Hibiscus Acid and Chromatographic Fractions from Hibiscus Sabdariffa Calyces: Antimicrobial Activity against Multidrug-Resistant Pathogenic Bacteria

The anti-microbial properties of acetone extracts from Hibiscus sabdariffa calyces, fractions isolated by silica gel chromatography and hibiscus acid purified from some of these fractions and additionally identified by nuclear magnetic resonance spectroscopy, mid-infrared spectroscopy and X-ray diffraction, were studied against both multidrug-resistant Salmonella strains and pathogenic Escherichia coli bacteria. Gel diffusion was used to determine the anti-microbial effects. The mode of action of hibiscus acid was determined by crystal violet assay. Hibiscus acid and 17 of the 25 chromatographic fractions obtained, displayed an anti-microbial effect against all bacterial strains tested. Hibiscus acid showed a greater anti-microbial effect than the acetone extract against most of the bacteria strains, while chromatographic fractions IX–XIV exerted the greatest anti-microbial effect against all bacteria. The minimum inhibitory concentration of the acetone extract was 7 mg/mL, and the minimum bactericidal concentration was 10 mg/mL, while the corresponding values for hibiscus acid were 4–7 and 7 mg/mL, respectively. The results of the crystal violet assay indicate that hibiscus acid alters membrane permeability. Hibiscus acid is a potential alternative to control multidrug-resistant bacteria. Due to its ready availability and easy extraction from H. sabdariffa, hibiscus acid is potentially useful in the food industries.


Introduction
Salmonella is one of the leading causes of foodborne diseases, and its infection (salmonellosis) is spread worldwide. Due to its prevalence, salmonellosis has become a public health burden, representing significant costs in many countries. A range of fresh fruit and vegetables, especially those eaten raw (lettuce, sprouts, melon and tomatoes), are implicated in Salmonella infection [1]. S. enterica subspecies enterica is composed of more than 1500 serotypes with some of great importance, such as S. Typhimurium and S. Enteritidis. S. enterica subsp. enterica is responsible for more than 99% of human salmonellosis and therefore it is widely studied [2].

Chromatographic Fractionation of Acetone Extract
Two hundred and thirty grams of dry acetone extract of H. sabdariffa calyces was separated by column chromatography. The dried extract was mixed with silica gel (Sigma-Aldrich, Toluca, México), previously activated at 120 • C for 1 h in a drying recirculation oven (Ambi-Hi-Low Chamber, Lab-Line, Jefferson, USA), at a ratio of 1:2. A glass chromatography column was filled with the silica gel-acetonic extract mixture. Different solvents (hexane, hexane-ethyl acetate, ethyl acetate, ethyl acetate-methanol and methanol) were used as the mobile phase to recover consecutive 100 mL fractions from the packed column. The fractions were concentrated on a rotary evaporator, placed in glass vials and analysed by thin-layer chromatography. The fractions whose components showed the same level of displacement in the plate were pooled and placed in an air recirculation oven at 40 • C to evaporate solvent residues. The anti-microbial activity of the chromatographic collections against multidrug-resistant pathogenic bacteria was determined.

Extraction of Hibiscus Acid
Two hundred and thirty grams of dry acetone extract of H. sabdariffa calyces was packed with silica gel in a chromatographic column, as described in Section 2.2. Hexane was used as the mobile phase to separate the oils in the extract, and 600 mL fractions were recovered in glass flasks. All the chromatographic fractions obtained were rotary-evaporated to remove the solvents and concentrate the separated compounds. After discarding most of the oils from the extract, the solvent mixture hexane-ethyl acetate (9:1 v/v) was used as the mobile phase to remove all residual oils. The mobile phase (8:2 v/v) passed through the packed column until some small crystals were observed in the rotary-evaporated fractions and it was then used at a ratio of 7:3 (v/v) to obtain well-defined crystals in the rotary-evaporated fractions. The crystals were analysed by thin-layer chromatography to determine their purity, re-crystallised using 7:3 (v/v) acetone-ethyl acetate in a separatory funnel and then stored for 24 h. Once the formation of crystals on the wall of the separation funnel was observed, the liquid was decanted, and the crystals were recovered. Finally, the residual acetone was removed in an air recirculation oven at 45 • C for 2 h.

