- Email: [email protected]
Drug solubilization mechanism of α-glucosyl stevia by NMR spectroscopy
International Journal of Pharmaceutics 465 (2014) 255–261
Contents lists available at ScienceDirect
International Journal of Pharmaceutics journal homepage: www.elsevier.com/locate/ijpharm
Drug solubilization mechanism of spectroscopy
a-glucosyl stevia by NMR
Junying Zhang a,b,1, Kenjirou Higashi a,1, Keisuke Ueda a , Kazunori Kadota c, Yuichi Tozuka c , Waree Limwikrant a,d , Keiji Yamamoto a , Kunikazu Moribe a, * a
Graduate School of Pharmaceutical Sciences, Chiba University, 1-8-1, Inohana, Chuo-ku, Chiba, 260-8675, Japan School of Traditional Chinese Medicine, China Pharmaceutical University, 24, Tongjiaxiang, Nanjing, 210009, China Osaka University of Pharmaceutical Sciences, 4-20-1, Nasahara, Takatsuki, Osaka, 569-1094, Japan d Department of Manufacturing Pharmacy, Faculty of Pharmacy, Mahidol University, 447 Sri Ayudhya Road, Ratchatewi, Bangkok, 10400, Thailand b c
A R T I C L E I N F O
A B S T R A C T
Article history: Received 12 October 2013 Received in revised form 6 January 2014 Accepted 28 January 2014 Available online 4 February 2014
We investigated the drug solubilization mechanism of a-glucosyl stevia (Stevia-G) which was synthesized from stevia (rebaudioside-A) by transglycosylation. 1H and 13C NMR peaks of Stevia-G in water were assigned by two-dimensional (2D) NMR experiments including 1H-1H correlation, 1H-13C heteronuclear multiple bond correlation, and 1H-13C heteronuclear multiple quantum coherence spectroscopies. The 1H and 13C peaks clearly showed the incorporation of two glucose units into rebaudioside-A to produce Stevia-G, supported by steviol glycoside and glucosyl residue assays. The concentration-dependent chemical shifts of Stevia-G protons correlated well with a mass-action law model, indicating the selfassociation of Stevia-G molecules in water. The critical micelle concentration (CMC) was 12.0 mg/mL at 37 C. The aggregation number was 2 below the CMC and 12 above the CMC. Dynamic light scattering and 2D 1H-1H nuclear Overhauser effect spectroscopy (NOESY) NMR experiments demonstrated that Stevia-G self-associated into micelles of a few nanometers in size with a core-shell structure, containing a kaurane diterpenoid-based hydrophobic core and a glucose-based shell. 2D 1H-1H NOESY NMR measurements also revealed that a poorly water-soluble drug, naringenin, was incorporated into the hydrophobic core of the Stevia-G micelle. The Stevia-G self-assembly behavior and micellar drug inclusion capacity can achieve significant enhancement in drug solubility. ã 2014 Elsevier B.V. All rights reserved.
Keywords: a-Glucosyl stevia Micelle Aggregation number Nuclear magnetic resonance Nuclear Overhauser effect spectroscopy Drug solubility improvement
1. Introduction More than 40% of the failures in new drug development is attributed to poor pharmaceutical properties, particularly water insolubility. The enhancement of drug solubility has been studied for many decades, and a number of techniques have been employed, e.g., the use of cyclodextrins (Higashi et al., 2009), solid dispersion systems (Dahlberg et al., 2010; Kojima et al., 2012; Sinha et al., 2010), nanoparticles (Zhang et al., 2012), and micelles (Mikhail and Allen, 2010). Recently, we have examined the feasibility of
Abbreviations: CAC, critical aggregation concentration; CMC, critical micelle concentration; COSY, correlation spectroscopy; [5_TD$IF]dmic, micelle chemical shifts; dmon, monomer chemical shift; DLS, dynamic light scattering; HMBC, heteronuclear multiple bond correlation spectroscopy; HMQC, heteronuclear multiple quantum coherence spectroscopy; NaTC, sodium taurocholate; NOESY, nuclear Overhauser effect spectroscopy; NRG, naringenin; SDS, sodium dodecyl sulfate; Stevia-G, a-glucosyl stevia; TSP, [6_TD$IF]3-(trimethylsilyl) propionic-2,2,3,3-d4 acid[7_TD$IF], sodium salt. * Corresponding author. Tel.: +81 43 226 2865; fax: +81 43 226 2867. E-mail address: [email protected] (K. Moribe). 1 [3_TD$IF]These authors contributed equally to this study. 0378-5173/$ – see front matter ã 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ijpharm.2014.01.035
transglycosylated food additives as new pharmaceutical excipients in order to improve the dissolution and bioavailability of poorly water-soluble drugs (Tozuka et al., 2010, 2012; Uchiyama et al., 2010a, 2012; Zhang et al., 2011). The transglycosylated food additives are synthesized by applying enzymatic transglycosylation to bioactive compounds using glycosyltransferases (Kittl and Withers, 2010; Kometani, 2010; Wang and Huang, 2009). Among these transglycosylated food additives, a-glucosyl stevia (Stevia-G, Fig. 1A), which is the enzymatically transglycosylated product of stevia via a-glycosyltransferase (Varuzhan et al., 2006), is promising as a new pharmaceutical excipient. Stevia is a herb belonging to the Compositae family, estimated to comprise 150–300 species (Grashoff et al., 1972; King and Robinson, 1967). Stevioside, rebaudioside-A, rebaudioside-C, and dulcoside-A are known as the major components. Stevia, which has no significant adverse effects, has been used for 20 years as a sweetener and sugar substitute. Since Stevia-G is sweeter than stevia, it can be useful in masking the bitter taste of drugs. We previously reported that spray-dried particles of a poorly water-soluble drug (flurbiprofen and probucol) and Stevia-G showed a significant enhancement in both the dissolution and absorption of the drug without any toxic
256
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
effects on Caco-2 cells (Uchiyama et al., 2010b). Interestingly, Stevia-G and surfactant formed a hybrid nanocomposite in water to improve drug dissolution (Uchiyama et al., 2012). It was demonstrated that a mixture of 0.1% sodium dodecyl sulfate (SDS) and 1% Stevia-G solution had little cytotoxicity to Caco-2 cells, whereas 0.1% SDS solution showed high toxicity. Fluorescence investigations in the previous study suggested a drug solubilization mechanism in which Stevia-G forms micellelike nanostructures in water and dissolves the drugs in these structures (Uchiyama et al., 2011). However, the solubilization mechanism at the molecular level was not yet understood. To understand the drug solubilization phenomenon in detail and efficiently design pharmaceutical formulations, it is necessary to clearly determine the solubilization mechanism. In recent decades, NMR has become a crucial method for chemical structure identification, and has been further applied in the conformational determination of self-assembled aggregates. For example, the static (chemical environment, degree of association, size, and shape) and the dynamic (molecular mobility, kinetics of aggregation, solubilization) properties of the aggregates are characterized by various NMR techniques (Emin et al., 2007). The proton chemical shift is sensitive to subtle changes in the local environment and used for detecting molecular association. Two-dimensional (2D) 1H-1H nuclear Overhauser effect spectroscopy (NOESY), based on the dipole-dipole interactions between nuclei in spatial proximity, is a powerful tool for studying the arrangement of aggregates (Denkova et al., 2009; Schedlbauer et al., 2009; Yang et al., 2009). In the present study, the chemical structure of Stevia-G, its detailed conformation in water, and the specific drug solubilization
[(Fig._1)TD$IG]
mechanism were investigated by NMR spectroscopy. The Stevia-G, which was the transglycosylation product of rebaudioside-A, had a purity over 98%. Rebaudioside-A (Fig. 1A) has the basic skeleton of a kaurane diterpenoid with three glucose moieties attached to the C13 hydroxyl and one glucose moiety in the form of an ester at C-19 (Steinmetz and Lin, 2009). 1D 1H and 13C NMR and 2D 1H-1H correlation spectroscopy (COSY), 1H-13C heteronuclear multiple bond correlation spectroscopy (HMBC), and 1H-13C heteronuclear multiple quantum coherence spectroscopy (HMQC) NMR experiments were employed to identify the chemical structure of SteviaG. The self-assembly process of Stevia-G in water was assessed by observing the chemical shift changes of stevia protons depending on the concentration. 2D 1H-1H NOESY NMR experiments were performed to evaluate the spatial structure of Stevia-G in water and the localization of a model drug, naringenin (NRG, Fig. 1B). 2. Materials and methods 2.1. Materials Stevia-G, synthesized from rebaudioside-A in high purity (over 98%), was a kind gift from the Toyo Sugar Refining Co., Ltd. (Tokyo, Japan). NRG was purchased from Tokyo Kasei Co., Ltd. (Tokyo, Japan) and used without further purification. Commercial deuterium oxide (D2O, 99.9%, Aldrich, St. Louis, MO, USA) was used as received. All other chemicals and solvents were of reagent grade. The chemical structures and atom numbering of Stevia-G and NRG are shown in Fig. 1. 2.2. Preparation of spray-dried powders Powders of NRG-loaded Stevia-G were prepared using a spraydrying method. To prepare samples by this method, NRG (500 mg) and Stevia-G (5 g) were dissolved in an ethanol/water solution (8:2 v/v). This solution was fed to a spray dryer (GS31; Yamato, Tokyo, Japan) at a rate of 10 mL/min, and sprayed into the chamber from a nozzle with a diameter of 406 mm at a pressure of 0.13 MPa. The inlet and outlet temperatures of the drying chamber were maintained at 120 and 70 C, respectively. All spray-dried powders were dried in a desiccator with blue silica gel under reduced pressure for 1 day before their physicochemical properties were tested. 2.3. Assay of steviol glycosides and glucosyl residues in Stevia-G The determination analysis followed those prescribed by Japan’s Specifications and Standards for Food Additives (8th Edition). Stevia-G (1.0 g) was dissolved in water (50 mL). The solution was transferred into a resin column (2.5 cm diameter) packed with acrylic acid ester resin (25 mL, XAD-7, Organo Co., Ltd., Japan). The solution was drained from the column at a rate of less than 3 mL/ min, and then the column was washed with water (250 mL) to remove the unreacted glucoses. The Stevia-G adsorbed on the column was eluted using 50% (v/v) ethanol (250 mL) at a flow rate below 3 mL/min. The eluted solution was evaporated to remove ethanol; then, glucoamylase was added in the solution to fully react. The obtained solution was assayed by HPLC to determine the content of steviol glycosides. The glucose content in the solution was determined using a glucose assay kit (Wako Pure Chemical Industries, Ltd., Japan). 2.4. Particle size analysis
Fig. 1. Chemical structures and atom numberings of (A) a-glucosyl stevia (Stevia-G) and (B) naringenin (NRG).
