Preparation and properties of a coniferin enantiomer

l-Coniferin (1L), which is an enantiomer of natural coniferin (d-coniferin (1D)), was prepared from l-glucose according to the conventional method for compound 1D. The reactivity of l-glucose and its derivatives was found to be almost same as that of the corresponding d-glucose and its derivatives during the preparation for compound 1L. Compound 1L showed resistance toward enzymatic hydrolysis by commercial β-glucosidase from Almond. However, unlike compound 1D, compound 1L was not transported across the membrane obtained from differentiating xylem of a hybrid poplar in the present assay. The result suggested for the first time that the d-/l-configuration of the glucose moiety of coniferin is an important factor affecting coniferin transport across the membrane.

Natural coniferin (hereafter “d-coniferin (1D)”) is a d-glucoside consisting of d-glucose and coniferyl alcohol. It is considered to be a form of coniferyl alcohol that functions in storage and transport in lignin biosynthesis [1,2,3,4]. There are three important pieces of information known about coniferin transport. First, compound 1D is found in the cambial sap of many tree species [1, 5]. Second, coniferin β-glucosidase is located in the cell wall of coniferous and broad-leaf trees [6, 7]. Third, the incorporation of radioactivity into cell wall lignin is observed when radiolabeled compound 1D is fed to a variety of plants [8]. Taken together, this suggests that compound 1D might be transported through the membrane to the cell wall and be subjected to enzymatic hydrolysis by β-glucosidase to form coniferyl alcohol for lignification. However, this hypothesis has not yet been proven and there is still debate as to whether coniferyl alcohol is transported in its free or glucoside form [3, 4].

Theoretically, there should be a non-natural coniferin, an enantiomer of compound 1D (hereafter “l-coniferin (1L)”), which consists of l-glucose and coniferyl alcohol as shown in Fig. 1, because there is an enantiomer of d-glucose (l-glucose). To the best of our knowledge, compound 1L was not found in nature. It is widely known that a pair of enantiomers has the same chemical and physical properties except for their optical properties such as specific rotation. In addition, l-glucoside is thought not to be hydrolyzed by β-glucosidase because of the substrate specificity of the enzyme [9], although the experimental data have not been found. Therefore, compound 1L might be useful as a transport tracer in an administration experiment, because it might be transported to the site of lignification to a similar degree as compound 1D, but would not be hydrolyzed by the enzyme as it accumulated. However, compound 1L has not been available, because there is no report about the preparation of compound 1L.

Natural coniferin (1D) and its enantiomer (1L)

Full size image

This research describes the first preparation of compound 1L and its properties including its enzymatic hydrolysis and membrane transport ability.

Materials

l-Glucose and β-glucosidase (from Almond, 121 U/mg) were purchased from Tokyo Chemical Industry Co. (Tokyo, Japan) and Oriental Yeast Co. (Tokyo, Japan), respectively. All other chemicals were purchased from commercial sources and used without further purification unless otherwise noted.

Measurements

¹H and ¹³C NMR spectra were recorded on a Varian INOVA300 MHz FT-NMR (300 MHz) spectrometer (Agilent Technologies, Santa Clara, CA, USA) using tetramethylsilane as an internal standard in DMSO-d₆ for compounds 1L and 1D or CDCl₃ for others. Chemical shift (δ) and coupling constant (J) are given in ppm (parts per million) and Hz, respectively. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) measurements were recorded on a Bruker MALDI-TOF MS REFLEX III (Bruker, Billerica, MA, USA) in the positive and linear ion modes. A nitrogen laser was used for the ionization of the samples. All spectra were obtained using 2,5-dihydroxybenzoic acid as a matrix. Specific rotations were recorded on a JASCO P-2200 polarimeter (JASCO, Hachioji, Japan) in H₂O for compounds 1L and 1D or CHCl₃ for others, and were determined as the average values of five measurements. Melting points (m.p.) were measured in a micro-melting point apparatus (Yanagimoto Seisakusho, Kyoto, Japan). UV–Vis spectra were recorded on a JASCO V-560 spectrophotometer (JASCO).

