TY - JOUR
AU - Hirabayashi, Jun
AB - Abstract Rare sugars are defined as monosaccharides with extremely low natural abundances. Their natural distributions and biological functions remain to be clarified. To establish a robust analytical system that can separate, identify and quantify rare sugars, 12 d-hexoses—including five rare aldohexoses and three rare ketohexoses—were labelled with 2-aminobenzamide (2-AB), and the fluorescently tagged monosaccharides were separated by high-performance liquid chromatography (HPLC). Because the ketohexoses were much less reactive than were the aldohexoses, the reaction conditions were optimized to achieve the maximum yields (>75%) for both aldohexoses and ketohexoses. The calibration curve determined for the rare ketohexose, d-psicose (Psi), was linear between 1 pmol and 1 µmol and had a correlation coefficient of 0.9999. Using the developed method, the psicose content in a leaf of Itea virginica, in which the presence of psicose has previously been reported, was found to be 2.7 mg psicose/g leaf. The result proved feasibility of the method even for natural products. Because the system is a comprehensive tool, it should help answer questions concerning the biosyntheses and functions of rare hexose sugars. aldohexose, 2-aminobenzamide, fluorescence labelling, ketohexose, rare sugar Rare sugars are defined as monosaccharides that exist in nature in extremely small amounts. They include the aldohexoses, d-/l-allose (All), d-/l-altrose (Alt), l-galactose (Gal), l-glucose (Glc), d-/l-gulose (Gul), d-/l-idose (Ido), l-mannose (Man) and d-/l-talose (Tal), and the ketohexoses, l-fructose (Fru), d-/l-psicose (Psi), d-/l-sorbose (Sor), and d-/l-tagatose (Tag). Izumori and colleagues have extensively studied the converting enzyme, d-tagatose 3-epimerase (1, 2), which has been used to systematically produce rare sugars (3–5). Therefore, rare sugars can now be considered to be research reagents, as their biological and physiological properties have been extensively characterized (6–10). Rare sugars are of commercial value in the food, pharmaceutical, and nutritional industries because they can be used as low calorie sweeteners—they are difficult to digest (6, 7)—as inhibitors of cancer cell proliferation (8), as inhibitors of microbial growth (9), and as memory enhancers (10). However, little is known about where and to what extent they are found in nature and in artificially prepared food products, primarily because there is no systematic method to detect and quantify them. Thus, this knowledge void inhibits understanding the biological significance(s) of rare sugars. Rare sugars are essentially absent in nature in that their sugar nucleotides are not involved in glycan biosynthesis. In fact, there are only few reports of naturally occurring rare sugars (11–13). These include the finding of Psi in the plant Itea virginica. Conversely, rare sugars have been found in certain artificially prepared food products probably as a result of a chemical conversion, i.e. the Lobry de Bruyn-Alberda van Ekenstein transformation, from the more abundant common sugars, e.g. d-Glc, d-Fru (14–16). Therefore, it remains to be clarified if rare sugars are essentially absent in nature or if they exist more widely than has been recognized to date. There has been no reliable method that not only identifies, but also quantifies, the compositions of monosaccharide mixtures that include rare sugars. Various techniques have been applied to detect rare sugars, including an HPLC system connected to a refractive index detector (4), a biosensor that uses enzyme reactions (17), and a high pH anion-exchange chromatography system coupled to a pulsed amperometric detector (16). However, none of these systems has achieved the necessary sensitivity, i.e. picomole level detection, required to identify naturally occurring rare sugars. Nakakita and colleagues (18) recently reported a highly sensitive method for the separation and identification of d-aldoses including the six rare sugars, Lyx, Tal, Gul, Ido, All and Alt, by combining pyridylamination of the sugars with conventional HPLC separation. Unfortunately, the method cannot be readily applied to ketoses, such as Psi, for which the natural occurrence has been of special interest (12, 13). Ketoses are pyridylaminated much less readily than are aldoses (19, 20). Among the monoamine-coupling methods targeting oligosaccharides, two fluorescence-labelling reagents, 2-aminopyridine (2-AP) (21) and 2-aminobenzamide (2-AB) (22) have widely been used, because they are useful for both the separation and the detailed structural analysis of oligosaccharides derived from glycoproteins (23, 24) and glycolipids (25, 26), as well as the free oligosaccharides (27, 28). However, such oligosaccharides are basically composed of aldoses (mostly aldohexoses), and thus, seldom contain ketoses. Therefore, the methods have only been occasionally attempted using ketoses (19, 20). An important point to keep in mind when developing a detection/separation system for ketoses is that unlike aldoses, two stereoisomeric R/S (at the C2 carbon) reaction products will be formed (Fig. 1). Fig. 1 View largeDownload slide The reaction of an aldohexose, (d-Glc; top) and of a ketohexose (d-Fru; bottom) with 2-AB. Note that when ketoses are subjected to monoamine-coupling and subsequent reduction, R/S isomers are generated at the C2 carbon (B). Whereas, only one product is produced for each of the aldohexoses (A). Fig. 1 View largeDownload slide The reaction of an aldohexose, (d-Glc; top) and