Enzymatic Esteriﬁcation as Potential Strategy to Enhance the Sorbic Acid Behavior as Food and Beverage Preservative

: Sorbic acid is the most commonly used preservative in the food industry. The antimicrobial inhibition of sorbic acid could be inﬂuenced by its lipophilic nature, which reduces its use in hydrophilic food formulations. Reactions between sorbic acid and glycerol catalyzed by lipases were studied in order to develop a novel sorbic acid derivate with a promising hydrophilic proﬁle. The esteriﬁcation reaction between sorbic acid and glycerol in a solvent-free system were performed with an immobilized lipase B from Candida antarctica (CALB). The glycerol sorbate product has been tested against S. griseus bacterium and Saccharomyces cerevisiae yeast. Results indicate that the esteriﬁcation of sorbic acid with glycerol does improve its antimicrobial properties against Saccharomyces cerevisie . The reported results demonstrate that esteriﬁcation can be used as a strategy to improve the antimicrobial activity of sorbic acid.


Introduction
Sorbic acid is a straight-chain alpha-beta-unsaturated fatty acid first isolated from the oil of unripe rowanberries in 1859 [1]. It acts as inhibitor of most molds, yeast, and some bacteria [2]. Sorbic acid, and especially its more soluble salts, known as sorbates, are used as preservative for food, animal feed, and cosmetic and pharmaceutical products [3]. The antimicrobial action of sorbates occurs at various stages of microbial life cycle (germination, outgrowth, and cell division) [4] and it may result in the alteration of the cell membrane, in the inhibition of transport systems and key enzymes, the creation of a proton flux into the cell, the inhibition of oxidative phosphorylation, or in a synergic effect of two or more of these factors [5,6]. The mode of action of sorbic acid on bacterial cells and spores has been reviewed by Sofos et al. [5], while York et al. reported the sorbic acid inhibition of the yeast Saccharomyces cerevisiae [6]. Sorbate acts as a competitive and reversible inhibitor of amino acid-induced germination [7], of several enzyme systems' activity (alcohol dehydrogenase, fumarase, anolase, aspartase, catalase, malate dehydrogenase, alfa-ketoglutarate dehydrogenase, succinic dehydrogenase, and ficin) [8], and of nutrient uptake [9]. The antifungal activity of sorbic acid seems to be related to an interference with the electrochemical membrane potential across the mitochondrial membranes [10]. Potassium sorbate is the most employed salified form of sorbic acid reported, leading to its application in several industrial sectors, in particular using chemical and enzymatic catalysts in the pharmaceutical and food industries. Ravelo et al. [10] and Tamayo et al. [12] have shown a successful strategy to perform esterification reactions catalyzed by lipase between benzoic acid or ibuprofen and glycerol in a solventless system [14]. Therefore, we aimed to exploit a solventless media strategy to perform the lipase-catalyzed esterification of sorbic acid with glycerol. The liquid nature of this short sugar alcohol allowed us to remove any organic solvent from our enzymatic synthesis strategy. Indeed, in this reaction, glycerol acts both as reagent and as a solvent.
Glycerol (propane-1,2,3-triol) is a polyol characterized by the presence of three hydroxylic groups, which confer to the compound the following properties: high hygroscopicity, high boiling point, and low vapor pressure. This nontoxic viscous liquid is widely used in chemical, cosmetic, pharmaceutical, and food industries [17]. Recently, glycerol availability grew simultaneously with the large increase in biodiesel production according to the international environmental laws (IEL). Glycerol is indeed a major byproduct of the biodiesel manufacturing process, representing approximately 10% w/w of biodiesel. Its consequent higher affordability allowed novel uses, different from the traditional direct ones applied in the food, cosmetic, and pharmaceutical fields. Thus, the use of glycerol as raw material paves the way for the strengthening of sustainable synthesis processes. [31]. The ability of glycerol and other polyhydric alcohols to confer stability to proteins is a widely recognized phenomenon [32]. Polyols, like glycerol, are frequently used as cosolvent for protein stabilization [33]. Glycerol is known to shift the native protein structure to more compact states [34], prevent the loss of enzymatic activity, increase the thermal unfolding temperature, and inhibit irreversible aggregation of proteins [35]. Moreover, glycerol has been reported as a nontoxic, biodegradable, and recyclable green solvent for high product yields and selectivity in catalysis and enzymatic catalysis [36,37].
In this paper, a monophasic solvent-free esterification of sorbic acid with glycerol has been carried out using a commercial immobilized enzyme Candida antarctica lipase type B (Figure 1), in order to improve water solubility of the resulting ester and test its antimicrobial activity in comparison with sorbic acid. In particular, glycerol can act as a hydrophilic moiety when combined with sorbic acid, with the concomitant effect of increasing polarity and, consequently, water solubility. To the best of the authors' knowledge, this is the first attempt concerning the enzymatic esterification of the sorbic acid with glycerol in a solventless system. A comparison in terms of antimicrobial activity of pure sorbic acid and its glycerol-ester against the yeast Saccharomyces cerevisiae and the bacteria Streptomyces griseus is also provided in this work. This yeast was previously reported as model in microbial stability of ready-to-drink beverages studies [6].

