Conformationally flexible ellagitannins: conformational analysis of davidiin and punicafolin by DFT-based ¹H–¹H coupling constant calculations

Abstract

Many ellagitannins with various conformations of their glucose moiety have been isolated from natural plant sources. Herein, we performed a conformational analysis using density functional theory (DFT) calculations of ¹H–¹H coupling constants. In solution, davidiin was found to exist as an equilibrium mixture of the B_O,3 (boat) and ¹C₄ (chair) conformations, whereas punicafolin is an equilibrium mixture of the ³S₁ (skew-boat) and ¹C₄ conformations. Their conformational equilibria vary depending on solvent and temperature. Such conformational flexibility may be important for the biosynthesis of ellagitannins with diverse structures.

---

Introduction

Hydrolyzable tannins, comprising gallotannins and ellagitannins, are a group of plant polyphenols possessing considerable structural diversity and various biological activities.¹ Most ellagitannins contain a glucose core esterified by acyl groups, primarily hexahydroxydiphenoyl (HHDP) and dehydrohexahydroxydiphenoyl (DHHDP) units, both derived from gallic acid.¹ Feldman and Quideau suggested that the DHHDP group is biosynthesized via intramolecular oxidative coupling of galloyl esters.² Furthermore, the DHHDP group can be converted to an HHDP group by chemical reduction,^2a,3 and our group recently demonstrated its reductive metabolism to the HHDP group in several plants.⁴ In addition, the DHHDP group can be formed by CuCl₂-mediated oxidation of galloyl ester derivatives in aqueous media,⁵ strongly indicating that DHHDP is the initial oxidative-coupling product of two galloyl groups during ellagitannin biosynthesis, after which reductive metabolism yields HHDP esters (Fig. 1a).^4,5

Fig. 1 (a) Plausible biosynthetic pathway of dehydrohexahydroxydiphenoyl (DHHDP) and hexahydroxydiphenoyl (HHDP) groups. (b) Structures of hydrolyzable tannins with various conformations. (c) Structures of davidiin (5), punicafolin (6), and corilagin (7).

Glucopyranose derivatives can adopt various conformations, including chair (C), boat (B), skew- or twist-boat (S), and envelope (E) (Fig. S1, ESI‡).⁶ Many ellagitannins are presumably biosynthesized from a common precursor, 1,2,3,4,6-penta-O-galloyl-β-D-glucose (1), which adopts a ⁴C₁ conformation,^3a,7 and several examples are shown in Fig. 1b. The glucopyranose unit of 1(β)-O-galloylpedunculagin (casuarictin), bearing 2,3-(S_a)-HHDP and 4,6-(S_a)-HHDP groups, exhibits the same ⁴C₁ conformation as 1.⁷ In contrast, geraniin (2), which contains a 3,6-(R_a)-HHDP and a 2,4-(R)-DHHDP group, adopts a ¹C₄ conformation with all substituents in axial orientations.^4c,8 Amariin (3), bearing 2,4-(R)- and 3,6-(R)-DHHDP groups, exists in an ^O,3B form,^4c while phyllanemblinin B (4), with a 2,4-(R_a)-HHDP group, adopts a ³S₁ conformation.⁹ However, the precise conformations of several ellagitannins, including davidiin (5) with a 1,6-(S_a)-HHDP group^3a,10 and punicafolin (6) with a 3,6-(R_a)-HHDP group,¹¹ remain unclear (Fig. 1c). Moreover, corilagin (7), an analogue of 6 lacking galloyl groups at C-2 and C-4 (Fig. 1c), is known to adopt different conformations depending on the solvent.¹² Notably, the conformations of 3,6-bridged glucopyranose derivatives vary with both substituent type and bridge length.¹³

To clarify the biosynthetic pathways and biological functions of ellagitannins, determining their precise conformations is essential. ¹H–¹H coupling constants (J_H,H) are highly informative for conformational analyses of glucose derivatives because they directly reflect the dihedral angles between vicinal protons. Therefore, in this study, we aimed to clarify the precise conformations of 5, 6, and related ellagitannins by comparing their experimental and DFT-calculated J_H,H values.

Results and discussion

Conformational analysis procedure

Ellagitannins possess many phenolic hydroxy groups, each capable of adopting different orientations, which complicates conformational analysis. To manage this complexity, we employed a multi-step approach:

(1) Initial conformational search. We conducted a broad conformational search on the molecular skeleton using the MMFF94 force field. This initial step accounted for ring-flipping in the glucopyranose moiety and macrocyclic rearrangements of the HHDP ester but did not consider the orientations of the phenolic hydroxy groups.

(2) Geometry optimization and classification. The candidate structures were optimized at the B3LYP/6-31G(d,p) level of theory. We then classified the resulting geometries according to their glucopyranose ring conformation. Within each conformation type, only the lowest-energy conformers were retained for further analysis.

(3) Refined search including hydroxy orientations. We next performed a secondary conformational search for each conformation type, explicitly considering different orientations of the phenolic hydroxy groups. Conformers within 6 kcal mol⁻¹ of the global minimum were re-optimized at the same DFT level, ensuring thorough sampling of relevant conformational space.

(4) Calculation of ¹H–¹H coupling constants (J_H,H). We computed J_H,H values at the B3LYP/6-31G(d,p)u+1s level, considering only the Fermi contact term.¹⁴ Conformers with a Boltzmann population above 1% were included in Boltzmann-weighted averaging to obtain final calculated J_H,H values in each group. The RMSD (root-mean-square deviation) between calculated and experimental J_H,H values using this basis set is reported as about 0.49 Hz.¹⁴ Hence, deviations below this threshold generally indicate good agreement. In some cases, signals observed as broad singlets in the ¹H NMR were treated as having J_H,H = 0 Hz, which can enlarge the apparent RMSD. Even if the deviation exceeds 0.49 Hz, comparing RMSD values remains helpful for identifying the most likely conformer(s).

By following these steps—systematic conformational searches, geometry optimizations, and direct J_H,H comparisons—we can characterize complex conformational equilibria in ellagitannins and related molecules.

Conformational analysis of phyllanemblinin B (4)

To validate the above approach, we first examined phyllanemblinin B (4), an ellagitannin with a relatively simple structure. An initial report suggested that it adopts a skew-boat conformation,^9b and Wakamori and co-workers later proposed a ³S₁ form based on experimental J_H,H values and MMFF calculations.^9a,15 Our present DFT analysis identified three possible conformations at the B3LYP/6-31G(d,p) level: ³S₁ (ΔG = 0.0 kcal mol⁻¹), B_1,4 (ΔG = +4.8 kcal mol⁻¹), and ¹C₄ (ΔG = +5.3 kcal mol⁻¹) (Fig. S46 and Table S9, ESI‡). The calculated J_H,H values for the ³S₁ conformer (lowest in free energy) agreed well with experiment (RMSD = 0.22 Hz),^9a supporting the conclusion that 4 most likely adopts a ³S₁ conformation (Table 1 and Tables S10–S12, ESI‡).¹⁶

Table 1 Experimental and calculated J_H,H values (Hz) for phyllanemblinin B (4)

	Exptl^a	Calcd^b
		^3 S 1	B 1,4	^1 C 4
a 500 MHz (ref. 9a). b Calculated at the B3LYP/6-31G(d,p)u+1s//B3LYP/6-31G(d,p) level. c RMSD (root mean square deviation) is calculated from the squared differences between the calculated and experimental JH,H values, with the final value being the square root of their average.
J 1,2	5.9	6.2	7.5	0.5
J 2,3	1.0	0.7	0.9	1.3
J 3,4	3.7	3.7	3.0	3.2
J 4,5	≈0	0.2	0.8	1.7
RMSD^c		0.22	0.96	2.87

---

Conformational analysis of davidiin (5)

Davidiin (5), which has a 1,6-(S_a)-HHDP and 2,3,4-trigalloyl substituents, has been isolated from Davidia involucrata,^3a,10,17Acer saccharum,¹⁸ and Persicaria capitata (syn. Polygonum capitatum).¹⁹ Various biological activities have been reported for this compound.^17,19a,20 Although a previous study assumed a skew-boat conformation for its glucopyranose ring,^3a subsequent literature has inconsistently depicted it as ¹C₄,^7b–d,f,213S₁,²² or B_O,3^17,19b,23 conformations (Fig. S2, ESI‡), leading to confusion. In particular, B_O,3 was proposed based on J_H,H data in acetone-d₆,^17,19b,23 but those J_H,H values differ from those of geraniin (2) (¹C₄)^8c or phyllanemblinin B (4) (³S₁)^9a (Table S1, ESI‡).