Infrared Spectroscopy with Attenuated Total Reflection
The crystals (Section 2.3) were ground in a mortar to reduce the particle size and analysed using a diamond-accessorised attenuated total reflection infrared spectrometer (Frontier, Perkin Elmer, Norwalk, CT, USA) at 25 ± 2 • C. Infrared spectra were recorded between 4000 and 400 cm −1 at a resolution of 4 cm −1 , and 64 spectra per sample were co-added to improve the sample-to-noise ratio.

X-ray Crystallography
The crystals were grown in aqueous acetone by slow evaporation. Diffraction data were measured on a Gemini CCD diffractometer (Oxford Diffraction Ltd., Abingdon, Oxfordshire, England) at room temperature using graphite-monochromated CuKα radiation (λ = 1.54184 Å) and processed using the CrysAlis program (version 1.171. 33.31, 2009; Oxford Diffraction Ltd., Abingdon, UK). The structure was solved using Olex2 [32]) and SHELXT [33] structure solution program using intrinsic phasing or direct methods and refined with the crystal structure refinement program SHELXL [34] using least-squares minimization.

Differential Scanning Calorimetry
The melting point of hibiscus acid was measured using a differential scanning calorimeter (Q2000, TA Instruments, New Castle, NJ, USA), previously calibrated with indium (onset temperature T o = 156.6 • C, transition enthalpy ∆H = 28.4 J/g, respectively. Five milligrams of hibiscus acid crystals were placed in an aluminum crucible, which was then sealed and heated from 25 to 250 • C at a rate of 5 • C/min. The transition temperatures and ∆H values were obtained directly using Universal Analysis software version 4.4A (TA Instruments).
2.5. Determination of the Anti-microbial Effect of Acetone Extract, Chromatographic Collections and Hibiscus Acid

Preparation of Test Solutions
Solutions of acetone extract, chromatographic collections and hibiscus acid were prepared at final concentrations of 100 mg/mL. Only distilled water was used to prepare the solutions of acetone extract and hibiscus acid. To obtain the solutions from the fraction collections, a mixture of distilled water and 20% Tween 80 (Sigma-Aldrich, Toluca, Mexico) was used.

Preparation of Bacterial Strains
The eight antibiotic-resistant strains were inoculated in 3 mL of tryptic soy broth (TSB; Bioxon, Becton Dickinson, Ciudad de México, Mexico) and incubated at 35 ± 2 • C for 18 h. The cultures were washed twice in sterile isotonic saline (0.85% NaCl; ISS) by centrifugation at 3500 rpm for 20 min, and the pellet was resuspended in ISS at approximately 10 9 colony forming units/mL (CFU/mL). Finally, a decimal dilution of these washed cultures was done with ISS to produce a final approximate concentration of 8 log CFU/mL [22,31].

Anti-microbial Activity of Acetone Extract, Chromatographic Collections and Hibiscus Acid
The gel diffusion technique with paper discs was used as follows: 100 µL washed bacterial cultures, from a concentration of 1 × 10 8 CFU/mL, were inoculated onto trypticase soy agar plates (Bioxon, Becton Dickinson) and distributed over the agar by the streak plate method. Sterilized paper discs (Whatman Grade 5, 6-mm diameter) were placed on the surface of the inoculated agar. Then, 20 µL aliquots containing acetonic extract, chromatographic collections and hibiscus acid, respectively, were placed on the paper disks (final dose per disk: 2 mg extract, chromatographic collection or hibiscus acid). ISS was used as a negative control. Treatments were performed in triplicate. The plates were incubated at 35 ± 2 • C for 24 h. For each treatment, the diameters (mm) of the resulting inhibition zones were measured and expressed as the average [31].