Preparation of l-coniferin (1L)

Compound 1L was prepared from l-glucose (2L) according to the conventional method for compound 1D [10] as shown in Fig. 2.

Synthetic route for d-conferin (1D) from d-glucose (2D) (a); l-conferin (1L) from l-glucose (2L) (b)

Full size image

l-Glucose pentaacetate (3L)

Compound 2L (5 g, 27.8 mmol) and CH₃COONa (2.5 g, 30.5 mmol) were suspended in Ac₂O (25 mL). The suspension was stirred at 80 °C for 3.5 h, cooled to ambient temperature and extracted with EtOAc. The organic layer was washed with a saturated NaHCO₃ solution, distilled water, and brine, dried over Na₂SO₄ and evaporated to give a colorless residue. The residue was recrystallized from EtOH to give compound 3L (9.63 g, 88.9% yield).

Compound 3L: \(\left[ \alpha \right]_{D}^{25}\) = − 20.7° (c = 1.03, in CHCl₃); ¹H and ¹³C NMR spectra data of compound 3L were in agreement with the published data of l-glucose pentaacetate [11]; MALDI-TOF MS: m/z calcd. for [M+Na]⁺ C₁₆H₂₂NaO₁₁: 413.34, found: 413.56.

2,3,4,6-Tetra-O-acetyl-α-l-glucopyranosyl bromide (4L)

Compound 3L (1000 mg, 2.56 mmol) was dissolved in CH₂Cl₂ (2 mL). 33%-HBr in AcOH (2.5 mL, 14.4 mmol) was added to the solution at 0 °C. After stirring the reaction solution at ambient temperature for 2 h, distilled water (10 mL) was added to the solution at 0 °C. The reaction mixture was extracted with EtOAc. The organic layer was washed with distilled water, a saturated NaHCO₃ solution, and brine, dried over Na₂SO₄ and evaporated to give a colorless residue. The residue was recrystallized from EtOH to give compound 4L (845 mg, 80.2% yield).

Compound 4L: \(\left[ \alpha \right]_{D}^{25}\) = − 194.3° (c = 1.04, in CHCl₃); ¹H NMR (CDCl₃): δ 6.62 (d, 1H, J = 4.2, H-1), 5.56 (t, 1H, J = 9.8, H-3), 5.16 (t, 1H, J = 9.8, H-4), 4.84 (dd, 1H, J = 9.8, 4.2, H-2), 4.33 (dd, 1H, J = 14.1, 4.2, H-6a), 4.30 (ddd, J = 9.8, 4.2, 1.5, H-5), 4.13 (dd, 1H, J = 14.1, 1.5, H-6b), 2.11, 2.10, 2.06, 2.04 (s, 3H, acetyl CH₃); ¹³C NMR (75 MHz, CDCl₃): δ 170.6, 170.0, 169.9, 169.6 (acetyl C=O), 86.7 (C-1), 72.2 (C-5), 70.7 (C-2), 70.3 (C-3), 67.3 (C-4), 61.0 (C-6), 20.8, 20.8, 20.7, 20.7 (acetyl CH₃); MALDI-TOF MS: m/z calcd. for [M+Na]⁺ C₁₄H₁₉BrNaO₉: 433.20, found: 434.99.

(4-Formyl-2-methoxy)phenyl 2,3,4,6-tetra-O-acetyl-β-l-glucopyranoside (5L)

Compound 4L: (700 mg, 1.65 mmol) and vanillin (250 mg, 1.65 mmol) were dissolved in quinoline (4.5 mL). Ag₂O (390 mg) was added to the solution at 0 °C. The reaction mixture was stirred at ambient temperature for 1.5 h, filtered through Celite (535RVS, Nacalai Tesque (Kyoto, Japan)), and extracted with EtOAc. The organic layer was washed with a 1 M HCl solution, a saturated NaHCO₃ solution, and brine, dried over Na₂SO₄ and evaporated to give a light brown residue. The residue was recrystallized from EtOH to give compound 5L (599 mg, 72.9% yield).

