1. Introduction

The steamed and dried rhizome of Gastrodia elata Blume (Orchidaceae), known as "tian ma" in Chinese, is one of the most important traditional Chinese medicines used for the treatment of headaches, migraine, dizziness, tetanus, epilepsy, neuralgia, paralysis, and other neuralgic and nervous disorders. It is also considered to have health benefits enhancing strength and virility and improving memory and blood circulation ^[1]. Pharmacological and chemical studies showed that 4-hydroxybenzyl analogues were the main active constituents ^[2-7]. However, the previous investigations were mainly carried out on ethanol or methanol extracts of the herbal medicine, which is inconsistent with a practical application by decocting the drug and formulations. Therefore, as part of a program to systematically study the chemical diversity of traditional Chinese medicines and their biological effects ^[8-24], an aqueous decoction of the G. elata rhizomes was investigated. We previously reported isolation and structure characterization of 49 compounds, including 14 new 4-hydroxybenzyl-substituted amino acids and glutathione derivatives, from the aqueous extract, as well as preliminary bioassays of those isolates ^[25-27] and pharmacological functions of a minor component N⁶ -(4-hydroxybenzyl)adenosine (NHBA) ^{[28, 29]}. Herein, we report the isolation and structure elucidation of an unusual natural product derived from ergothioneine, named gastrolatathioneine (1) (Fig. 1), as well as its plausible biosynthetic pathway of the plant collaborated with a symbiotic fungus.

2. Experimental 2.1. General experimental procedures

Optical rotations were measured using a Rudolph Research Autopol III polarimeter (Rudolph Research Analytical, East Sussex, UK). The UV spectrum was measured on a Cary 300 spectrometer (Agilent Technologies, California, USA). The CD spectrum was measured on a JASCO J-81′ spectropolarimeter (JASCO, Tokyo, Japan). The IR spectrum was recorded on a Nicolet 5700 FT-IR microscope transmission spectrometer (Thermo Electron Corporation, Madison, WI, USA). 1D- and 2D-NMR spectra were obtained at 500 MHz for ¹H and 125 MHz for ¹³C, respectively, on INOVA 500 MHz spectrometer (Varian Associates Inc., Palo Alto, CA, USA), with solvent peaks serving as references. ESIMS and HR-ESIMS data were measured with a QSTAR Elite LC/MS/MS spectrometer (Applied Biosystem Mds Sciex, CA, USA) and an Agilent 6520 Accurate-Mass Q-TOFL CMS spectrometers (Agilent Technologies, Ltd., Santa Clara, CA, USA), respectively. Column chromatography was performed with silica gel (200-300 mesh, Qingdao Marine Chemical Inc., Qingdao, China) and Sephadex LH-2′ (Pharmacia Biotech AB, Uppsala, Sweden). Glass precoated silica gel GF254 plates were used for TLC. Spots were visualized under UV light or by spraying with 5% H₂SO₄ in EtOH, followed by heating.

2.2. Plant material

The rhizomes of Gastrodia elata were collected at the plantation field Xiao Cao Ba, Zhaotong County of Yunnan province, China, in December 2009. Plant identification was verified by Mr. Lin Ma (Institute of Materia Medica, Beijing 100050, China). A voucher specimen (No. ID-S-2384) was deposited at the herbarium of Institute of Materia Medica.

2.3. Extraction and isolation

The steamed and air-dried G. elata rhizomes (50 kg) were pulverized and extracted with H₂O (150 L, 3 × 1 h) under ambient temperature and ultrasonication. The aqueous extracts were combined and evaporated under reduced pressure to yield a concentrated solution (50 L), which was loaded on a macroporous adsorbent resin (HPD-1″, 30 kg) column (20 × 200 cm), and eluted successively with H₂O (50 L), 30% EtOH (150 L), 50% EtOH (120 L), and 95% EtOH (80 L) to yield four corresponding fractions A-D. After removing the solvent under reduced pressure, fraction C (1.9 kg) was chromatographed over MCI gel (CHP 20P, 10 L), with successive elution using H₂O (30 L), 30% EtOH (70 L), 50% EtOH (70 L), 95% EtOH (30 L), and Me₂CO (20 L), to afford fractions C1-C5. Fraction C3 (237 g) was subjected to CC over silica gel, eluting with a gradient of increasing MeOH concentration (0-100%) in EtOAc followed by 30% EtOH, to yield fractions C3- 1-C3-5 based on TLC analysis. Fraction C3-5 (20 g) was chromatographed over silica gel (CH3Cl-MeOH 20:1 - 1:1) to give C3-5-1-C3-5-4, of which C3-5-1 (1.2 g) was further separated by silica gel CC (ethyl acetate-MeOH 5:1) to afford 1 (120 mg, R_f = 0.6).