Our DFT analysis revealed three possible conformer classes for 5 at B3LYP/6-31G(d,p): ¹C₄ (ΔG = 0.0 kcal mol⁻¹), ^1,4B–¹S₅ (an intermediate between ^1,4B and ¹S₅, ΔG = +2.8 kcal mol⁻¹), and B_O,3 (ΔG = +4.4 kcal mol⁻¹)²⁴ as summarized in Fig. 2 and Table S13 (ESI‡). None of these three alone reproduces the experimental J_H,H data in acetone-d₆ (Table 2 and Tables S18–S20, ESI‡), implying that 5 does not adopt a single dominant conformation. However, a mixture of the B_O,3 and ¹C₄ conformers, weight-averaged at 60 : 40, gave the best fit with the experimental data (RMSD = 0.36 Hz) (Table 2). This ratio is approximate, as uncertainties in the DFT calculations and the inherent ∼0.49 Hz RMSD in the J_H,H calculations limit the precision of population estimates. Indeed, similarly good agreements were obtained for slightly different ratios (Table S21, ESI‡). Nevertheless, our conformational analysis procedure exhaustively identifies all feasible conformers of 5, and by selecting the combination and population ratio that minimize the discrepancy (RMSD) between calculated and observed J_H,H values, we can perform a relatively reliable conformational analysis. Accordingly, as for 5 in acetone-d₆, the 60 : 40 ratio is presented here as a representative best fit rather than an exact measurement. Moreover, including a small fraction of ^1,4B–¹S₅ (e.g., B_O,3/¹C₄/^1,4B–¹S₅ = 55 : 40 : 5) improved the fit further (RMSD = 0.15 Hz) (Table 2), suggesting a minor population of ^1,4B–¹S₅. Moreover, the coupling constants J_1,2, J_2,3, and J_3,4 increase at lower temperature (Table S3, ESI‡), pointing to an increased B_O,3 population relative to ¹C₄.

Fig. 2 Three representative conformations of davidiin (5) optimized at the B3LYP/6-31G(d,p) level.

Table 2 Experimental and calculated J_H,H values (Hz) for davidiin (5)

	Exptl^a				Calcd^d
	acetone-d6^b	CD3OD^b	DMSO-d6^b	D2O^c	^1 C 4	^1,4 B–^1S5	B O,3	B O,3/^1C4 (60 : 40)	B O,3/^1C4/^1,4B–^1S5 (55 : 40 : 5)
a 500 MHz. b 20 °C. c 80 °C. d Calculated at the B3LYP/6-31G(d,p)u+1s//B3LYP/6-31G(d,p) level. e RMSD (root mean square deviation) was calculated as the square root of the average of the squared differences between the calculated and experimental JH,H values.
J 1,2	2.9	2.9	4.0	≈0	0.8	2.2	4.1	2.8	2.7
J 2,3	7.3	7.1	10.2	≈0	2.1	0.3	11.5	7.7	7.2
J 3,4	6.7	6.4	8.9	≈0	2.2	7.4	9.7	6.7	6.6
J 4,5	2.6	2.5	3.5	≈0	0.5	11.5	3.3	2.2	2.6
J 5,6a	5.3	5.1	4.7	5.2	5.4	4.9	4.8	5.1	5.1
J 5,6b	12.1	11.8	11.9	12.1	12.7	0.2	12.6	12.7	12.0
RMSD (acetone-d6)^e					3.08	5.88	2.22	0.36	0.15
RMSD (CD3OD)^e					2.96	5.78	2.36	0.48	0.17
RMSD (DMSO-d6)^e					4.69	6.28	0.71	−	−
RMSD (D2O)^e					1.32	6.69	6.53	−	−

---

¹H NMR spectra of 5 in CD₃OD, DMSO-d₆, and D₂O (Fig. 3a) revealed further solvent-dependent shifts in J_H,H. In CD₃OD, the J_H,H values closely match those in acetone-d₆, whereas in DMSO-d₆, J_1,2, J_2,3, J_3,4, and J_4,5 are larger, consistent with a predominantly B_O,3 form (RMSD = 0.71 Hz, Table 2). By contrast, in D₂O at 20 °C, the ¹H NMR signals are largely broad, reflecting slow exchange among multiple conformers. Heating to 80 °C sharpens the signals, and the small J_1,2, J_2,3, and J_3,4 (observed as broad singlets) indicate a ¹C₄ form (RMSD = 1.32 Hz) (Fig. 3b and Table 2).

Fig. 3 (a) ¹H NMR spectra of the glucopyranose moiety of davidiin (5) in various solvents at 20 °C and 500 MHz. (b) Temperature-dependent ¹H NMR spectra of 5 in D₂O at 500 MHz. The numbered signals in each spectrum correspond to the proton positions (H-1–6) of the glucopyranose moiety.

When solvent effects were incorporated via the polarizable continuum model (PCM),²⁵¹C₄ remained the lowest free-energy conformer in all solvents as in the gas phase, though the free-energy gaps became smaller (Table S13, ESI‡). However, using the solvation model based on density (SMD),²⁶—often better for polar and flexible molecules including intramolecular hydrogen bonding²⁷—led to B_O,3 as the most stable form in all solvents (Table S14, ESI‡). Because PCM can overestimate intramolecular hydrogen-bond stabilization,^27,28 it may favor ¹C₄ (containing intramolecular hydrogen bonds between the HHDP and 3-galloyl groups and between 2- and 4-galloyl groups) and ^1,4B–¹S₅ (containing intramolecular hydrogen bonds between HHDP and 3-galloyl groups) (Fig. 2), whereas in reality, solute–solvent interactions may diminish these intramolecular interactions especially in DMSO-d₆.

In principle, we can estimate conformer populations from computed (free) energies by Boltzmann weighting. However, as shown above, free-energy calculations alone—even with solvent models—cannot precisely predict the conformational behavior of 5 across various solvents. In contrast, direct comparison of experimental and calculated J_H,H values can robustly identify the specific conformer. Even if the J_H,H values calculated for each possible conformation individually fail to match the experimental data, one can compute averaged values for various combinations and population ratios of multiple conformations, compare their RMSDs, and the best-fit averaged values provide an approximate conformer distribution—even for a highly flexible molecule.²⁹

Including dispersion corrections via B3LYP-D3(BJ)^30–32 strongly stabilized ¹C₄ (ΔG = 0.0 kcal mol⁻¹) over B_O,3 (ΔG = +13.7 kcal mol⁻¹) and ^1,4B–¹S₅ (ΔG = +11.7 kcal mol⁻¹) (Table S13, ESI‡). The optimized ¹C₄ geometry at this level shows an intramolecular stacking arrangement between the HHDP and 3-galloyl groups (Fig. S53, ESI‡), plausibly favored in aqueous solvents (D₂O) via π–π and hydrophobic effects. However, in DMSO-d₆, strong solute–solvent hydrogen bonding may disrupt these intramolecular interactions, shifting equilibrium toward B_O,3. Altogether, 5 can adopt B_O,3, ¹C₄, and/or ^1,4B–¹S₅ conformations, with the equilibrium distribution changing according to solvent and temperature (Table 3).