Synthesis of Glycerol Ester of Sorbic Acid
Sorbic Acid (100 mg), pure glycerol (10 mL), and 40 mg of immobilized CALB were placed inside a 20 mL capped vial. The reaction solution was mixed using a magnetic stirrer hot plate at 55 °C and 720 rpm. The use of a crosshead magnetic stirrer was employed to avoid enzyme support disruption. At the end of 24 h of reaction, samples were stocked at −20 °C, aiming to inhibit the enzymatic activity. A negative control of the reaction was prepared without the use of lipase. Ravelo et al. reported kinetic analysis about the solvent-free synthesis of ibuprofen monoglyceride exploiting an immobilized form of CALB [38]. In order to determine the amount of the enzyme, a range between 1 and 5 g/L has been tested. Samples were withdrawn for every concentration and analyzed by JASCO HPLC modular system equipped with reverse phase column (Synergi 4 um Hydro-RP 80 Å − 250 × 4.6 mm), refractive index (model RI-4030) and UV/vis detector (model UV-4070); 30 °C, mobile phase 90:10 MeOH/H2O (pH 2.2), 0.8 mL/min. The conversion yield has been calculated using the following equation:

Synthesis of Glycerol Ester of Sorbic Acid
Sorbic Acid (100 mg), pure glycerol (10 mL), and 40 mg of immobilized CALB were placed inside a 20 mL capped vial. The reaction solution was mixed using a magnetic stirrer hot plate at 55 • C and 720 rpm. The use of a crosshead magnetic stirrer was employed to avoid enzyme support disruption. At the end of 24 h of reaction, samples were stocked at −20 • C, aiming to inhibit the enzymatic activity. A negative control of the reaction was prepared without the use of lipase. Ravelo et al. reported kinetic analysis about the solvent-free synthesis of ibuprofen monoglyceride exploiting an immobilized form of CALB [38]. In order to determine the amount of the enzyme, a range between 1 and 5 g/L has been tested. Samples were withdrawn for every concentration and analyzed by JASCO HPLC modular system equipped with reverse phase column (Synergi 4 µm Hydro-RP 80 Å − 250 × 4.6 mm), refractive index (model RI-4030) and UV/vis detector (model UV-4070); 30 • C, mobile phase 90:10 MeOH/H 2 O (pH 2.2), 0.8 mL/min. The conversion yield has been calculated using the following equation: where A MG means area monoglyceride, and A S means area of sorbic acid.

Thin Layer Chromatography (TLC)
Thin layer chromatography was performed to monitor the reaction process. 100 µL of the reaction mixtures were diluted in 8 mL of acetone and 1.9 mL of distilled water. This step was performed because of the glycerol high viscosity, which made difficult its sowing on TLC plate. The standard sorbic acid sample has been suspended in methanol (1 mg/mL −1 ). TLC (Silica gel 60, 5 × 10 cm, Merck, Germany) plates were analyzed with the elution system ethyl acetate/hexane/acetic acid 60: 35:5 (v/v/v). Starting from the conditioning of the TLC chamber, the mobile phase has been prepared in 10 mL final volume with this proportion: 6.5 mL ethyl acetate, 3.5 mL hexane, and 0.5 mL acetic acid. The compound Rf of sorbic acid was 0.72, while glycerol, more hydrophilic, showed a retention factor (Rf) of 0.23. In these conditions, the Rf of the glycerol ester of sorbic acid was 0.34. This Rf characterizes a molecule with a higher affinity for the stationary phase of TLC compared to the sorbic acid. Indeed, the presence of two hydroxyl groups in the ester of sorbic acid decreases its logP, and, consequently, raises its hydrophilicity and water solubility.

