATCC17802 (8.27%, 1.50-fold), S. aureus ATCC25923 (10.34%, 1.24-fold), and S. aureus ATCC8095 (12.20%, 1.19-fold) (p < 0.05). Increasing treatment time, the cell surface hydrophobicity was further increased. After the 4 h treatment, the cell surface hydrophobicity was the most significantly increased (11.97%, 1.97-fold) in the V. parahaemolyticus B5-29 treatment group. The highest increase (15.96%, 2.63-fold) was also observed in V. parahaemolyticus B5-29, after treatment for 6 h. The results indicated that Fragment 1 from P. kleiniana Wight et Arn can significantly increase the cell surface hydrophobicity of both Gram-negative V. parahemolyticus and Gram-positive S. aureus pathogens.

Cell Membrane Fluidity
Cell membrane is a natural barrier to prevent extracellular substances from freely entering the cell [27]. In this study, as shown in Figure 2B, when compared to the control groups, the membrane fluidity of V. parahaemolyticus B5-29, S. aureus ATCC25923, and S. aureus ATCC8095 did not change significantly after treatment with Fragment 1 (1 × MIC) for 2 h and 4 h. However, a significant decrease (1.16-fold, 1.25-fold, and 1.24-fold) was observed in these three treatment groups after treatment for 6 h, respectively (p < 0.05). In addition, a significant decrease in cell membrane fluidity was only found in V. parahaemolyticus ATCC17802 after treatment for 4 h (1.16-fold) and 6 h (1.24-fold), respectively (p < 0.05). These results indicated that Fragment 1 from P. kleiniana Wight et Arn can significantly reduce the cell membrane fluidity of both Gram-negative V. parahemolyticus and Gram-positive S. aureus pathogens.

Cell Membrane Permeability
β-galactosidase is a macromolecular protein naturally found in the interior of cells that can hydrolyze the substrate o-nitrophenyl-β-D-galactopyranosi (ONPG) to galactose and o-nitrophenol in yellow. If the inner membrane of bacterial cells is damaged, ONPG will quickly enter the cell [28]. In this study, the ONPG was used as a probe to assess whether the bacterial inner membrane is damaged. As illustrated in Figure 3D, the inner cell membrane permeability of S. aureus ATCC8095 did not change significantly after treatment with Fragment 1 (1 × MIC) from P. kleiniana Wight et Arn for 2 h (p > 0.05); conversely, significant increases were observed in V. parahaemolyticus B5-29, V. parahaemolyticus ATCC17802, and S. aureus ATCC25923 treatment groups (1.15-fold, 1.18-fold, and 1.04-fold), respectively (p < 0.05). After being treated for 4 h, the highest increase was

Cell Membrane Fluidity
Cell membrane is a natural barrier to prevent extracellular substances from freely entering the cell [27]. In this study, as shown in Figure 2B, when compared to the control groups, the membrane fluidity of V. parahaemolyticus B5-29, S. aureus ATCC25923, and S. aureus ATCC8095 did not change significantly after treatment with Fragment 1 (1× MIC) for 2 h and 4 h. However, a significant decrease (1.16-fold, 1.25-fold, and 1.24-fold) was observed in these three treatment groups after treatment for 6 h, respectively (p < 0.05). In addition, a significant decrease in cell membrane fluidity was only found in V. parahaemolyticus ATCC17802 after treatment for 4 h (1.16-fold) and 6 h (1.24-fold), respectively (p < 0.05). These results indicated that Fragment 1 from P. kleiniana Wight et Arn can significantly reduce the cell membrane fluidity of both Gram-negative V. parahemolyticus and Gram-positive S. aureus pathogens.