Table 3 Conformations of davidiin (5), punicafolin (6), and corilagin (7) in various solvents, as assigned from experimental (20 °C) and calculated J_H,H values

	acetone-d6	CD3OD	DMSO-d6	D2O
a 80 °C. b D2O/DMSO-d6 (9 : 1).
Davidiin (5)	B O,3/^1C4 (60 : 40) [or BO,3/^1C4/^1,4B–^1S5 (55 : 40 : 5)]	B O,3/^1C4 (60 : 40) [or BO,3/^1C4/^1,4B–^1S5 (55 : 40 : 5)]	B O,3	^1 C 4
Punicafolin (6)	^3 S 1/^1C4 (65 : 35)	^3 S 1/^1C4 (55 : 45)	^3 S 1/^1C4 (90 : 10) [or ^3S1]	^1 C 4
Corilagin (7)	^3 S 1/^1C4 (10 : 90)	^3 S 1/^1C4 (10 : 90)	^3 S 1	^1 C 4

---

Conformational analysis of punicafolin (6)

Punicafolin (6), bearing a 3,6-(R_a)-HHDP and 1,2,4-trigalloyl groups, has been isolated from Punica granatum,¹¹Mallotus japonicus,³³Euphorbia helioscopia,³⁴Macaranga tanarius,^3c and Phyllanthus emblica.^9b It shows multiple biological activities.³⁵ Early studies proposed a ¹C₄ or skew-boat conformation in acetone-d₆,¹¹ whereas a B_1,4 form was suggested in DMSO-d₆ based on J_H,H.³⁶ Reviews often depict it as ¹C₄,^7d,f and a DFT-based study reported ¹C₄ as most stable.³⁷ Nevertheless, the experimental J_H,H values in acetone-d₆ do not match those of geraniin (2), a well-established ¹C₄-type ellagitannin (Table S2, ESI‡).

Our DFT-based analysis identified two major conformer families for 6: ¹C₄ (ΔG = 0.0 kcal mol⁻¹) and ³S₁ (ΔG = +2.6 kcal mol⁻¹) (B3LYP/6-31G(d,p)) (Fig. 4 and Table S22, ESI‡). Neither conformer alone reproduces the experimental J_H,H values in acetone-d₆, but a ∼65 : 35 mixture of ³S₁ and ¹C₄ reproduces the observed data (RMSD = 0.78 Hz) (Table 4 and Tables S26–S28, ESI‡). Moreover, J_1,2 and J_2,3 increase upon cooling (Table S4, ESI‡), indicating a larger ³S₁ population at lower temperatures. Solvent effects are also pronounced: in CD₃OD, 6 exists as ³S₁/¹C₄ in ∼55 : 45 (RMSD = 0.61 Hz), whereas in D₂O/DMSO-d₆ (9 : 1),³⁸ it predominantly adopts ¹C₄ (Fig. 5a and Table 4). In DMSO-d₆, focusing on the J_1,2 values, the calculated value for the ³S₁/¹C₄ (90 : 10) mixture closely matched the experimental value (J_1,2 (Hz): exptl, 7.1 Hz; calcd for ³S₁/¹C₄ (90 : 10), 7.0 Hz; calcd for ³S₁, 7.7 Hz). However, RMSD analysis indicated that the calculation for the pure ³S₁ conformation alone showed a slightly better overall fit (RMSD for ³S₁/¹C₄ (90 : 10): 1.09 Hz; ³S₁: 1.04 Hz) (Table 4).

Fig. 4 Two representative conformations of punicafolin (6) optimized at the B3LYP/6-31G(d,p) level.

Table 4 Experimental and calculated J_H,H values (Hz) for punicafolin (6)

	Exptl^a				Calcd^b
	acetone-d6	CD3OD	DMSO-d6	D2O/DMSO-d6 (9 : 1)	^1 C 4	^3 S 1	^3 S 1/^1C4 (55 : 45)	^3 S 1/^1C4 (65 : 35)	^3 S 1/^1C4 (90 : 10)
a 500 MHz, 20 °C. b Calculated at the B3LYP/6-31G(d,p)u+1s//B3LYP/6-31G(d,p) level. c Not resolved. d RMSD (root mean square deviation) was calculated as the square root of the average of the squared differences between the calculated and experimental JH,H values.
J 1,2	5.1	4.5	7.1	≈0	0.7	7.7	4.5	5.2	7.0
J 2,3	≈0	≈0	≈0	≈0	1.8	0.4	1.0	0.9	0.6
J 3,4	3.3	3.2	3.2	3.1	2.7	3.3	3.0	3.1	3.3
J 4,5	≈0	≈0	≈0	≈0	1.0	0.3	0.6	0.5	0.4
J 5,6a	7.0	—^c	5.6	7.5	8.1	8.1	8.1	8.1	8.1
J 5,6b	8.0	—^c	7.8	—^c	11.1	8.0	9.4	9.1	8.3
RMSD (acetone-d6)^d					2.41	1.14	0.92	0.78	—
RMSD (CD3OD)^d					2.18	1.60	0.61	0.64	—
RMSD (DMSO-d6)^d					3.23	1.04	—	—	1.09
RMSD (D2O/DMSO-d6 (9 : 1))^d					1.12	3.84	—	—	—

---

Fig. 5 ¹H NMR spectra of the glucopyranose moiety of (a) punicafolin (6) and (b) corilagin (7) in various solvents (20 °C, 500 MHz). Numbered signals correspond to H-1–6 of the glucopyranose ring.

Geometry optimizations using PCM or SMD solvation models lowered the free-energy differences but did not accurately predict the experimental distributions (Tables S22 and S23, ESI‡). Incorporating dispersion corrections (B3LYP-D3(BJ)) strongly favored ¹C₄ (ΔG = 0.0 kcal mol⁻¹) over ³S₁ (ΔG = +11.7 kcal mol⁻¹) (Tables S22 and S23, ESI‡), presumably reflecting intramolecular interactions between the HHDP and 1-galloyl groups as well as between 2- and 4-galloyl groups (Fig. S63, ESI‡). In D₂O/DMSO-d₆ (90 : 10), such intramolecular interactions may be strongly stabilized by π–π and hydrophobic effects, favoring ¹C₄. By contrast, the strong solute–solvent interactions in pure DMSO-d₆ can diminish these intramolecular effects, favoring ³S₁. Overall, 6 interconverts between ³S₁ and ¹C₄, with the ratio dependent on both solvent and temperature.

Conformational analysis of corilagin (7)

Corilagin (7), a desgalloyl analogue of 6, is known for high conformational flexibility. Previous reports described an intermediate between B_1,4 and ^O,3B (or ¹C₄) in DMSO-d₆ and a slightly perturbed ^O,3B (or ¹C₄) form in acetone-d₆;^12a,b its J_H,H values in DMSO-d₆ also vary with temperature.^12a Other studies suggested a ¹C₄ form with possible flattening near C-1 and O-5,^3a or a B_1,4 form³⁶ in DMSO-d₆; and a ¹C₄ form in acetone-d₆^8a or in CD₃OD.^8c Various computational approaches have yielded divergent results: a model compound study using molecular mechanics (PIMM91) suggested ¹C₄,³⁹ whereas MM2 calculations indicated a skew-boat conformation,^40,41 and semiempirical PM3 calculations suggested B_1,4.⁴² More recently, a combined analysis of electronic/vibrational circular dichroism, J_H,H, and DFT indicated that 7 exists as ³S₁ and ¹C₄ in DMSO-d₆ and CD₃OD, respectively.^12c,43