Purification and Spectroscopic Characterization of Glycerol Ester of Sorbic Acid
Before proceeding with the purification, the entire volume of initial sample was washed three times with 10 mL ethyl acetate, in order to remove the excess of unreacted glycerol and extract the ester into the organic solvent. Once the ethyl acetate has been removed with a rotary evaporator, the separation of the glycerol ester was achieved by glass column chromatography. Silica gel column was prepared with the following eluent solution: acetate/hexane/acetic acid 60:35:5 (v/v/v). The eluted fractions were collected, the solvent was removed using a rotatory evaporator, and the rest was analyzed by NMR and uHPLC. The glass column chromatography allowed for obtaining 101 mg of the product of interest, with a conversion yield of 61%. The purified molecule obtained was analyzed by IR and NMR (provided as Supplementary Materials). IR: 3394.89; 2933.42; 1703.14; 1643.86; 1243.02; 1141.21; 997.14 cm −1 .
The NMR (acquired with a 400-MHz Varian Gemini spectrometer; Varian, Palo Alto, CA, USA) sample was prepared by dissolving it in deuterated methanol (1 mL). 1 H-and 13 C-NMR spectra showed that the compound structure was the expected one. Because of the previous purification of the ester, the presence of adequate signals of glycerol (δ = 4.1-4.4 ppm) and sorbic acid (δ = 1.85 ppm) indicates a successful esterification. NMR Spectra showed the following peaks; 1

Analytical uHPLC-MS Method
uHPLC-MS identification was carried out using an Waters Acquity UPLC equipped with a ZQ 2000 ESI mass spectrometry (Waters, Milford, MA, USA) and Mass Link software (Waters, Milford, MA, USA). The 2.6 µm Kinetex 50 × 4.6 mm C18 column was chosen to perform the analysis. The mobile phases used were water with 0.1% formic acid as solvent A and acetonitrile with 0.1% formic acid as solvent B. The liquid chromatography ran in a gradient condition from 100% H 2 O at t 0 to 100% acetonitrile at t 5 (5 min) under a flow rate of 0.3 mL/min. The column operated at a stationary temperature of 40 • C. Temperature, nebulizer pressure, and flow rate of drying gas (N 2 ) were, respectively, 230 • C, 35 psi and 10 L/min. The further operation parameters were 1200 V for nozzle voltage and 2500 V for the capillary voltage. Mass spectra were tracked in a mass-to-charge (m/z) ratio range of 150-500 in positive ion detection mode.

Evaluation of Antibacterial and Antifungal Activity
The antibacterial and antifungal activity of the glycerol ester of sorbic acid was assessed by disc diffusion and minimum inhibitory concentration methods. The disk diffusion (DD) technique allows the identification of several microorganisms' sensibilities against pharmaceutical compounds, and it is adequate when the resistance mechanism occurs owing to the antimicrobial agent destruction by the microorganism [39]. The bacterial species was Streptomyces griseus while the choice of the fungal species fell on Saccharomyces cerevisiae. The cultures were maintained in the laboratory as agar slants on a suitable culture medium: i.e., GYM (Glucose-Yeast extract-Malt extract) for Streptomyces griseus, and YMB (Yeast Mannitol Broth) for the yeast. In order to evaluate biological activity, mother cultures of each microorganism were set up in Petri dishes and grown until confluence. Then, appropriate agar medium (GYM or YMB) and different quantities of the tested compounds were added in other Petri dishes. The final concentrations reached are shown in Table 1. Three replicates were used for each concentration. 100 µL of the suspension containing 10 6 colony-forming units (CFU)/mL of microbial cells were spread on Petri dishes with appropriate agar-based growth medium (GYM or YMB). The compounds were solubilized using a 50% ethanol solution. A negative control was prepared by adding equal quantities of this solution. The agar surface was first inoculated with the specific test microorganism. Then, filter paper disks of 5 mm diameter were slightly soaked using 10 µL of drug derivative solution and placed on the agar surface. Clotrimazole [1 mg/mL] and chloramphenicol [2 mg/mL] were added at the positive controls, while the disks with solvent were made as negative controls. Cultured plates were kept in an inverted position during incubation at 28 • C for 6 days. The areas where the growth microorganism was stopped by the antibacterial agents are called inhibition zones. In order to determine the antibacterial activity, the diameters of the inhibition zones for the test organism were measured in millimeters (mm), including disk diameter of 5 mm, and compared to the negative controls. To measure the diameter of the inhibition zone, the software ImageJ has been employed as image analyzer. Digital images of the Petri dishes were analyzed after calibrating the software. The calibration step correlates pixels of a reference in the image with the real word distance measurement (mm). Inhibition zones were selected, and the diameter of the selection were displayed in the "results" window of the software.
Moreover, minimum inhibitory concentration (MICs) evaluations were performed. MIC were commonly defined as the lowest concentration of compound that completely inhibited (MIC 100 ) or reduced to 50% (MIC 50 ) clearly visible the microbial growth after the whole period of incubation, which was 6 days. These studies of the newly formed glycerol esters of sorbic acid will be improved with tests of aqueous solubility and stability at different pH values.