The volumetric particle size distribution was determined by a dynamic light scattering (DLS) method using a Microtrac UPA1 (Nikkiso Co., Ltd., Japan). The mean particle size was the average
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
value of three times repetition. The detection range of the UPA was 0.8 nm–6.5 mm. 2.5. NMR measurements All NMR spectra were acquired on an ECX-400 spectrometer (9.4 Tesla, JEOL Resonance, Tokyo, Japan) using a 1H/X probe (JEOL Resonance, Tokyo, Japan) at 37 C. The samples dissolved in D2O were put into 5 mm NMR sample tubes and equilibrated at 37 C for ca. 30 min prior to the NMR experiments. Chemical shifts were referenced to the internal signal of 0.05% 3-(trimethylsilyl) propionic-2,2,3,3-d4 acid, sodium salt (TSP) at 0.0 ppm. Resonance assignments for Stevia-G were made according to 2D 1H-1H COSY, 1 H-13C HMBC, and 1H-13C HMQC NMR spectra (Figs. S1–4) and reference data (Chaturvedula and Prakash, 2011a; Chaturvedula and Prakash, 2011b; Steinmetz and Lin, 2009). The peaks of NRG were assigned based on prior reports (Ficarra et al., 2002; Saveyn et al., 2009; Wade et al., 1990). The 2D 1H-1H NOESY NMR spectra were recorded with 1024 data points in the t2 time domain, 256 t1increments, and 32 scans. A mixing time of 0.5 s and an acquisition time of 0.15 s were used. A sine apodization function was applied in both dimensions before Fourier transformation. The spectra were zero-filled in the t1 dimension to give a 1024 1024 data matrix in the frequency domain. 3. Results and discussion 3.1. Chemical structure identification of Stevia-G
[(Fig._2)TD$IG]
The Stevia-G used in this study is the transglycosylation product of rebaudioside-A. To understand the solubilization mechanism of Stevia-G, there is a need to clarify the number and positions of the
Fig. 2. One-dimensional (1D) (A) 1H NMR and (B) transglycosylation.
13
257
glucosyl units attached to the rebaudioside-A during the transglycosylation process. Fig. 2 shows the 1D 1H and 13C NMR spectra of a Stevia-G solution (30 mg/mL). The NMR peaks of Stevia-G were assigned according to the 2D 1H-1H COSY, 1H-13C HMQC, and 1H-13C HMBC NMR spectra (Figs. S1–4), as well as previous reports (Chaturvedula and Prakash, 2011a; Chaturvedula and Prakash, 2011b; Chaturvedula and Prakash, 2011c; Steinmetz and Lin, 2009). The 1D 1H NMR spectrum of Stevia-G (Fig. 2A) showed the presence of six anomeric protons at d 4.75, 4.81, 4.88, 5.39, 5.42, and 5.48 ppm. The corresponding anomeric carbons at d 98.7, 105.0, 104.9, 102.6, 102.4, and 96.8 ppm were observed in the 1D 13C NMR spectrum (Fig. 2B). These NMR results indicated the presence of six glucose units in the Stevia-G structure. Since rebaudioside-A has four glucose units in its structure (Steinmetz and Lin, 2009), two additional glucose units were added to stevia during the transglycosylation process. An assay of steviol glycosides and glucosyl residues in Stevia-G was performed to confirm the number of glucose units introduced into rebaudioside-A. This result showed that Stevia-G had 72.05% steviol glycosides and 25.26% glucosyl residues. The number of introduced glucose units was calculated to be two by taking into account the molecular weight of rebaudioside-A (967) and glucose unit (162); therefore, two additional glucose units were attached to rebaudioside-A upon the formation of Stevia-G. Rebaudioside-A has the basic skeleton of a kaurane diterpenoid with three glucose moieties attached to the C-13 hydroxyl and one glucose moiety in the form of an ester at C-19 (Steinmetz and Lin, 2009). The intermolecular transglycosylation reaction occurs exclusively or mainly at the C4-hydroxyl group of the non-reducing end glucose residue because of the acceptor specificity (Chiba et al., 1975; Suzuki and Suzuki, 1991). This 1,4-a-transglycosylation is commonly applied for the preparation of stevia glucosylated
C NMR spectra of Stevia-G at 30 mg/mL in D2O at 37 C. 1000 00 and 1000 000 are derived from the glucoses transferred by
258
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
[(Fig._4)TD$IG]
products (Fukunaga et al., 1989), and the Stevia-G in this study was prepared following this method. Therefore, the transglycosylation reaction could occur at the free C40 -, C4000 -, or C40000 -hydroxyl groups in the rebaudioside-A structure (Fig. 1A). Further experiments including 2D 1H-1H COSY, 2D 1H-1H total correlation spectroscopy, 1 H-13C HMBC, and 1H-13C HMQC NMR measurements were performed. However, extreme signal overlap in the glucose region in both the 1H and 13C NMR spectra made it difficult to identify the specific hydroxyl group at which the two additional glucosyl units were attached.
3.2. Nanostructure formed by Stevia-G in water The proton chemical shift (ppm) is sensitive to changes in the molecular environment and is effective in detecting molecular associations (Okano et al., 1997; Saveyn et al., 2009; Shimizu et al., 2003). Fig. 3 shows the 1D 1H NMR spectra of Stevia-G over gradual concentration increases (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, and 50 mg/mL). The chemical shifts of all the protons changed with increasing concentration, indicating the intermolecular interactions during the Stevia-G aggregation process. The NMR chemical shift variation corresponding to different nuclei has been previously evaluated for critical micelle concentration (CMC) or critical aggregation concentration (CAC) detection (Andrade-Dias et al., 2007; Okano et al., 1997). This method can also apply to the investigation of large molecular weight polymer micelles (Emin et al., 2007). In this study, the CMC was used as the index parameter. Below the CMC, the protons are surrounded by solvent molecules and the observed chemical shift, dobs, is the chemical shift of the monomer, dmon. Above the CMC, when the exchange of monomers and micelles in the solution is fast on the NMR time scale, dobs is described as the weighted average of the monomer (dmon) and the micelle chemical shifts (dmic). If the monomer concentration is constant above the CMC, the observed chemical shift is described using the mass-action law model (Eq. (1)) (Andrade-Dias et al., 2007; Farías et al., 2009; Okano et al., 1997), as follows, where C is the total concentration.
Fig. 4. Plots of (A) dobs against 1/C and (B) log (C(dmon dobs) against log (C(dobs dmic) for the H-20 peak from the 1H NMR spectra in Fig. 3.
dobs ¼ dmic
CMC ðdmic dmon Þ C
(1)
Fig. 4A presents plots of dobs at the H-20 peak as a function of the reciprocal of the total Stevia-G concentration. The approximation straight line of the experiment data fitted to the theoretical mode. The intersection of the two lines corresponds to 1/C = 1/CMC. The CMC of Stevia-G was calculated to be approximately 12 mg/mL at 37 C. This value was slightly lower than previously obtained results (16 mg/mL) by surface tension and fluorescence measurements (Uchiyama et al., 2011). The difference may be attributed to the different measurement principle.
[(Fig._3)TD$IG]
Fig. 3. 1H NMR spectra of Stevia-G solution at concentrations of 1–50 mg/mL recorded in D2O at 37 C.
