Investigation of improved rebaudioside D solubility and the characteristics of an erythritol/rebaudioside D/fructose ternary complex

Abstract

We developed a ternary system using an erythritol/fructose melt to form a complex in which rebaudioside D (RebD) was dispersed. This ternary complex improved the solubility of RebD in water compared to RebD alone. In addition, the complex was observed to have a taste closer to sucrose. Scanning electron microscopy indicated that the ternary complex has finer surface crevices, and it exhibited nonuniform dissolution in a dissolution test. Examination of RebD dispersion in the solid by differential scanning calorimetry suggested that erythritol and fructose might have a solid solution composition. We confirmed that the solubility of RebD was improved in the multicomponent system as well as in the binary system. This useful finding will enable the production of a single composition with various advantages such as improved solubility and taste modifications; furthermore, it can achieve operational efficiency in an industrial-level manufacturing process.

Introduction

The leaves of Stevia rebaudiana Bertoni have long been used as a sweet herbⁱ⁾ (Khiraoui et al., 2017). Steviol glycosides are thought to be the primary zero-calorie sweetener components of the leaves, having a higher sweetness intensity compared to sucrose (Khiraoui et al., 2017). They are currently used in beverages, snacks, desserts, and tabletop sweeteners. Furthermore, stevia leaves have a wide variety of functionality effects (Wölwer-Rieck, 2012; Prata et al., 2017; Samuel et al., 2018).

According to the Food and Agriculture Organization of the United Nations 2017ⁱⁱ⁾, stevia leaves contain more than fifty types of steviol glycosides, and novel steviol glycosides are reported yearly. Among them, RebA and stevioside, which contain high ratios of steviol glycosides (approximately 4–13% w/w and 2–4% w/w, respectively; Momtazi-Borojeni et al., 2017) have been commercialized. Their sweetness intensity is 200–300 times greater than sucrose (Prakash et al., 2014), and both are widely used as alternative natural sweeteners because they contain no calories. However, compared to sugar, they have a less preferable flavor, especially with respect to bitterness and aftertaste (Hellfritsch et al., 2012; Prakash et al., 2014).

In recent years, sweeteners containing certain high-purity components such as RebD and RebM have been reported (Perrier et al., 2018; Prakash et al., 2014). In these components, one or two glucose units are conjugated with RebA, respectively, and the taste of these formulations is superior to both RebA and stevioside; the taste is reported to be close to sugar (Abelyan et al., 2011; Prakash et al., 2014).

However, RebD is insoluble in water, which presents a problem in the beverage production process, where a thick liquid syrup is prepared prior to adjustment to the final liquid by water dilution. For example, when producing a final product with a 10 sweetness/sucrose equivalent value (SEV) and making a 5-fold syrup, the SEV required in the upstream process is 50 SEV. When calculating a SEV of 5, for example, this indicates a 5% sucrose solution by weight, SEV being an equivalent value for the sweetness of sucrose. According to our previous study, the solubility of RebD in water is ∼600 mg/L at room temperature (Urai et al., 2020) and the sweetness intensity of RebD is 300 times that of sucrose. Therefore, if a 50-SEV syrup is required, the amount of RebD to be dissolved in water is ∼1 700 mg/L, having to be tripled to maximize the solubility of RebD. Because the solubility of RebD is insufficient for syrup making, improvement of the solubility is required.

Various approaches have been used to improve the solubility of RebD, which mainly involve the processing of RebD separately before combining it with other components (Babu et al., 2011; Pang et al., 2015; Upreti et al., 2011). Separate RebD processing can involve the well-known approach of investigating crystalline polymorphs (de Paula et al., 2013), which are compounds with different crystal structures resulting from different interactions between the same molecules. There are two RebD polymorphs: form α and form β (Urai et al., 2020). Moreover, solubility can be improved by increasing the amount of amorphous compound, i.e., a compound that has no regular crystal structure (Babu et al., 2011; Uchiyama et al., 2018). The creation of amorphous RebD by spray-drying or freeze-drying can improve the solubility by 0.3–5.0% in water and 0.1–2.5% at room temperature (Markosyan et al., 2011). However, amorphous substances generally have poor stability because there is a risk of dislocation to a crystalline form during long-term storage (Gupta et al., 2004).