Compound 5L: \(\left[ \alpha \right]_{D}^{25}\) = + 42.9° (c = 0.53, in CHCl₃); ¹H NMR (300 MHz, CDCl₃): δ 9.90 (s, 1H, CHO), 7.44 (d, 1H, J = 2.1, H-2′), 7.42 (dd, 1H, J = 8.7, 2.1, H-6′), 7.22 (d, 1H, J = 8.7, H-5′), 5.37–5.27 (m, 2H, H-2, H-3), 5.22–5.15 (m, 1H, H-4), 5.11 (d, 1H, J = 7.8, H-1), 4.28 (dd, 1H, J = 12.3, 5.4, 1H, H-6a), 4.19 (dd, 1H, J = 12.3, 2.7, H-6b), 3.90 (s, 3H, OCH₃), 3.85 (ddd, 1H, J = 9.8, 5.4, 2.7, H-5), 2.08, 2.08, 2.05, 2.05 (s, 3H, acetyl CH₃); ¹³C NMR (75 MHz, CDCl₃): δ 191.1 (C-α), 170.7, 170.4, 169.5, 169.4 (acetyl C=O), 151.2 (C-4′), 151.1 (C-3′), 133.0 (C-1′), 125.5 (C-6′), 118.3 (C-5′), 110.8 (C-2′), 99.9 (C-1), 72.5 (C-3), 72.4 (C-5), 71.1 (C-2), 68.4 (C-4), 62.0 (C-6), 56.2 (OCH₃), 20.9, 20.9, 20.8, 20.8 (acetyl CH₃); MALDI-TOF MS: m/z calcd. for [M+Na]⁺ C₂₂H₂₆NaO₁₂: 505.44, found: 505.35.

(4-Ethoxycarbonyl-2-methoxy)phenyl 2,3,4,6-tetra-O-acetyl-β-l-glucopyranoside (6L)

Compound 5L (1000 mg, 2.07 mmol) and ethyl malonic acid (520 mg, 3.96 mmol) were dissolved in pyridine (16.5 mL). After the addition of piperidine (0.275 mL, 2.78 mmol), the reaction solution was stirred at 100 °C for 1.5 h and concentrated by azeotrope distillation with EtOH to give a colorless residue. The residue was recrystallized from EtOH to give compound 6L (979 mg, 85.9% yield).

Compound 6L: \(\left[ \alpha \right]_{D}^{25}\) = + 27.2° (c = 0.56, in CHCl₃); ¹H NMR (300 MHz, CDCl₃): δ 7.62 (d, 1H, J = 16.2, H-α), 7.10 (d, 1H, J = 8.4, H-5′), 7.07 (d, 1H, J = 1.5, H-2′), 7.05 (dd, 1H, J = 8.4, 1.5, H-6′), 6.35 (d, 1H, J = 16.2, H-β), 5.40–5.26 (m, 2H, H-2, H-3), 5.20–5.13 (m, 1H, H-4), 5.02–5.00 (m, 1H, H-1), 4.28 (dd, 1H, J = 12.5, 5.1, H-6a), 4.26 (dd, 2H, J = 14.1, 7.2, CH₂), 4.17 (dd, 1H, J = 12.5, 2.4, H-6b), 3.85 (s, 3H, OCH₃), 3.80 (ddd, 1H, J = 10.2, 5.1, 2.4, H-5), 2.08, 2.05, 2.05, 2.04 (s, 3H, acetyl CH₃), 1.34 (t, 3H, J = 7.2, ethyl CH₃); ¹³C NMR (75 MHz, CDCl₃): δ 170.5, 170.3, 169.4, 169.3 (acetyl C=O), 166.9 (C-ɤ), 150.7 (C-3′), 147.7 (C-4′), 143.9 (C-α), 131.0 (C-1′), 121.6 (C-6′), 119.5 (C-5′), 117.6 (C-β), 111.3 (C-2′), 100.3 (C-1), 72.4 (C-3), 72.0 (C-5), 71.0 (C-2), 68.3 (C-4), 61.8 (C-6), 60.5 (ethyl CH₂), 56.0 (OCH₃), 20.7, 20.6, 20.6, 20.6 (acetyl CH₃), 14.3 (ethyl CH₃); MALDI-TOF MS: m/z calcd. for [M+Na]⁺ C₂₆H₃₂NaO₁₃: 575.53, found: 575.57.