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
[(Fig._5)TD$IG]
259
The chemical shift variation was further employed to calculate the aggregation number of Stevia-G. The single-step equilibrium model, in which molecules exist as either monomers or a single type of aggregate with a fixed aggregation number, was applied. The aggregation equilibrium can be expressed as in Eq. (2) K ¼ ½An ½An
(2)
In the above equation, K is the equilibrium constant, n is the aggregation number, and A and An are the monomeric or aggregated molecules. Combined with Eq. (1) (Okano et al.,1997), Eq. (2) can be rearranged and represented by Eq. (3) (Luchetti, 2000). logðCðdmon dobs ÞÞ ¼ nlogðCðdobs dmic ÞÞ þ logðnKÞ þð1 nÞlogðdmon dobs Þ
(3)
A plot of log [C(dmon dobs)] as a function of log [C(dobs dmic)] consisted of two straight lines—one above and the other below the CMC. The slope yielded the aggregation number n. Fig. 4B shows the plot for the H-20 peak of Stevia-G according to Eq. (3). Two straight lines, whose intersection represents the CMC, indicate the presence of two different equilibria. Below the CMC, the aggregation number was 2, indicating dimer formation for Stevia-G. Above the CMC, the Stevia-G aggregation number increased to 12. Since the plots for other proton peaks (data not shown) were similar to that for the H20 peak, a single type of Stevia-G aggregation above the CMC was proposed. We reported previously that the amount of indomethacin dissolved from a physical mixture of indomethacin/Stevia-G was proportional to the amount of Stevia-G loaded, suggesting the solubilization capacity of Stevia-G even below the CMC (Uchiyama et al., 2011). Li and McGown (1994, 1993) reported the aggregation and solubilization of the bile salt sodium taurocholate (NaTC) in water. The dimer of NaTC was able to solubilize a drug by forming a “sandwich model”, in which the drug was solubilized between upper and lower NaTC molecules. Thus, a dimeric structure for Stevia-G in water can explain why drug solubilization occurred even below the CMC. The size of the Stevia-G micelles was determined by the DLS method. The particle size increased from ca. 1.8 nm to 3 nm as a function of the concentration (Fig. 5), confirming the formation of aggregates. The increased particle size was in accordance with the increased aggregation number. To further investigate the Stevia-G conformation in water, 2D 1 H-1H NOESY NMR measurements of Stevia-G solution at 30 mg/mL were performed (Fig. 6). The NOE cross peaks are the result of crossrelaxation between neighboring protons that are spatially close to each other (roughly, less than 5 Å). The NOESY spectrum provided a number of correlation peaks representing not only the intramolecular interactions, but also the intermolecular interactions with another Stevia-G molecule. However, there were fewer NOE crosspeaks between the basic skeleton and the glucose group of Stevia-G. The Stevia-G micelle could have a core–shell structure in which the basic kaurane diterpenoid skeleton was segregated from the aqueous exterior, forming an inner core encased by a shell of glucose units. Conventional core-forming segments such as alkyl or aryl chains form cores through a combination of intermolecular forces (Kataoka et al., 2001). Similarly, the kaurane diterpenoid skeleton of Stevia-G showed appreciable self-associating behavior driven by hydrophobic interactions, and was stabilized by hydrophilic glucose units. Notably, the methyl protons (H-18, 20) and olefinic proton (H-17) showed some NOE correlations with the glucose protons. These protons, projected from the basic skeleton, could possibly be located proximate to the surface of the hydrophobic core. The NOE correlation of the anomeric protons (H-10 and H-100 ) with the basic skeleton protons arose because these
Fig. 5. Particle size distributions of Stevia-G depending on the concentration from 5–50 mg/mL measured by the dynamic light scattering (DLS) method.
anomeric protons connecting to the basic skeleton could also be spatially close to the surface of the hydrophobic core. 3.3. Solubilization mechanism of poorly water-soluble drugs Fig. 7 shows the 1D 1H NMR spectra of untreated Stevia-G and NRG-loaded Stevia-G in D2O at 37 C. The NRG peaks were clearly detected in the spectrum (Fig. 7B). In general, solution NMR cannot be used to analyze molecules with low mobility, e.g., crystals, because of signal broadening. The solution NMR properties suggested that the NRG in Stevia-G solution was in a molecularly dispersed state rather than the crystal state. The solubility of NRG in Stevia-G solution was determined at 37 C from the integration values of the 1H peaks using 0.05% TSP as an internal standard. The concentration of NRG in Stevia-G solution was calculated as 1.78 mg/mL, which was almost 36 times higher than the solubility of untreated NRG at 0.05 mg/mL. The enhancement in the apparent NRG solubility was because of the incorporation of the NRG into the Stevia-G micelle. NRG-loaded Stevia-G exhibited upfield shifts in the proton peaks of the kaurane diterpenoid skeleton compared
260
[(Fig._6)TD$IG]
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
[(Fig._8)TD$IG]
Fig. 8. Two-dimensional 1H-1H NOESY NMR spectrum of NRG-loaded Stevia-G at a Stevia-G concentration of 30 mg/mL in D2O at 37 C. Fig. 6. Two-dimensional (2D) 1H-1H nuclear Overhauser effect spectroscopy (NOESY) NMR spectrum of Stevia-G solution at 30 mg/mL in D2O at 37 C.
with the untreated Stevia-G, whereas no significant change was observed in peaks of the glucose protons (Fig. 7). This upfield shift demonstrated the interaction of the kaurane diterpenoid skeleton of Stevia-G with NRG. Two-dimensional 1H-1H NOESY NMR spectroscopy was used to investigate the location of the NRG in the Stevia-G micelle. Fig. 8 shows the 2D 1H-1H NOESY NMR spectrum of NRG-loaded Stevia-G in D2O. The key NOE cross-peaks used to deduce the possible interaction sites between the Stevia-G and NRG are highlighted. These new cross-peaks were not observed in the NOESY spectrum of the pure micelle of Stevia-G without the drug (Fig. 6). The aromatic protons of the NRG (H-B0 , H-F0, H-H, H-F, H-C0 and H-E0 ) exhibited NOE correlations with those of the kaurane diterpenoid skeleton (H-18 and H-20) in Stevia-G. Additionally, the H-H and H-F of NRG exhibited cross-peaks with H-14 (1), H-15 (1), and H-17 of Stevia-G. Meanwhile, no NOE cross-peaks between the NRG protons and the glucose units of Stevia-G were found. The observed NOE interactions clearly represented the interaction of NRG with the hydrophobic core of the Stevia-G micelle. All NMR observations
[(Fig._7)TD$IG]
undoubtedly confirmed that the solubilized NRG mainly accumulated and was stabilized in the Stevia-G hydrophobic core. 4. Conclusions The drug solubilization mechanism of Stevia-G, which was synthesized from rebaudioside-A, was successfully characterized by NMR spectroscopy. The peak assignments of the 1H and 13C peaks of Stevia-G in water were achieved by 2D NMR spectroscopic techniques such as 1H-1H COSY, 1H-1H HMBC, and 1H-1H HMQC. The number of anomeric peaks of Stevia-G in the 1D 1H and 13C spectra indicated that two glucose units were attached to rebaudioside-A during the transglycosylation process. The assay of steviol glycosides and glucosyl residues confirmed the introduction of these two glucoses into rebaudioside-A. The concentration-dependent variation of proton chemical shifts revealed the self-association of Stevia-G in water. The CMC of Stevia-G in water at 37 C was determined to be around 12 mg/mL. The aggregation number was 2 below the CMC and 12 above the CMC. Dimer formation in Stevia-G could enable the solubilization of drugs even below the CMC. The mean particle size of the Stevia-G micelle determined by DLS measurement was around 3 nm. 2D 1H-1H NOESY NMR experiments indicated that the Stevia-G micelle consisted of a kaurane diterpenoid skeleton core and a glucose shell. NRG was effectively trapped into the hydrophobic core of the Stevia-G micelle, which significantly enhanced the apparent solubility of NRG. The obtained results provide a better understanding of the drug solubilization mechanism of Stevia-G. New pharmaceutical formulations using Stevia-G can be designed efficiently based on the knowledge presented here. Acknowledgements
Fig. 7. One-dimensional 1H NMR spectra of (A) untreated Stevia-G and (B) NRGloaded Stevia-G at a Stevia-G concentration of 30 mg/mL in D2O at 37 C.
This study was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (Monbukagakusho), Japan (25504018). We thank Toyo Sugar Refining Co., Ltd. for the kind gift of a-glycosyl transferase-treated stevia.