Gastrolatathioneine (1): Colorless prisms (MeOH-DMF-H₂O, 3:1:0.5), m.p. 216 - 218 ℃; [α]²⁰ _D +142.9 (c 0.12, H₂O); UV (MeOH) λ_max (log ε) 200 (3.88), 228 (4.07), 270 (3.81) nm; CD (MeOH) 214 (Δε +0.94), 230 (Δε -1.3), 259 (Δε +2.7) nm; IR nmax 3172, 3023, 2961, 2916, 2818, 2746, 2684, 1641, 1612, 1597, 1515, 1442, 1413, 1366, 1295, 1269, 1236, 1169, 1105, 1078, 973, 955, 931, 908, 865, 829 cm^-1; ¹H NMR (DMSO-d6, 500 MHz) data, see Table 1; ¹³C NMR (DMSO-d₆, 125 MHz) data, see Table 1; (+)-ESIMS m/z 442 [M + H ]⁺, 464 [M + Na ]⁺, 480 [M + K ]⁺; (+)-HRESIMS m/z 442.1802 [M + H]⁺ (calcd. for C₂₃H₂₈N₃O₄S, 442.1795), 464.1615 [M + Na ]⁺ (calcd. for C₂₃H₂₇N₃O₄SNa, 464.1614).

2.4. ECD Calculation of 1

Conformational analysis of 1 was carried out via Monte Carlo searching with the MMFF94 molecular mechanics force field using the Spartan 10 software ^[30]. The lowest energy conformers having relative energies within 2 kcal/mol were optimized with the Gaussian 09 program ^[31], which was further optimized at the B3LYP/6-31 g (d, p) level. Conductor-like polarizable continuum model (CPCM) was adopted to consider a solvent effect using the dielectric constant of MeOH (ε = 32.6). The energies, oscillator strengths, and rotational strengths of the first 54 electronic excitations were calculated using the TDDFT methodology at the B3LYP/6-311G (2d, 2p) level. Among the B3LYP/6-31G+ (d, p) reoptimized 12 conformers, finally eight conformers showed relative Gibbs free energies (ΔG) under 2 kcal/mol (Fig. S1 in Supporting information), which were chosen for further ECD spectra simulation. The ECD spectrum of the conformer was simulated using a Gaussian function with a half-bandwidth of 0.30 eV. To get the final spectrum of 1, the simulated spectra of the 8 lowest energy conformers were averaged according to the Boltzmann distribution theory and their relative Gibbs free energy (ΔG). The corresponding theoretical ECD spectrum of the R-stereoisomer was depicted by inverting that of 1. All quantum computations were performed using Gaussian 09 program package, on an IBM cluster machine located at the High Performance Computing Center of Peking Union Medical College.

2.5. X-ray crystallography of 1

C₂₃H₂₇N₃O₄S(MeOH, M = 473.58, orthorhombic, P212121, a = 8.7361 (11)Å , b = 9.9339 (12)Å , c = 27.352 (3)Å , a = β = γ = 908, V = 2373.7 (5)Å ³, Z = 4, D_calcd = 1.325 g/cm³, 4389 reflections independent, 2461 reflections observed [I ≤ 2((I)], R1 = 0.0689, wR2 = 0.1620, S = 1.444. The data were collected on a Bruker P4 diffractometer with Mo Ka radiation ((= 0.71073Å ) by using the (scan technique (2.8° < 2(< 51.0°). The crystal structure was solved by direct methods using SHELXTL ^[32], and all nonhydrogen atoms were refined anisotropically using the leastsquares method. All hydrogen atoms were positioned by geometric calculations. Crystallographic data for the structure of 1 have been deposited with the Cambridge Crystallographic Data Center as supplementary publication CCDC 969999. Copies of these data can be obtained free of charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB21EZ, UK; fax: +44 1223 336 033; or e-mail: deposit@ccdc.cam.ac.uk).