In a preliminary study, we attempted to improve the solubility of RebD by dissolving RebD in an erythritol melt. This method aimed to improve RebD solubility in water by dispersing RebD into the solid phase of erythritol after solidification from a melt. The solubility was successfully increased approximately 6-fold to 4 000 mg/L. At the same time, the dissolution rate was also shown to be significantly improved compared to RebD alone. However, in the beverage manufacturing industry, it is common to add a variety of components to adjust the taste, flavor, and acidity and to prevent foaming. In addition, during manufacturing, it is desirable to combine a number of raw materials into one agent for operational efficiency. However, erythritol is known to have eutectic and solid-solution-system characteristics depending on the components it is mixed with (Kollau et al., 2018; Gunasekara et al., 2018). This means that the addition of other ingredients can cause a change in the structure of the composition, resulting in a change to the dissolution behavior. In the field of lipid crystallization, which uses melt crystallization technology, it has been reported that additional components can cause changes in the original crystal networks, resulting in changes in the properties of compositions (Daels et al., 2018; Sato, 2001). Therefore, the main objective of this study was to dissolve RebD in a solid of the multiple melt components (erythritol and fructose) and characterize the solubility potential of RebD.

Materials and Methods

Materials Rebaudioside D (95% purity w/w; Jining Renewal & Joint International, Shandong, China), erythritol (> 98% purity w/w; Mitsubishi-Chemical Foods Corp., Tokyo, Japan), and fructose (> 95% purity w/w; Nacalai Tesque, Kyoto, Japan) were used in this study. Liquid chromatography-mass spectrometry (LCMS-8050; Shimadzu Corp., Kyoto, Japan) was performed using RebA, RebB, RebC, RebF, stevioside, steviolbioside, dulcoside A, and rubusoside (FUJIFILM Wako Pure Chemical Corp., Japan); RebD, RebM, and RebN (ChromaDex Corp., USA); and RebG and RebI (Carbosynth Ltd., Berkshire, UK) standards.

Binary complexes of erythritol/RebD were produced using the following method. After melting 4 g of erythritol as a solvent at > 175 °C in an oil bath, 100 mg of RebD was added to the solvent and stirred for 2 min. The solvent and the dissolved RebD were then transferred to a preheated (60 °C) oil bath. Complexes of erythritol/RebD/fructose were produced by preparing 3.2 g of erythritol and 100 mg of RebD as above and then adding 0.8 g of fructose 1.5 min after the RebD was visually completely dissolved. After stirring for 0.5 min, the glass vials containing the molten complexes were transferred to a preheated (60 °C) oil bath. After the molten liquids had solidified, the complexes (2.4% w/w RebD) were used for the following solubility experiments.

Evaluation of solubility

Measurement of soluble RebD in water RebD, a physical mixture of erythritol/RebD/fructose, and the molten ternary complex were powdered and added to water at room temperature (RT), and adjusted to a RebD concentration of ∼4 000 mg/L (4 × 10⁻³ g/m³) for the calculation. After 5 min of stirring, a sample of the suspension was membrane-filtered (0.45 µm PTFE membrane filter; TOSC Japan Ltd., Tokyo, Japan). The amount of solubilized RebD was measured using LCMS (See Fig. 1).

Fig. 1.

The confirmation of RebD concentration of ternary complex (erythritol/RebD/fructose) compared to RebD alone, physical mixture of ternary components and binary complex of erythritol/RebD.

Investigation of erythritol/RebD/fructose complex dissolution rate The prepared binary and ternary complexes were suspended in water to make a ∼4 000 mg/L (4 × 10⁻³ g/m³) RebD solution and then stirred at 150 rpm at RT. As a control, the same procedure was performed for RebD alone. Samples were obtained at different times and filtered using a 0.45-µm PTFE membrane filter. The RebD concentration in the permeate was measured using LCMS (See Fig. 2).

Fig. 2.

The Dissolution rate of a complex of erythritol/RebD/fructose compared to RebD and binary complex.