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ijpharm.2014.01.035. References Andrade-Dias, C., Lima, S., Teixeira-Dias, J.J., 2007. From simple amphiphilic to surfactant behavior: analysis of 1H NMR chemical shift variations. Journal of Colloid and Interface Science 316, 31–36. Chaturvedula, V.S.P., Prakash, I., 2011a. Diterpene glycosides from Stevia rebaudiana. Journal of Medicinal Plants Research 5, 4838–4842. Chaturvedula, V.S.P., Prakash, I., 2011b. A new diterpene glycoside from Stevia rebaudiana venkata sai. Molecules 16, 2937–2943. Chaturvedula, V.S.P., Prakash, I., 2011c. Isolation, NMR spectral assignments and partial synthesis of rebaudioside B; a sweet component of Stevia rebaudiana. International Journal of Biological and Medical Research 2, 1090–1093. Chiba, S., Okada, S., Kitahata, S., Shimomura, T., 1975. A new trisaccharide, 2a-maltosyl-glucose, synthesized by the transglycosylation reaction of cyclodextrin glycosyltransferase. Agricultural and Biological Chemistry 39, 2353–2357. Dahlberg, C., Millqvist-Fureby, A., Schuleit, M., Furo, I., 2010. Relationships between solid dispersion preparation process, particle size and drug release – an NMR and NMR microimaging study. European Journal of Pharmaceutics and Biopharmaceutics 76, 311–319. Denkova, P.S., Van Lokeren, L., Willem, R., 2009. Mixed micelles of Triton X-100, sodium dodecyl dioxyethylene sulfate, and synperonic l61 investigated by NOESY and diffusion ordered NMR spectroscopy. Journal of Physical Chemistry B 113, 6703–6709. Emin, S.M., Denkova, P.S., Papazova, K.I., Dushkin, C.D., Adachi, E., 2007. Study of reverse micelles of di-isobutyl-phenoxy-ethoxy-ethyl-dimethyl-benzyl ammonium methacrylate in benzene by nuclear magnetic resonance spectroscopy. Journal of Colloid and Interface Science 305, 133–141. Farías, T., de Ménorval, L.C., Zajac, J., Riveraa, A., 2009. Solubilization of drugs by cationic surfactants micelles: conductivity and 1H NMR experiments. Colloids and Surfaces A: Physicochemical Engineering Aspects 345, 51–57. Ficarra, R., Tommasini, S., Raneri, D., Calabro, M.L., Di Bella, M.R., Rustichelli, C., Gamberini, M.C., Ficarra, P., 2002. Study of flavonoids/beta-cyclodextrins inclusion complexes by NMR, FT-IR, DSC, X-ray investigation. Journal of Pharmaceutical and Biomedical Analysis 29, 1005–1014. Fukunaga, Y., Miyata, T., Nakayasu, N., Mizutani, K., Kasai, R., Tanaka, D., 1989. Enzymic transglucosylation products of stevioside: separation and sweetnessevaluation. Agricultural and Biological Chemistry 53, 1603–1607. Grashoff, L.J., Bierner, W.M., Northington, K.D., 1972. Chromosome numbers in north and central American compositae. Brittonia 24, 379–394. Higashi, K., Ideura, S., Waraya, H., Moribe, K., Yamamoto, K., 2009. Incorporation of salicylic acid molecules into the intermolecular spaces of g-cyclodextrinpolypseudorotaxane. Crystal Growth and Design 9, 4243–4246. Kataoka, K., Harada, A., Nagasaki, Y., 2001. Block copolymer micelles for drug delivery: design, characterization and biological significance. Advanced Drug Delivery Reviews 47, 113–131. King, D., Robinson, E., 1967. Aspects of the development of Moniliformis dubius. Journal of Parasitology 53, 142–149. Kittl, R., Withers, S.G., 2010. New approaches to enzymatic glycoside synthesis through directed evolution. Carbohydrate Research 345, 1272–1279. Kojima, T., Higashi, K., Suzuki, T., Tomono, K., Moribe, K., Yamamoto, K., 2012. Stabilization of a supersaturated solution of mefenamic acid from a solid dispersion with EUDRAGIT1 EPO. Pharmaceutical Research 29, 2777–2791. Kometani, T., 2010. Production of novel bioactive compounds by enzymes, and their application to food. Pure and Applied Chemistry 82, 269–287. Li, G., McGown, L.B., 1993. A new approach to polydispersity studies of sodium taurocholate and sodium taurodeoxycholate aggregates using dynamic fluorescence anisotropy. Journal of Physics and Chemistry 97, 6745–6752. Li, G., McGown, L.B., 1994. Model for bile salt micellization and solubilization from studies of a “polydisperse” array of fluorescent probes and molecular modeling. Journal of Physics and Chemistry 98, 13711–13719.
261
Luchetti, L., 2000. NMR investigation on the various aggregates formed by a gemini chiral surfactant. Langmuir 16, 161–165. Mikhail, A.S., Allen, C., 2010. Poly(ethylene glycol)-b-poly(e-caprolactone) micelles containing chemically conjugated and physically entrapped docetaxel: synthesis, characterization, and the influence of the drug on micelle morphology. Biomacromolecules 11, 1273–1280. Okano, L.T., Seoud, O.A.E., Halstead, T.K., 1997. A proton NMR study on aggregation of cationic surfactants in water: effects of the structure of the headgroup. Colloid and Polymer Science 275, 138–145. Saveyn, P., Cocquyt, E., Zhu, W., Sinnaeve, D., Haustraete, K., Martins, J.C., Van der Meeren, P., 2009. Solubilization of flurbiprofen within non-ionic tween 20 surfactant micelles: a 19F and 1H NMR study. Physical Chemistry Chemical Physics 11, 5462–5468. Schedlbauer, A., Coudevylle, N., Auer, R., Kloiber, K., Tollinger, M., Konrat, R., 2009. Autocorrelation analysis of NOESY data provides residue compactness for folded and unfolded proteins. Journal of the Americal Chemical Society 131, 6038–6039. Shimizu, S., Pires, P.A.R., Fish, H., Halstead, T.K., Seoud, O.A.E., 2003. Proton and carbon-13 NMR study of the aggregation of benzyl-(2-acylaminoethyl) dimethylammonium chloride surfactants in D2O. Physical Chemistry Chemical Physics 5, 3489–3497. Sinha, S., Ali, M., Baboota, S., Ahuja, A., Kumar, A., Ali, J., 2010. Solid dispersion as an approach for bioavailability enhancement of poorly water-soluble drug ritonavir. AAPS PharmSciTech 11, 518–527. Steinmetz, W.E., Lin, A., 2009. NMR studies of the conformation of the natural sweetener rebaudioside A. Carbohydrate Research 344, 2533–2538. Suzuki, Y., Suzuki, K., 1991. Enzymatic formation of 4G-a-D-glucopyranosyl-rutin. Agricultural and Biological Chemistry 55, 181–187. Tozuka, Y., Kishi, J., Takeuchi, H., 2010. Anomalous dissolution property enhancement of naringenin from spray-dried particles with a-glucosylhesperidin. Advanced Powder Technology 21, 305–309. Tozuka, Y., Higashi, K., Morita, T., Nishiwaka, M., Uchiyama, H., Zhang, J., Moribe, K., Nishiwaka, K., Takeuchi, H., Yamamoto, K., 2012. Transglycosylated rutin-specific non-surface-active nanostructure affects absorption enhancement of flurbiprofen. European Journal of Pharmaceutics and Biopharmaceutics 82, 120–126. Uchiyama, H., Tozuka, Y., Imono, M., Takeuchi, H., 2010a. Improvement of dissolution and absorption properties of poorly water-soluble drug by preparing spray-dried powders with alpha-glucosyl hesperidin. International Journal of Pharmaceutics 392, 101–106. Uchiyama, H., Tozuka, Y., Imono, M., Takeuchi, H., 2010b. Transglycosylated stevia and hesperidin as pharmaceutical excipients: dramatic improvement in drug dissolution and bioavailability. European Journal of Pharmaceutics and Biopharmaceutics 76, 238–244. Uchiyama, H., Tozuka, Y., Asamato, F., Takeuchi, H., 2011. Fluorescence investigation of a specific structure formed by aggregation of transglycosylated stevias: solubilizing effect of poorly water-soluble drugs. European Journal of Pharmaceutical Sciences 43, 71–77. Uchiyama, H., Tozuka, Y., Nishikawa, M., Takeuchi, H., 2012. Nanocomposite formation between alpha-glucosyl stevia and surfactant improves the dissolution profile of poorly water-soluble drug. International Journal of Pharmaceutics 30, 183–186. Varuzhan, H.A., Vahe, T.G., Avetik, A.M., Mariam, O.A., Lidia, A.A., 2006. Extraction, separation and modification of sweet glycosides from the stevia rebaudiana plant. United States Patent 7838044. Wade, K.E., Wilson, I.D., Troke, J.A., Nicholson, J.K., 1990. 19F and 1H magnetic resonance strategies for metabolic studies on fluorinated xenobiotics: application to flurbiprofen [2-(2-fluoro-4-biphenylyl)propionic acid]. Journal of Pharmaceutical and Biomedical Analysis 8, 401–410. Wang, L.X., Huang, W., 2009. Enzymatic transglycosylation for glycoconjugate synthesis. Current Opinion in Chemical Biology 13, 592–600. Yang, W., Li, Y., Cheng, Y., Wu, Q., Wen, L., Xu, T., 2009. Evaluation of phenylbutazone and poly(amidoamine) dendrimers interactions by a combination of solubility, 2D-NOESY NMR, and isothermal titration calorimetry studies. Journal of Pharmaceutical Sciences 98, 1075–1085. Zhang, J., Higashi, K., Limwikrant, W., Moribe, K., Yamamoto, K., 2012. Molecularlevel characterization of probucol nanocrystal in water by in situ solid-state NMR spectroscopy. International Journal of Pharmaceutics 423, 571–576. Zhang, J., Tozuka, Y., Uchiyama, H., Higashi, K., Moribe, K., Takeuchi, H., Yamamoto, K., 2011. NMR investigation of a novel excipient, a-glucosylhesperidin, as a suitable solubilizing agent for poorly water-soluble drugs. Journal of Pharmaceutical Sciences 100, 4421–4431.