Minimum Inhibitory Concentration and Minimum Bactericidal Concentration
The broth macrodilution method [37] was used to obtain the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC). Tubes were prepared with TSB containing acetonic extract or hibiscus acid at concentrations of 1-100 mg/mL. The tubes were inoculated with a final suspension of microorganisms at 1 × 10 5 CFU/mL (from the culture washed in ISS at a concentration of 1 × 10 9 CFU/mL, 2 decimal dilutions were made in TSB, and from the last dilution 10 µL were taken and inoculated in a tube containing 990 µL to have a final concentration of 1 × 10 5 CFU/mL) and incubated at 37 • C for 24 h. The MIC was the lowest concentration of acetone extract or hibiscus acid to inhibit bacterial growth without turbidity in the tubes. To assess the MBC, TSB tubes containing the lowest concentrate ions of extract or hibiscus acid and no turbidity were inoculated into trypticase soy agar using the pour plate technique and incubated at 35 • C for 24-48 h. The MBC was defined as the lowest concentration of acetone extract or hibiscus acid that showed no colony growth in TSB.

Measurement of Permeability with Crystal Violet
Alteration of membrane permeability was detected by crystal violet assay exactly as described by Devil et al. 2010 [38]. Briefly, one hundred µL of S. Typhimurium C65 and enterohemorrhagic E. coli EHEC A were inoculated in TSB and incubated at 37 • C for 6 h. The bacterial suspensions were centrifuged at 10,000 rpm for 20 min. The supernatant was discarded and the cell pellets were washed twice with 0.5 mM potassium phosphate buffer solution (PBS). The bacterial cell suspension was prepared by re-suspending the cell pellet in PBS. The washed bacterial cell suspensions were incubated with different concentration of hibiscus acid at minimum sub-inhibitory concentration (MSIC; 1.25 mg/mL and 1.75 mg/mL for enterohemorrhagic E. coli EHEC A and S. Typhimurium C65, respectively), MIC (5 mg/mL and 7 mg/mL for enterohemorrhagic E. coli EHEC A and S. Typhimurium C65, respectively), 10× MIC (50 mg/mL and 70 mg/mL, for enterohemorrhagic E. coli EHEC A and S. Typhimurium C65, respectively), MBC (7 mg/mL) and ethylenediaminetetraacetic acid (EDTA, positive control, 0.25 M) at 37 • C for 60 min. Control samples were prepared similarly without treatment and EDTA (0.25 M) was used as a positive control. The cells were harvested (10,000 rpm for 5 min) and suspended in PBS containing crystal violet (10 µg/mL). The cell suspension was then incubated (10 min at 37 • C) and centrifuged (10,000 rpm for 5 min). The optical density (OD) 590 of the supernatant was measured using a UV-VIS spectrophotometer (Thermo Scientific, Nanodrop, Verona, Wisconsin, USA). The OD value of crystal violet solution was considered to be 100% excluded. The OD of the supernatant of the normal untreated cell was used as blank. The percentage of crystal violet uptake for all samples was calculated using the following formula: % uptake of crystal violet = (OD Value of sample)/(OD Value of CV solution) × 100 (1)

Statistical Analysis
Significant differences (p < 0.05) between treatments were calculated by analysis of variance and Tukey's test using SPSS Statistics 20 (IBM Corp., Armonk, NY, USA).

Anti-microbial Activity of Acetonic Extract of Hibiscus Sabdariffa
A total of 4.6 g of dry acetonic extract was obtained per 100 g of dehydrated H. sabdariffa calyces. The dry extract had anti-microbial activity against the eight multidrug-resistant Salmonella and pathogenic E. coli strains, while the radial inhibition zone on the culture medium varied from 9.8 to 12.6 mm. These results agree with those previously reported on the anti-microbial effect of extracts obtained from H. sabdariffa calyces [18].