(4-(3-Hydroxy-2-propenyl)-2-methoxy)phenyl β-l-glucopyranoside (l-coniferin) (1L)

Compound 6L (900 mg, 1.63 mmol) was dissolved in toluene (22 mL). The 1.01 mol/L of DIBAL-H in toluene (15 mL, 15.1 mmol) was added dropwise to the solution at 0 °C for 10 min. The reaction mixture was stirred at 0 °C for 1 h and EtOH (20 mL) was added slowly to the mixture. The reaction mixture was stirred at 0 °C for another 30 min, concentrated and filtered with hot water. The filtrate was concentrated by azeotrope distillation with EtOH to give a colorless residue. The residue was purified by preparative thin-layer chromatography using a silica gel plate (silica gel 60F₂₅₄, 2-mm thickness, Merck, Darmstadt, Germany) developed with 20% MeOH/CH₂Cl₂ (v/v) and recrystallized from water three times to give compound 1L (295 mg, 49.1% yield).

Compound 1L: m.p.: 179–182 °C; \(\left[ \alpha \right]_{D}^{25}\) = + 65.8° (c = 0.5, in CHCl₃); ¹H NMR (300 MHz, DMSO-d₆): δ 7.06 (d, 1H, J = 1.8, H-2′), 7.02 (d, 1H, J = 8.4, H-5′), 6.90 (dd, 1H, J = 8.4, 1.8, H-6′), 6.48 (d, 1H, J = 15.9, H-α), 6.28 (dt, 1H, J = 15.9, 5.1, H-β), 5.25 (d, 1H, J = 4.8, OH-3), 5.10 (d, 1H, J = 4.8, OH-2), 5.04 (d, 1H, J = 4.8, OH-4), 4.89 (d, 1H, J = 8.1, H-1), 4.85 (d, 1H, J = 5.4, OH-γ), 4.57 (t, 1H, J = 5.4, OH-6), 4.10 (dd, 1H, J = 5.1, 1.2, H-ɤa), 4.09 (dd, 1H, J = 5.1, 1.2, H-ɤb), 3.79 (s, 3H, OCH₃), 3.68–3.65 (m, 1H, H-6a), 3.52–3.17 (m, 5H, H-2, H-3, H-4, H-5, H-6b); ¹³C NMR (75 MHz, DMSO-d₆): δ 149.4 (C-3′), 146.4 (C-4′), 131.4 (C-1′), 129.4 (C-β), 128.9 (C-α), 119.5 (C-6′), 115.6 (C-5′), 110.2 (C-2′), 100.4 (C-1), 77.4 (C-3), 77.3 (C-5), 73.7 (C-2), 70.7 (C-4), 62.1 (C-ɤ), 61.1 (C-6), 56.0 (OCH₃); MALDI-TOF MS: m/z calcd. for [M+Na]⁺ C₁₆H₂₂NaO₈: 365.34, found: 365.30.