Sorbic Acid Monoglyceride Production
In this work, we aimed to employ a monophasic system to reach an effective esterification reaction between sorbic acid and glycerol in order to develop a novel sorbic acid derivate with a promising hydrophilic profile. Glycerol has proved to be a good candidate for the solubilization of sorbic acid in our experimental conditions. The viscosity of glycerol, together with the choice of a crosshead magnetic stirrer and the right stirring speed, allowed the enzyme to remain highly stable and perfectly separable from the medium. Glycerol viscosity and the selection of the right magnetic stirrer permit minimizing the contact surface between the reactor wall and the stirrer, thus avoiding particle degradation and, in consequence, ensuring the physical integrity of the enzyme support.
The effect of enzyme concentration on the conversion of sorbic acid has been evaluated through reverse-phase HPLC and allowed us to set up the enzyme concentration to 4 g L −1 . As shown in Figure 2, the conversion yield increased with the amount of enzyme following a hyperbolic trend typic in the esterification reactions lipase catalyzed. At higher enzyme concentrations, more active sites are present for substrate binding. Therefore, the reaction rate increases. No improvement is observed after 4 g L −1 , so these values were selected as the most adequate enzyme concentration.
stirrer permit minimizing the contact surface between the reactor wall and the stirrer, thus avoiding particle degradation and, in consequence, ensuring the physical integrity of the enzyme support.
The effect of enzyme concentration on the conversion of sorbic acid has been evaluated through reverse-phase HPLC and allowed us to set up the enzyme concentration to 4 g L −1 . As shown in Figure  2, the conversion yield increased with the amount of enzyme following a hyperbolic trend typic in the esterification reactions lipase catalyzed. At higher enzyme concentrations, more active sites are present for substrate binding. Therefore, the reaction rate increases. No improvement is observed after 4 g L −1 , so these values were selected as the most adequate enzyme concentration. It has been previously observed that migration of compounds such as glycerol, 2-hydroxyethyl benzoate, 2-hydroxyethyl sorbate, and sorbic acid occurs from immobilized forms of CALB, as they are used as stabilizers [40]. In this work, due to the minute concentration in which these compounds are present in CALB and, in some cases (glycerol, sorbic acid), the presence of much higher concentrations of such compounds in the reaction liquid, their effect on the reaction can be considered negligible. In any case, the chromatographic analysis here performed did not report the presence of most of these compounds, despite the fact that they have a very similar LogP to that of the compounds here studied.
Water activity has been previously reported in literature for this kind of glycerol esterification [41]. In fact, a small amount of water (about 10% v/v) can be an important strategic advantage. Indeed, water can enhance mass transfer by reducing the viscosity of glycerol [38] and enhance the biocatalyst flexibility, resulting in better esterification activity [38,42].