Cell Membrane Permeability
β-galactosidase is a macromolecular protein naturally found in the interior of cells that can hydrolyze the substrate o-nitrophenyl-β-D-galactopyranosi (ONPG) to galactose and o-nitrophenol in yellow. If the inner membrane of bacterial cells is damaged, ONPG will quickly enter the cell [28]. In this study, the ONPG was used as a probe to assess whether the bacterial inner membrane is damaged. As illustrated in Figure 3D, the inner cell membrane permeability of S. aureus ATCC8095 did not change significantly after treatment with Fragment 1 (1× MIC) from P. kleiniana Wight et Arn for 2 h (p > 0.05); conversely, significant increases were observed in V. parahaemolyticus B5-29, V. parahaemolyticus ATCC17802, and S. aureus ATCC25923 treatment groups (1.15-fold, 1.18-fold, and 1.04-fold), respectively (p < 0.05). After being treated for 4 h, the highest increase was found in V. parahaemolyticus B5-29 (1.22-fold). After treatment for 6 h, significant increases were also observed in V. parahaemolyticus B5-29, V. parahaemolyticus ATCC17802, S. aureus ATCC25923, and S. aureus ATCC8095 (1.20-fold, 1.17-fold, 1.07-fold, and 1.08-fold), respectively (p < 0.05). These results indicated that Fragment 1 from P. kleiniana Wight et Arn can significantly increase the inner cell membrane permeability of both Gram-negative V. parahemolyticus and Gram-positive S. aureus pathogens. Outer membrane permeability was assessed by measuring the uptake of a hydrophobic fluorescent probe N-phenyl-1-naphthylamine (NPN) [29]. The outer membrane permeability increased significantly in the four treatment groups, after being treated with Fragment 1 for 2 h (1.38-fold to 1.66-fold) (p < 0.01), and 4 h (1.77-fold to 2.72-fold), respectively (p < 0.001) ( Figure 2C). The highest increase was found in V. parahaemolyticus ATCC17802 (2.70-fold), after treatment for 6 h. These results indicated that  Outer membrane permeability was assessed by measuring the uptake of a hydrophobic fluorescent probe N-phenyl-1-naphthylamine (NPN) [29]. The outer membrane permeability increased significantly in the four treatment groups, after being treated with Fragment 1 for 2 h (1.38-fold to 1.66-fold) (p < 0.01), and 4 h (1.77-fold to 2.72-fold), respectively (p < 0.001) ( Figure 2C). The highest increase was found in V. parahaemolyticus ATCC17802 (2.70-fold), after treatment for 6 h. These results indicated that Fragment 1 from P. kleiniana  Gram-negative V. parahemolyticus and Gram-positive S. aureus pathogens. Antibacterial compounds (e.g., flavonoids) in Fragment 1 from P. kleiniana Wight et Arn may have interacted with lipid components of the bacterial cell membrane. The disorder in lipid chains resulted in changed permeability and fluidity of the bacterial cell membrane [30]. The compounds may also have interacted with the bacterial cell surface proteins, leading to the altered nanomechanical properties, which consequently changed cell surface hydrophobicity and fluidity [31]. The two common pathogens V. parahemolyticus and S. aureus were chosen for further analysis in this study. The former is the leading sea foodborne pathogen worldwide [20], while the latter leads to food poisoning, as well as human and animal infections [23].

Bacterial Cell Surface Structure Changes Triggered by Fragment 1 from P. kleiniana Wight et Arn
Based on the obtained results in this study, the representative Gram-negative V. parahaemolyticus ATCC17802 and Gram-positive S. aureus ATCC25923 strains were chosen for further scanning electron microscope (SEM) analysis. As shown in Figure 4, the cells of V. parahaemolyticus ATCC17802 were intact in shape with a flat surface, showing a typical rod-like structure, while those of S. aureus ATCC25923 were also intact and clear, showing a typical spherical structure. In remarkable contrast to the control groups, the bacterial morphological structures were altered to varying degrees in the treatment groups triggered by Fragment 1 (1× MIC) for different times.
For the Gram-negative V. parahaemolyticus ATCC17802, its cell surface was slightly shrunken after being treated with Fragment 1 for 2 h. After 4 h of treatment, the cell surface was more wrinkled and was slightly depressed, the cell membrane was folded and some contents were exuded. After 6 h of the treatment, the cells were severely deformed and crumpled, with a large amount of content leaked.
For the Gram-positive S. aureus ATCC25923, its cell surface was rough and slightly wrinkled, but certain cells were depressed, with a small amount of content leaked after the treatment for 2 h. Upon the increased treatment time (4 h), more cells were obviously wrinkled and deformed with the irregularly spherical, and more content leaked out. The cell morphological structure was seriously damaged after being treated for 6 h.