of a ketohexose (d-Fru; bottom) with 2-AB. Note that when ketoses are subjected to monoamine-coupling and subsequent reduction, R/S isomers are generated at the C2 carbon (B). Whereas, only one product is produced for each of the aldohexoses (A). For the work reported herein, the d-aldohexoses, Glc, Man, All, Alt, Gul, Ido, Gal and Tal, and the ketohexoses, Fru, Psi, Tag and Sor, were coupled with 2-AB, but not 2-AP, and the 2-AB sugars were separated using various modes of liquid chromatography: gel filtration, reversed-phase chromatography, and normal-phase chromatography. To validate the reproducibility and sensitivity of the separation system, the types and the amounts of hexoses contained in a leaf of I. virginica were determined, and the rare sugar, Psi, was found. Although Psi has been found in the leaves of I. virginica previously (12, 13), ours is the first study to quantify the amount of Psi present. Materials and Methods Materials d-All, d-Tal and 2-AB were purchased from Tokyo Chemical Industry (Tokyo, Japan). d-Alt, d-Gul and d-Ido were obtained from Funakoshi (Tokyo, Japan). d-Fru, sodium cyanotrihydroborate (NaBH3CN), acetic acid and dimethyl sulphoxide (DMSO) were obtained from Wako Pure Chemical Industries (Osaka, Japan). d-Glc, d-Man, d-Gal and a Cosmosil 5C18-AR column (6 × 250 mm) were obtained from Nakarai Co., Ltd (Kyoto, Japan). d-Psi, d-Tag and d-Sor were obtained from Fushimi Pharmaceutical Co., Ltd (Kagawa, Japan). The 2-AB Glucose Homopolymer Standard was purchased from Cosmo Bio Co., Ltd. (Tokyo, Japan). The TSKgel HW-40F resin was obtained from Tosoh (Tokyo, Japan). An Inertsil NH2 column (4.6 × 150 mm) was obtained from GL Science Inc. (Tokyo, Japan). An Ultrasphere ODS column (2.0 × 500 mm) was obtained from Beckman Coulter (Fullerton, CA, USA). Fluorescent labelling The previously established method that was used to derivatize oligosaccharides with 2-AB was modified to accommodate the lesser reactivities of the ketoses (19, 20). Briefly, sugars (both aldoses and ketoses) were dissolved into 20 µl of a DMSO:acetic acid (7:3, v/v) solution containing various amounts (1–15 mg) of 2-AB and 1.0 M NaBH3CN. The mixtures were held at various temperatures (between 65°C and 90°C) for 3–36 h. After 20 µl of methanol and 40 µl of toluene were added to a reaction mixture, a stream of N2 gas was blown over it, and the solution was heated to 60°C for 10 min in vacuo in a Palstation model 4000 apparatus (Takara Biomedicals, Kyoto, Japan). By repeating this procedure three times, excess 2-AB was removed almost completely. Residual amounts of unwanted materials were removed by gel filtration over a column of TSK gel HW-40F (7 × 55 mm) equilibrated with 10 mM ammonium acetate (pH 6.0). The fluorescence intensity of each chromatographic fraction was measured using excitation and emission wavelengths of 320 nm and 420 nm, respectively. Fractions containing fluorescent substances were pooled, and were further analysed by high performance liquid chromatography (HPLC) (see below). High performance liquid chromatography A Shimadzu Prominence HGE-Fluorescence system was used throughout this study. 2-AB monosaccharide fractions were first separated by size-fractionation HPLC over Inertsil NH2 using a 0.2% formic acid–acetonitrile gradient eluent at a flow rate of 0.6 ml/min at 30°C. After injection of a sample solution, the proportion of the eluent containing acetonitrile, 0.2% formic acid was changed linearly from 0% to 20% (v) in 1 min, to 27.5% (v) in 10 min and to 55% (v) in 43 min. The fluorescence of the 2-AB monosaccharides was detected using an excitation wavelength of 320 nm and an emission wavelength of 420 nm. A second separation followed that was performed using Cosmosil 5C18-AR reversed-phase HPLC with the isocratic solvent system, a: 50 mM acetic acid, 0.2% (v/v) acetonitrile. When necessary, an additional separation of 2-AB-aldohexoses was performed using Cosmosil 5C18-AR and the isocratic solvent system, b: 0.1 M citric acid, 0.1% (v/v) acetonitrile, Additional separations of 2-AB-ketohexoses were preformed using Cosmosil 5C18-AR and the isocratic solvent system, c: water, 0.1% (v/v) trifluoroacetic acid, 0.5% (v/v) acetonitrile. For the Cosmosil 5C18-AR chromatographies, the flow rates and the column temperatures were always 1.0 ml/min and 40°C, respectively. The elution positions of the 2-AB monosaccharides were detected fluorescently as described above. Quantification of 2-AB monosaccharides Each of the 2-AB monosaccharides was quantified on the basis of its peak area in comparison with that of the authentic standard (2-AB Glucose Homopolymer) that had been chromatographed under the same HPLC conditions over Inertsil NH2. The 2-AB monosaccharides thus quantified were also used as standards for reversed-phase HPLC analysis. Determination of hexose contents in natural products Itea virginica was a generous gift from Dr Takeshita (Fushimi Pharmaceutical Co., Ltd, Takamatsu, Japan). The leaf (25 mg, fresh weight) was cut into small pieces and ground in a mortar with a pestle under distilled water (0.2 ml). After the insoluble material was removed by centrifugation, methanol–chloroform (1:1, v/v) treatment removed the lipids. The extract was dried, and reacted with 2-AB. Each 2-AB hexose contained in the leaf extract was quantified on the basis of its peak area in comparison with that of the corresponding standard separated under the same reversed-phase HPLC conditions. Results Fluorescent labelling of ketohexoses with 2-AP Bonn and colleagues have previously attempted