Preparation of d-coniferin (1D)

Compound 1D was also prepared by the conventional method as shown in Fig. 2a [10]. Compound 3D: \(\left[ \alpha \right]_{D}^{25}\) = + 21.4° (c = 1.34, in CHCl₃); 4D: \(\left[ \alpha \right]_{D}^{25}\) = + 193.5° (c = 1.13, in CHCl₃), 5D: \(\left[ \alpha \right]_{D}^{25}\) = − 43.7° (c = 0.52, in CHCl₃), 6D: \(\left[ \alpha \right]_{D}^{25}\) = − 26.6° (c = 0.56, in CHCl₃), 1D: m.p. 181–183 °C (literature: 183–185 °C [10], 185–188 °C [12]); \(\left[ \alpha \right]_{D}^{25}\) = − 64.8° (c = 0.43, in CHCl₃).

Enzymatic hydrolysis of d- and l-coniferin (1D and 1L) by β-glucosidase

An aqueous solution of β-glucosidase (6 × 10⁻⁵ g/L, 5 mL) was added to an aqueous solution of coniferin (0.3 mmol/L, 10 mL). The reaction solution was stirred at 40 °C. An aliquot of the reaction solution (1 mL) was taken out at the prescribed time and poured into distilled water (3 mL) and then 0.1 mL of a 2.5 mM NaOH aqueous solution was immediately added. The mixed solution was subjected to UV–Vis measurement. The amount of coniferyl alcohol formed was determined using a calibration curve of coniferyl alcohol at 311 nm.

Transport assay of d- and l-coniferin (1D and 1L)

The preparation of microsomal fractions from differentiating xylem of poplar (Populus sieboldii × P. grandidentata) and the transport assay were carried out according to the methods used in previous studies [3, 13]. Uptake of coniferin by membrane vesicles was measured at 28 °C for 20 min in 100 μL of reaction mixture [50 mM HEPES–KOH (pH 7.5), 5 mM Mg/ATP, 50 μM substrate and membrane vesicles (ca. 10 μg protein)], unless otherwise stated. Data are reported as technical replicates.

Preparation of d- and l-coniferin (1D and 1L)

Compound 1L was prepared from compound 2L according to the conventional method for compound 1D [10] as shown in Fig. 2. Compound 1D was also prepared as a control compound. All the reaction steps [(a) acetylation, (b) bromination, (c) glycosidation, (d) Knoevenagel condensation, (e) reduction with DIBAL-H], proceeded smoothly to afford the final compound 6L. Indeed, the reactivity of the l-glucose/l-glucose derivative (compounds 2L to 6L) was found to be almost the same as the d-glucose/d-glucose derivative (compounds 2D to 6D) during the preparation of compound 1L. Reports of similar reactivities among l-glucose and d-glucose derivatives are known in the literature for the synthesis of digitoxigenin glycoside [14]. All products (Compounds 1L, 3L to 6L) were characterized by acquiring their ¹H and ¹³C NMR and MALDI-TOF MS spectra, and their specific rotation. The ¹H and ¹³C NMR spectra of all the l-glucose derivatives had the same peak pattern as those of the d-glucose derivatives, but the sign of the specific rotation of the l-glucose derivatives was opposite to that of the specific rotation of the d-glucose derivatives (Additional file 1: Figs. S1–S10).

As an example, the ¹H and ¹³C NMR spectra of compounds 1L and 1D are shown in Fig. 3. The spectra of compound 1L were the same as those of compound 1D. The value of [M+Na]⁺ for compound 1L also corresponded to its calculated value from the MALDI-TOF–MS data of compound 1L. The specific rotations of compounds 1L and 1D were + 65.8° and − 64.8°, respectively. These results clearly indicate that compound 1L was an enantiomer of compound 1D.

¹H NMR spectra of compounds 1L (a) and 1D (b); ¹³C NMR spectra of compounds 1L (a′) and 1D (b′)

Full size image

Enzymatic hydrolysis of d- and l-coniferin (1D and 1L) by β-glucosidase

The enzymatic hydrolysis of compounds 1L and 1D by the commercial β-glucosidase from Almond was evaluated based on the formation of coniferyl alcohol in the reaction. The determination of the coniferyl alcohol formed in the reaction mixture was performed by a UV–Vis method (detection wavelength: 311 nm), because the peak for coniferyl alcohol in an alkali solution at 311 nm did not overlap with any peaks from coniferin or the β-glucosidase (Fig. 4a). The percent hydrolysis of compounds 1L and 1D is shown in Fig. 4b. The percent hydrolysis of compound 1D increased with an increase of time and reached almost 100% after 24 h. By contrast, the hydrolysis of compound 1L was minimal with only 5.4% hydrolyzed after 24 h. The results clearly show that compound 1L was resistant to enzymatic hydrolysis by the β-glucosidase, although further investigation (for example, the experiments using other β-glucosidase) is required.