Bacterial Cell Surface Structure Changes Triggered by Fragment 1 from P. kleiniana Wight et Arn
Based on the obtained results in this study, the representative Gram-negative V. parahaemolyticus ATCC17802 and Gram-positive S. aureus ATCC25923 strains were chosen for further scanning electron microscope (SEM) analysis. As shown in Figure 4, the cells of V. parahaemolyticus ATCC17802 were intact in shape with a flat surface, showing a typical rod-like structure, while those of S. aureus ATCC25923 were also intact and clear, showing a typical spherical structure. In remarkable contrast to the control groups, the bacterial morphological structures were altered to varying degrees in the treatment groups triggered by Fragment 1 (1 × MIC) for different times.     Table 3, a total of 66 different compounds were identified. The highest relative percentage of the compounds was D-maltose (6.77%), followed by oxymorphone (6.29%), rutin (6.29%), D-proline (5.41%), and L-proline (5.41%). In addition, alkaloids, flavonoids, phenols, sesquiterpenoids, fatty acyls, and organic acids were also detected ( Table 3).
Highly concentrated sugar solutions, such as the D-maltose identified in this study, are known to be effective antimicrobial agents [32]. Previous research has indicated that the antibacterial activity of phenanthrenes and derivatives, such as the oxymorphone identified in this study, was primarily related to the destruction of the bacterial cell wall structure [33]. Plant extracts contain a large number of bioactive compounds, mainly polyphenols including flavonoids and phenolic compounds. Flavonoids, such as the rutin identified in this study, could exert antibacterial activity via damaging the cytoplasmic membrane, inhibiting energy metabolism and synthesis of nucleic acids [34]. Tao et al. also reported the major compounds of the TFP were 3-O-methylducheside A, naringenin, rutin and quercetin [9,13]. Phenols, such as the p-octopamine identified in this study, are potent antibacterial agents against both Gram-positive and Gram-negative bacteria via the disruption of the bacterial membrane, leading to bacterial lysis and leakage of intracellular contents [35]. Indole alkaloids, such as the indole identified in this study, possess not only intriguing structural features but also biological/pharmacological activities e.g., antimicrobial activity [36]. Additionally, amino acids and its derivatives, such as the D-proline, L-proline, glutamic acid, 5-aminovaleric acid, lysine, pipecolic acid, and L-valine identified in this study, are a kind of antibacterial agent with the advantages of being not easily drug-resistant, and having low toxicity or harmless metabolites [37]. Approximately 13.07% (580 of 4436 genes) of V. parahaemolyticus ATCC17802 genes were differentially expressed in the treatment group, as compared to the control group. Of these, 238 DEGs showed higher transcriptional levels (fold change ≥ 2.0), whereas 342 DEGs were significantly down-regulated (fold change ≤ 0.5) (p < 0.05). Sixteen significantly altered metabolic pathways were identified in V. parahaemolyticus ATCC 17802, including the citrate cycle; glyoxylate and dicarboxylate metabolism; fatty acid degradation; glycine, serine, and threonine metabolism; oxidative phosphorylation; pyruvate metabolism; propanoate metabolism; beta-Lactam resistance; ABC transporters; two-component system; alanine, aspartate, and glutamate metabolism; phosphotransferase system (PTS); butanoate metabolism; lysine degradation; quorum sensing (QS); and nitrogen metabolism ( Figure 5, Table 4).   In the citrate cycle, all the DEGs (n = 14) were significantly repressed (0.146-fold to 0.35-fold) (p < 0.05) in V. parahaemolyticus ATCC17802 after treatment by Fragment 1 from P. kleiniana Wight et Arn. For instance, the DEGs (sucABCD, WU75_19785 and WU75_19790, WU75_19795, and WU75_19800), encoding a 2-oxoglutarate dehydrogenase, a dihydrolipoamide succinyltransferase, and succinyl-CoA synthetase subunits alpha and beta, respectively, were highly inhibited (0.146-fold, 0.133-fold, 0.134-fold, and 0.16-fold) (p < 0.05). Moreover, the DEGs (sdhABCD, WU75_19775, WU75_19780, WU75_19765, and WU75_19770) encoding a succinate dehydrogenase were also highly repressed (0.144-fold to 0.199-fold) (p < 0.05), which links two essential energy-producing processes, the citrate cycle and oxidative phosphorylation [38]. The inhibited key enzymes in the citrate cycle highlighted inactive energy production in V. parahaemolyticus ATCC17802 triggered by Fragment 1.   In the propanoate metabolism, all the DEGs (n = 2) were significantly inhibited (0.402-fold to 0.435-fold) in the V. parahaemolyticus ATCC17802 treatment group (p < 0.05). For example, the DEG (prpC, WU75_15770) encoding a 2-methylcitrate synthase was significantly inhibited (0.435-fold) (p < 0.05). It has been reported that the strategic inhibition of organic acid catabolism in P. aeruginosa through inhibition of PrpC activity may be a potent mechanism to halt the growth of this pathogen [39].