Anti-microbial Activity of Chromatographic Collections against Pathogenic Bacteria
The acetone extract was separated by column chromatography into 903 fractions using different solvent mixtures (Table 1). Fractions displaying the same or similar thin-layer chromatogram were pooled together. Among these 25 collections (I-XXV; Table 2), 14 were anti-microbial against all multidrug-resistant Salmonella and pathogenic E. coli strains tested (Table 3), while three showed an effect against some of the Salmonella and pathogenic E. coli strains (Table 3). Collection VI was only effective against Salmonella C1 and C65, and collection VII, only against Salmonella C65, respectively. In contrast, collection XXII was not anti-microbial against E. coli C558, E. coli C636 and enteroinvasive E. coli MAC B (Table 3). Finally, collections I, II, III, IV, V, VIII, XXIV and XXV had no anti-microbial effect against any pathogenic bacteria. In general, statistically significant differences were observed between the effects produced by some collections and within collections against different pathogenic strains (Table 3).  Table 3. Anti-microbial effect of chromatographic fraction collections from Hibiscus sabdariffa calyx acetonic extract against eight multidrug-resistant Salmonella and pathogenic Escherichia coli bacteria.
In the current study, collections IX, X, XI, XIII and XIV showed greater anti-microbial activity than the rest. These collections were obtained with the polarities of the following mixtures: 70% hexane-30% ethyl acetate, 60% hexane-40% ethyl acetate and 50% hexane-50% ethyl acetate (Tables 1 and 2). It is important to note that defined crystals were formed in the pooled fractions IX, X and XI, which were collected using 70% hexane-30% ethyl acetate as the mobile phase (Tables 1 and 2).
Since the preliminary NMR analysis of the crystals suggested the presence of hibiscus acid and other compounds, a second chromatographic separation was completed using another sample of dry acetonic extract (230 g) from H. sabdariffa to obtain a higher concentration of crystals for purification. The aim was to confirm the presence and anti-microbial activity of hibiscus acid by different structural analysis techniques. The second column chromatography separating the acetonic extract yielded presumptive crystals of hibiscus acid.

1 H NMR Spectrum
The 1 H NMR spectrum of presumptive crystals of hibiscus acid corresponded to that of the molecular structure of hibiscus acid (Figure 1). Most of the proton signals ( 1 H) appeared between δ H 0 and 12. The signal observed at δ H 2.05 corresponds to the acetone-d 6   In an NMR (400 MHz) analysis of the structure of hibiscus acid determined using acetone-d6 to dissolve the crystals, Ibnusaud et al. [42] [44]. In this context, the parameters published by Ibnusaud et al. [42] and Rasheed et al. [45] are most similar to those obtained in the present work ( Figure 1).

Infrared Spectroscopy
The infrared spectrum of the crystals prepared from the H. sabdariffa acetone extract contained signals at 3410 cm -1 (OH groups), 1797 cm -1 (ester groups) and 1742 cm -1 (C=O stretching) (Figure 2). These values corresponded strongly to those reported by Ibnusaud et al. [42], which were 3400, 1790 and 1735 cm -1 . In an NMR (400 MHz) analysis of the structure of hibiscus acid determined using acetone-d 6 to dissolve the crystals, Ibnusaud et al. [42] [44]. In this context, the parameters published by Ibnusaud et al. [42] and Rasheed et al. [45] are most similar to those obtained in the present work ( Figure 1).

X-ray Crystallography
The molecular structure of hibiscus acid (Figure 3), was confirmed by X-ray diffraction which was solvated with a water molecule through a hydrogen bond (O6-H6•••O8, distance = 1.841 Å, ∡ O6-H6•••O8 = 162.62°) (Figure 3). Hibiscus acid is a five-membered lactone ring, with four carbon atoms and one oxygen atom. C3 (sp2) has a double-bonded oxygen atom, C1 an OH group and a COOH group, and C2 a COOH group, respectively.  The crystallographic details and refined structure of hibiscus acid are provided in Table 4 and Tables S1-S5. Analogous X-ray crystallographic data of hibiscus acid, albeit attached to a dimethylsulphoxide molecule, were presented by Zheoat et al. [44]. Figure 4 shows the possible hydrogen bond donors of the hibiscus acid molecule, in which the interactions that can be established with other molecules of hibiscus acid and water molecules are observed. Table 4. X-ray spectroscopy details of crystal data and structure refinement parameters of hibiscus acid isolated from Hibiscus sabdariffa calyx acetonic extract. The crystallographic details and refined structure of hibiscus acid are provided in Table 4 and Tables S1-S5. Analogous X-ray crystallographic data of hibiscus acid, albeit attached to a dimethylsulphoxide molecule, were presented by Zheoat et al. [44]. Figure 4 shows the possible hydrogen bond donors of the hibiscus acid molecule, in which the interactions that can be established with other molecules of hibiscus acid and water molecules are observed.  Table 4. X-ray spectroscopy details of crystal data and structure refinement parameters of hibiscus acid isolated from Hibiscus sabdariffa calyx acetonic extract.