Contents lists available at ScienceDirect
International Journal of Pharmaceutics journal homepage: www.elsevier.com/locate/ijpharm
Drug solubilization mechanism of spectroscopy
a-glucosyl stevia by NMR
Junying Zhang a,b,1, Kenjirou Higashi a,1, Keisuke Ueda a , Kazunori Kadota c, Yuichi Tozuka c , Waree Limwikrant a,d , Keiji Yamamoto a , Kunikazu Moribe a, * a
Graduate School of Pharmaceutical Sciences, Chiba University, 1-8-1, Inohana, Chuo-ku, Chiba, 260-8675, Japan School of Traditional Chinese Medicine, China Pharmaceutical University, 24, Tongjiaxiang, Nanjing, 210009, China Osaka University of Pharmaceutical Sciences, 4-20-1, Nasahara, Takatsuki, Osaka, 569-1094, Japan d Department of Manufacturing Pharmacy, Faculty of Pharmacy, Mahidol University, 447 Sri Ayudhya Road, Ratchatewi, Bangkok, 10400, Thailand b c
A R T I C L E I N F O
A B S T R A C T
Article history: Received 12 October 2013 Received in revised form 6 January 2014 Accepted 28 January 2014 Available online 4 February 2014
We investigated the drug solubilization mechanism of a-glucosyl stevia (Stevia-G) which was synthesized from stevia (rebaudioside-A) by transglycosylation. 1H and 13C NMR peaks of Stevia-G in water were assigned by two-dimensional (2D) NMR experiments including 1H-1H correlation, 1H-13C heteronuclear multiple bond correlation, and 1H-13C heteronuclear multiple quantum coherence spectroscopies. The 1H and 13C peaks clearly showed the incorporation of two glucose units into rebaudioside-A to produce Stevia-G, supported by steviol glycoside and glucosyl residue assays. The concentration-dependent chemical shifts of Stevia-G protons correlated well with a mass-action law model, indicating the selfassociation of Stevia-G molecules in water. The critical micelle concentration (CMC) was 12.0 mg/mL at 37 C. The aggregation number was 2 below the CMC and 12 above the CMC. Dynamic light scattering and 2D 1H-1H nuclear Overhauser effect spectroscopy (NOESY) NMR experiments demonstrated that Stevia-G self-associated into micelles of a few nanometers in size with a core-shell structure, containing a kaurane diterpenoid-based hydrophobic core and a glucose-based shell. 2D 1H-1H NOESY NMR measurements also revealed that a poorly water-soluble drug, naringenin, was incorporated into the hydrophobic core of the Stevia-G micelle. The Stevia-G self-assembly behavior and micellar drug inclusion capacity can achieve significant enhancement in drug solubility. ã 2014 Elsevier B.V. All rights reserved.
Keywords: a-Glucosyl stevia Micelle Aggregation number Nuclear magnetic resonance Nuclear Overhauser effect spectroscopy Drug solubility improvement
1. Introduction More than 40% of the failures in new drug development is attributed to poor pharmaceutical properties, particularly water insolubility. The enhancement of drug solubility has been studied for many decades, and a number of techniques have been employed, e.g., the use of cyclodextrins (Higashi et al., 2009), solid dispersion systems (Dahlberg et al., 2010; Kojima et al., 2012; Sinha et al., 2010), nanoparticles (Zhang et al., 2012), and micelles (Mikhail and Allen, 2010). Recently, we have examined the feasibility of
Abbreviations: CAC, critical aggregation concentration; CMC, critical micelle concentration; COSY, correlation spectroscopy; [5_TD$IF]dmic, micelle chemical shifts; dmon, monomer chemical shift; DLS, dynamic light scattering; HMBC, heteronuclear multiple bond correlation spectroscopy; HMQC, heteronuclear multiple quantum coherence spectroscopy; NaTC, sodium taurocholate; NOESY, nuclear Overhauser effect spectroscopy; NRG, naringenin; SDS, sodium dodecyl sulfate; Stevia-G, a-glucosyl stevia; TSP, [6_TD$IF]3-(trimethylsilyl) propionic-2,2,3,3-d4 acid[7_TD$IF], sodium salt. * Corresponding author. Tel.: +81 43 226 2865; fax: +81 43 226 2867. E-mail address: [email protected] (K. Moribe). 1 [3_TD$IF]These authors contributed equally to this study. 0378-5173/$ – see front matter ã 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ijpharm.2014.01.035
transglycosylated food additives as new pharmaceutical excipients in order to improve the dissolution and bioavailability of poorly water-soluble drugs (Tozuka et al., 2010, 2012; Uchiyama et al., 2010a, 2012; Zhang et al., 2011). The transglycosylated food additives are synthesized by applying enzymatic transglycosylation to bioactive compounds using glycosyltransferases (Kittl and Withers, 2010; Kometani, 2010; Wang and Huang, 2009). Among these transglycosylated food additives, a-glucosyl stevia (Stevia-G, Fig. 1A), which is the enzymatically transglycosylated product of stevia via a-glycosyltransferase (Varuzhan et al., 2006), is promising as a new pharmaceutical excipient. Stevia is a herb belonging to the Compositae family, estimated to comprise 150–300 species (Grashoff et al., 1972; King and Robinson, 1967). Stevioside, rebaudioside-A, rebaudioside-C, and dulcoside-A are known as the major components. Stevia, which has no significant adverse effects, has been used for 20 years as a sweetener and sugar substitute. Since Stevia-G is sweeter than stevia, it can be useful in masking the bitter taste of drugs. We previously reported that spray-dried particles of a poorly water-soluble drug (flurbiprofen and probucol) and Stevia-G showed a significant enhancement in both the dissolution and absorption of the drug without any toxic
256
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
effects on Caco-2 cells (Uchiyama et al., 2010b). Interestingly, Stevia-G and surfactant formed a hybrid nanocomposite in water to improve drug dissolution (Uchiyama et al., 2012). It was demonstrated that a mixture of 0.1% sodium dodecyl sulfate (SDS) and 1% Stevia-G solution had little cytotoxicity to Caco-2 cells, whereas 0.1% SDS solution showed high toxicity. Fluorescence investigations in the previous study suggested a drug solubilization mechanism in which Stevia-G forms micellelike nanostructures in water and dissolves the drugs in these structures (Uchiyama et al., 2011). However, the solubilization mechanism at the molecular level was not yet understood. To understand the drug solubilization phenomenon in detail and efficiently design pharmaceutical formulations, it is necessary to clearly determine the solubilization mechanism. In recent decades, NMR has become a crucial method for chemical structure identification, and has been further applied in the conformational determination of self-assembled aggregates. For example, the static (chemical environment, degree of association, size, and shape) and the dynamic (molecular mobility, kinetics of aggregation, solubilization) properties of the aggregates are characterized by various NMR techniques (Emin et al., 2007). The proton chemical shift is sensitive to subtle changes in the local environment and used for detecting molecular association. Two-dimensional (2D) 1H-1H nuclear Overhauser effect spectroscopy (NOESY), based on the dipole-dipole interactions between nuclei in spatial proximity, is a powerful tool for studying the arrangement of aggregates (Denkova et al., 2009; Schedlbauer et al., 2009; Yang et al., 2009). In the present study, the chemical structure of Stevia-G, its detailed conformation in water, and the specific drug solubilization
[(Fig._1)TD$IG]
mechanism were investigated by NMR spectroscopy. The Stevia-G, which was the transglycosylation product of rebaudioside-A, had a purity over 98%. Rebaudioside-A (Fig. 1A) has the basic skeleton of a kaurane diterpenoid with three glucose moieties attached to the C13 hydroxyl and one glucose moiety in the form of an ester at C-19 (Steinmetz and Lin, 2009). 1D 1H and 13C NMR and 2D 1H-1H correlation spectroscopy (COSY), 1H-13C heteronuclear multiple bond correlation spectroscopy (HMBC), and 1H-13C heteronuclear multiple quantum coherence spectroscopy (HMQC) NMR experiments were employed to identify the chemical structure of SteviaG. The self-assembly process of Stevia-G in water was assessed by observing the chemical shift changes of stevia protons depending on the concentration. 2D 1H-1H NOESY NMR experiments were performed to evaluate the spatial structure of Stevia-G in water and the localization of a model drug, naringenin (NRG, Fig. 1B). 2. Materials and methods 2.1. Materials Stevia-G, synthesized from rebaudioside-A in high purity (over 98%), was a kind gift from the Toyo Sugar Refining Co., Ltd. (Tokyo, Japan). NRG was purchased from Tokyo Kasei Co., Ltd. (Tokyo, Japan) and used without further purification. Commercial deuterium oxide (D2O, 99.9%, Aldrich, St. Louis, MO, USA) was used as received. All other chemicals and solvents were of reagent grade. The chemical structures and atom numbering of Stevia-G and NRG are shown in Fig. 1. 2.2. Preparation of spray-dried powders Powders of NRG-loaded Stevia-G were prepared using a spraydrying method. To prepare samples by this method, NRG (500 mg) and Stevia-G (5 g) were dissolved in an ethanol/water solution (8:2 v/v). This solution was fed to a spray dryer (GS31; Yamato, Tokyo, Japan) at a rate of 10 mL/min, and sprayed into the chamber from a nozzle with a diameter of 406 mm at a pressure of 0.13 MPa. The inlet and outlet temperatures of the drying chamber were maintained at 120 and 70 C, respectively. All spray-dried powders were dried in a desiccator with blue silica gel under reduced pressure for 1 day before their physicochemical properties were tested. 2.3. Assay of steviol glycosides and glucosyl residues in Stevia-G The determination analysis followed those prescribed by Japan’s Specifications and Standards for Food Additives (8th Edition). Stevia-G (1.0 g) was dissolved in water (50 mL). The solution was transferred into a resin column (2.5 cm diameter) packed with acrylic acid ester resin (25 mL, XAD-7, Organo Co., Ltd., Japan). The solution was drained from the column at a rate of less than 3 mL/ min, and then the column was washed with water (250 mL) to remove the unreacted glucoses. The Stevia-G adsorbed on the column was eluted using 50% (v/v) ethanol (250 mL) at a flow rate below 3 mL/min. The eluted solution was evaporated to remove ethanol; then, glucoamylase was added in the solution to fully react. The obtained solution was assayed by HPLC to determine the content of steviol glycosides. The glucose content in the solution was determined using a glucose assay kit (Wako Pure Chemical Industries, Ltd., Japan). 2.4. Particle size analysis
Fig. 1. Chemical structures and atom numberings of (A) a-glucosyl stevia (Stevia-G) and (B) naringenin (NRG).