After IR and NMR, further proof of the effective lipase-catalyzed esterification has been obtained by uHPLC-MS analysis. The result from the reverse phase separation showed that glycerol sorbate eluted at 0.94 min (Figure 3; peak No 1). Because of the reaction conditions (55 °C, 24 h), no 2monoacylglycerol (2-MAG) regioisomer as by-product seems to have formed. This interesting aspect remains to be investigated and will be the subject of further studies. Curiously, no 1,3-diester was found. Indeed, this compound would be more lipophilic than the sorbic acid monoglyceride and would be held back by the c18 column, producing a peak at retention times higher than the monoglyceride. It has been previously observed that migration of compounds such as glycerol, 2-hydroxyethyl benzoate, 2-hydroxyethyl sorbate, and sorbic acid occurs from immobilized forms of CALB, as they are used as stabilizers [40]. In this work, due to the minute concentration in which these compounds are present in CALB and, in some cases (glycerol, sorbic acid), the presence of much higher concentrations of such compounds in the reaction liquid, their effect on the reaction can be considered negligible. In any case, the chromatographic analysis here performed did not report the presence of most of these compounds, despite the fact that they have a very similar LogP to that of the compounds here studied.
Water activity has been previously reported in literature for this kind of glycerol esterification [41]. In fact, a small amount of water (about 10% v/v) can be an important strategic advantage. Indeed, water can enhance mass transfer by reducing the viscosity of glycerol [38] and enhance the biocatalyst flexibility, resulting in better esterification activity [38,42].
After IR and NMR, further proof of the effective lipase-catalyzed esterification has been obtained by uHPLC-MS analysis. The result from the reverse phase separation showed that glycerol sorbate eluted at 0.94 min (Figure 3; peak No 1). Because of the reaction conditions (55 • C, 24 h), no 2-monoacylglycerol (2-MAG) regioisomer as by-product seems to have formed. This interesting aspect remains to be investigated and will be the subject of further studies. Curiously, no 1,3-diester was found. Indeed, this compound would be more lipophilic than the sorbic acid monoglyceride and would be held back by the c18 column, producing a peak at retention times higher than the monoglyceride. Electrospray ionization (ESI) was performed on the peaks of the HPLC product separations to obtain mass spectra. The results from the mass spectrometry confirmed that the expected Electrospray ionization (ESI) was performed on the peaks of the HPLC product separations to obtain mass spectra. The results from the mass spectrometry confirmed that the expected esterification reaction occurs between the carboxylic acid group and oxydril group of glycerol. The newly formed product mass prediction was m/z 186. The fragmentation pattern showed most high peaks at m/z 187 (peak No 2) and 169 (peak No 3), identifying the glycerol sorbate. Peaks appearing at an m/z ratio of 187 and 169 represent the ionized [M + H] + and [M − OH] − . To the best of our knowledge, this enzymatic esterification of sorbic acid with glycerol is reported for the first time. Until now, CALB lipase has been proposed as biocatalysts for a similar esterification reaction between benzoic acid and glycerol [19]. This strategy was demonstrated to be an successful way to minimize the effect of water in the esterification/hydrolysis activity equilibrium of the lipase. In fact, a reaction environment completely made of glycerol creates a huge disequilibrium of stoichiometry ratio between alcohol (134 mmoles) and acid (0.89 mmoles) that boosts the esterification reaction, avoids the hydrolysis due to the small amount of water produced (0.89 mmoles at complete sorbic acid conversion into the sorbate), and favors the production of the monoglyceride of the acid, avoiding the formation of di-and triglycerides. Thus, the excess of glycerol is one of the advantages of this esterification approach. This polyol excess shifts the equilibrium of formation of the ester bond towards ester and water products. Moreover, the excess glycerol could easily be recovered and used again as solvent and reagent in batch processing or recycled in a continuous process.

