Edited by: Haifeng Zhao, South China University of Technology, China
Reviewed by: Wenbin Zhang, Jiangnan University, China; Farshad Darvishi, Alzahra University, Iran
This article was submitted to Industrial Biotechnology, a section of the journal Frontiers in Bioengineering and Biotechnology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Considering the growing demand for natural, unprocessed foods, it is evident that natural sweeteners are also very popular, especially for individuals suffering from diabetes or obesity. However, steviol glycoside-containing products are not so well received by a large number of consumers, due to their lingering bitter aftertaste. Although the exact correlation between glycoside structure and taste has not been elucidated, a study by
Another promising enzyme group regarding the modification of steviol glycosides is glucansucrases (EC 2.1.4.-). Glucansucrases are found only in lactic acid bacteria, and they are able to introduce α- 1,6 and α- 1,3 glycosidic bonds to steviol glycosides, with high yields (
The modification of stevioside can also be achieved by the action of the native enzyme present in stevia plant, UDP-glucosyltransferase (EC 2.4.1.-). These enzymes introduce β-glycosidic bonds to the steviol glycosides, which makes them ideal candidates for their biocatalytic modification, since these bonds are not hydrolyzed in the human body. However, the need for UDP- activated donors greatly increases the cost of the process. Nonetheless, a few interesting efforts were described, to develop a biocatalytic system for the
Regarding hydrolases, although they have been used extensively in the literature for various transglycosylation reactions, studies reporting the transglycosylation of steviol glycosides are sparse. β-Glucosidases and β-galactosidases have been used mainly for the hydrolysis of stevioside, aiming at the production of less glucosylated products (
In the present study, two novel glycosyl hydrolases were employed for the transglycosylation of stevioside and rebA.
Cellobiohydrolase I (CBH1) and endoglucanase 7 (EG7) were purchased from Megazyme (Bray, Co. Wicklow, Ireland). Avicel was purchased from Merck KGaA, Darmstadt, Germany. Stevioside (>85%) and Rebaudioside A (>98%) were purchased from TCI America (Boston, MA, United States). All other chemicals were purchased from Merck KGaA, Darmstadt, Germany, and they were of the highest purity available.
The acid whey that was used for the enzymatic production of GOS was provided by Delta Foods SA and it was obtained from production processes of Greek strained fat-free yogurt. The composition was as follows; total protein, 0.2% (w/w); fat, 0.07% (w/w); TSS 4.95% (w/w); ash, 0.70% (w/w). The initial lactose concentration was 3.4% (w/v), the initial monosaccharides (glucose and galactose) concentration was 0.79% (w/v) and the pH of the whey was 4.3 ± 0.25. In order to concentrate the lactose contained in the whey, a Hei-VAP value digital rotary evaporator (Heidolph Instruments GmbH & CO. KG) was used. The temperature of the water bath and the rotation speed were set to 52°C and 120 rpm respectively. After concentration, the final lactose concentration was 9.3% (w/v).
The transglycosylation of stevioside and RebA was performed in 1 mL reactions, in an Eppendorf thermomixer, under vigorous stirring.
The acid whey preparation was also utilized for transgalactosylation reactions, with lactose concentration of 9.3% (w/v).
For all reactions, 0.5 mL aliquots were removed from the reaction mixture at selected time intervals. The aliquots were immediately boiled for 10 min, filtered using 0.22 μm PVDFL filters and diluted suitably prior to analysis.
An LC-20AD HPLC system (Shimadzu) was employed for the analysis of the reaction products using a C18 CC 250/4.6 Nucleosil 100-5 column (Macherey-Nagel). The mobile phase consisted of 70% (v/v) phosphate buffer 10 mM pH 2.6 and 30% (v/v) acetonitrile. After sample injection (25 μL), chromatographic separation of steviol glycosides was performed with isocratic elution, at 1 mL min–1, while monitoring was performed at 210 nm using a UV–vis ProStar 335 Diode Array detector (Agilent Technologies). Stevioside and rebA concentrations were determined against a suitable calibration curve.