Taste evaluation A taste evaluation was undertaken to confirm the effect of the erythritol/RebD/fructose ternary complex. RebA and RebD were prepared (according to the Materials section), and sucrose was used as a control. The binary and ternary complexes, containing ∼3.6% RebD, were prepared as solids (according to the Materials section). Each complex and RebA were adjusted to 300 mg/L in solution by adding water in terms of RebA or RebD concentration. A 9 SEV sucrose solution was prepared as control. The taste evaluation was performed by five well-trained panelists. Results were scored on a scale of 0–6 based on sweetness onset, sweetness lingering, bitterness, bitterness lingering, and an overall evaluation in comparison with the control, for which the score was a score of 3. A score of 6 meant that the sweetness increases quickly, sweetness lingering is short, bitterness intensity is low, bitterness lingering is short, the overall taste satisfaction is high (See Fig. 3).

Fig. 3.

Taste evaluation of ternary complex of erythritol/RebD/fructose.

SEM imaging Scanning electron microscopy (SEM) was performed to observe the surface structure of the complexes. The samples were attached to double-sided conductive tape and fixed on the microscope mount. The dried samples were coated with gold to reduce the charge effect. The experiments were performed using a JSM-6510 microscope (JEOL Ltd., Tokyo, Japan) with an acceleration voltage of 10 kV. For erythritol, the surface of the solid was observed after it had been fused at 175 °C and subsequently solidified (see Fig. 4).

Fig. 4.

The surface structure of erythritol/RebD/fructose complex.

After preparing the erythritol/RebD and erythritol/RebD/fructose complex samples, they were roughly powdered using a pestle. Approximately 1 g of sample was added to water and stirred at 400 rpm. After 60 s and 120 s, the solid was obtained from the solution by filtration using a 0.45 µm membrane filter. The solids were then dried for more than 24 h at 60 °C to remove any liquid. The surfaces of the samples were observed via SEM (for details of the procedure, see Fig. 5A).

Fig. 5.

(A) The procedure of sample preparation for SEM; (B) the surface of complex of erythritol/RebD/fructose.

Confirmation of the dispersion of steviol glycosides and RebD in the solid phase A melt in which RebD was dissolved was prepared by the method described in the Materials section, which was then poured into a metal vessel having a diameter of 7.0 cm and preheated at 60 °C to prepare the complexes. Then, the complex was photographed from above (See Fig. 6A). The complex shown in Fig. 6A was divided into “outside” and “inside” as shown in the schematic diagram (see Fig. 6B), and each surface was shaved using a knife. The steviol glycosides (i.e., RebA, RebB, RebC, RebD, RebE, RebF, RebG, RebM, RebN, RebI, stevioside, steviolbioside, dulcoside A, and rubusoside) and RebD in the samples were quantified using LCMS, and the ratio of both contained in the solid phase was calculated. Finally, the change in dispersibility was confirmed by determining how much the content of the target component changed for the “outside” compared to the “inside”. Dispersibility was expressed as follows:   

Fig. 6.

(A) Appearance of binary and ternary complex and photographing direction; (B) the schematic diagram classifying “inside” and “outside” in the sample; Dispersibility of SGs (C) and RebD (D).

where SGs represents a steviol glycoside. The lower the dispersibility value, the better the steviol glycoside or RebD was dispersed in the solid.

Characteristics of the molten mixed solvent, erythritol, and fructose It was confirmed that in the ternary complex, the solid-phase dispersion of RebD was relatively enhanced compared to that in the binary system. To investigate the dispersion factors, the characteristics of the erythritol/fructose solid used as the solvent for RebD were confirmed. Shifts in the endothermic peak, along with changes in the ratios of fructose to erythritol, were measured using differential scanning calorimetry (DSC). The complexes were prepared in line with the method described in the Materials section, except that RebD was not added and the ratio of fructose to erythritol changed from 10 to 35%. The measurement conditions were as follows: temperature range, ambient temperature to 135–160 °C; rate of temperature increase, 5 K/min; nitrogen flow rate, 50 mL/min; and aluminum sealed sample pans.