Antimicrobial Activity
Two models were chosen to test the antimicrobial activity of the ester; the bacterium Streptomyces griseus, and the yeast Saccharomyces cerevisiae. Disc diffusion (DD) and minimal inhibitory concentration assays (MIC) were performed. Cultures of a bacterium (S. griseus) and a yeast (S. cerevisiae), with a cell density of 7 × 10 6 CFU/mL, were exposed to different concentrations of the preservative and its ester ( Table 1).
The disc diameters of the inhibition zone are shown in Table 2. The antimicrobial activity of the ester was compared to sorbic acid, the currently used antimicrobial agents often used as preservatives in food object of this research. The antimicrobial activity of potassium sorbate, the potassium salt of sorbic acid, was also assessed because of its enhanced solubility compared to sorbic acid. This parameter is the one that identifies potassium sorbate as the current choice of microbial control in foods. The antifungal effect of sorbic acid, as well as glycerol sorbate, against S. cerevisiae was clearly confirmed by the disk diffusion experiments. The diameters of the inhibition zone of the newly synthesized glycerol ester of sorbic acid were 11.5 mm for S. cerevisiae, while sorbic acid caused an inhibition zone of 10 mm. YMB was pH 6.5. Due to its pKa (4.76), potassium sorbate was partially charged. However, its antimicrobial activity was better than the sorbic acid ester. For this reason, no tests at different pH ranges have been conducted. The yeast growth inhibition of the ester was higher than the not esterified sorbic acid. Therefore, esterification could be an interesting strategy to increase antimicrobial activity of food preservatives. This strategy will be further investigated by designing other esters with increased water solubility and, consequently, antimicrobial activity. According to the results presented in Table 2, it can be noted that glycerol sorbate as well as sorbic acid showed a good inhibitory effect against the yeast specie tested but not for the bacterium. However, we found that glycerol sorbate, as well as sorbic acid, does not inhibit growth of the Gram-positive bacterium S. griseus. This is probably due to the more complex cell wall structure of Gram-positive bacteria, which has additional lipopolysaccharides on the outer surface that generally reduces the ability of most antibiotics and extracts to penetrate the bacterial cells [43]. Furthermore, it has been reported how some microorganisms, for example Acetobacter species, are able to degrade sorbic acid using it as carbon source [2].
The MIC values of glycerol sorbate on S. cerevisiae were on the range of 0.14 mg/mL (Table 3). Natural products presenting MICs of ≤1.00 mg/mL −1 are normally considerate noteworthy in terms of antimicrobial activity [44]. With this assumption, although potassium sorbate is the compound with the lowest MIC, the antimicrobial activity of the sorbic acid ester in this work results to be good for the control of the yeast S. cerevisiae. Moreover, the growth inhibition of the glycerol ester of sorbic acid resulted to be better compared to the not-esterified food preservative sorbic acid. Nevertheless, additional safety information regarding this ester is required to obtain approval as food, beverage, or cosmetic preservative. As widely discussed in the literature, health information will need to investigate the effects of these additives and the consequences of their long-term dietary consumption. Indeed, though the actuals salts benzoate and sorbate are widely used compounds for food and beverage preservation, there continues to be a concern about their complete safety. Cell culture studies and model organisms have shown some issues. Benzoate and sorbate are reported to results in chromosome aberrations in cultured human lymphocytes. Thus, exposure to these agents should be evaluated in terms of downsizing [45]. About this, the positive aspects reported in the literature regarding the use of monoglycerides as preservatives in food and drinks can be of great interest. Monoacylglycerols, like monocaprin and monolaurin, used as fruit juice preservatives, suppressed or prevented the growth of filamentous fungi in vitro [46]. Monoglycerides also inhibited the germination of bacteria like Bacillus cereus and Clostridium botulinum [47]. Model membrane-based biophysical measurement techniques showed that fatty acid and monoglycerides as antimicrobial destabilize phospholipid membranes, causing several direct and indirect inhibitory effects [48]. Therefore, the importance of developing monoglycerides-based preservatives, as well as deeply testing their effects on human safety, turns out to be very essential.

Conclusions
The employment of sorbic acid as a preservative in foods consisting of water is limited by its hydrophobic nature. In the present study, solventless lipase-catalyzed esterification of sorbic acid and glycerol was performed to obtain glycerol sorbate. The immobilized form of C. antarctica lipase B has proven to be a suitable catalyst to efficiently esterify glycerol and sorbic acid in reaction environment with water close to zero. To the best of our knowledge, 2,3-dihydroxypropyl-sorbate (glycerol sorbate) has been characterized ( 1 H-, 13 C-NMR and IR) for the first time. MS spectroscopy assays confirmed the effective esterification of sorbic acid. Lately, the antimicrobial activity of glycerol sorbate against S. griseus and S. cerevisiae has been evaluated, giving encouraging results for the control of the fungal model. With this study, we aimed to provide a scalable way to design synthesis production processes of bioactive compounds.