to label Fru and Sor with p-aminobenzoic acid (2-AA). They found that the yields of the 2-AA ketoses were considerably lower than that of d-Glc labeled using the same conditions (19, 20). Therefore, it was obvious that the possible means of labeling the ketoses and the reaction conditions should first be studied in detail and then optimized. At first, we tried to use 2-AP as the fluorescent reagent. Its reaction with the four d-ketohexoses, Fru, Psi, Tag, Sor, was carried according to the general procedure used previously for oligosaccharides (21). Unexpectedly, considerable amounts of side-reaction products were obtained. Figure 2 shows a typical Ultrasphere ODS chromatogram obtained for the reaction products of d-Sor. The last two eluting peaks (3 and 4) contained PA-Ido and PA-Gul, respectively; whereas, the first two were considered to correspond to the C2 carbon R/S-isomers of pyridylaminated d-Sor (Fig. 1). Fluorescence intensities of the four peaks indicated that the reaction yield for PA-Sor was about 45%. The low yield can be attributed to a base-catalyzed, ketose–aldose transformation, i.e. the Lobry de Bruyn-Alberda van Ekenstein transformation (29–31). In general, this transformation proceeds via an equilibrium between ketoses and aldoses, and thus, product yields depend on the relative thermodynamic stabilities of the individual PA-saccharides. Since pyridine behaves as both a solvent and a catalyst (32–34), it is quite possible that 2-AP catalyzed the transformation of Sor to the corresponding aldoses, i.e. Gul and Ido, before the monoamine-coupling reaction proceeded. Similar results were found for other three ketoses, i.e. Fru, Psi and Tag. If this is the case, it would be difficult to produce PA-ketohexoses quantitatively. Fig. 2 View largeDownload slide Separation of pyridylaminated d-Sor and its derivatives by reversed-phase HPLC.d-Sor (100 nmol) was labelled with 2-AP, and an aliquot (1/10,000 volume of the reaction mixture) containing the PA-derivatives was eluted over Ultrasphere ODS (2.0 × 500 mm column) with 0.2 M sodium acetate (pH 4.0), 1% (v/v) acetonitrile at a flow rate of 1.0 ml/min. The PA-monosaccharides were detected using an excitation wavelength of 320 nm and an emission wavelength of 400 nm. Theoretically, a monoamine-coupling should produce the S/R stereoisomers at the C2 carbon. The corresponding two diasteriomers were found in peaks 1 and 2 of the chromatogram. Additionally, substantial amounts of by-products, produced possibly as a result of a base-catalysed transformation as described in the text, were generated. For reference, the elution positions of the two standard PA-aldohexoses, PA-Ido (3) and PA-Gul (4), are shown. Fig. 2 View largeDownload slide Separation of pyridylaminated d-Sor and its derivatives by reversed-phase HPLC.d-Sor (100 nmol) was labelled with 2-AP, and an aliquot (1/10,000 volume of the reaction mixture) containing the PA-derivatives was eluted over Ultrasphere ODS (2.0 × 500 mm column) with 0.2 M sodium acetate (pH 4.0), 1% (v/v) acetonitrile at a flow rate of 1.0 ml/min. The PA-monosaccharides were detected using an excitation wavelength of 320 nm and an emission wavelength of 400 nm. Theoretically, a monoamine-coupling should produce the S/R stereoisomers at the C2 carbon. The corresponding two diasteriomers were found in peaks 1 and 2 of the chromatogram. Additionally, substantial amounts of by-products, produced possibly as a result of a base-catalysed transformation as described in the text, were generated. For reference, the elution positions of the two standard PA-aldohexoses, PA-Ido (3) and PA-Gul (4), are shown. Fluorescent labelling of ketohexoses with 2-AB We then attempted the monoamine coupling using another common reagent, 2-AB, which lacks a base-catalyst moiety. According to the reaction conditions reported by Bigge and colleagues (22), each of the eight d-aldohexoses (Glc, Man, All, Alt, Gal, Tal, Ido and Gul; 10 nmol) and the four d-ketohexoses (Fru, Psi, Tag and Sor; 10 nmol) was dissolved in 20 µl of a DMSO: acetic acid (7:3, v/v) solution containing 1 mg 2-AB and 1.0 M NaBH3CN. Each mixture was reacted at 65°C for 3 h. After the coupling reaction, excess reagents were removed as follows. Most of the 2-AB was removed using a stream of nitrogen gas, and then the remainder was removed by TSKgel HW-40F chromatography (23). The gel filtration procedure completely separated the 2-AB monosaccharides from the residual 2-AB and other low molecular weight by-products (Fig. 3). The 2-AB ketohexoses were eluted at fractions 20–24 (fraction volume, 1 ml), while the 2-AB aldohexoses were eluted at fractions 26–30. Notably, the peak areas of the 2-AB ketohexoses were smaller than those of the 2-AB aldohexoses (for d-Glc and d-Fru, see Fig. 3). To confirm these observations, the 2-AB monosaccharide fractions were subjected to size-fractionation HPLC using Inertsil NH2 (chromatogram not shown). As expected, the yields of the 2-AB ketohexoses were much lower (1–5%) than those of the 2-AB aldohexoses (80–90%). (For the latter, the less than complete yields may be attributed to the chromatography procedure itself). As no significant amounts of contaminating 2-AB aldohexoses were formed when the ketohexoses were reacted, it seemed worthwhile to further explore the 2-AB reaction. Fig. 3 View largeDownload slide Gel filtration chromatograms of d-Glc (A) and d-Fru (B) labelled with 2-AB using the conditions of Bigge and colleagues (19). The eluent for the TSK gel HW-40F chromatography (7 × 500 mm column) was 10 mM ammonium