Experimental Data
Empirical Formula C6H6O7  H2O

Hibiscus Acid Melting Point by Differential Scanning Calorimetry
The differential scanning calorimetry of the crystals of hibiscus acid showed a T o of 186.87 • C, peak temperature of 190.61 • C and a final temperature of 194.15 • C, while the fusion enthalpy (∆H) was 146.7 J/g. Contrastingly, Ibnusaud et al. [37] reported a melting point of 182 ºC for hibiscus acid, which differs from the current results, possibly because of either the equipment or the technique used by the authors.

Anti-microbial Effect of Hibiscus Acid
Hibiscus acid demonstrated an anti-microbial effect against all multidrug-resistant Salmonella and pathogenic E. coli strains ( Table 5). In general, the anti-microbial effect of hibiscus acid was higher (p < 0.05) than that of the acetonic extract (Table 5). Previous reports attributed the anti-microbial activity of H. sabdariffa calyces to compounds, such as protocatechuic acid and anthocyanins, in the plant [14,21,23,46]. However, no information regarding the isolation and identification of anti-microbial compounds obtained directly from H. sabdariffa calyces has been published until now. It should be noted that while several publications described the anti-microbial effect of H. sabdariffa calyx extracts obtained with solvents of varying polarities (including acetonic extract) when tested against different pathogenic bacteria, no published article describes or suggests that hibiscus acid or its derivatives are anti-microbial. Furthermore, there is no proof of their efficacy in the control and elimination of multidrug-resistant pathogenic bacteria. In other words, this document constitutes the first report on the anti-microbial activity of hibiscus acid, even against multidrug-resistant pathogenic bacteria. The MIC and the MBC of hibiscus acid were determined to assess its potential use as an anti-microbial agent in the industry.

Determination of the MIC and MBC of the Acetone Extract and Hibiscus Acid
The MICs and MBCs of the acetonic extract and hibiscus acid obtained from H. sabdariffa calyces were determined using the eight multidrug-resistant pathogenic strains. For hibiscus acid, the MIC values were 4-7 mg/mL, while the MBC range was 5-7 mg/mL ( Table 6). The MIC of the acetonic extract was 7 mg/mL for all the pathogenic strains, while the MBC was 10 mg/mL for most of the pathogenic bacteria (Table 6). Abdallah [42] evaluated the MIC and MBC of the H. sabdariffa calyces methanolic extract by broth dilution using five multidrug-resistant Acinetobacter baumannii strains and obtained MIC and MBC values of 25-50 and 50-100 mg/mL, respectively. These MIC and MBC values are higher than those obtained with the acetonic extract of H. sabdariffa (Table 6).  [27].
As mentioned above, protocatechuic acid is a compound that was reported to be present in the calyces of H. sabdariffa and is likely to be anti-microbial [21,26,27]. However, there is no information available in the literature regarding the isolation, characterization or concentration of the protocatechuic acid in H. sabdariffa calyces. Protocatechuic acid is widely distributed in a variety of plants [46]. Commercial protocatechuic acid showed anti-microbial activity against S. aureus, Klebsiella pneumoniae, Pseudomonas aeruginosa and A. baumannii, with MIC values of 8, 16, 24 and 16 mg/mL, respectively [21]. In comparison, however, Chao and Yin [26] recorded much lower MIC values (24-44 µg/mL) for the protocatechuic acid against S. Typhimurium, E. coli, L. monocytogenes, S. aureus and B. cereus. Since both studies used a pure commercial compound, the difference in MIC values was expected to be small, especially because they included a bacterial strain of the same genus and species (S. aureus). In this study, hibiscus acid showed MIC values (Table 6) within the limits of those reported for protocatechuic acid [26,47]).
Finally, the anti-microbial effect of the acetone extract and hibiscus acid from H. sabdariffa calyces was determined as bactericidal or bacteriostatic. A compound is considered bactericidal when the MBC/MIC ratio is ≤4 and bacteriostatic when this ratio is >4 [48]. Both the acetone extract and the hibiscus acid were bactericidal against all eight multidrug-resistant pathogenic strains (Table 6). Abdallah [49] also reported the bactericidal activity of methanolic extract from H. sabdariffa calyces (MBC/MIC 1-2 mg/mL) against A. baumannii strains.