The mono-, di- and oligosaccharides contained in the aliquots were separated with an HPLC system, using a Microsorb-MV 100 NH2 (250 mm × 4.6 mm) column equipped with a Polaris 5 NH2 MetaGuard guard column (Agilent Technologies Sales & Services GmbH & Co. KG, Santa Clara, CA, United States). The eluent used for the separation was 75% (v/v) acetonitrile in water and the flow rate was set to 1.0 mL min–1. The sugars were detected using a Shimadzu RID-10A refractive index detector (Hewllet-Packard Company, Palo Alto, CA, United States). The identification and quantification of the sugars were performed using galactose, lactose and raffinose as standard compounds for mono-, di- and oligosaccharides, respectively (
In order to tentatively identify the steviol glycosides produced during the reactions, accurate data were obtained using Liquid Chromatography coupled to High Resolution Mass Spectrometry (LC-HRMS). An Ultrahigh-Performance Liquid Chromatography (UHPLC) system with a HPG-3400 pump (Dionex Ultimate 3000 RSLC, Thermo Fischer Scientific, Dreieich, Germany) interfaced to a Quadrupole Time-of-Flight (QTOF) mass spectrometer (Maxis Impact, Bruker Daltonics, Bremen, Germany) was employed.
A Hydrophilic Interaction Liquid Chromatography (HILIC) system was applied for the chromatographic separation of the examined transglycosylation products. An ACQUITY BEH Amide chromatographic column (2.1 × 100 mm, 1.7 μm) from Waters (Dublin, Ireland) was used, preceded by a guard column of the same material, thermostated at 40°C. The mobile phases used for the analysis consisted of deionized water (solvent A) and acetonitrile/water 95/5 (solvent B) both containing 1 mM ammonium formate and 0.01% formic acid. A gradient elution program was performed, started with 100% B (0–2 min), decreasing to 5% within 10 min (t:12 min) and kept constant for the following 5 min (t:17 min). Restorage of initial conditions was carried out within 0.1 min and a post-run of 8 min was programmed to re- equilibrate the column between analyses, resulting to a total analysis time of 25 min. The constant flow rate was set at 0.2 mL min–1 while the injection volume was 5 μL. Samples from enzymatic reactions were subjected to serial dilution 100-fold in ACN/water (80:20) prior to analysis.
HRMS analysis was achieved by a QTOF system equipped with an Electrospray Ionization source (ESI), operating in negative ionization mode, with the following ionization parameters: capillary voltage 2500 V (PI) and 3500 (NI); end plate offset, 500 V; nebulizer pressure 2 bar; drying gas 8 L min–1 and dry temperature 200°C. QTOF MS system was operated in broadband collision- induced dissociation (bbCID) acquisition mode, providing MS and MS/MS spectra within the same injection using two different collision energies. At low collision energy (4 eV), MS spectra were acquired and at high collision energy (25 eV), fragmentation is taking place at the collision cell resulting in MS/MS spectra. Recording of full scan mass spectra for each sample was elaborated in a mass range of 50–1000 m/z, setting a scan rate of 2 Hz. For the accuracy of the obtained mass spectra, MS external calibration was performed using as calibrant 10 mM sodium formate diluted in a mixture of water/isopropanol (1:1). The theoretical exact masses of calibration ions with formulas Na (NaCOOH) 1-14 in the range of 40-1500 Da were used for calibration. Additional internal calibration was also performed by a calibrant injection at the beginning of each chromatogram within 0.1-0.25 min.
Data acquisition was carried out with the HyStar and Compass software (Bruker Daltonics, Bremen, Germany). Mass spectra interpretation was performed by Data Analysis 4.4 software packages (Bruker Daltonics, Bremen, Germany). Since no reference standards were available for all the transglycosylated products, a suspect screening approach was applied for compounds’ detection and identification. An in-house suspect database was built including the all the possible transglycosylated products. Molecular formulas of the suspect analytes, pseudomolecular ions [M-H]– and possible adduct formations ([M + HCOOH-H]– etc.) occurring during the ionization processing were included in the database.
Full-scan chromatogram of each injected sample was screened for all the suspect products. Exploiting Data Analysis software, Extracted Ion Chromatograms (EIC) were created for all detected mass features and the obtained peaks were evaluated according to specific detection and identification criteria (peak area, intensity, mass accuracy, isotopic fitting and MS/MS fragmentation). Based on our previous study, for the accurate evaluation of the resulting chromatographic data, strict chromatographic parameters were set: peak area threshold was set >800, intensity threshold >200 and a signal to noise ratio at 3 (
Data analysis was performed with SigmaPlot v12.5 software (Systat Software, Inc., San Jose, CA, United States). Error bars represent the standard deviation of the mean value.
The β-galactosidase
Transglycosylation of steviol glycosides by
The conversion of both stevioside and RebA started almost immediately, and the lowest concentration was measured after 4 h of reaction in the case of stevioside, and after 24 h of reaction for RebA. After that, the concentration of both glycosides started to rise slowly, indicating a certain degree of hydrolysis of the transglycosylated products. The maximum conversion of stevioside and RebA was 27.7 ± 1.4% and 31.8 ± 0.5% respectively.