acetate (pH 6.0), and a flow rate of 0.1 ml/min was used. The fluorescence intensity of each 1 ml fraction was measured. The fractions marked by the bars were pooled. Peak 1 contains 2-AB Glc, and peak 2 contains 2-AB Fru. Fig. 3 View largeDownload slide Gel filtration chromatograms of d-Glc (A) and d-Fru (B) labelled with 2-AB using the conditions of Bigge and colleagues (19). The eluent for the TSK gel HW-40F chromatography (7 × 500 mm column) was 10 mM ammonium acetate (pH 6.0), and a flow rate of 0.1 ml/min was used. The fluorescence intensity of each 1 ml fraction was measured. The fractions marked by the bars were pooled. Peak 1 contains 2-AB Glc, and peak 2 contains 2-AB Fru. To optimize the reaction conditions, the following parameters were examined: the 2-AB concentration, the reaction temperature, and the reaction period. As a solvent system for 2-AB coupling, 70% DMSO:30% acetic acid (v/v) was used throughout (22). The concentration of NaBH3CN was fixed at 1.0 M. The monosaccharides, d-Glc and d-Psi, were chosen as a representative aldohexose and ketohexose, respectively. The results of the surveys are summarized in Fig. 4A–C, and the following conclusions were drawn: At least 10 mg of 2-AB per reaction (reaction volume, 20 µl) was required to label Psi; whereas, 1 mg labelled >80% of the Glc present (Fig. 4A). Thus, the 2-AB concentration was set at 10 mg/reaction (500 mg/ml). The optimum temperature was 65°C for both d-Glc and d-Psi; although the yield of 2-AB Psi was more sensitive to temperature (Fig. 4B). Thus, the reaction temperature was set at 65°C. At least 18 h were necessary to achieve the maximum yield of 2-AB Psi (>70%), while after 3 h >85% of the d-Glc had been labelled (Fig. 4C). Thus, the reaction time was set at 18 h. Fig. 4 View largeDownload slide Surveys of the experimental conditions for reaction of 2-AB with d-Glc (filled triangle) and with d-Psi (filled square). The amounts of the 2-AB monosaccharides were determined using the peak areas found for the Inertsil NH2 chromatographies. The yields were investigated as functions of (A) 2-AB concentration, (B) reaction temperature and (C) reaction period. Each data point is the average of three separate experiments and the bars show the standard errors. Fig. 4 View largeDownload slide Surveys of the experimental conditions for reaction of 2-AB with d-Glc (filled triangle) and with d-Psi (filled square). The amounts of the 2-AB monosaccharides were determined using the peak areas found for the Inertsil NH2 chromatographies. The yields were investigated as functions of (A) 2-AB concentration, (B) reaction temperature and (C) reaction period. Each data point is the average of three separate experiments and the bars show the standard errors. Validation of the optimized conditions Using the optimized conditions, the eight d-aldohexoses and the four d-ketohexoses were each labelled with 2-AB. Each sugar (10 nmol) was dissolved in 20 µl of the reaction solution (prepared by dissolving 75 mg 2-AB and 9 mg NaBH3CN in 70 µl DMSO/30 µl acetic acid to make up a 150 µl solution), and the solution was incubated at 65°C for 18 h. At the end of the reaction time, excess reagents were removed by a nitrogen flush and gel filtration, and the 2-AB sugars were chromatographed over Inertsil NH2. The yields of all of the 2-AB monosaccharides were referenced to the 2-AB Glucose Homopolymer Standard because authentic standards for 2-AB aldoses and 2-AB ketoses did not exist. Using the optimized reaction conditions, the apparent yields of the 2-AB monosaccharides were each at least 75% (Table I). However, the apparent yields of the 2-AB monosaccharides differed significantly. The apparent yields of 2-AB Alt and 2-AB Gul exceeded 100%, possibly because their relative fluorescent intensities are significantly greater than that of the standard. Conversely, the yield of 2-AB Psi was relatively low. It is known that the reducing terminals of monosaccharides significantly affect the fluorescent intensities of the label (35). Therefore, for each 2-AB monosaccharide, its absolute fluorescent intensity should be determined so that it can be precisely quantified. Table I. Yields of 2-AB monosaccharides under the optimized conditionsa. 2-AB aldose  Yield (%)  2-AB ketose  Yield (%)  2-AB Glc  87  2-AB Fru  90  2-AB Man  88  2-AB Psi  75  2-AB All  89  2-AB Sor  91  2-AB Alt  130  2-AB Tag  79  2-AB Ido  90      2-AB Gul  120      2-AB Gal  80      2-AB Tal  93      2-AB aldose  Yield (%)  2-AB ketose  Yield (%)  2-AB Glc  87  2-AB Fru  90  2-AB Man  88  2-AB Psi  75  2-AB All  89  2-AB Sor  91  2-AB Alt  130  2-AB Tag  79  2-AB Ido  90      2-AB Gul  120      2-AB Gal  80      2-AB Tal  93      aYield was determined from fluorescence intensity of each 2-AB monosaccharide in comparison with that of 2-AB Glc on the assumption that their fluorescence intensities are the same as that of 2-AB Glc. View Large Calibration curves were next obtained for d-Glc and d-Psi (Fig. 5). The d-Psi calibration curve was linear between 1 pmol and 1 µmol and had a correlation coefficient (R2) of 0.9999; whereas, the d-Glc calibration curve was linear between 1 pmol and 10 µmol and had a correlation coefficient of 0.9999. For the Inertsil NH2 chromatographies, the detection limit was ∼0.1 pmol, and the signal-to-noise ratio was approximately 3. Fig. 5 View largeDownload slide Calibration curves obtained for (A) d-Glc and (B) d-Psi. Each data point is the average of three separate experiments and the bars show the standard errors. Fig. 5 View largeDownload slide Calibration