UV–Vis spectra of coniferin and coniferyl alcohol in an alkali solution (a); hydrolysis of compounds 1D and 1L by β-glucosidase (b)

Full size image

Transport assay of d- and l-coniferin (1D and 1L)

ATP-dependent compound 1D transport in the lignifying tissues of woody plants was suggested from the transport assay results of compound 1D [3]. That is, strong transport activity of compound 1D was observed in the presence of ATP, whereas weak transport activity was observed in the absence of ATP. Compound 1L was also tested in the transport assay. Figure 5 shows the uptake of compounds 1L and 1D into crude microsomal membrane vesicles derived from plasma membrane, vacuolar membrane, Golgi membrane and other endomembranes obtained from differentiating xylem of hybrid poplar. Surprisingly, only weak uptake activity was observed for compound 1L in the presence of ATP and no uptake activity was observed in the absence of ATP. Thus, there was a significant difference of uptake activity between compounds 1L and 1D in the presence of ATP, although compound 1L is an enantiomer of compound 1D. Previous study indicated that transport of compound 1D is involved in vacuolar type H⁺-ATPase which is localized not only vacuolar membrane but also endoplasmic reticulum, Golgi apparatus, and other endomembrane systems [3]. Although it is indistinct whether transport activity of compound 1D shows transport into vacuole and/or other endomembrane systems, compound 1L was not transported across the crude microsomal membranes containing plasma membrane, vacuolar membrane, Golgi membrane and other endomembranes obtained from differentiating xylem of hybrid poplar. Therefore, the d-/l-configuration of the sugar moiety of coniferin was strictly recognized during the coniferin transport process across the membrane in the present assay.

Uptake of compounds 1D and 1L into membrane vesicles in the presence and absence of ATP over 20 min

Full size image

Compound 1L was successfully prepared according to the conventional method for compound 1D. Compound 1L was not found to be useful as a transport tracer in an administration experiment, although compound 1L was resistant toward enzymatic hydrolysis by commercial β-glucosidase from Almond. However, the d-/l-configuration of the sugar moiety of coniferin was found to be a factor that affected coniferin transport across the membrane. This new finding might provide supporting evidence that compound 1D is a transport form of coniferyl alcohol.

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

MALDI-TOF MS:

matrix-assisted laser desorption/ionization time-of-flight mass spectrometry

m.p.:

melting point

The authors thank to JSPS Grant-in-Aid for Challenging Research 18K19230.

This work was supported by a JSPS Grant-in-Aid for Challenging Research 18K19230.

Affiliations

1. Division of Forest and Biomaterials Science, Graduate School of Agriculture, Kyoto University, Kitashirakawa oiwake-cho, Sakyo-ku, Kyoto, 606-8502, Japan
   • Haruka Maeda
   • , Keiji Takabe
   • , Hiroshi Kamitakahara
   •  & Toshiyuki Takano
2. Department of Forest and Environmental Sciences, Faculty of Agriculture, University of Miyazaki, 1-1 Gakuenkibanadai-nishi, Miyazaki, 889-2192, Japan
   • Taku Tsuyama

Contributions

HM contributed to preparation of d- and l-coniferin and their enzymatic hydrolysis experiments. HK supported to HM’s experiments. TT and KT contributed to transport assay. TT (corresponding author) designed this study and wrote this paper. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Toshiyuki Takano.

Competing interests

The authors declare that they have no competing interests.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Reprints and Permissions