In the present work, the experimental J_H,H values of 6 and 7 in DMSO-d₆ and D₂O are quite similar, implying a ³S₁/¹C₄ mixture for both (Fig. 5b and 6 and Table 5). In acetone-d₆ and CD₃OD, however, J_1,2 of 7 (acetone-d₆: 1.8 Hz; CD₃OD: 2.0 Hz) is smaller than that of 6 (acetone-d₆: 5.1 Hz; CD₃OD: 4.5 Hz) (Tables 4 and 5), suggesting a larger ¹C₄ population in 7. Our search located four possible conformers for 7: ¹C₄ (ΔG = 0.0 kcal mol⁻¹), ^O,3B (ΔG = +2.9 kcal mol⁻¹), ³S₁–^O,3B (ΔG = +3.1 kcal mol⁻¹), and ³S₁ (ΔG = +4.4 kcal mol⁻¹) (Fig. S66 and Table S29, ESI‡). The ^O,3B and ³S₁–^O,3B conformers gave poor fits to the observed J_5,6, indicating negligible contributions (Table 5 and Tables S34–S37, ESI‡). For ¹C₄ and ³S₁, the calculated J_H,H values resemble those of 6 except that J_1,2 in ³S₁ differs between 6 (7.7 Hz) and 7 (4.6 Hz) (Tables 4 and 5). In the lowest-energy conformer of 7, the 2-hydroxy group forms an intramolecular hydrogen bond with the 1-galloyl carbonyl oxygen, altering the H-1/H-2 dihedral angle (6: 159.4°; 7: 144.8°) (Fig. 4 and Fig. S66, Table S49, ESI‡). In polar solvents, that 2-hydroxy group may preferentially hydrogen-bond to the solvent, and the calculated J_1,2 values of the ³S₁ conformers of 7 with the absence of the intramolecular hydrogen bonding was 7.4 Hz, which is practically the same as those of 6 (7.7 Hz) (Table 5 and Table S38, Fig. S66, ESI‡). The best agreement with experiment shows that 7 exists as a ³S₁/¹C₄ (10 : 90) mixture in acetone-d₆ (RMSD = 1.63 Hz) and CD₃OD (RMSD = 0.89 Hz), predominantly ³S₁ in DMSO-d₆ (RMSD = 1.37 Hz), and predominantly ¹C₄ in D₂O (RMSD = 1.51 Hz) (Tables 3, 5 and Table S39, ESI‡).

Fig. 6 Two conformation types of corilagin (7).

Table 5 Experimental and calculated J_H,H values (Hz) for corilagin (7)

	Exptl^a				Calcd^b
	acetone-d6	CD3OD	DMSO-d6	D2O	^1 C 4	^O,3 B	^3 S 1–^O,3B	^3 S 1	^3 S 1/^1C4 (10 : 90)^d
a 500 MHz, 20 °C. b Calculated at the B3LYP/6-31G(d,p)u+1s//B3LYP/6-31G(d,p) level. c Not resolved. d The values in parentheses show the results of conformers without intramolecular hydrogen bond between the 2-hydroxy and 1-galloyl carbonyl groups. e RMSD (root mean square deviation) was calculated as the square root of the average of the squared differences between the calculated and experimental JH,H values. For this calculation, the computed JH,H values of the ^3S1 conformers that do not form an intramolecular hydrogen bond between the 2-hydroxy and 1-galloyl carbonyl groups were used.
J 1,2	1.8	2.0	7.2	≈0	1.2	1.7	4.1	4.6 (7.4)	1.6 (1.9)
J 2,3	≈0	3.5	≈0	≈0	2.5	2.6	1.5	0.7 (0.4)	2.3 (2.3)
J 3,4	≈0	1.6	≈0	3.1	3.0	3.0	3.7	4.2 (3.2)	3.1 (3.1)
J 4,5	≈0	≈0	≈0	≈0	1.1	3.0	1.1	0.2 (0.3)	1.0 (1.0)
J 5,6a	10.9	10.9	7.9	—^c	10.8	0.5	0.2	6.9 (9.3)	10.4 (10.7)
J 5,6b	8.1	8.0	8.7	—^c	8.4	4.2	6.3	8.4 (7.9)	8.4 (8.3)
RMSD (acetone-d6)^e					1.69	4.61	4.20	2.73	1.63
RMSD (CD3OD)^e					0.91	4.34	4.00	2.70	0.89
RMSD (DMSO-d6)^e					3.08	4.65	3.91	1.37	—
RMSD (D2O)^e					1.51	2.17	2.25	3.71	—

---

Overall, the ¹C₄ population in 7 is higher than in 6 (Table 3). For 6, the distance between O-2 and O-4 in the glucopyranose ring is smaller in ¹C₄ (2.9 Å) than in ³S₁ (3.2 Å), creating a larger steric clash between its 2- and 4-galloyl groups in the ¹C₄ form. Consequently, 6 has a stronger tendency to adopt the ³S₁ form, with less steric hindrance, than 7 does.

Conformational analysis of macaranganin (8)

Macaranganin (8), a diastereomer of 6 bearing a 3,6-(S_a)-HHDP group, was isolated from Macaranga tanarius.^3c A recent DFT study assigned 8 as ⁵S₁.³⁷ In our analysis, four possible conformations were found: ³S₁ (ΔG = 0.0 kcal mol⁻¹), ⁵S₁–B_1,4 (ΔG = +0.6 kcal mol⁻¹), ⁵S₁–⁵E (ΔG = +2.5 kcal mol⁻¹), and ⁵S₁ (ΔG = +4.5 kcal mol⁻¹) at B3LYP/6-31G(d,p) (Fig. S70 and Table S40, ESI‡). PCM solvation narrowed the free-energy gaps further: ³S₁ (ΔG = 0.00 kcal mol⁻¹), ⁵S₁–B_1,4 (ΔG = +0.19 kcal mol⁻¹), ⁵S₁–⁵E (ΔG = +0.90 kcal mol⁻¹), and ⁵S₁ (ΔG = +0.73 kcal mol⁻¹) (Table S40, ESI‡). The reported J_H,H values in acetone-d₆/D₂O³⁷ aligned best with ⁵S₁ (RMSD = 0.74 Hz) (Table 6 and Tables S45–S48, ESI‡), confirming that 8 primarily adopts ⁵S₁ (Fig. 7). Because we did not have access to an authentic sample of 8, its conformation in other solvents remains uncertain.

Table 6 Experimental and calculated J_H,H values (Hz) for macaranganin (8)

	Exptl^a^b	Calcd^c
	acetone-d6 + D2O	^3 S 1	^5 S 1–B1,4	^5 E	^5 S 1
a 400 MHz. b Ref. 37. c Calculated at the B3LYP/6-31G(d,p)u+1s//B3LYP/6-31G(d,p) level. d RMSD (root mean square deviation) was calculated as the square root of the average of the squared differences between the calculated and experimental JH,H values.
J 1,2	9.1	7.1	5.6	7.6	9.3
J 2,3	6.1	0.3	1.5	3.8	6.9
J 3,4	≈0	2.7	2.3	1.3	0.8
J 4,5	≈0	0.7	2.1	1.7	1.1
J 5,6a	2.9	4.2	4.9	4.4	3.4
J 5,6b	≈0	0.4	0.3	0.4	0.8
RMSD^d		2.78	2.81	1.57	0.74

---

Fig. 7 The ⁵S₁ conformation of macaranganin (8) optimized at the B3LYP/6-31G(d,p) level.

Biosynthetic considerations of ¹C₄-type ellagitannins

The conformational flexibility revealed here clarifies how ¹C₄-type ellagitannins can be biosynthesized from 1 (⁴C₁) (Scheme 1a). For example, carpinusin (9)⁴⁴ and helioscopinin A (10),^19b,34 both bearing a 1,6-(S_a)-HHDP group—and each locked in the ¹C₄ conformation by an additional 2,4-bridged DHHDP group—may arise from a conformational isomerization of the glucopyranose ring from ⁴C₁ to ¹C₄. The conformationally flexible 5, generated by oxidative coupling of the 1,6-galloyl groups in 1, can adopt a B_O,3 form that is relatively close to ⁴C₁. However, in aqueous environments, 5 shifts to ¹C₄, setting the stage for further oxidative coupling between the 2-and 4-galloyl groups to yield 9 and 10. Similarly, 6 can form from 1via oxidative coupling of its 3- and 6-galloyl groups, since the ³S₁ conformer of 6 is relatively close to ⁴C₁. In water, 6 also exists in ¹C₄, leading to geraniin (2) and granatin B (11),^3a,11,45 which are locked as ¹C₄ by the 2,4-bridged DHHDP group.