curves obtained for (A) d-Glc and (B) d-Psi. Each data point is the average of three separate experiments and the bars show the standard errors. Systematization of the HPLC separation of the 2-AB hexoses Recently (18), we labeled a series of d-aldohexoses with 2-AP, and the PA-saccharides were separated by HPLC. For this report, d-ketohexoses, in addition to d-aldohexoses, were labelled with 2-AB, with the goal of constructing a universal separation system. We first used the separation conditions described by Chiba and colleagues (36) who separated 2-AB Glc, 2-AB Man, and 2-AB Gal over Cosmosil 5C18-AR. For this study, a mixture of the eight d-aldohexoses (Glc, Man, All, Alt, Gul, Ido, Gal, and Tal; 1 nmol each) or the four d-ketohexoses (Fru, Psi, Sor and Tag; 1 nmol each) was labelled with 2-AB, gel filtrated and eluted over Cosmosil 5C18-AR, equilibrated with solvent system a (Fig. 6A and B). The 2-AB aldohexoses eluted with retention times between 44.9 and 63.0 min (Fig. 6A). On the other hand, the 2-AB ketohexoses eluted much more rapidly, with retention times between 32.8 and 39.3 min (Fig. 6B). Because two diasteriomers are generated from each ketohexose (Figs 1 and 2), their chromatographic separations are more difficult. Using solvent system a, the following 2-AB monosaccharides migrated together: (i) Tal (3), All (4) and Ido (5) eluted at 49.6 min (Fig. 6A); (ii) Tag2 (10) and Psi1 (11) eluted at 35.1 min; (iii) Sor2 (13) and Fru1 (14) eluted at 37.6 min; and (iv) Fru2 (15) and Psi2 (16) eluted at 39.3 min (Fig. 6B). (The subscripts 1 and 2 reference the diasteriomers of the 2-AB ketohexoses and their elution positions). Fig. 6 View largeDownload slide Separation of (A, C) the 2-AB aldohexoses and (B, D–F) the 2-AB ketohexoses by Cosmosil 5C18-AR chromatography. For (A) and (B) the eluent was solvent system a; For (C), the eluent was solvent system b; For (D–F), the eluent was solvent system c. A mixture of (A) d-aldohexoses (Glc, Man, All, Alt, Gul, Ido, Gal, Tal; 1 nmol each) and a mixture (B) of d-ketohexoses (Fru, Psi, Sor, Tag; 1 nmol each) were each labeled with 2-AB, and the two 2-AB derivative s mixtures (50 pmol each) were individually eluted over Cosmosil 5C18-AR with solvent system a. The co-eluted 2-AB monosaccharide fractions (marked by bars) were pooled, and chromatographed over Cosmosil 5C18-AR using solvent system b (C, Tal/All/Ido) or solvent system c (D; Psi1/Tag2: E; Fru1/Sor2: F; Fru2/Psi2). For each chromatography, the 2-AB monosaccharides were separated isocratically at a flow rate of 1 ml/min. For reference, elution positions of all standard 2-AB hexoses are indicated with arrowheads; 1, 2-AB Gal; 2, 2-AB Man; 3, 2-AB Tal; 4, 2-AB All; 5, 2-AB Ido; 6, 2-AB Glc; 7, 2-AB Gul; 8, 2-AB Alt; 9, 2-AB Tag1; 10, 2-AB Tag2; 11, 2-AB Psi1; 12, 2-AB Sor1; 13, 2-AB Sor2; 14, 2-AB Fru1; 15, 2-AB Fru2; 16, 2-AB Psi2. For each ketohexose, two peaks corresponding to the two diasteriomers with R/S chirality at the C2 carbon are generated (also see Fig. 2). Because it is not known which peak contains the S or the R isomer, the individual isomers are labeled with the subscripts, l and 2, which identify their order of elution. Fig. 6 View largeDownload slide Separation of (A, C) the 2-AB aldohexoses and (B, D–F) the 2-AB ketohexoses by Cosmosil 5C18-AR chromatography. For (A) and (B) the eluent was solvent system a; For (C), the eluent was solvent system b; For (D–F), the eluent was solvent system c. A mixture of (A) d-aldohexoses (Glc, Man, All, Alt, Gul, Ido, Gal, Tal; 1 nmol each) and a mixture (B) of d-ketohexoses (Fru, Psi, Sor, Tag; 1 nmol each) were each labeled with 2-AB, and the two 2-AB derivative s mixtures (50 pmol each) were individually eluted over Cosmosil 5C18-AR with solvent system a. The co-eluted 2-AB monosaccharide fractions (marked by bars) were pooled, and chromatographed over Cosmosil 5C18-AR using solvent system b (C, Tal/All/Ido) or solvent system c (D; Psi1/Tag2: E; Fru1/Sor2: F; Fru2/Psi2). For each chromatography, the 2-AB monosaccharides were separated isocratically at a flow rate of 1 ml/min. For reference, elution positions of all standard 2-AB hexoses are indicated with arrowheads; 1, 2-AB Gal; 2, 2-AB Man; 3, 2-AB Tal; 4, 2-AB All; 5, 2-AB Ido; 6, 2-AB Glc; 7, 2-AB Gul; 8, 2-AB Alt; 9, 2-AB Tag1; 10, 2-AB Tag2; 11, 2-AB Psi1; 12, 2-AB Sor1; 13, 2-AB Sor2; 14, 2-AB Fru1; 15, 2-AB Fru2; 16, 2-AB Psi2. For each ketohexose, two peaks corresponding to the two diasteriomers with R/S chirality at the C2 carbon are generated (also see Fig. 2). Because it is not known which peak contains the S or the R isomer, the individual isomers are labeled with the subscripts, l and 2, which identify their order of elution. To separate completely the co-eluted 2-AB hexoses, we then tried Cosmosil 5C18-AR chromatography in conjunction with solvent system b (Fig. 6C). Each mixture containing the co-eluted 2-AB derivatives, was individually applied to the column and eluted. Using this system and the 2-AB aldohexoses (Fig. 6C), 2-AB Tal (3) and 2-AB All (4) were partially separated from each other (50% baseline resolution) with retention times of 68.9 min and 70.7 min, respectively; whereas, 2-AB Ido (5) was cleanly separated with a retention time of 65.6 min. Thus, the 2-AB d-aldohexoses were successfully separated after two chromatographic rounds. However, for the three pairs of ketohexose derivatives (Psi1/Tag2, Fru1/Sor2 and Fru2/Psi2), solvent system b was unsatisfactory because Psi1/Tag2 still eluted together, even though Fru1/Sor2 and Fru2/Psi2 were separable (data not shown). Thus, the ability of solvent system c, to separate the 2-AB ketohexoses was assessed. Using this solvent system, each of the 2-AB ketohexose pairs were separated (Fig. 6D–F). Taken together, the results clearly demonstrated that the 2-AB hexoses could be separated chromatographically by using gel filtration to separate the aldohexoses from the ketohexoses and then using reversed-phase HPLC to separate the individual aldohexoses and the individual ketohexoses. In both cases, the first separation involves Cosmosil 5C18-AR chromatography with solvent system a and then is followed by Cosmosil 5C18-AR chromatography with solvent system b for the aldohexoses or solvent system c for the ketohexoses, respectively. A flowchart for the separation is shown in Fig. 7, and the HPLC elution times of the 2-AB hexoses are summarized in Table II. Fig. 7 View largeDownload slide Schematic flowchart for the separation of the 2-AB hexoses. The sugar numbers 1–16 correspond to those used in Fig. 6. Fig. 7 View largeDownload slide Schematic flowchart for the separation of the 2-AB hexoses. The sugar numbers 1–16 correspond to those used in Fig. 6. Table II. Elution Times of the 2-AB Hexoses Upon Reversed-phase HPLC, Cosmosil 5C18-AR Chromatographya. Solvent systemb  Aldohexoses   Ketohexoses   Gal  Man  Tal  All  Ido  Glc  Gul  Alt  Tag1  Tag2  Psi1  Sor1  Sor2  Fru1  Fru2  Psi2  a  44.9  48.0  49.6  49.6  49.6  51.4  55.3  63.0  32.8  35.1  35.1  36.6  37.6  37.6  39.3  39.3  b  61.5  68.3  68.9  70.7  65.6  68.5  76.9  87.9  41.9  42.8  42.8  45.3  46.6  48.2  48.4  50.3  c  20.9  22.3  23.2  23.2  23.2  24.9  26.2  26.2  23.3  31.2  28.6  30.0  30.0  26.3  32.8  29.1  Solvent systemb  Aldohexoses   Ketohexoses   Gal  Man  Tal  All  Ido  Glc  Gul  Alt  Tag1  Tag2  Psi1  Sor1  Sor2  Fru1  Fru2  Psi2  a  44.9  48.0  49.6  49.6  49.6  51.4  55.3  63.0  32.8  35.1  35.1  36.6  37.6  37.6  39.3  39.3  b  61.5  68.3  68.9  70.7  65.6  68.5  76.9  87.9  41.9  42.8  42.8  45.3  46.6  48.2  48.4  50.3  c  20.9  22.3  23.2  23.2  23.2  24.9  26.2  26.2  23.3  31.2  28.6  30.0  30.0  26.3  32.8  29.1  aElution times in minutes. ba, 50 mM acetic acid, 0.2% (v/v) acetonitrile; b, 0.1 M citric acid, 0.1% (v/v) acetonitrile; c, water, 0.1% (v/v) trifluoroacetic acid, 0.5% (v/v) acetonitrile. View Large Practical application: assessment of the hexose content in a leaf of I. virginica It has been reported that Itea plants contain d-Psi in their leaves and flowers (12, 13). Because Psi is thought to occur unmodified in nature, we measured the amount of free Psi contained in 25 mg of a fresh I. virginica leaf. Monosaccharides were extracted from the leaf, labelled with 2-AB, and the 2-AB hexoses were quantified using Cosmosil 5C18-AR reversed-phase HPLC (Fig. 8). The Psi content was determined to be 2.7 mg/g leaf (wet weight), and no other rare sugars were identified. To our surprise, the Glc content was much lower (0.6 mg/g) than that of Psi. Considering the natural abundance of these monosaccharides, the amount of Psi found in the leaf is extremely large. The common monosaccharides, Gal and Fru were also detected; although, their amounts were relatively low (0.03 mg/g and 0.1 mg/g, respectively). Our results are consistent with the previous report by Hough and colleagues (12), who detected Psi using paper chromatography. Notably, that method did not detect the less abundant Fru and Glc even when 6 g (fresh weight) of leaves was assayed. Thus, we are now able to measure a series of hexoses contained in both natural and artificial products. Although, its practical application regarding other rare sugars remains to be tested. Fig. 8 View largeDownload slide Separation of 2-AB hexoses extracted from a leaf of I. virginica. The monosaccharides extracted and labelled with 2-AB were separated by Cosmosil 5C18-AR chromatography using solvent system a. The elution positions of standard 2AB monosaccharides are indicated with arrowheads: 1, 2-AB Psi1; 2, 2-AB Fru1; 3, 2-AB Psi2, 2-AB Fru2; 4, 2-AB Gal; 5, 2-AB Man; 6, 2-AB All; 7, 2-AB Glc. Fig. 8 View largeDownload slide Separation of 2-AB hexoses extracted from a leaf of I. virginica. The monosaccharides extracted and labelled with 2-AB were separated by Cosmosil 5C18-AR chromatography using solvent system a. The elution positions of standard 2AB monosaccharides are indicated with arrowheads: 1, 2-AB Psi1; 2, 2-AB Fru1; 3, 2-AB Psi2, 2-AB Fru2; 4, 2-AB Gal; 5, 2-AB Man; 6, 2-AB All; 7, 2-AB Glc. Discussion For this study, we developed a comprehensive analytical HPLC system for the separation, identification, and quantification of common and rare hexoses. With the exception of our recent report (18), no practical methods to detect and quantify naturally occurring rare monosaccharides exist. Moreover, although various monoamine-coupling methods have been taken employed for oligosaccharide analyses, they have never been used to detect and quantify ketoses. Our first attempt, using 2-AP, was unsuccessful, not only because 2-AP reacted poorly with the ketohexoses, but also because it catalyzed keto-aldose isomerization in accordance with the Lobry de Bruyn-Alberda van Ekenstein transformation. Conversely, reaction of the monoamine-coupling reagent, 2-AB, with the ketohexoses proceeded without the undesirable isomerization; although, it was necessary to optimize the reaction conditions to generate 2-AB ketoses in large yields. Our optimized reaction conditions for 2-AB labeling of hexoses are as follows: 10 mg 2-AB/20 µl reaction mixture, a reaction temperature of 