The volumetric particle size distribution was determined by a dynamic light scattering (DLS) method using a Microtrac UPA1 (Nikkiso Co., Ltd., Japan). The mean particle size was the average
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
value of three times repetition. The detection range of the UPA was 0.8 nm–6.5 mm. 2.5. NMR measurements All NMR spectra were acquired on an ECX-400 spectrometer (9.4 Tesla, JEOL Resonance, Tokyo, Japan) using a 1H/X probe (JEOL Resonance, Tokyo, Japan) at 37 C. The samples dissolved in D2O were put into 5 mm NMR sample tubes and equilibrated at 37 C for ca. 30 min prior to the NMR experiments. Chemical shifts were referenced to the internal signal of 0.05% 3-(trimethylsilyl) propionic-2,2,3,3-d4 acid, sodium salt (TSP) at 0.0 ppm. Resonance assignments for Stevia-G were made according to 2D 1H-1H COSY, 1 H-13C HMBC, and 1H-13C HMQC NMR spectra (Figs. S1–4) and reference data (Chaturvedula and Prakash, 2011a; Chaturvedula and Prakash, 2011b; Steinmetz and Lin, 2009). The peaks of NRG were assigned based on prior reports (Ficarra et al., 2002; Saveyn et al., 2009; Wade et al., 1990). The 2D 1H-1H NOESY NMR spectra were recorded with 1024 data points in the t2 time domain, 256 t1increments, and 32 scans. A mixing time of 0.5 s and an acquisition time of 0.15 s were used. A sine apodization function was applied in both dimensions before Fourier transformation. The spectra were zero-filled in the t1 dimension to give a 1024 1024 data matrix in the frequency domain. 3. Results and discussion 3.1. Chemical structure identification of Stevia-G
[(Fig._2)TD$IG]
The Stevia-G used in this study is the transglycosylation product of rebaudioside-A. To understand the solubilization mechanism of Stevia-G, there is a need to clarify the number and positions of the
Fig. 2. One-dimensional (1D) (A) 1H NMR and (B) transglycosylation.
13
257
glucosyl units attached to the rebaudioside-A during the transglycosylation process. Fig. 2 shows the 1D 1H and 13C NMR spectra of a Stevia-G solution (30 mg/mL). The NMR peaks of Stevia-G were assigned according to the 2D 1H-1H COSY, 1H-13C HMQC, and 1H-13C HMBC NMR spectra (Figs. S1–4), as well as previous reports (Chaturvedula and Prakash, 2011a; Chaturvedula and Prakash, 2011b; Chaturvedula and Prakash, 2011c; Steinmetz and Lin, 2009). The 1D 1H NMR spectrum of Stevia-G (Fig. 2A) showed the presence of six anomeric protons at d 4.75, 4.81, 4.88, 5.39, 5.42, and 5.48 ppm. The corresponding anomeric carbons at d 98.7, 105.0, 104.9, 102.6, 102.4, and 96.8 ppm were observed in the 1D 13C NMR spectrum (Fig. 2B). These NMR results indicated the presence of six glucose units in the Stevia-G structure. Since rebaudioside-A has four glucose units in its structure (Steinmetz and Lin, 2009), two additional glucose units were added to stevia during the transglycosylation process. An assay of steviol glycosides and glucosyl residues in Stevia-G was performed to confirm the number of glucose units introduced into rebaudioside-A. This result showed that Stevia-G had 72.05% steviol glycosides and 25.26% glucosyl residues. The number of introduced glucose units was calculated to be two by taking into account the molecular weight of rebaudioside-A (967) and glucose unit (162); therefore, two additional glucose units were attached to rebaudioside-A upon the formation of Stevia-G. Rebaudioside-A has the basic skeleton of a kaurane diterpenoid with three glucose moieties attached to the C-13 hydroxyl and one glucose moiety in the form of an ester at C-19 (Steinmetz and Lin, 2009). The intermolecular transglycosylation reaction occurs exclusively or mainly at the C4-hydroxyl group of the non-reducing end glucose residue because of the acceptor specificity (Chiba et al., 1975; Suzuki and Suzuki, 1991). This 1,4-a-transglycosylation is commonly applied for the preparation of stevia glucosylated
C NMR spectra of Stevia-G at 30 mg/mL in D2O at 37 C. 1000 00 and 1000 000 are derived from the glucoses transferred by
258
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
[(Fig._4)TD$IG]
products (Fukunaga et al., 1989), and the Stevia-G in this study was prepared following this method. Therefore, the transglycosylation reaction could occur at the free C40 -, C4000 -, or C40000 -hydroxyl groups in the rebaudioside-A structure (Fig. 1A). Further experiments including 2D 1H-1H COSY, 2D 1H-1H total correlation spectroscopy, 1 H-13C HMBC, and 1H-13C HMQC NMR measurements were performed. However, extreme signal overlap in the glucose region in both the 1H and 13C NMR spectra made it difficult to identify the specific hydroxyl group at which the two additional glucosyl units were attached.
3.2. Nanostructure formed by Stevia-G in water The proton chemical shift (ppm) is sensitive to changes in the molecular environment and is effective in detecting molecular associations (Okano et al., 1997; Saveyn et al., 2009; Shimizu et al., 2003). Fig. 3 shows the 1D 1H NMR spectra of Stevia-G over gradual concentration increases (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, and 50 mg/mL). The chemical shifts of all the protons changed with increasing concentration, indicating the intermolecular interactions during the Stevia-G aggregation process. The NMR chemical shift variation corresponding to different nuclei has been previously evaluated for critical micelle concentration (CMC) or critical aggregation concentration (CAC) detection (Andrade-Dias et al., 2007; Okano et al., 1997). This method can also apply to the investigation of large molecular weight polymer micelles (Emin et al., 2007). In this study, the CMC was used as the index parameter. Below the CMC, the protons are surrounded by solvent molecules and the observed chemical shift, dobs, is the chemical shift of the monomer, dmon. Above the CMC, when the exchange of monomers and micelles in the solution is fast on the NMR time scale, dobs is described as the weighted average of the monomer (dmon) and the micelle chemical shifts (dmic). If the monomer concentration is constant above the CMC, the observed chemical shift is described using the mass-action law model (Eq. (1)) (Andrade-Dias et al., 2007; Farías et al., 2009; Okano et al., 1997), as follows, where C is the total concentration.
Fig. 4. Plots of (A) dobs against 1/C and (B) log (C(dmon dobs) against log (C(dobs dmic) for the H-20 peak from the 1H NMR spectra in Fig. 3.
dobs ¼ dmic
CMC ðdmic dmon Þ C
(1)
Fig. 4A presents plots of dobs at the H-20 peak as a function of the reciprocal of the total Stevia-G concentration. The approximation straight line of the experiment data fitted to the theoretical mode. The intersection of the two lines corresponds to 1/C = 1/CMC. The CMC of Stevia-G was calculated to be approximately 12 mg/mL at 37 C. This value was slightly lower than previously obtained results (16 mg/mL) by surface tension and fluorescence measurements (Uchiyama et al., 2011). The difference may be attributed to the different measurement principle.
[(Fig._3)TD$IG]
Fig. 3. 1H NMR spectra of Stevia-G solution at concentrations of 1–50 mg/mL recorded in D2O at 37 C.