Scheme 1 Possible biosynthetic pathways for ¹C₄-locked ellagitannins 2 and 9–11 from 1,2,3,4,6-penta-O-galloyl-β-D-glucose (1). (a) Via the conformationally flexible intermediates 5 and 6. (b) Via terchebin (12). HHDP: hexahydroxydiphenoyl; DHHDP: dehydrohexahydroxydiphenoyl.

Another possible biosynthetic route to ¹C₄-type ellagitannins involves reductive metabolism of amariin (3) and/or isoamariin (15), both bearing 2,4-(R)- and 3,6-(R)-DHHDP groups in an ^O,3B conformation,^4c,46 to give geraniin (2) (Scheme 1b). The precursors 3 and 15 might themselves arise from 1via terchebin (12), which has a 2,4-(R)-DHHDP group.^45a Although the conformation of 12 remains uninvestigated, it was suggested to adopt an intermediate between B_1,4 and ^O,3B (or ¹C₄) in DMSO-d₆.^12a We suspect that 12 is also flexible. Likewise, 9 might form via oxidation of 12 into putative intermediates 13 and/or 14, followed by reduction to 9.

Conclusions

By comparing experimentally measured and DFT-calculated J_H,H values, we elucidated the conformational preferences of davidiin (5), bearing a 1,6-(S_a)-HHDP group, and punicafolin (6), bearing a 3,6-(R_a)-HHDP group. Both display conformationally flexibility in solution, with populations influenced by solvent and temperature. A similar analysis was conducted for corilagin (7), a well-known example of solvent-dependent conformational switching. Notably, it is difficult to predict the exact conformations of such flexible ellagitannins solely from computed relative free energies, even when solvent effects are included under various models. Although several previous computational studies on 6 and 7 focused on identifying single global minima, our results emphasize that mixtures of conformers must be considered. In contrast, our approach—direct comparison of experimental and calculated J_H,H values—enables more reliable identification of the conformational ensembles. Recently, Auer and co-workers demonstrated the value of J_H,H calculations for detecting mixtures of conformers in xylopyranoside derivatives,⁴⁷ and our findings align with their conclusion that J_H,H-based methods can quantify multiple conformers more sensitively than relative free energies or NMR chemical shifts alone.⁴⁷

From a biosynthetic perspective, conformationally flexible ellagitannins such as 5 and 6 likely play key roles in the formation of ¹C₄-locked ellagitannins. Their flexibility may also underlie the diverse biological activities of ellagitannins. Although several in silico docking studies on 5–7 have been reported,⁴⁸ incorporating the conformational equilibria described here could improve future bioinformatics research on these compounds. Moreover, ellagitannins are appealing synthetic targets because of their structural variety and complexity.⁴⁹ Indeed, total syntheses of 4–8 have been reported,^9a,22a,37,50 and understanding the solution conformations presented here could facilitate more efficient synthetic design of these and related molecules, including biomimetic approaches.

Author contributions

Y. M. and T. T. designed the research; Y. M., M. I., C. O. and T. T. conducted the experiments; Y. M., M. I. and C. O. performed the computational studies; all authors discussed the results; Y. M. wrote the manuscript with feedback from all authors.

Data availability

The experimental and calculated data supporting the results are provided in the ESI.‡

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was partly supported by JSPS KAKENHI Grant Number JP20K07102. This work was the result of using research equipment shared in the MEXT Project for promoting public utilization of advanced research infrastructure (Program for supporting introduction of the new sharing system) Grant Number JPMXS0422500320. The computation was partly carried out using the computer resource offered under the category of General Projects by the Research Institute for Information Technology, Kyushu University. The authors would like to thank Enago (www.enago.jp) for the English language review.