65°C, and a reaction time of 18 hours. The detection limit was approximately 0.1 pmol for both aldohexoses and ketohexoses, while signal-to-noise ratio was approximately 3. Since the fluorescent intensity of a 2-AB monosaccharide significantly depends on the structure of the sugar, an accurate molar fluorescence coefficient needs to be determined for each sugar. Because the present HPLC system does not discriminate between the R/S-isomers of the 2-AB ketoses or between their d/l-enantiomers, other methods must be developed to solve these issues. We expect that NMR spectroscopy and bio-affinity techniques can be used to discriminate between the R/S-isomers and the d/l-enantiomers. Our system should be capable of detecting and quantifying rare ketoses and aldoses in both natural and artificial products. As an application of this method, we could identify free Psi using a relatively small amount of a natural material, i.e. 25 mg of a fresh I. virginica leaf. The determined Psi content was 2.7 mg/g leaf (wet weight), while no other rare sugars were identified. Compared with the Glc content (0.6 mg/g), this value is apparently high. At the moment, however, it is not clear why the leaf should contain such a high amount of Psi as its biological function(s) is not known. Using this methodology, other types of monosaccharides, such as ketopentoses, could also be targeted. If the locations, quantities, and biosynthetic pathways, of rare sugars could be determined, it should be possible to begin to understand their biological functions. This would pioneer a novel, attractive field in glycoscience, especially considering the recent findings that some rare sugars, e.g. d-Psi, are biologically active (6–10). In this regard, rare sugar studies have just begun. Funding City Area Program (Development stage) Takamatsu, Ministry of Education, Culture, Sports, Science and Technology (MEXT, partial). Conflict of interest None declared. References 1 Itoh H,  Okaya H,  Khan AR,  Tajima S,  Hayakawa S,  Izumori K.  Purification and characterization of D-Tagatose 3-epimerase from Pseudomonas sp. ST-24,  Biosci. Biotech. Biochem. ,  1994, vol.  58 (pg.  2168- 2171) Google Scholar CrossRef Search ADS   2 Takeshita K,  Ishida Y,  Takada G,  Izumori K.  Direct production of allitol from D-fructose by a coupling reaction using D-tagatose 3-epimerase, ribitol dehydrogenase and formate dehydrogenase,  J. Biosci. Bioeng. ,  2000, vol.  90 (pg.  545- 548) Google Scholar CrossRef Search ADS PubMed  3 Takeshita K,  Suga A,  Takada G,  Izumori K.  Mass production of D-psicose from D-fructose by a continuous bioreactor system using immobilized D-tagatose 3-epimerase,  J. Biosci. Bioeng. ,  2000, vol.  90 (pg.  453- 455) Google Scholar CrossRef Search ADS PubMed  4 Izumori K.  Bioproduction strategies for rare hexose sugars,  Naturwissenschaften. ,  2002, vol.  89 (pg.  120- 124) Google Scholar CrossRef Search ADS PubMed  5 Granström TB,  Takata G,  Tokuda M,  Izumori K.  Izumoring: a novel and complete strategy for bioproduction of rare sugars,  J. Biosci. Bioeng. ,  2004, vol.  97 (pg.  89- 94) Google Scholar CrossRef Search ADS PubMed  6 Matsuo T,  Suzuki H,  Hashiguchi M,  Izumori K.  D-psicose is a rare sugar that provides no energy to growing rats,  J. Nutr. Sci. Vitaminol. ,  2002, vol.  48 (pg.  77- 80) Google Scholar CrossRef Search ADS PubMed  7 Livesey G,  Brown JC.  D-tagatose is a bulk sweetener with zero energy determined in rats,  J. Nutr. ,  1996, vol.  126 (pg.  1601- 1609) Google Scholar PubMed  8 Sui L,  Dong Y,  Watanebe Y,  Yamaguchi F,  Hatano N,  Tsukamoto I,  Izumori K,  Tokuda M.  The inhibitory effect and possible mechanisms of D-allose on cancer cell proliferation,  Int. J. Oncol. ,  2005, vol.  27 (pg.  907- 912) Google Scholar PubMed  9 Bautista DA,  Pegg RB,  Shang PJ.  Effect of L-glucose and D-tagatose on bacterial growth in media and a cooked cured ham product,  J. Food Prot. ,  2000, vol.  63 (pg.  71- 77) Google Scholar PubMed  10 Lawson CJ,  Homewood J,  Taylor A.  The effects of L-glucose on memory in mice are modulated by peripherally acting cholinergic drugs,  J. Neurobiol. Learn. Mem. ,  2002, vol.  77 (pg.  17- 28) Google Scholar CrossRef Search ADS   11 Miller BS,  Swan T.  Chromatographic analyses of the free amino-acids, organic acids and sugars in wheat plant extracts,  J. Sci. Food Agric. ,  1960, vol.  11 (pg.  344- 348) Google Scholar CrossRef Search ADS   12 Hough L,  Stacey BE.  The occurrence of D-ribohexulose in Itea ilicifolia, Itea virginica, and Itea yunnanensis,  Phytochemistry ,  1963, vol.  2 (pg.  315- 320) Google Scholar CrossRef Search ADS   13 Hough L,  Stacey BE.  Biosynthesis and metabolism of allitol and D-allulose in Itea plants; incorporation of 14CO2,  Phytochemistry ,  1966, vol.  5 (pg.  215- 222) Google Scholar CrossRef Search ADS   14 Binkley WW.  The fate of cane juice simple sugars during molasses formation IV. Probable conversion of D-fructose to D-psicose,  Int. Sugar J. ,  1963, vol.  65 (pg.  105- 106) 15 Luger A,  Steinhart T.  Carbohydrates in steam treated coffee,  ASIC ,  1995, vol.  Vol. 16  Kyoto Colloque(pg.  366- 371) 16 Oshima H,  Kimura I,  Izumori K.  Psicose contents in various food products and its origin,  Food Sci. Technol. Res. ,  2006, vol.  12 (pg.  137- 143) Google Scholar CrossRef Search ADS   17 Miyanishi N,  Sato N,  Nakakita S,  Sumiyoshi W,  Morimoto K,  Okuma H,  Tokuda M,  Izumori K,  Watanabe E,  Hirabayashi J.  