In the glycine, serine, and threonine metabolism, all the DEGs (n = 15) were significantly inhibited (0.113-fold to 0.495-fold) in V. parahaemolyticus ATCC17802 (p < 0.05). For example, the DEGs (ectBAC, WU75_16140, WU75_16145, and WU75_16135), encoding a diaminobutyrate-2-oxoglutarate aminotransferase, a 2% 2C4-diaminobutyric acid acetyltransferase, and an ectoine synthase, which are involved in the synthesis of ectoine that is commonly found in halophilic and halotolerant microorganisms to maintain cell osmotic balance [42]. Additionally, in the alanine, aspartate, and glutamate metabolism, ten of the thirteen DEGs were significantly down-regulated (0.037-fold to 0.466-fold) in V. parahaemolyticus ATCC17802 as well (p < 0.05). Conversely, the DEGs (ansAB, WU75_20915, and WU75_01110) were up-regulated (2.141-fold and 2.718-fold) (p < 0.05), which encoded a cytoplasmic asparaginase I and a L-asparaginase II. The asparaginase I is required for bacterial growth on asparagine as the sole nitrogen source [43], while asparaginases are important in maintaining nitrogen balance and the levels of amino acids within cells [43]. These results indicated that the amino acid synthesis was inhibited in V. parahaemolyticus ATCC17802 mediated by Fragment 1.
For the ABC transporters, 29 of the 35 DEGs were significantly down-regulated (0.106-fold to 0.491-fold) in V. parahaemolyticus ATCC17802 (p < 0.05). Of these, the DEGs (proVXW, WU75_10380, WU75_10390, and WU75_10385), encoding a choline ABC transporter ATP-binding protein, a choline ABC transporter substrate-binding protein, and a choline ABC transporter permease subunit that are responsible for the choline transport, were all significantly repressed (0.106-fold to 0.138-fold). The DEGs (oppABCDF, WU75_12765, WU75_12770, WU75_12775, WU75_12780, and WU75_12785) encoding a peptide ABC transporter substrate-binding protein, an oligopeptide transporter permease, a peptide ABC transporter permease, an oligopeptide transporter ATP-binding component, and a peptide ABC transporter ATP-binding protein, respectively, were all highly repressed (0.172-fold and 0.214-fold). Additionally, the DEGs (yejABE, WU75_13090, WU75_07210, WU75_07220, and WU75_07215) encoding a diguanylate cyclase, an ABC transporter permease subunit, and a peptide ABC transporter permease, respectively, were highly repressed as well (0.151-fold and 0.220-fold). The ABC transporter YejABEF is required for resistance to antimicrobial peptides and virulence of Brucella melitensis [44]. These results indicated that the inhibited ABC transporters likely led to the repressed substance transport and harmful substances discharged in V. parahaemolyticus ATCC17802.
In contrast, in the PTS, nine of the eleven DEGs were significantly up-regulated (2.36-fold to 6.946-fold) in the V. parahaemolyticus ATCC17802 treatment group (p < 0.05).
In the two-component system, 19 DEGs were significantly inhibited (0.186-fold to 0.491-fold), whereas 9 DEGs were significantly enhanced (2.068-fold to 26.5-fold) (p < 0.05). The two-component system is one of the primary pathways by which bacteria adapt to environmental stresses [48]. For instance, the DEGs (cpxAR, WU75_18570, and WU75_18575) encoding a two-component sensor protein and a transcriptional regulator were strongly up-regulated (10.981-fold and 26.500-fold) (p < 0.05). The CpxAR is a key modulator of capsule export that facilitates Actinobacillus pleuropneumoniae survival in the host [49]. It also regulates cell membrane permeability and efflux pump activity and induces multidrug resistance (MDR) in Salmonella enteritidis [50].
Additionally, in the beta-lactam resistance, all the DEGs (acrAB, WU75_09925, WU75_09315, and WU75_09310) were strongly up-regulated (6.699-fold to 40.366-fold) in the V. parahaemolyticus ATCC17802 treatment group (p < 0.05), which encoded a multidrug efflux resistance nodulation division (RND) transporter periplasmic adaptor subunit and a multidrug transporter. The RND family efflux pumps, including the major pump AcrAB-TolC, are important mediators of intrinsic and evolved antibiotic resistance [51].
Taken together, these results indicated that Fragment 1 from P. kleiniana Wight et Arn can significantly change sixteen metabolic pathways in the Gram-negative V. parahaemolyticus ATCC17802, which consequently led to repressed substance transporting, energy production, and protein translation, but enhanced stringent response, and harmful substance discharging, and thereby cell death.