Measurement of Permeability with Crystal Violet
Hydrophobic crystal violet is known to display weak penetration of the outer membrane but on the contrary, it is found to penetrate cells with impaired cell membranes thus, crystal violet assay may be employed for the detection of membrane damage [38,50].
The uptake of crystal violet by enterohemorrhagic E. coli EHEC A was 9% in the absence of hibiscus acid, but increased to 40%, 57% and 90% after MIC, MBC, 10× MIC hibiscus acid treatments, respectively ( Figure 5). The uptake of crystal violet by Salmonella C65 was 8% in the absence of hibiscus acid, but increased to 66%, 68% and 82% after MIC, MBC, 10× MIC hibiscus acid treatments, respectively ( Figure 6). Minimum sub-inhibitory concentration (MSIC) of hibiscus acid showed no effect, which reveals that it did not alter the membrane permeability in both pathogenic bacteria (Figures 5 and 6).   The results of the crystal violet absorption assay indicate that hibiscus acid alters membrane permeability of enterohemorrhagic E. coli EHEC A and Salmonella C65.
The effect of hibiscus acid on outer membrane permeability was shown by the uptake of the dye crystal violet. Crystal violet penetrates the outer membrane poorly, but it easily enters when the membrane is damaged. A significant enhancement in the uptake of crystal violet was observed in enterohemorrhagic E. coli EHEC A and Salmonella C65 treated with hibiscus acid when compared to control cells. This shows that hibiscus acid alters membrane permeability and makes the cells perpermeable to solutes. Furthermore, EDTA also significantly increased the uptake of crystal violet into the cells (Figures 5 and 6). In Gram-negative bacteria, EDTA induces outer membrane permeabilization and cell lysis [51].

Conclusions
The present work showed that hibiscus acid is one of the compounds responsible for the anti-microbial effect of H. sabdariffa calyces. It was found at a level of 1.3% (13 g/kg) in dried H. sabdariffa calyces, and 28.3% in the dry acetonic extract from H. sabdariffa calyces, respectively. Due to its relatively high concentration, hibiscus acid is likely one of the main bactericidal compounds in H. sabdariffa calyces, although other anti-microbial compounds yet to be reported may also contribute to this effect. Both the hibiscus acid and the acetonic extract from H. sabdariffa constitute a potential alternative in the control of multidrug-resistant pathogenic bacteria, such as Salmonella and E. coli pathotypes. In addition, the hibiscus acid from H. sabdariffa calyces is potentially useful in the food industries given its relative abundance and availability. Finally, further research is needed to identify other anti-microbial compounds in H. sabdariffa and their mechanisms of action against bacteria. In addition, hibiscus acid affected membrane permeability of enterohemorrhagic E. coli EHEC A and Salmonella C65.  Funding: This study was financially supported through the project "Development of innovative strategies and products that contribute to the prevention and control of emerging diseases of national importance", as part of the Program for Higher-Education Teacher Development (Programa para el Desarrollo Profesional Docente para el Tipo Superior, PRODEP), funded by the Secretary of Public Education of Mexico.