Development of an amperometric flow analysis sensor for specific detection of D-psicose,  Biosens. Bioelectron. ,  2008, vol.  23 (pg.  1347- 1352) Google Scholar CrossRef Search ADS PubMed  18 Nakakita S,  Hasehira K,  Hosokawa T,  Tokuda M,  Izumori K,  Takekawa K,  Hirabayashi J.  Development of a high-sensitivity chromatographic separation system for pyridylaminated aldopentoses and aldohexoses,  J. Chromatogra. A. ,  2009, vol.  1216 (pg.  5112- 5115) Google Scholar CrossRef Search ADS   19 Grill E,  Huber C,  Oefner P,  Vorndran A,  Bonn G.  Capillary zone electrophoresis of p-aminobenzoic acid derivatives of aldoses, ketoses and uronic acids,  Electrophoresis ,  1993, vol.  14 (pg.  1004- 1010) Google Scholar CrossRef Search ADS PubMed  20 Schwaiger H,  Oefner PJ,  Huber C,  Grill E,  Bonn G.  Capillary zone electrophoresis and micellar electrokinetic chromatography of 4-aminobenzonitrile carbohydrate derivatives,  Electrophoresis ,  1994, vol.  15 (pg.  941- 952) Google Scholar CrossRef Search ADS PubMed  21 Hase S,  Ikenaka T,  Matsushima Y.  Structure analyses of oligosaccharides by tagging of the reducing end sugars with a fluorescent compound,  Biochem. Biophys. Res. Commun. ,  1978, vol.  85 (pg.  257- 263) Google Scholar CrossRef Search ADS PubMed  22 Bigge JC,  Patel TP,  Bruce JA,  Goulding PN,  Charles SM,  Parekh RB.  Nonselective and efficient fluorescent labeling of glycans using 2-amino benzamide and anthranilic acid,  Anal. Biochem. ,  1995, vol.  230 (pg.  229- 238) Google Scholar CrossRef Search ADS PubMed  23 Kuraya N,  Hase S.  Release of O-linked sugar chains from glycoproteins with anhydrous hydrazine and pyridylamination of the sugar chains with improved reaction conditions,  J. Biochem. ,  1992, vol.  112 (pg.  122- 126) Google Scholar PubMed  24 Klausen NK,  Bayne S,  Palm L.  Analysis of the site-specific asparagine-linked glycosylation of recombinant human coagulation factor VIIa by glycosidase digestions, liquid chromatography, and mass spectrometry,  Mol. Biotechnol. ,  1998, vol.  9 (pg.  195- 204) Google Scholar CrossRef Search ADS PubMed  25 Ohara K,  Sano M,  Kondo A,  Kato I.  Two-dimensional mapping by high-performance liquid chromatography of pyridylamino oligosaccharides from various glycosphingolipids,  J. Chromatogr. ,  1991, vol.  586 (pg.  35- 41) Google Scholar CrossRef Search ADS PubMed  26 Wing DR,  Garner B,  Hunnam V,  Reinkensmeier G,  Andersson U,  Harvey DJ,  Dwek RA,  Platt FM,  Butters TD.  High-performance liquid chromatography analysis of ganglioside carbohydrates at the picomole level after ceramide glycanase digestion and fluorescent labeling with 2-aminobenzamide, Anal,  Biochem. ,  2001, vol.  298 (pg.  207- 217) 27 Suzuki T,  Matsuo I,  Totani K,  Funayama S,  Seino J,  Taniguchi N,  Ito Y,  Hase S.  Dual-gradient high-performance liquid chromatography for identification of cytosolic high-mannose-type free glycans,  Anal. Biochem. ,  2008, vol.  381 (pg.  224- 232) Google Scholar CrossRef Search ADS PubMed  28 Urashima T,  Odaka G,  Asakuma S,  Uemura Y,  Goto K,  Senda A,  Saito T,  Fukuda K,  Messer M,  Oftedal OT.  Chemical characterization of oligosaccharides in chimpanzee, bonobo, gorilla, orangutan, and siamang milk or colostrum,  Glycobiology ,  2009, vol.  19 (pg.  499- 508) Google Scholar CrossRef Search ADS PubMed  29 Lobry de Bruyn CA,  van Ekenstein WA.  Action des alkalis sur les sucres, II,  Recl. Trav. Chim. Pays-Bas. ,  1895, vol.  14 (pg.  203- 216) Google Scholar CrossRef Search ADS   30 Lobry de Bruyn CA,  van Ekenstein WA.  Action des alkalis sur les sucres, IV,  Recl. Trav. Chim. Pays-Bas. ,  1897, vol.  16 (pg.  257- 261) Google Scholar CrossRef Search ADS   31 Speck JC.  The Lobry de Bruyn-Alberda van Ekenstein transformation,  Adv. Carbohyd. Chem. ,  1958, vol.  13 (pg.  63- 103) 32 Fischer HOL,  Taube G,  Baer E.  Über den krystallisierten Glycerinaldehyd und seine Umwandlung in Dioxy-aceton,  Ber. ,  1927, vol.  60 (pg.  479- 485) 33 Ekeberg D,  Morgenlie S,  Stenstrøm Y.  Isomerisation of aldoses in pyridine in the presence of aluminium oxide,  Carbohydr. Res. ,  2005, vol.  340 (pg.  373- 377) Google Scholar CrossRef Search ADS PubMed  34 Ekeberg D,  Morgenlie S,  Stenstrøm Y.  Aldose-ketose interconversion in pyridine in the presence of aluminium oxide,  Carbohydr. Res. ,  2007, vol.  342 (pg.  1992- 1997) Google Scholar CrossRef Search ADS PubMed  35 Hase S.  Analysis of sugar chains by pyridylamination,  Methods Mol. Biol. ,  1993, vol.  14 (pg.  69- 80) Google Scholar PubMed  36 Chiba A,  Matsumura K,  Yamada H,  Inazu T,  Shimizu T,  Kusunoki S,  Kanazawa I,  Kobata A,  Endo T.  Structures of sialylated O-linked oligosaccharides of bovine peripheral nerve alpha-dystroglycan. The role of a novel O-mannosyl-type oligosaccharide in the binding of alpha-dystroglycan with laminin,  J. Biol. Chem. ,  1997, vol.  272 (pg.  122- 126) Abbreviations Abbreviations All allose Alt altrose 2-AB 2-aminobenzamide 2-AP 2-aminopyridine Fru fructose Gal galactose Glc glucose Gul gulose HPLC high performance liquid chromatography Ido idose Man mannose PA- pyridylaminated Psi psicose Sor sorbose Tag tagtose Tal talose © The Authors 2009. Published by Oxford University Press on behalf of the Japanese Biochemical Society. All rights reserved
TI - A comprehensive HPLC analytical system for the identification and quantification of hexoses that employs 2-aminobenzamide coupling
JF - The Journal of Biochemistry
DO - 10.1093/jb/mvp199
DA - 2010-03-26
UR - https://www.deepdyve.com/lp/oxford-university-press/a-comprehensive-hplc-analytical-system-for-the-identification-and-IdZcT8Gs7E
SP - 501
EP - 509
VL - 147
IS - 4
DP - DeepDyve
ER -