Notes and references

1. (a) T. Okuda, T. Yoshida and T. Hatano, Hydrolyzable Tannins and Related Polyphenols, in Fortschritte der Chemie Organischer Naturstoffe/Progress in the Chemistry of Organic Natural Products, ed. W. Herz, G. W. Kirby, R. E. Moore, W. Steglich and C. Tamm, Springer, Vienna, vol. 66, 1995, pp. 1–117 DOI:10.1007/978-3-7091-9363-1_1; (b) Chemistry and Biology of Ellagitannins: An Underestimated Class of Bioactive Plant Polyphenols, ed. S. Quideau, World Scientific Publishing, 2009 DOI:10.1142/6795; (c) T. Okuda and H. Ito, Tannins of constant structure in medicinal and food plants—Hydrolyzable tannins and polyphenols related to tannins, Molecules, 2011, 16, 2191–2217; (d) H. Yamada, S. Wakamori, T. Hirokane, K. Ikeuchi and S. Matsumoto, Structural revisions in natural ellagitannins, Molecules, 2018, 23, 1901.
2. (a) S. Quideau and K. S. Feldman, Ellagitannin chemistry. The first synthesis of dehydrohexahydroxydiphenoate esters from oxidative coupling of unetherified methyl gallate, J. Org. Chem., 1997, 62, 8809–8813; (b) K. S. Feldman, Recent progress in ellagitannin chemistry, Phytochemistry, 2005, 66, 1984–2000.
3. (a) E. A. Haddock, R. K. Gupta and E. Haslam, The metabolism of gallic acid and hexahydroxydiphenic acid in plants. Part 3. Esters of (R)- and (S)-hexahydroxydiphenic acid and dehydrohexahydroxydiphenic acid with D-glucopyranose (¹C₄ and related conformations), J. Chem. Soc., Perkin Trans. 1, 1982, 2535–2545; (b) G. Nonaka, Y. Matsumoto, I. Nishioka, M. Nishizawa and T. Yamagishi, Trapain, a new hydrolyzable tannin from Trapa Japonica FLEROV, Chem. Pharm. Bull., 1981, 29, 1184–1187; (c) J.-H. Lin, G. Nonaka and I. Nishioka, Tannins and related compounds. XCIV. Isolation and characterization of seven new hydrolyzable tannins from the leaves of Macaranga tanarius (L.) MUELL. et ARG, Chem. Pharm. Bull., 1990, 38, 1218–1223.
4. (a) T. Tanaka, Reactions of ellagitannins related to their metabolism in higher plants, in Recent Advances in Polyphenol Research, ed. J.-P. Salminen, K. Wähälä, V. Freitas and S. Quideau, Wiley-Blackwell, vol. 8, 2023, pp 347–368 DOI:10.1002/9781119844792.ch12; (b) D. Kojima, K. Shimizu, K. Aritake, M. Era, Y. Matsuo, Y. Saito, T. Tanaka and G. Nonaka, Highly oxidized ellagitannins of Carpinus japonica and their oxidation–reduction disproportionation, J. Nat. Prod., 2020, 83, 3424–3434; (c) M. Era, Y. Matsuo, Y. Saito and T. Tanaka, Production of ellagitannin hexahydroxydiphenoyl ester by spontaneous reduction of dehydrohexa-hydroxydiphenoyl ester, Molecules, 2020, 25, 1051; (d) H. Wakamatsu, S. Tanaka, Y. Matsuo, Y. Saito, K. Nishida and T. Tanaka, Reductive metabolism of ellagitannins in the young leaves of Castanopsis sieboldii, Molecules, 2019, 24, 4279.
5. T. Yamashita, Y. Matsuo, Y. Saito and T. Tanaka, Formation of dehydrohexahydroxydiphenoyl esters by oxidative coupling of galloyl esters in an aqueous medium involved in ellagitannin biosynthesis, Chem. Asian J., 2021, 16, 1735–1740.
6. (a) H. B. Mayes, L. J. Broadbelt and G. T. Beckham, How sugars pucker: Electronic structure calculations map the kinetic landscape of five biologically paramount monosaccharides and their implications for enzymatic catalysis, J. Am. Chem. Soc., 2014, 136, 1008–1022; (b) X. Biarnés, A. Ardèvol, A. Planas, C. Rovira, A. Laio and M. Parrinello, The conformational free energy landscape of β-D-glucopyranose. Implications for substrate preactivation in β-glucoside hydrolases, J. Am. Chem. Soc., 2007, 129, 10686–10693; (c) M. Appell, G. Strati, J. L. Willett and F. A. Momany, B3LYP/6-311++G** study of α- and β-D-glucopyranose and 1,5-anhydro-D-glucitol: ⁴C₁ and ¹C₄ chairs, ^3,OB and B_3,O boats, and skew-boat conformations, Carbohydr. Res., 2004, 339, 537–551; (d) IUPAC-IUB Joint Commission on Biochemical Nomenclature (JCBN), Conformational nomenclature for five- and six-membered ring forms of monosaccharides and their derivatives, Pure Appl. Chem., 1981, 53, 1901–1905.
7. (a) R. K. Gupta, S. M. K. Al-Shafi, K. Layden and E. Haslam, The metabolism of gallic acid and hexahydroxydiphenic acid in plants. Part 2. Esters of (S)-hexahydroxydiphenic acid with D-glucopyranose (⁴C₁), J. Chem. Soc., Perkin Trans. 1, 1982, 2525–2534; (b) E. Haslam, Secondary metabolism – fact and fiction, Nat. Prod. Rep., 1986, 3, 217–249; (c) C. M. Spencer, Y. Cai, R. Martin, S. H. Gaffney, P. N. Goulding, D. Magnolato, T. H. Lilley and E. Haslam, Polyphenol complexation—some thoughts and observations, Phytochemistry, 1988, 27, 2397–2409; (d) E. Haslam and Y. Cai, Plant polyphenols (vegetable tannins): gallic acid metabolism, Nat. Prod. Rep., 1994, 11, 41–66; (e) E. Haslam, Taste, bitterness and astringency, Practical Polyphenolics: From structure to molecular recognition and physiological action, Cambridge University Press, 1998, pp. 178–225; (f) R. F. Helm, L. Zhentian, T. Ranatunga, J. Jervis and T. Elder, Toward understanding monomeric ellagitannin biosynthesis, in Plant Polyphenols 2: Chemistry, Biology, Pharmacology, Ecology, ed. G. G. Gross, R. W. Hemingway, T. Yoshida and S. J. Branham, Springer, 1999, vol. 66, pp 83–99 DOI:10.1007/978-1-4615-4139-4_5; (g) P. Grundhöfer, R. Niemetz, G. Schilling and G. G. Gross, Biosynthesis and subcellular distribution of hydrolyzable tannins, Phytochemistry, 2001, 57, 915–927.
8. (a) T. Hatano, T. Yoshida, T. Shingu and T. Okuda, ¹³C Nuclear magnetic resonance spectra of hydrolyzable tannins. III. Tannins having ¹C₄ glucose and C-glucosidic linkage, Chem. Pharm. Bull., 1988, 36, 3849–3856; (b) P. Luger, M. Weber, S. Kashino, Y. Amakura, T. Yoshida, T. Okuda, G. Beurskens and Z. Dauter, Structure of the tannin geraniin based on conventional X-ray data at 295 K and on synchrotron data at 293 and 120 K, Acta Crystallogr., Sect. B, 1998, B54, 687–694; (c) Y. Sudjaroen, W. E. Hull, G. Erben, G. Würtele, S. Changbumrung, C. M. Ulrich and R. W. Owen, Isolation and characterization of ellagitannins as the major polyphenolic components of Longan (Dimocarpus longan Lour) seeds, Phytochemistry, 2012, 77, 226–237.
9. (a) S. Matsumoto, S. Wakamori, K. Nishii, T. Tanaka and H. Yamada, Total synthesis of phyllanemblinin B, Synlett, 2020, 31, 1389–1393; (b) Y.-J. Zhang, T. Abe, T. Tanaka, C.-R. Yang and I. Kouno, Phyllanemblinins A−F, new ellagitannins from Phyllanthus emblica, J. Nat. Prod., 2001, 64, 1527–1532.
10. C. M. Spencer, Y. Cai, R. Martin, T. H. Lilley and E. Haslam, The metabolism of gallic acid and hexahydroxydiophenic acid in higher plants part 4; Polyphenol interactions part 3. Spectroscopic and physical properties of esters of gallic acid and (S)-hexahydroxydiphenic acid with D-glucopyranose (⁴C₁), J. Chem. Soc., Perkin Trans. 2, 1990, 651–660.
11. T. Tanaka, G. Nonaka and I. Nishioka, Punicafolin, an ellagitannin from the leaves of Punica granatum, Phytochemistry, 1985, 24, 2075–2078.
12. (a) J. C. Jochims, G. Taigel and O. T. Schmidt, Über natürliche gerbstoffe, XLI. Protonenresonanz-spektren und konformationsbestimmung einiger natürlicher gerbstoffe, Justus Liebigs Ann. Chem., 1968, 717, 169–185; (b) M. K. Seikel and W. E. Hillis, Hydrolysable tannins of Eucalyptus delegatensis wood, Phytochemistry, 1970, 9, 1115–1128; (c) R. F. Sprenger, S. S. Thomasi, A. G. Ferreira, Q. B. Cass and J. M. Batista Junior, Solution-state conformations of natural products from chiroptical spectroscopy: the case of isocorilagin, Org. Biomol. Chem., 2016, 14, 3369–3375.