3. Results and discussion

Compound 1 was obtained as colorless prisms (MeOH-DMFH₂O, 3:1:0.5) with α_D²⁰ +142.9 (c 0.12, H₂O). Its IR spectrum showed absorptions for hydroxy (3172 cm^-1), carboxylic ion (1641 cm^-1) ^[33], and aromatic ring (1612, 1597, and 1515 cm^-1) functional groups. The molecular formula of 1 was determined as C₂₃H₂₇N₃O₄S by HRESIMS at m/z 442.1802 [M + H ]⁺ (calcd. for C₂₃H₂₈N₃O₄S, 442.1795), combined with the NMR data (Table 1). The ¹H NMR spectrum of 1 exhibited resonances attributable to two nonequivalent p-hydroxybenzyl units at δ_H 7.15 (d, 2H, J = 9.0 Hz, H-2″ and H-6″), 7.12 (d, 2H, J = 8.5 Hz, H-2′′′ and H-6′′′), 6.73 (d, 2H, J = 9.0 Hz, H-3″ and H-5″), 6.71 (d, 2H, J = 8.5 Hz, H-3′′′ and H-5′′′), 5.14 and 5.04 (each d, 1H, J = 14.5 Hz, H-7″a and H-7″b), and 5.39 and 5.29 (each d, 1H, J = 15.5 Hz, H-7′′′a and H-7′′′b); an isolated trisubstituted double bond at δ_H 6.91 (s, H- 50); one nitrogen-bearing methine at δ_H 3.61 (dd, J = 10.5 and 2.5 Hz, H-2) connected to a methylene at δH 2.94 (dd, J = 15.0 and 2.5 Hz, H-3a) and 2.85 (dd, J = 15.0 and 10.5 Hz, H-3b); and three equivalent nitrogen-bearing methyl groups at δ_H 3.06 [s, 9H, N(CH₃)₃]. The ¹³C NMR and DEPT spectra displayed signals (Table 1) corresponding to the above units and two additional quaternary carbonyls attributable to carboxylic (δ_C 166.3, C-1) and thiocarbonyl (δ_C 162.4, C-2′) groups ^[33]. As compared with those of the previously isolated constituents from "tian ma" ^[25-27], these spectroscopic data, especially the chemical shifts of the methylene units for the two p-hydroxybenzyls, suggest that 1 possesses an unusual and unsaturated sulfur-containing hetereocylic parent structure with substitution of the two p-hydroxybenzyls at two nitrogen atoms, which was further elucidated by 2D NMR data. The proton resonances and corresponding protonbearing carbon resonances in the NMR spectra were unequivocally assigned by the HSQC spectroscopic data analysis. In the 1H-1H COSY spectrum of 1, cross-peaks of H-2/H2-3, H-2″(H-6″)/H-3″(H-5″), and H-2′′′(H-6′′′)/H-3′′′(H-5′′′) confirmed the presence of a vicinal homonuclear coupling system and the two 4-hydroxybenzyl units (Fig. 2, thick lines). In the gHMBC spectrum, two- and three-bond correlations from H-2 to C-1 and C-3; from H2-3 to C-1 and C-2; and from the methyl protons to C-2, together with the chemical shifts of these proton and carbon resonances, demonstrate that there is a 3-substituted 2-(trimethylammino)propanoate moiety in 1. The HMBC correlations from H-5′ to C-2′, C-4′, and C-7″; from H2-700 to C-1″ , C-2′ , C-2′0/C-6″ , and C-5′; and from H2-7000 to C-1′′′, C-2′ , C-2′00/C-6′′′, and C-4′ , along with their chemical shifts, demonstrated the presence of a 40-substituted 10, 30-bis(4-hydroxybenzyl)- 10H-imidazole-20(30H)-thione moiety. In addition, the HMBC correlations from H-2 to C-4′; from H2-3 to C-4′ and C-5′; and from H-5′ to C-3; combined with the molecular composition, revealed that the two moieties are connected via a C-3-C-4′ bond. Accordingly, the planar structure of 1 is elucidated as an unusual ergothioneine derivative ^[33]. Because 1 is an optical active compound containing only one chiral center identical with that of L-ergothioneine (2) ^[33], similarity of the positive specific rotations suggests that the 1 has the same S-configuration as 2 ^[33]. This was supported by comparing the experimental circular dichroism (CD) spectrum of 1 with the electronic CD (ECD) spectra predicted from the quantum-mechanical time-dependent density functional theory (TDDFT) calculations ^[34] of 1 and the Rstereoisomer, wherein the calculated ECD spectrum of 1 was in good agreement with the experimental CD spectrum (Fig. 3). To further confirm the structure assignment, efforts were made to crystallize 1 in a variety of solvents, and finally a single crystal suitable for X-ray diffraction was obtained in a solvent mixture of MeOH-DMF-H₂O (3:1:0.5). Subsequent single-crystal X-ray crystallographic analysis gave an expected result, an ORTEP drawing with atomnumbering shownin Fig. 4. In the crystal, 1was assembled with MeOH in a 1:1 molar ratio, which is in consistent with appearance of the signals due to MeOH in the NMR spectra.Therefore, the structureof compound1wasdeterminedas (+)-L-N, Nbis( 4-hydroxybenzyl)ergothioneine and named gastrolatathioneine.