Quantitative analysis of the steviol glycosides The quantity of steviol glycosides in the solid or soluble RebD in the permeate was analyzed using LCMS as an electronic detection system operated at RT with the following settings: column, Shim-pack XR-ODS II [(2.0 mm (i.d.) × 150 mm (L); Shimadzu Corp., Kyoto, Japan)]; mobile-phase composition of 30% acetonitrile/70% Milli-Q water with 0.1% formic acid; flow rate, 0.34 ml/min. Quantification of RebA, RebB, RebC, RebD, RebE, RebF, RebG, RebM, RebN, RebI, stevioside, steviolbioside, dulcoside A, and rubusoside was performed using a standard calibration curve.

Results and Discussion

Potentiality of the erythritol/RebD/fructose complex for water solubility To clarify the difference in solubility with RebD alone and compare the solubility with the binary system, the solubility of the three-component system was confirmed. Inclusion of 2.4% RebD in the total weight of the three components erythritol, RebD, and fructose in the ternary complex increased the solubility of RebD in water to greater than 4 × 10⁻³ g/m³, which was similar to the binary system of erythritol/RebD (see Fig. 1). This favorably contrasts with the solubility of RebD alone (0.83 × 10⁻³ g/m³). Although the dissolution rate of the ternary system was slightly higher than that of the binary system during the initial dissolution stage, they were comparatively the same (see Fig. 2). Both systems, therefore, could significantly improve RebD solubility compared to RebD alone.

Taste evaluation of the ternary complex Sensory evaluation was performed to examine the effect of the ternary complex on taste (see Fig. 3). As reported in previous studies, this study showed that RebD may have a taste closer to sucrose than RebA (Hellfritsch et al., 2012; Prakash et al., 2014). In terms of the overall evaluation, the binary and ternary complexes were observed to taste more like sucrose than RebD. In particular, the ternary complex tended to show improved bitterness characteristics compared to the binary complex. Therefore, the ternary complex was shown to have both excellent solubility and taste properties.

SEM imaging of the surface structure of the ternary complex To confirm whether the binary and ternary complexes had a similar surface structure, the surface of each composition type was examined using SEM. The binary complex exhibited many surface crevices compared to erythritol and RebD. Similarly, the surface of the ternary complex also exhibited crevices, but they were finer than those observed on the binary complex (see Fig. 4). However, this structural difference would have little influence on the dissolution behavior.

Confirmation of the ternary complex dissolution behavior To further understand the dissolution behavior, the solids were sampled during a dissolution test, with the dried surfaces then observed using SEM (Fig. 5A). The binary and ternary complexes exhibited completely different dissolution behaviors, as shown in Fig. 5B. In contrast to the binary complex, the surface of the ternary complex showed nonuniform dissolution.

Dispersion of steviol glycosides and RebD in the complexes Given that the binary and ternary systems were shown to have different solidification (see Fig. 5B) and that changes in solidification are thought to affect the solid dispersion of RebD, the dispersibility of the RebD dissolved in the erythritol/fructose melt (i.e., the ternary complex) was investigated. The prepared ternary complex (see Fig. 6A) was divided into “outside” and “inside” as shown in the schematic diagram of Fig. 6B, and the ratio of RebD included in each was confirmed. In addition, because RebD shows some decomposition behavior at high temperatures, the localization of the other steviol glycosides, including RebD, was also confirmed. For comparison, the same operation was performed for the binary complex. This confirmed that the steviol glycosides contained in the ternary complex were better dispersed than those in the binary complex (see Fig. 6C). Furthermore, the dispersibility of the RebD contained in the ternary complex was much higher than for the binary complex (see Fig. 6D). This implies that the steviol glycosides and RebD were well-dispersed in the solid phase of the ternary complex.