The Major Changed Metabolic Pathways in S. aureus ATCC25923
Approximately 7.3% (196 of 2672 genes) of S. aureus ATCC25923 genes were differentially expressed in the treatment group, as compared to the control group. Of these, 156 DEGs showed higher transcriptional levels (fold changes ≥ 2.0), whereas 40 DEGs were significantly down-regulated (fold changes ≤ 0.5) (p < 0.05). Based on the comparative transcriptomic analysis, seven significantly altered metabolic pathways were identified in S. aureus ATCC25923, including the two-component system; nitrogen metabolism; riboflavin metabolism; arginine and proline metabolism; atrazine degradation; alanine, aspartate and glutamate metabolism; and pyrimidine metabolism ( Figure 6, Table 5).
In the arginine and proline metabolism, all the DEGs (n = 4) were significantly downregulated at the transcription levels (0.109-fold to 0.461-fold) in S. aureus ATCC25923 (p < 0.05). The arginine metabolism converts L-arginine to urea and L-ornithine, which are further metabolized into proline and polyamides that drive collagen synthesis and bioenergetic pathways critical for cell proliferation, respectively [52]. For instance, the DEG (rocF, KQ76_11235) encoding an arginase was significantly down-regulated (0.461-fold) (p < 0.05), and was associated with the ability of Helicobacter pylori to establish chronic infections [53].
All the DEGs (n = 4) in the riboflavin metabolism were also significantly inhibited (ribBADEH, 0.3734-fold to 0.480-fold) (p < 0.05). In this pathway, the redox cofactors flavin mononucleotide and flavin adenine dinucleotide and their precursor riboflavin play important roles in many cellular processes, such as respiration, DNA repair, biosyntheses of heme groups, cofactors and nucleotides, fatty acid beta-oxidation, and bioluminescence [54].
Bacteria use two-component signal transduction systems to elicit adaptive responses to environmental changes [55]. In this study, seven DEGs in the two-component system were significantly up-regulated (2.117-fold to 28.924-fold) in S. aureus ATCC25923 (p < 0.05). For instance, the DEGs (agrB, KQ76_10520; and graS, KQ76_03245) encoding histidine kinases were significantly up-regulated by 2.565-fold and 2.989-fold, respectively (p < 0.05). The accessory gene regulator (agr) quorum-sensing system contributes to its pathogenicity of S. aureus [56]. GraS, the sensor histidine kinase of the GraXRS system, has been suggested to directly activate the response regulator ArlR [53]. Loss of the ArlR alone impairs the ability of S. aureus to respond to host-imposed manganese starvation and glucose limitation [57].   Interestingly, expression of all the DEGs (n = 7) in the nitrogen metabolism was significantly increased at the transcription level (3.529-fold to 10.404-fold) in S. aureus ATCC25923 (p < 0.05). The seven DEGs (nirBD, narHIJZT) were all involved in nitrate reduction [58][59][60]. Of these, the NirD (KQ76_12515) was a small subunit of cytoplasmic NADH-dependent nitrite reductase complex NirBD [61,62]. Over-expression of nirD limits RelA-dependent accumulation of guanosine 5 -triphosphate 3 -diphosphate ((p)ppGpp) in vivo and can prevent activation of the stringent response during amino acid starvation in E. coli [62].
In the alanine, aspartate, and glutamate metabolism, two DEGs (carBA, KQ76_05770 and KQ76_05765) encoding carbamoyl phosphate synthase were significantly up-regulated (2.154-fold and 3.084-fold) in S. aureus ATCC25923 (p < 0.05). The interface residues located near the CarB region of carboxy phosphate synthetic domain plays a key role in carbamoyl phosphate synthetase, aspartate transcarbamoylase, and dihydroorotase (CAD) complex regulation in the pyrimidine biosynthesis [63]. Correspondingly, in the pyrimidine metabolism, four DEGs (pyrBCR, KQ76_05755, KQ76_05760, and KQ76_05745) were also significantly up-regulated (2.968-fold to 3.213-fold) (p < 0.05), and encoded an aspartate carbamoyltransferase, a dihydroorotase, and a phosphoribosyl transferase, respectively. The pyrimidines are involved in the synthesis of DNA, RNA, lipids, and carbohydrates. The pyrimidine metabolism is involved in the synthesis, degradation, salvage, interconversion, and transport of these compounds [64]. Taken together, these results indicate that Fragment 1 from P. kleiniana Wight et Arn can significantly influence seven metabolic pathways in the Gran-positive S. aureus ATCC25923. Of these, the two-component system, alanine, aspartate and glutamate metabolism, and nitrogen metabolism were also changed in the Gram-negative V. parahaemolyticus ATCC17802, which led to the enhanced regulation of stringent response in the two pathogens. On the other hand, we also found distinct transcriptomic profiles between the Gram-positive and Gram-negative pathogens triggered by Fragment 1. For example, consistent with the results obtained from the cell structure analysis, V. parahaemolyticus ATCC17802 was more sensitive to Fragment 1 treatment, as more metabolic pathways were altered, such as the citrate cycle, glyoxylate and dicarboxylate metabolism, fatty acid degradation, glycine, serine and threonine metabolism, oxidative phosphorylation, pyruvate metabolism, propanoate metabolism, beta-lactam resistance, ABC transporters, PTS, butanoate metabolism, lysine degradation, and QS, which resulted in cell destruction and even death.
In addition, to validate the transcriptome data, we tested 16 representative DEGs (Table S1) via reverse transcription real time-quantitative PCR (RT-qPCR) analysis, and the resulting data were generally correlated with those yielded from the transcriptome analysis (Table S2).

Bacterial Strains and Culture Conditions
The bacterial strains and culture media used in this study are listed in Table S3. Vibrio strains and non-Vibrio strains were incubated as described in our recent studies [ 19.90" E), Guizhou Province, China, in October of 2021. Bioactive substances were extracted from the samples using the methanol and chloroform method described in our recent reports [15,16,66]. Briefly, aliquot of a 500 g of the whole plant sample was lyophilized, pulverised, powded, sonicated, and then filtered and collected for the secondary extraction. The methanol and chloroform phases were separated and then concentrated using the Rotary Evaporator (IKA, Staufen, Germany) [15,16].

Antimicrobial Susceptibility Assay
The susceptibility of the bacterial strains (Table S3) to the extracts from P. kleiniana Wight et Arn were determined according to the standard method issued by the Clinical and Laboratory Standards Institute, USA (CLSI, M100-S23, 2018). The antibacterial activity was defined as described previously [15,16]. Broth dilution testing (microdilution) (CLSI, M100-S18, 2018) was used to determine MICs of the extracts. The MIC was defined as described previously [15,16].

Bacterial Cell Surface Hydrophobicity and Membrane Fluidity Assays
The cell surface hydrophobicity was measured according to the method of Cui et al. [68]. The cell membrane fluidity was measured according to the method of Kuhry et al. [69], using the 1,6-diphenyl-1,3,5-hexatriene (DPH, Sangon, Shanghai, China).