Gastrolatathioneine (1) represents the first N, N-disubstituted ergothioneine derivative in nature though several S-substituted ergothioneine-containing natural products were reported ^{[33, 35]}. The parent structure 2 is a histidine betaine derivative with a tautomeric thiol/thione group at C-2 of the imidazole ring, which was initially isolated in 1909 from ergot and has been detected in various organisms, including plants, fungi, bacteria, animals, and humans. However, it is believed that plants obtain 2 from the soil while the animals and humans obtain 2 from their diets. Especially, animal and human bodies enrich 2 in specific tissues and cells, such as liver, kidney, ocular lens, central nervous system, seminal fluid, and red blood cells ^[36]. Due to its likely ubiquitous existence and unusual enrichment in specific tissues and cells, this compound continuously attracts interest of scientists in chemistry and biology. Although a cation transporter (OCTN1) with high specificity for 2 was reported to be responsible for the distribution difference, exact functions of 2 in microbial and human tissues are still unknown except for its unusual redox properties to benefit health ^[36-38]. Recently, genes encoding five enzymes [e.g. a g-glutamyl cysteine synthetase (EgtA), a formylglycine-generating enzyme (FGE)-like protein (EgtB or/and OvoA), a glutamine amidotransferase (EgtC), an S-adenosyl methionine (SAM)-dependent methyltransferase (EgtD), and a pyridoxal 5-phosphate (PLP)- dependent (EgtE)] and a role of each enzyme in the biosynthesis of 2 were reported ^{[39, 40]}. A plausible biosynthetic pathway of 1 is proposed on the basis of a metabolic post-modification of 2 (Scheme 1). However, since the genes encoding the enzymes (EgtD and EgtE) that catalyze key biosynthetic reactions of 2 are absent in higher eukaryotes, but common in actinobacteria, cyanobacteria, pezizomycotina, and basidiomycota, as well as in numerous bacteroidetes and proteobacteria, ^[39], the holomycotrophic plant G. elata must obtain 2 from a symbiotic fungus to synthesize 1.

As the plant may not produce the biosynthetic precursor, whether1 isan artifact formedinthe experimental procedurewould be an important question to be answered. To support the natural origin of 1, a recollection sample of "tian ma" (∼1 g) was soaked in water (30 ml) at room temperature overnight to afford a crude extract, which was analyzed by HPLC-ESIMS. The result clearly indicated the presence of 1 with t_R 17.9 min and m/z 442.2 in the freshly prepared crude extract (Fig. S29 in Supporting information).

In the preliminary in vitro assays against serum deprivationinduced PC12 cell damage, against Fe²⁺ -cysteine induced rat liver microsomal lipid peroxidation, and against scavenging 1, 1- diphenyl-2-picrylhydrazyl (DPPH) radical, as well as inhibitory activity against several human cancer cell lines and HIV-1 replication ^{[26, 27]}, compound 1 was inactive at 10 µmol/L. Because the biogenetical precursor 2 is a well-established antioxidant playing fundamental physiological roles in cellular redox processes ^[36-38], loss of the activity of 1 suggests that a tautomerization of the thiol/thione group in 2 may be responsible for the antioxidative properties, which is prohibited by substitution of 4-hydroxybenzyl at the two nitrogen atoms of the imidazole ring in 1.

4. Conclusion

A new N, N-disubstituted derivative of ergothioneine, named gastrolatathioneine (1), was isolated from an aqueous extract of the rhizomes of G. elata. Its structure was unambiguously established by extensive spectroscopic data analysis and confirmed by X-ray crystallography. Compound 1 is the first ergothioneine derivative isolated and structurally characterized from a plant extract since all the previously reported ergothioneine- derived natural products are from fungi and animal ^{[33, 35, 36, 41]}. Because G. elata is a holomycotrophic plant living on a symbiotic mycorrhizal fungus Armillaria mellea ^[42], the absence of the biosynthesis genes of ergothioneine in the higher eukaryotes indicates that 1 is co-produced by the plant and symbiotic fungus, e.g. the later synthesizes, and provides the precursor ergothioneine for the former to make 1. This, together with the previous characterization of various 4-hydroxybenzyl analogues and 4-hydroxybenzyl-substituted metabolites, including 4-hydroxybenzyl-substituted amino acids, glutathione derivatives, and nucleosides ^[25-27], suggests the presence of unique biosynthetic pathways and associated enzymes to produce and transfer the 4-hydroxybenzyl unit to the diverse substrates. Although 1 was inactive in the assays carried out in this study, the unique structure adds a new member to the few natural products derived from the abnormal amino acid ergothioneine, and provides a new entity for synthesis and further biological evaluation on other biological models including animal models in future. In particular, the proposed biosynthetic pathway provides an important clue for investigation of a metabolite-associated relationship between the plant and symbiotic fungus, as well as biosynthesis of the unique structure.

Acknowledgments

Financial support from the National Natural Science Foundation of China (Nos. 81502942, 81522050, 30825044), and the National Science and Technology Project of China (Nos. 2012ZX09301002- 002, 2011ZX09307-002-01) is acknowledged.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cclet.2016.06.040