Characteristics of the solvent used in the experiment To examine possible factors contributing to the observed difference in RebD dissolution behavior, we examined the characteristics of erythritol/fructose used as the solvent for the production of the ternary complex. DSC was used to observe endothermic heat absorption peaks to confirm the properties of the erythritol/fructose complex (see Fig. 7). Upon blending 20% fructose, a shift in the endothermic peak was observed with an unconfirmed second peak. However, upon blending 11.7, 27.7, and 34.4% fructose, one or two peaks were confirmed at different temperatures. Incidentally, it was not possible to completely solidify the complex with further addition of fructose (data not shown). This indicates that the erythritol/fructose complex may have the characteristics of a solid solution, and in that sense, it is considered that RebD may show greater dispersion in the solid of the melt component having a solid solution composition.

Fig. 7.

The characteristic of complex of erythritol and fructose.

Fructose ratio (%) in weight is 11.1% (a), 20.0% (b), 27.7% (c), and 34.4% (d).

Conclusions

The effect of adding fructose, which is widely used in beverages, to molten erythritol on the solubility of RebD was examined, as solid-phase properties have been reported to change depending on the components added to erythritol. In addition to improving the solubility compared to RebD alone, we confirmed that the erythritol/RebD/fructose ternary complex exhibited increased RebD dissolution and rate of dissolution, and was equivalent to that of the erythritol/RebD binary system. We also observed that the ternary complex had a taste closer to sucrose. However, the dissolution behavior of the ternary complex was found to differ from that of the binary complex. In particular, the dissolution tendency of the ternary complex exhibited nonuniform dissolution, and the RebD was well-dispersed in the complex. Furthermore, DSC suggested that the erythritol/fructose complex had a solid solution feature. Although further analysis of the RebD dispersion in a crystal structure and examination of control methods are required in the future, the ternary complex is superior to RebD alone in solubility and taste, and has the same effect as the binary complex. Therefore, for industrial manufacturing, the ternary system would be more useful in terms of the optimization of product characteristics, such as solubility, taste, and production efficiency.

Supplement data

A Sweetness onset

B Sweet lingering

C Overall evaluation

D Bitterness

E Bitter lingering

Supplement data

Taste evaluation score of each taste factor compared to Sucrose (Score: 3).

(A) Sweetness onset; (B) Sweet lingering; (C) Overall evaluation; (D) Bitterness; (E) Bitter lingering.