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
[(Fig._5)TD$IG]
259
The chemical shift variation was further employed to calculate the aggregation number of Stevia-G. The single-step equilibrium model, in which molecules exist as either monomers or a single type of aggregate with a fixed aggregation number, was applied. The aggregation equilibrium can be expressed as in Eq. (2) K ¼ ½An ½An
(2)
In the above equation, K is the equilibrium constant, n is the aggregation number, and A and An are the monomeric or aggregated molecules. Combined with Eq. (1) (Okano et al.,1997), Eq. (2) can be rearranged and represented by Eq. (3) (Luchetti, 2000). logðCðdmon dobs ÞÞ ¼ nlogðCðdobs dmic ÞÞ þ logðnKÞ þð1 nÞlogðdmon dobs Þ
(3)
A plot of log [C(dmon dobs)] as a function of log [C(dobs dmic)] consisted of two straight lines—one above and the other below the CMC. The slope yielded the aggregation number n. Fig. 4B shows the plot for the H-20 peak of Stevia-G according to Eq. (3). Two straight lines, whose intersection represents the CMC, indicate the presence of two different equilibria. Below the CMC, the aggregation number was 2, indicating dimer formation for Stevia-G. Above the CMC, the Stevia-G aggregation number increased to 12. Since the plots for other proton peaks (data not shown) were similar to that for the H20 peak, a single type of Stevia-G aggregation above the CMC was proposed. We reported previously that the amount of indomethacin dissolved from a physical mixture of indomethacin/Stevia-G was proportional to the amount of Stevia-G loaded, suggesting the solubilization capacity of Stevia-G even below the CMC (Uchiyama et al., 2011). Li and McGown (1994, 1993) reported the aggregation and solubilization of the bile salt sodium taurocholate (NaTC) in water. The dimer of NaTC was able to solubilize a drug by forming a “sandwich model”, in which the drug was solubilized between upper and lower NaTC molecules. Thus, a dimeric structure for Stevia-G in water can explain why drug solubilization occurred even below the CMC. The size of the Stevia-G micelles was determined by the DLS method. The particle size increased from ca. 1.8 nm to 3 nm as a function of the concentration (Fig. 5), confirming the formation of aggregates. The increased particle size was in accordance with the increased aggregation number. To further investigate the Stevia-G conformation in water, 2D 1 H-1H NOESY NMR measurements of Stevia-G solution at 30 mg/mL were performed (Fig. 6). The NOE cross peaks are the result of crossrelaxation between neighboring protons that are spatially close to each other (roughly, less than 5 Å). The NOESY spectrum provided a number of correlation peaks representing not only the intramolecular interactions, but also the intermolecular interactions with another Stevia-G molecule. However, there were fewer NOE crosspeaks between the basic skeleton and the glucose group of Stevia-G. The Stevia-G micelle could have a core–shell structure in which the basic kaurane diterpenoid skeleton was segregated from the aqueous exterior, forming an inner core encased by a shell of glucose units. Conventional core-forming segments such as alkyl or aryl chains form cores through a combination of intermolecular forces (Kataoka et al., 2001). Similarly, the kaurane diterpenoid skeleton of Stevia-G showed appreciable self-associating behavior driven by hydrophobic interactions, and was stabilized by hydrophilic glucose units. Notably, the methyl protons (H-18, 20) and olefinic proton (H-17) showed some NOE correlations with the glucose protons. These protons, projected from the basic skeleton, could possibly be located proximate to the surface of the hydrophobic core. The NOE correlation of the anomeric protons (H-10 and H-100 ) with the basic skeleton protons arose because these
Fig. 5. Particle size distributions of Stevia-G depending on the concentration from 5–50 mg/mL measured by the dynamic light scattering (DLS) method.
anomeric protons connecting to the basic skeleton could also be spatially close to the surface of the hydrophobic core. 3.3. Solubilization mechanism of poorly water-soluble drugs Fig. 7 shows the 1D 1H NMR spectra of untreated Stevia-G and NRG-loaded Stevia-G in D2O at 37 C. The NRG peaks were clearly detected in the spectrum (Fig. 7B). In general, solution NMR cannot be used to analyze molecules with low mobility, e.g., crystals, because of signal broadening. The solution NMR properties suggested that the NRG in Stevia-G solution was in a molecularly dispersed state rather than the crystal state. The solubility of NRG in Stevia-G solution was determined at 37 C from the integration values of the 1H peaks using 0.05% TSP as an internal standard. The concentration of NRG in Stevia-G solution was calculated as 1.78 mg/mL, which was almost 36 times higher than the solubility of untreated NRG at 0.05 mg/mL. The enhancement in the apparent NRG solubility was because of the incorporation of the NRG into the Stevia-G micelle. NRG-loaded Stevia-G exhibited upfield shifts in the proton peaks of the kaurane diterpenoid skeleton compared
260
[(Fig._6)TD$IG]
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
[(Fig._8)TD$IG]
Fig. 8. Two-dimensional 1H-1H NOESY NMR spectrum of NRG-loaded Stevia-G at a Stevia-G concentration of 30 mg/mL in D2O at 37 C. Fig. 6. Two-dimensional (2D) 1H-1H nuclear Overhauser effect spectroscopy (NOESY) NMR spectrum of Stevia-G solution at 30 mg/mL in D2O at 37 C.
with the untreated Stevia-G, whereas no significant change was observed in peaks of the glucose protons (Fig. 7). This upfield shift demonstrated the interaction of the kaurane diterpenoid skeleton of Stevia-G with NRG. Two-dimensional 1H-1H NOESY NMR spectroscopy was used to investigate the location of the NRG in the Stevia-G micelle. Fig. 8 shows the 2D 1H-1H NOESY NMR spectrum of NRG-loaded Stevia-G in D2O. The key NOE cross-peaks used to deduce the possible interaction sites between the Stevia-G and NRG are highlighted. These new cross-peaks were not observed in the NOESY spectrum of the pure micelle of Stevia-G without the drug (Fig. 6). The aromatic protons of the NRG (H-B0 , H-F0, H-H, H-F, H-C0 and H-E0 ) exhibited NOE correlations with those of the kaurane diterpenoid skeleton (H-18 and H-20) in Stevia-G. Additionally, the H-H and H-F of NRG exhibited cross-peaks with H-14 (1), H-15 (1), and H-17 of Stevia-G. Meanwhile, no NOE cross-peaks between the NRG protons and the glucose units of Stevia-G were found. The observed NOE interactions clearly represented the interaction of NRG with the hydrophobic core of the Stevia-G micelle. All NMR observations
[(Fig._7)TD$IG]
undoubtedly confirmed that the solubilized NRG mainly accumulated and was stabilized in the Stevia-G hydrophobic core. 4. Conclusions The drug solubilization mechanism of Stevia-G, which was synthesized from rebaudioside-A, was successfully characterized by NMR spectroscopy. The peak assignments of the 1H and 13C peaks of Stevia-G in water were achieved by 2D NMR spectroscopic techniques such as 1H-1H COSY, 1H-1H HMBC, and 1H-1H HMQC. The number of anomeric peaks of Stevia-G in the 1D 1H and 13C spectra indicated that two glucose units were attached to rebaudioside-A during the transglycosylation process. The assay of steviol glycosides and glucosyl residues confirmed the introduction of these two glucoses into rebaudioside-A. The concentration-dependent variation of proton chemical shifts revealed the self-association of Stevia-G in water. The CMC of Stevia-G in water at 37 C was determined to be around 12 mg/mL. The aggregation number was 2 below the CMC and 12 above the CMC. Dimer formation in Stevia-G could enable the solubilization of drugs even below the CMC. The mean particle size of the Stevia-G micelle determined by DLS measurement was around 3 nm. 2D 1H-1H NOESY NMR experiments indicated that the Stevia-G micelle consisted of a kaurane diterpenoid skeleton core and a glucose shell. NRG was effectively trapped into the hydrophobic core of the Stevia-G micelle, which significantly enhanced the apparent solubility of NRG. The obtained results provide a better understanding of the drug solubilization mechanism of Stevia-G. New pharmaceutical formulations using Stevia-G can be designed efficiently based on the knowledge presented here. Acknowledgements
Fig. 7. One-dimensional 1H NMR spectra of (A) untreated Stevia-G and (B) NRGloaded Stevia-G at a Stevia-G concentration of 30 mg/mL in D2O at 37 C.
This study was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (Monbukagakusho), Japan (25504018). We thank Toyo Sugar Refining Co., Ltd. for the kind gift of a-glycosyl transferase-treated stevia.