13. (a) D. Ikuta, Y. Hirata, S. Wakamori, H. Shimada, Y. Tomabechi, Y. Kawasaki, K. Ikeuchi, T. Hagimori, S. Matsumoto and H. Yamada, Conformationally supple glucose monomers enable synthesis of the smallest cyclodextrins, Science, 2019, 364, 674–677; (b) M. Heuckendorff, C. M. Pedersen and M. Bols, Conformationally armed 3,6-tethered glycosyl donors: Synthesis, conformation, reactivity, and selectivity, J. Org. Chem., 2013, 78, 7234–7248.
14. T. Bally and P. R. Rablen, Quantum-chemical simulation of ¹H NMR Spectra. 2. Comparison of DFT-based procedures for computing proton–proton coupling constants in organic molecules, J. Org. Chem., 2011, 76, 4818–4830.
15. In ref. 9a, although 4 bears a 2,4-(R_a)-HHDP group, the Supporting Information figure incorrectly shows a 2,4-(S_a)-HHDP configuration.
16. Because the hydroxy group at C-6 of the glucopyranose moiety in 4 is not involved in an HHDP group, this site is more conformationally flexible than in 5–8. Consequently, the conformation around C-6 cannot be determined with sufficient accuracy, and the calculated J_5,6 values were therefore omitted from Table 1.
17. Y. Shimozu, Y. Kimura, A. Esumi, H. Aoyama, T. Kuroda, H. Sakagami and T. Hatano, Ellagitannins of Davidia involucrata. I. Structure of davicratinic acid A and effects of Davidia tannins on drug-resistant bacteria and human oral squamous cell carcinomas, Molecules, 2017, 22, 470.
18. T. Hatano, S. Hattori, Y. Ikeda, T. Shingu and T. Okuda, Gallotannins having a 1,5-anhydro-D-glucitol core and some ellagitannins from Acer species, Chem. Pharm. Bull., 1990, 38, 1902–1905.
19. (a) J. Fu, J.-Y. Ma, X.-F. Zhang, Y. Wang, R. Feng, Y.-C. Chen, X.-S. Tan, Y.-Y. Zhang, Y.-P. Sun, Y. Zhou, C. Ma, C.-Y. He, Z.-X. Zhao and X.-W. Du, Identification of metabolites of FR429, a potential antitumor ellagitannin, transformed by rat intestinal bacteria in vitro, based on liquid chromatography–ion trap-time of flight mass spectrometry analysis, J. Pharm. Biomed. Anal., 2012, 71, 162–167; (b) Y.-Q. Li, M. Kitaoka, J. Takayoshi, Y.-F. Wang, Y. Matsuo, Y. Saito, Y.-L. Huang, D.-P. Li, G. Nonaka, Z.-H. Jiang and T. Tanaka, Ellagitannins and oligomeric proanthocyanidins of three Polygonaceous plants, Molecules, 2021, 26, 337.
20. (a) M. Zhu, J. D. Phillipson, P. M. Greengrass, N. E. Bowery and Y. Cai, Plant polyphenols: Biologically active compounds or non-selective binders to protein?, Phytochemistry, 1997, 44, 441–447; (b) Y. Wang, J. Ma, S. C. Chow, C. H. Li, Z. Xiao, R. Feng, J. Fu and Y. Chen, A potential antitumor ellagitannin, davidiin, inhibited hepatocellular tumor growth by targeting EZH2, Tumor Biol., 2014, 35, 205–212; (c) M. Takemoto, Y. Kawamura, M. Hirohama, Y. Yamaguchi, H. Handa, H. Saitoh, Y. Nakao, M. Kawada, K. Khalid, H. Koshino, K. Kimura, A. Ito and M. Yoshida, Inhibition of protein SUMOylation by davidiin, an ellagitannin from Davidia involucrata, J. Antibiot., 2014, 67, 335–338.
21. (a) S. Quideau and K. S. Feldman, Ellagitannin chemistry, Chem. Rev., 1996, 96, 475–503; (b) X. Su, G. L. Thomas, W. R. J. D. Galloway, D. S. Surry, R. J. Spandl and D. R. Spring, Synthesis of biaryl-containing medium-ring systems by organocuprate oxidation: Applications in the total synthesis of ellagitannin natural products, Synthesis, 2009, 3880–3896.
22. (a) Y. Kasai, N. Michihata, H. Nishimura, T. Hirokane and H. Yamada, Total synthesis of (+)-davidiin, Angew. Chem., Int. Ed., 2012, 51, 8026–8029; (b) S. Ashibe, K. Ikeuchi, Y. Kume, S. Wakamori, Y. Ueno, T. Iwashita and H. Yamada, Non-enzymatic oxidation of a pentagalloylglucose analogue into members of the ellagitannin family, Angew. Chem., Int. Ed., 2017, 56, 15402–15406.
23. Ref. 17 describes the conformation of 5 as a skew-boat-type.
24. Strictly speaking, the B_O,3 conformer of 5 is an intermediate between B_O,3 and ²S_O. However, we designate it as B_O,3 here because its geometry is closer to B_O,3.
25. J. Tomasi, B. Mennucci and R. Cammi, Quantum mechanical continuum solvation models, Chem. Rev., 2005, 105, 2999–3093.
26. A. V. Marenich, C. J. Cramer and D. G. Truhlar, Universal solvation model based on solute electron density and on a continuum model of the solvent defined by the bulk dielectric constant and atomic surface tensions, J. Phys. Chem. B, 2009, 113, 6378–6396.
27. (a) M. M. Zanardi, M. O. Marcarino and A. M. Sarotti, Redefining the impact of Boltzmann analysis in the stereochemical assignment of polar and flexible molecules by NMR calculations, Org. Lett., 2020, 22, 52–56; (b) E. A. Del Vigo, C. Marino and C. A. Stortz, Exhaustive rotamer search of the ⁴C₁ conformation of α- and β-D-galactopyranose, Carbohydr. Res., 2017, 448, 136–147.
28. M. M. Zanardi, M. A. Sortino and A. M. Sarotti, On the effect of intramolecular H-bonding in the configurational assessment of polyhydroxylated compounds with computational methods. The hyacinthacines case, Carbohydr. Res., 2019, 474, 72–79.
29. In each main conformer class, the computed relative free energies may be inaccurate owing to over- or underestimation of noncovalent interactions. Nonetheless, the dihedral angles among vicinal protons in the glucopyranose ring remain relatively constant within each class. Thus, Boltzmann weighting of J_H,H values within each class can still yield reasonable estimates.
30. (a) S. Grimme, S. Ehrlich and L. Goerigk, Effect of the damping function in dispersion corrected density functional theory, J. Comput. Chem., 2011, 32, 1456–1465; (b) S. Grimme, J. Antony, S. Ehrlich and H. Krieg, A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu, J. Chem. Phys., 2010, 132, 154104.
31. A. Hayasaka, K. Hashimoto, K. Konno, K. Tanaka and M. Hashimoto, Isolation, identification, and DFT-based conformational analysis of sesquikarahanadienone and its congeners from freshwater Dothideomycetes Neohelicascus aquaticus KT4120, Bull. Chem. Soc. Jpn., 2022, 95, 833–845.
32. Hashimoto and co-workers reported that dispersion-corrected functionals (e.g., ωB97X-D, B3LYP-D3) can overestimate certain van der Waals interactions in flexible molecules compared to classical functionals (e.g., B3LYP).³¹.
33. R. Saijo, G. Nonaka and I. Nishioka, Tannins and related compounds. LXXXIV. Isolation and characterization of five new hydrolyzable tannins from the bark of Mallotus japonicus, Chem. Pharm. Bull., 1989, 37, 2063–2070.
34. S.-H. Lee, T. Tanaka, G. Nonaka and I. Nishioka, Tannins and related compounds. XCV. Isolation and characterization of helioscopinins and helioscopins, four new hydrolyzable tannins from Euphorbia helioscopia L., Chem. Pharm. Bull., 1990, 38, 1518–1523.
35. (a) Y. Kashiwada, G. Nonaka, I. Nishioka, K. J.-H. Lee, I. Bori, Y. Fukushima, K. F. Bastow and K.-H. Lee, Tannins as potent inhibitors of DNA topoisomerase II in vitro, J. Pharm. Sci., 1993, 82, 487–492; (b) S. Tanimura, R. Kadomoto, T. Tanaka, Y.-J. Zhang, I. Kouno and M. Kohno, Suppression of tumor cell invasiveness by hydrolyzable tannins (plant polyphenols) via the inhibition of matrix metalloproteinase-2/-9 activity, Biochem. Biophys. Res. Commun., 2005, 330, 1306–1313; (c) M. Xu, H.-T. Zhu, R.-R. Cheng, D. Wang, C.-R. Yang, T. Tanaka, I. Kouno and Y.-J. Zhang, Antioxidant and hyaluronidase inhibitory activities of diverse phenolics in Phyllanthus emblica, Nat. Prod. Res., 2016, 30, 2726–2729; (d) J. Lee, S.-H. Lee, K. R. Min, K.-S. Lee, J.-S. Ro, J.-C. Ryu and Y. Kim, Inhibitory effects of hydrolyzable tannins on Ca²⁺-activated hyaluronidase, Planta Med., 1993, 59, 381–382; (e) Y. Kashiwada, L. Huang, L. M. Ballas, J. B. Jiang, W. P. Janzen and K.-H. Lee, New hexahydroxybiphenyl derivatives as inhibitors of protein kinase C, J. Med. Chem., 1994, 37, 195–200.