References

•  Abelyan,  V.,  Markosyan,  A., and  Abelyan,  L. (2011). High-purity rebaudiosideD. US patent No 20110092684, Apr. 21.
•  Allen,  B. C. and  Isserow,  S. (1957). Segregation at the eutectic temperature. Acta Mater., 5, 465-472.
•  Babu,  N. J. and  Nangia,  A. (2011). Solubility advantage of amorphous drugs and pharmaceutical cocrystals. Cryst. Growth Des., 11, 2662-2679.
•  Daels,  E.,  Goderis,  B.,  Steen,  L., and  Foubert,  I. (2018). Influence of adding a commercial phytosterol ester mixture on the ‘equilibrium’ crystallization behavior of palm oil. Food Structure, 11, 1-8.
•  Gunasekara,  S. N.,  Chiu,  J. N. W.,  Martin,  V., and  Hedströmb,  P. (2018). The experimental phase diagram study of the binary polyols system erythritol-xylitol. Sol. Energ. Mat. Sol. C., 174, 248-262.
•  Gupta,  P.,  Chawla,  G., and  Bansal,  A. K. (2004). Physical stability and solubility advantage from amorphous celecoxib: the role of thermodynamic quantities and molecular mobility. Mol. Pharm., 1, 406-413.
•  Hellfritsch,  C.,  Brockhoff,  A.,  Stähler,  F.,  Meyerhof,  W., and  Hafmann,  T. (2012). Human psychometric and taste receptor responses to steviol glycosides. J. Agric. Food Chem., 60, 6782-6793.
•  Khiraoui,  A.,  Hasib,  A.,  Faiz,  C. A.,  Amchra,  F.,  Bakha,  M., and  Boulli,  A. (2017). Stevia Rebaudiana Bertoni (Honey Leaf): A magnificent natural bio-sweetener, biochemical composition, nutritional and therapeutic values. J. Nat. Sci. Res., 7, 75-85.
•  Kollau,  L. J. B. M.,  Vis,  M.,  van den Bruinhorst,  A.,  Esteves,  A. C. C., and  Tuinier,  R. (2018). Quantification of the liquid window of deep eutectic solvents. Chem. Commun., 54, 13351-13354.
•  Markosyan,  A. and  Purkayastha,  S. (2011). Highly soluble rebaudioside d. EP patent No 2651960, Dec. 12.
•  Momtazi-Borojeni,  A. A.,  Esmaeili,  S. A.,  Abdollahi,  E., and  Sahebkar,  A. (2017). A Review on the pharmacology and toxicology of steviol glycosides extracted from Stevia rebaudiana. Curr. Pharm. Des., 23, 1616-1622.
•  Pang,  S.,  Ma,  C.,  Zhang,  N., and  He,  L. (2015). Investigation of the solubility enhancement mechanism of rebaudioside D using a solid dispersion technique with potassium sorbate as a carrier. Food Chem., 174, 564-570.
•  de Paula,  K.,  Camí,  G. E.,  Brusau,  E. V.,  Narda,  G. E., and  Ellena,  J. (2013). Mebendazole Mesylate Monohydrate: A New Route to Improve the Solubility of Mebendazole Polymorphs. J. Pharm. Sci., 102, 3528-3538.
•  Perrier,  J. D.,  Mihalov,  J. J., and  Carlson,  S. J. (2018). FDA regulatory approach to steviol glycosides. Food Chem. Toxicol., 122, 132-142.
•  Prakash,  I.,  Markosyan,  A., and  Bunders,  C. (2014). Development of next generation stevia sweetener: rebaudioside M. Foods, 3, 162-175.
•  Prata,  C.,  Zambonin,  L.,  Rizzo,  B.,  Maraldi,  T.,  Angeloni,  C.,  Sega,  F. V. D.,  Fiorentini,  D., and  Hrelia,  S. (2017). Glycosides from stevia rebaudiana bertoni possess insulinmimetic and antioxidant activities in rat cardiac fibroblasts. Oxid. Med. Cell. Longev., Article ID 3724545, 13 pages.
•  Samuel,  P.,  Ayoob,  K. T.,  Magnuson,  B. A., and  Wölwer-Rieck,  U. (2018). Stevia leaf to stevia sweetener: exploring its science, benefits, and future potential. J. Nutr., 148, 1186S-1205S.
•  Sato,  K. (2001). Crystallization behaviour of fats and lipids - a review. Chem. Eng. Sci., 56, 2255-2265.
•  Uchiyama,  H.,  Wada,  Y.,  Takamatsu,  M.,  Kadota,  K., and  Tozuka,  Y. (2018). Improved solubility of quercetin by preparing amorphous solid with transglycosylated rutin and isoquercitrin. Environ. Control. Biol., 4, 161-165.
•  Upreti,  M.,  Strassburger,  K.,  Chen,  Y. L.,  Wu,  S., and  Prakash,  I. (2011). Solubility enhancement of steviol glycosides and characterization of their inclusion complexes with gamma-cyclodextrin. Int. J. Mol. Sci., 12, 7529-7553.
•  Urai,  S.,  Kudo,  S. and  Takiyama,  H. (2020). Development of rebaudioside D polymorphs with improved solubility. Food Sci. Technol. Res., 26, 17-23.
•  Wölwer-Rieck,  U. (2012). The Leaves of stevia rebaudiana (bertoni), their constituents and the analyses thereof: a review. J. Agric. Food Chem., 60, 886-895.
• Food and Agriculture Organization of the United Nations, (2017). Steviol glycosides, 84th JECFA - Chemical and Technical Assessment (CTA).
• i) https://www.researchgate.net/profile/Jaime_Teixeira_Da_Silva/publication/283348088_Stevia_Stevia_rebaudiana_Bertoni_Futuristic_View_of_the_Sweeter_Side_of_Life/links/5635d44b08ae88cf81bc03c6/Stevia-Stevia-rebaudiana-Bertoni-Futuristic-View-of-the-Sweeter-Side-of-Life.pdf?origin=publication_detail (Apr. 18, 2020).
• ii) http://www.fao.org/3/BU609en/bu609en.pdf (Apr. 18, 2020).