Cell Membrane Permeability Analysis
Cell outer membrane permeability was measured according to the method of Wang et al. [70], with the NPN solution (Sangon, Shanghai, China). The inner membrane permeability was measured according to the method of Huang et al. [71], with the ONPG solution (Sangon, Shanghai, China).

Scanning Electron Microscope (SEM) Assay
The preparation of the samples for the SEM analysis was performed using the method described in our recent reports [15,16,72]. The samples were observed using the Scanning Electron Microscope (Tescan Mira 3 XH, Tescan, Brno, Czech Republic, 5.0 kV, 30,000×).

Illumina RNA Sequencing
The bacterial cell culture at the mid-LGP was treated with Fragment 1 (1× MIC) from P. kleiniana Wight et Arn for 6 h, and then collected via centrifugation for the total RNA preparation [15,16,72]. Three independently prepared RNA samples for each strain were subjected for the Illumina RNA sequencing analysis, using Illumina HiSeq 2500 platform (Illumina, Santiago, CA, USA) [72].

RT-qPCR Assay
The RT-qPCR assay was performed according to the method described in our recent reports [15,16,72]. The oligonucleotide primers were designed (Table S1), and synthesized via Sangon (Shanghai, China).

Data Analysis
The DEGs were analyzed as described in our recent reports [15,16,72]. All tests were carried out in triplicate. The data were analyzed using the SPSS statistical analysis software version 17.0 (SPSS Inc., Armonk, NY, USA). One-way analysis of variance (ANOVA) was performed using the least-significant difference (LSD) method and homogeneity of variance test. There was no significant difference between the control and the treatment groups if the generalized p-values were more than 0.05; conversely, there was significant difference if p-values were less than 0.05.

Conclusions
In this study, the methanol-phase extract from P. kleiniana Wight et Arn showed an inhibition rate of 68.18% against 22 species of common pathogenic bacteria. The methanolphase extraction inhibited the growth of one species of Gram-positive S. aureus, and 14 species of Gram-negative bacteria, including B. cereus, E. cloacae, E. coli, P. aeruginosa, S. typhimurium 1, S. dysenteriae, S. flexneri, S. flexneri, S. sonnei, V. alginolyticus, V. cholerae, V. fluvialis, V. mimicus, V. parahemolyticus, and V. vulnificus strains. This extract was further purified using the Prep-HPLC, and three separated fragments were obtained. Fragment 1 significantly increased bacterial cell surface hydrophobicity and membrane permeability and decreased membrane fluidity, disrupting the cell integrity of the Grampositive and Gram-negative bacteria such as S. aureus ATCC25923, S. aureus ATCC8095, V. parahaemolyticus ATCC17802, and V. parahaemolyticus B5-29. The MIC values of Fragment 1 ranged from 6.25 mg/mL to 50 mg/mL. A total of 66 different compounds in Fragment 1 were identified. The highest relative percentage of the compounds was D-maltose (6.77%), followed by oxymorphone (6.29%), rutin (6.29%), D-proline (5.41%), and L-proline (5.41%). Highly concentrated sugar solutions, such as the D-maltose identified in Fragment 1, are known to be effective antimicrobial agents. The identified oxymorphone and rutin could exert antibacterial activity via damaging the bacterial cell wall and cytoplasmic membrane, respectively. Multiple cellular metabolic pathways altered by Fragment 1 in the representative Gram-negative V. parahaemolyticus ATCC17802 and Gram-positive S. aureus ATCC25923 pathogens after treatment with Fragment 1 (1× MIC) for 6 h (p < 0.05). These results indicated that the energy supply and protein translation of the tested strains was inhibited, the signal transduction was blocked, and the ability to pump foreign harmful substances was reduced, leading to cell death. Overall, the results of this study demonstrate that Fragment 1 from P. kleiniana Wight et Arn is a promising candidate for antibacterial medicine and food preservatives.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/foods12081640/s1, Table S1: The oligonucleotide primers designed and used in the RT-qPCR assay; Table S2: The relative expression of representative DEGs by the RT-qPCR assay; Table S3: The bacterial strains and media used in this study; Figure S1