36. M. A. M. Nawwar, S. A. M. Hussein and I. Merfort, NMR spectral analysis of polyphenols from Punica granatum, Phytochemistry, 1994, 36, 793–798.
37. H. Shibayama, Y. Ueda, T. Tanaka and T. Kawabata, Seven-step stereodivergent total syntheses of punicafolin and macaranganin, J. Am. Chem. Soc., 2021, 143, 1428–1434.
38. Because 6 is poorly soluble in D₂O, we used D₂O/DMSO-d₆ (9:1) instead.
39. S. Immel and K. Khanbabaee, Atropdiastereoisomers of ellagitannin model compounds: configuration, conformation, and relative stability of D-glucose diphenoyl derivatives, Tetrahedron: Asymmetry, 2000, 11, 2495–2507.
40. X. Li, J. Liu, B. Chen, Y. Chen, W. Dai, Y. Li and M. Zhu, Covalent bridging of corilagin improves antiferroptosis activity: Comparison with 1,3,6-tri-O-galloyl-β-D-glucopyranose, ACS Med. Chem. Lett., 2020, 11, 2232–2237.
41. In ref. 40, 7 is depicted with a 3,6-(S_a)-HHDP group, whereas its actual structure bears a 3,6-(R_a)-HHDP group.
42. R. Gaudreault, T. G. M. van de Ven and M. A. Whitehead, Molecular modeling of poly(ethylene oxide) model cofactors; 1,3,6-tri-O-galloyl-β-D-glucose and corilagin, J. Mol. Model., 2002, 8, 73–80.
43. In ref. 12c, 7 was reported to exist as an equilibrium mixture of B_1,4 and ³S₁ in DMSO-d₆, although the ³S₁ form was erroneously labeled as “^OS₅”. Notably, both conformers illustrated in that reference closely resemble the ³S₁ conformation.
44. G. Nonaka, M. Akazawa and I. Nishioka, Two new ellagitannin metabolites, carpinusin and carpinusnin from Carpinus laxiflora, Heterocycles, 1992, 33, 597–606.
45. (a) T. Okuda, T. Hatano, H. Nitta and R. Fujii, Hydrolysable tannins having enantiomeric dehydrohexahydroxydiphenoyl group: Revised structure of terchebin and structure of granatin B, Tetrahedron Lett., 1980, 21, 4361–4364; (b) T. Tanaka, G. Nonaka and I. Nishioka, Tannins and related compounds. C. Reaction of dehydrohexahydroxydiphenic acid esters with bases, and its application to the structure determination of pomegranate tannins, granatins A and B, Chem. Pharm. Bull., 1990, 38, 2424–2428; (c) W. E. Steinmetz, NMR assignment and characterization of proton exchange of the ellagitannin granatin B, Magn. Reson. Chem., 2010, 48, 565–570.
46. L. Y. Foo, Amariin, a di-dehydrohexahydroxydiphenoyl hydrolysable tannin from Phyllanthus amarus, Phytochemistry, 1993, 33, 487–491.
47. S. O. Jaeschke, T. K. Lindhorst and A. Auer, Between two chairs: Combination of theory and experiment for the determination of the conformational dynamics of xylosides, Chem. Eur. J., 2022, 28, e202201544.
48. (a) J.-Y. Ma, X. Zhou, J. Fu, T. Hu, P. M. Y. Or, R. Feng, C.-Y. He, W.-J. Chen, X. Zhang, Y. Chen, Y. Wang and J. H. K. Yeung, Metabolite profiling analysis of FR429, an ellagitannin purified from Polygonum capitatum, in rat and human liver microsomes, cytosol and rat primary hepatocytes in vitro, Chem. Biol. Interact., 2014, 220, 33–40; (b) D. P. Yeggoni, A. Rachamallu and R. Subramanyam, A comparative binding mechanism between human serum albumin and α-1-acid glycoprotein with corilagin: biophysical and computational approach, RSC Adv., 2016, 6, 40225–40237; (c) M. Sobeh, M. F. Mahmoud, R. A. Hasan, M. A. O. Abdelfattah, S. Osman, H. Rashid, A. M. El-Shazly and M. Wink, Chemical composition, antioxidant and hepatoprotective activities of methanol extracts from leaves of Terminalia bellirica and Terminalia sericea (Combretaceae), PeerJ, 2019, 7, e6322; (d) Y. Yang, X. Wang, Y. Gao and X. Niu, Insight into the dual inhibition mechanism of corilagin against MRSA serine/threonine phosphatase (Stp1) by molecular modeling, ACS Omega, 2020, 5, 32959–32968; (e) V. Binette, S. Côté, M. Haddad, P. T. Nguyen, S. Bélanger, S. Bourgault, C. Ramassamy, R. Gaudreault and N. Mousseau, Corilagin and 1,3,6-tri-O-galloy-β-D-glucose: potential inhibitors of SARS-CoV-2 variants, Phys. Chem. Chem. Phys., 2021, 23, 14873–14888; (f) L. J. Yang, R. H. Chen, S. Hamdoun, P. Coghi, J. P. L. Ng, D. W. Zhang, X. Guo, C. Xia, B. Y. K. Law and V. K. W. Wong, Corilagin prevents SARS-CoV-2 infection by targeting RBD-ACE2 binding, Phytomedicine, 2021, 87, 153591; (g) Q. Li, D. Yi, X. Lei, J. Zhao, Y. Zhang, X. Cui, X. Xiao, T. Jiao, X. Dong, X. Zhao, H. Zeng, C. Liang, L. Ren, F. Guo, X. Li, J. Wang and S. Cen, Corilagin inhibits SARS-CoV-2 replication by targeting viral RNA-dependent RNA polymerase, Acta Pharm. Sin. B, 2021, 11, 1555–1567; (h) Q. Sheng, X. Hou, N. Wang, M. Liu, H. Zhu, X. Deng, X. Liang and G. Chi, Corilagin: A novel antivirulence strategy to alleviate Streptococcus pneumoniae infection by diminishing pneumolysin oligomers, Molecules, 2022, 27, 5063; (i) S. Pradeep, S. M. Patil, C. Dharmashekara, A. Jain, R. Ramu, P. S. Shirahatti, S. P. Mandal, P. Reddy, C. Srinivasa, S. S. Patil, J. Ortega-Castro, J. Frau, N. Flores-Holgúın, C. Shivamallu, S. P. Kollur and D. Glossman-Mitnik, Molecular insights into the in silico discovery of corilagin from Terminalia chebula as a potential dual inhibitor of SARS-CoV-2 structural proteins, J. Biomol. Struct. Dyn., 2023, 41, 10869–10884.
49. (a) K. S. Feldman, K. Sahasrabudhe, S. Quideau, K. L. Hunter and M. D. Lawlor, Prospects and progress in ellagitannin synthesis, in Plant Polyphenols 2: Chemistry, Biology, Pharmacology, Ecology, ed. G. G. Gross, R. W. Hemingway, T. Yoshida and S. J. Branham, Springer, Basic Life Sciences, vol. 66, 1999, pp 101–125 DOI:10.1007/978-1-4615-4139-4_6; (b) K. Khanbabaee and T. van Ree, Strategies for the synthesis of ellagitannins, Synthesis, 2001, 1585–1610; (c) T. Tanaka, I. Kouno and G. Nonaka, Biomimetic synthesis and related reactions of ellagitannins, in Biomimetic Organic Synthesis, ed. E. Poupon and B. Nay, Wiley-VCH, 2011, vol. 2, pp. 637–675 DOI:10.1002/9783527634606.ch17; (d) L. Pouységu, D. Deffieux, G. Malik, A. Natangelo and S. Quideau, Synthesis of ellagitannin natural products, Nat. Prod. Rep., 2011, 28, 853–874; (e) H. Yamada, T. Hirokane, N. Asakura, Y. Kasai and K. Nagao, Strategies and methods for the total synthesis of ellagitannins, Curr. Org. Chem., 2012, 16, 578–604; (f) H. Yamada, T. Hirokane, K. Ikeuchi and S. Wakamori, Fundamental methods in ellagitannin synthesis, Nat. Prod. Commun., 2017, 12, 1351–1358.
50. (a) H. Yamada, K. Nagao, K. Dokei, Y. Kasai and N. Michihata, Total synthesis of (−)-corilagin, J. Am. Chem. Soc., 2008, 130, 7566–7567; (b) K. Yamashita, Y. Kume, S. Ashibe, C. A. D. Puspita, K. Tanigawa, N. Michihata, S. Wakamori, K. Ikeuchi and H. Yamada, Total synthesis of mallotusinin, Chem. Eur. J., 2020, 26, 16408–16421.

---

Footnotes

† A previous version of this manuscript was deposited on a preprint server (DOI: https://doi.org/10.26434/chemrxiv-2022-wrlxz-v4).

‡ Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5nj01278c

§ These authors contributed equally.

---