J. Zhang et al. / International Journal of Pharmaceutics 465 (2014) 255–261
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ijpharm.2014.01.035. References Andrade-Dias, C., Lima, S., Teixeira-Dias, J.J., 2007. From simple amphiphilic to surfactant behavior: analysis of 1H NMR chemical shift variations. Journal of Colloid and Interface Science 316, 31–36. Chaturvedula, V.S.P., Prakash, I., 2011a. Diterpene glycosides from Stevia rebaudiana. Journal of Medicinal Plants Research 5, 4838–4842. Chaturvedula, V.S.P., Prakash, I., 2011b. A new diterpene glycoside from Stevia rebaudiana venkata sai. Molecules 16, 2937–2943. Chaturvedula, V.S.P., Prakash, I., 2011c. Isolation, NMR spectral assignments and partial synthesis of rebaudioside B; a sweet component of Stevia rebaudiana. International Journal of Biological and Medical Research 2, 1090–1093. Chiba, S., Okada, S., Kitahata, S., Shimomura, T., 1975. A new trisaccharide, 2a-maltosyl-glucose, synthesized by the transglycosylation reaction of cyclodextrin glycosyltransferase. Agricultural and Biological Chemistry 39, 2353–2357. Dahlberg, C., Millqvist-Fureby, A., Schuleit, M., Furo, I., 2010. Relationships between solid dispersion preparation process, particle size and drug release – an NMR and NMR microimaging study. European Journal of Pharmaceutics and Biopharmaceutics 76, 311–319. Denkova, P.S., Van Lokeren, L., Willem, R., 2009. Mixed micelles of Triton X-100, sodium dodecyl dioxyethylene sulfate, and synperonic l61 investigated by NOESY and diffusion ordered NMR spectroscopy. Journal of Physical Chemistry B 113, 6703–6709. Emin, S.M., Denkova, P.S., Papazova, K.I., Dushkin, C.D., Adachi, E., 2007. Study of reverse micelles of di-isobutyl-phenoxy-ethoxy-ethyl-dimethyl-benzyl ammonium methacrylate in benzene by nuclear magnetic resonance spectroscopy. Journal of Colloid and Interface Science 305, 133–141. Farías, T., de Ménorval, L.C., Zajac, J., Riveraa, A., 2009. Solubilization of drugs by cationic surfactants micelles: conductivity and 1H NMR experiments. Colloids and Surfaces A: Physicochemical Engineering Aspects 345, 51–57. Ficarra, R., Tommasini, S., Raneri, D., Calabro, M.L., Di Bella, M.R., Rustichelli, C., Gamberini, M.C., Ficarra, P., 2002. Study of flavonoids/beta-cyclodextrins inclusion complexes by NMR, FT-IR, DSC, X-ray investigation. Journal of Pharmaceutical and Biomedical Analysis 29, 1005–1014. Fukunaga, Y., Miyata, T., Nakayasu, N., Mizutani, K., Kasai, R., Tanaka, D., 1989. Enzymic transglucosylation products of stevioside: separation and sweetnessevaluation. Agricultural and Biological Chemistry 53, 1603–1607. Grashoff, L.J., Bierner, W.M., Northington, K.D., 1972. Chromosome numbers in north and central American compositae. Brittonia 24, 379–394. Higashi, K., Ideura, S., Waraya, H., Moribe, K., Yamamoto, K., 2009. Incorporation of salicylic acid molecules into the intermolecular spaces of g-cyclodextrinpolypseudorotaxane. Crystal Growth and Design 9, 4243–4246. Kataoka, K., Harada, A., Nagasaki, Y., 2001. Block copolymer micelles for drug delivery: design, characterization and biological significance. Advanced Drug Delivery Reviews 47, 113–131. King, D., Robinson, E., 1967. Aspects of the development of Moniliformis dubius. Journal of Parasitology 53, 142–149. Kittl, R., Withers, S.G., 2010. New approaches to enzymatic glycoside synthesis through directed evolution. Carbohydrate Research 345, 1272–1279. Kojima, T., Higashi, K., Suzuki, T., Tomono, K., Moribe, K., Yamamoto, K., 2012. Stabilization of a supersaturated solution of mefenamic acid from a solid dispersion with EUDRAGIT1 EPO. Pharmaceutical Research 29, 2777–2791. Kometani, T., 2010. Production of novel bioactive compounds by enzymes, and their application to food. Pure and Applied Chemistry 82, 269–287. Li, G., McGown, L.B., 1993. A new approach to polydispersity studies of sodium taurocholate and sodium taurodeoxycholate aggregates using dynamic fluorescence anisotropy. Journal of Physics and Chemistry 97, 6745–6752. Li, G., McGown, L.B., 1994. Model for bile salt micellization and solubilization from studies of a “polydisperse” array of fluorescent probes and molecular modeling. Journal of Physics and Chemistry 98, 13711–13719.
261
Luchetti, L., 2000. NMR investigation on the various aggregates formed by a gemini chiral surfactant. Langmuir 16, 161–165. Mikhail, A.S., Allen, C., 2010. Poly(ethylene glycol)-b-poly(e-caprolactone) micelles containing chemically conjugated and physically entrapped docetaxel: synthesis, characterization, and the influence of the drug on micelle morphology. Biomacromolecules 11, 1273–1280. Okano, L.T., Seoud, O.A.E., Halstead, T.K., 1997. A proton NMR study on aggregation of cationic surfactants in water: effects of the structure of the headgroup. Colloid and Polymer Science 275, 138–145. Saveyn, P., Cocquyt, E., Zhu, W., Sinnaeve, D., Haustraete, K., Martins, J.C., Van der Meeren, P., 2009. Solubilization of flurbiprofen within non-ionic tween 20 surfactant micelles: a 19F and 1H NMR study. Physical Chemistry Chemical Physics 11, 5462–5468. Schedlbauer, A., Coudevylle, N., Auer, R., Kloiber, K., Tollinger, M., Konrat, R., 2009. Autocorrelation analysis of NOESY data provides residue compactness for folded and unfolded proteins. Journal of the Americal Chemical Society 131, 6038–6039. Shimizu, S., Pires, P.A.R., Fish, H., Halstead, T.K., Seoud, O.A.E., 2003. Proton and carbon-13 NMR study of the aggregation of benzyl-(2-acylaminoethyl) dimethylammonium chloride surfactants in D2O. Physical Chemistry Chemical Physics 5, 3489–3497. Sinha, S., Ali, M., Baboota, S., Ahuja, A., Kumar, A., Ali, J., 2010. Solid dispersion as an approach for bioavailability enhancement of poorly water-soluble drug ritonavir. AAPS PharmSciTech 11, 518–527. Steinmetz, W.E., Lin, A., 2009. NMR studies of the conformation of the natural sweetener rebaudioside A. Carbohydrate Research 344, 2533–2538. Suzuki, Y., Suzuki, K., 1991. Enzymatic formation of 4G-a-D-glucopyranosyl-rutin. Agricultural and Biological Chemistry 55, 181–187. Tozuka, Y., Kishi, J., Takeuchi, H., 2010. Anomalous dissolution property enhancement of naringenin from spray-dried particles with a-glucosylhesperidin. Advanced Powder Technology 21, 305–309. Tozuka, Y., Higashi, K., Morita, T., Nishiwaka, M., Uchiyama, H., Zhang, J., Moribe, K., Nishiwaka, K., Takeuchi, H., Yamamoto, K., 2012. Transglycosylated rutin-specific non-surface-active nanostructure affects absorption enhancement of flurbiprofen. European Journal of Pharmaceutics and Biopharmaceutics 82, 120–126. Uchiyama, H., Tozuka, Y., Imono, M., Takeuchi, H., 2010a. Improvement of dissolution and absorption properties of poorly water-soluble drug by preparing spray-dried powders with alpha-glucosyl hesperidin. International Journal of Pharmaceutics 392, 101–106. Uchiyama, H., Tozuka, Y., Imono, M., Takeuchi, H., 2010b. Transglycosylated stevia and hesperidin as pharmaceutical excipients: dramatic improvement in drug dissolution and bioavailability. European Journal of Pharmaceutics and Biopharmaceutics 76, 238–244. Uchiyama, H., Tozuka, Y., Asamato, F., Takeuchi, H., 2011. Fluorescence investigation of a specific structure formed by aggregation of transglycosylated stevias: solubilizing effect of poorly water-soluble drugs. European Journal of Pharmaceutical Sciences 43, 71–77. Uchiyama, H., Tozuka, Y., Nishikawa, M., Takeuchi, H., 2012. Nanocomposite formation between alpha-glucosyl stevia and surfactant improves the dissolution profile of poorly water-soluble drug. International Journal of Pharmaceutics 30, 183–186. Varuzhan, H.A., Vahe, T.G., Avetik, A.M., Mariam, O.A., Lidia, A.A., 2006. Extraction, separation and modification of sweet glycosides from the stevia rebaudiana plant. United States Patent 7838044. Wade, K.E., Wilson, I.D., Troke, J.A., Nicholson, J.K., 1990. 19F and 1H magnetic resonance strategies for metabolic studies on fluorinated xenobiotics: application to flurbiprofen [2-(2-fluoro-4-biphenylyl)propionic acid]. Journal of Pharmaceutical and Biomedical Analysis 8, 401–410. Wang, L.X., Huang, W., 2009. Enzymatic transglycosylation for glycoconjugate synthesis. Current Opinion in Chemical Biology 13, 592–600. Yang, W., Li, Y., Cheng, Y., Wu, Q., Wen, L., Xu, T., 2009. Evaluation of phenylbutazone and poly(amidoamine) dendrimers interactions by a combination of solubility, 2D-NOESY NMR, and isothermal titration calorimetry studies. Journal of Pharmaceutical Sciences 98, 1075–1085. Zhang, J., Higashi, K., Limwikrant, W., Moribe, K., Yamamoto, K., 2012. Molecularlevel characterization of probucol nanocrystal in water by in situ solid-state NMR spectroscopy. International Journal of Pharmaceutics 423, 571–576. Zhang, J., Tozuka, Y., Uchiyama, H., Higashi, K., Moribe, K., Takeuchi, H., Yamamoto, K., 2011. NMR investigation of a novel excipient, a-glucosylhesperidin, as a suitable solubilizing agent for poorly water-soluble drugs. Journal of Pharmaceutical Sciences 100, 4421–4431.