Considering the above HRMS-based workflow, during both transglycosylation reactions, different glycosylated products were detected and identified. The majority of the identified products were detected in two ionized forms ([M-H]– and [M + HCOOH-H]–), enhancing the identification points. The most abundant precursor ion was found to be the formate adduct [M + HCOOH-H]–. During the transglycosylation reaction using stevioside as acceptor, two different chromatographic peaks were detected in 6.75 and 7.22 min, respectively, corresponding to two different mono-glycosylated products with the same molecular formula (C44H70O23). Three di-glycosylated products (C50H80O28) were also detected, corresponding to retention times 7.28, 7.46, and 7.71 min. Respectively, during conversion of RebA two chromatographic peaks were detected, corresponding to the formation of mono- glycosylated products with molecular formula (C50H80O28). Chromatograms and annotated MS and MS/MS spectra - corresponding to the detected products are provided in
During RebA transglycosylation, the concentration of donor lactose was also monitored, and the results are shown in
The transglycosylation reaction was also tested in different concentrations of steviol glycosides, and the results are shown in
Conversion of different concentrations of
The β-glucosidase
Transglycosylation of steviol glycosides by
During the reaction with stevioside as the acceptor, three extra peaks appeared, corresponding to shorter retention times (5.6, 7.7, and 8.3 min), while for RebA two extra peaks appeared (8.4 and 12.8 min). The conversion of stevioside and RebA was monitored during the course of the reaction, and the results are shown in
In the case of RebA transglycosylation reaction, the concentration of donor cellobiose was also monitored, and the results are shown in
The transglycosylation reaction was also tested in different concentrations of steviol glycosides, and the results are shown in
Conversion of different concentrations of
For stevioside transglycosylation, the optimum conversion (34.6 ± 0.6%) was observed again for 24 h reaction, but in lower starting concentration, 2.5 g L–1, while the maximum concentration of consumed stevioside was observed in 10 g L–1 after 24 h reaction (2.3 ± 0.04 g L–1). For RebA transglycosylation, the optimum conversion (25.6 ± 1.2%) was observed for 2.5 g L–1, after 24 h of reaction, but the maximum concentration of consumed rebA was observed in 7.5 g L–1 after 48 h reaction (2.3 ± 0.4 g L–1).
The next step was to use low-cost industrial byproducts as the source of donor sugars for the transglycosylation of steviol glycosides. In the case of
For
In the present work, two glycosyl hydrolases, a β-galactosidase
All reactions were monitored for 72 h, in order to obtain a complete time profile of the reaction. The reactions were performed with different starting concentrations of acceptors, stevioside and rebA, in order to increase the conversion yields. Moreover, the reactions were also performed using low-cost industrial byproducts as donor sugars, acid whey as a source of lactose for
Regarding the transglycosylation performed by
β-Galactosidases were previously used for the deglycosylation of stevioside to yield less glycosylated products. For example, β-galactosidase from
β-Glucosidases also have been used widely for the hydrolysis of stevioside to less glycosylated products. Microbial β-glucosidases have been described in the literature, with the ability to cleave the β-1,2 glycosidic bond (
Regarding
Aside from the transglycosylation of steviol glycosides, and the simultaneous hydrolysis of their respective donor sugars, both hydrolases were also found to transglycosylate their donors, lactose and cellobiose, to the respective oligosaccharides. The transgalactosylating activity of
As mentioned earlier, many different enzyme activities have been used in previous studies for the modification of steviol glycosides. Many studies report the transglycosylation with CGtases, with starch or other sugars as donors. Product yields exceed 80% with the use of such systems (
Another advantage of the enzyme systems proposed here, is the possibility of exploiting low-cost industrial byproducts, as shown by our results. Specifically, for β-galactosidase, the use of acid whey as sugar donor resulted in similar yields with pure lactose. The development an enzymatic process for stevioside modification with acid whey as the donor, could be implicated on-site in already existing strained yogurt production plants, simultaneously valorizing the liquid byproducts, and producing a high-quality non-caloric sweetener as an additive to the dairy products produced on the same plant. The ultimate goal of the design and implementation of such bioprocesses is the compliance of existing plants and production lines with the principles of circular economy.
Regarding the
In the present work, a β-galactosidase
The original contributions presented in the study are included in the article/
AZ designed and performed experiments, analyzed data and wrote the original draft, with input from all authors. KC and AK performed experiments and analyses. NT supervised and designed experiments. ET designed and supervised the study and wrote the manuscript. All authors read and approved the final version of the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Delta Foods SA is gratefully acknowledged for providing the acid whey wastewater. Athanasios Limnaios is acknowledged for the preparation of concentrated acid whey wastewater.
The Supplementary Material for this article can be found online at: