Acutifoliside, a novel benzoic acid glycoside from Salix acutifolia

Abstract Ultra high-performance liquid chromatography–mass spectrometry (UHPLC–MS) profiling of a polar solvent extract of juvenile stem tissue of Salix acutifolia Willd. identified a range of phenolic metabolites. Salicortin, 1, a well-known salicinoid, was the major compound present and the study identified young stem tissue of this species as a potential source of this compound for future studies. Three further known metabolites (salicin 2, catechin 3 and tremuloidin 4) were also present. The UHPLC–MS analysis also revealed the presence of a further, less polar, unknown compound, which was isolated via HPLC peak collection. The structure was elucidated by high-resolution mass spectroscopic analysis, 1- and 2-D NMR analysis and chemical derivatisation and was shown to be a novel benzoic acid glycoside 5, which we have named as acutifoliside.

or the long-leaved violet willow and is native to Russia and eastern Asia. It is a deciduous and reaches a height of up to 10 m. Previously, Zapesochnaya et al. (2002) reported the isolation and structures of 14 phenolic substances from bark of S. acutifolia Willd. (S. acutifolia). In this article, we describe the identification of four known compounds (1-4) and the isolation and structural elucidation of a novel benzoic acid glycoside 5 that is present in juvenile stem tissue of this plant (Figure 1).

Results and discussion
Ultra high-performance liquid chromatography-mass spectrometry (UHPLC-MS) profiling was carried out, in negative-ion mode, on replicated polar solvent extracts of S.acutifolia stem tissue to determine the major components of the mixture. The total ion chromatogram ( Figure S1) showed that the UHPLC-MS profile was dominated by a very large peak appearing at 20.64 min with m/z ions at 423 [M−H] − and 469 [M−H + HCOOH] − . MS-MS fragmentation was consistent with the known salicinoid, salicortin, 1, and its UHPLC-MS retention time; MS and 1 H NMR data agreed with an authenticated reference standard. Salicortin is a wellknown member of the salicinoid family and is widespread in the Salicaceae (Thieme 1964(Thieme , 1967Pearl & Darling 1969-1971Boeckler et al. 2011). Salicortin content is influenced by a number of factors including genotype, tissue type, nutritional status and plant age and levels of 0.8% (Clausen et al. 1989), 1.2-1.4% (Massad et al. 2014) and 3-6% (Lindroth et al. 2007) have all been reported in stem material of the Populus species. Of note in this work is the abundance (13% of dry stem weight (d.w.)) of this molecule in the juvenile stem tissue from this particular willow species. Often, complex mixtures of different salicinoids can be found in extracts of Salix species (Boeckler et al. 2011). In this case, young stems of S. acutifolia provide a source from which salicortin can be isolated in reasonable concentration and purity, e.g. for biosynthetic studies or onward chemical synthesis.
Additional smaller peaks were also evident in the UHPLC-MS total ion chromatogram at 12.14, 16.83 and 25.43 min and showed m/z ions at 331, 289 and 435, respectively. each of these peaks corresponded to known metabolites and was compared to data from authentic standards. The peak at 12.14 min corresponded to the known compound salicin 2 and m/z 331 arises from the formate adduct of this molecule. Identification of the peak at 16.83 min with [M−H] − 289 and comparison of its NMR data to those of an authentic standard confirmed it to be catechin 3, a well-known plant flavan-3-ol. The peak at 25.43 min was identified as tremuloidin 4 whereby m/z 435 corresponded to the formate adduct. Compounds 2-4 were present at levels of 2.0, 0.17 and 0.21% d.w., respectively. Compounds 1-4 were also present in S. acutifolia leaf tissues, and their structural assignments were confirmed after larger scale isolation from this tissue on the basis of comparisons of their 1 H-NMR spectra to those of authentic compounds, isolated from other Salix species and also by comparison to literature values (Lindroth et al. 1987;Reichardt et al. 1992) Compound 5, retention time 26.60 min ( Figure S1), was shown to be present in profiles of both juvenile stem and leaf tissue ( Figure S2) and was purified from the latter due to the availability of higher quantities of this tissue type. extraction of dried, powdered leaf tissue was carried out with a water:methanol (8:2) mixture, and compound 5 was purified by repeated injections into an analytical HPLC system. The UHPLC-MS data in negative-ion mode ( Figure S3), of the HPLC-purified compound, showed a single peak at 26.60 min whose mass spectrum had a single ion at m/z 419. An accurate mass of 419.0982 corresponded to [M−H] − of a compound with the molecular formula of C 20 H 20 O 10 . The 1 H NMR (Table S1 and Figure S4) and 1 H-1 H COSY spectra ( Figure S5) of 5 indicated signals of three-proton spin systems, which were attributable to three distinct structural units. Proton signals of an AA′BB′C spin system ( 3 J = 8.3, 7.5 Hz; 4 J = 1.2 Hz) at δ 7.99, δ 7.54 and δ 7.70 indicated a mono-substituted aromatic ring, and these signals corresponded to H-2″/6″, H-3″/5″ and H-4″, respectively ( Figure 1). HSQC data (figure S6) allowed assignment of the associated 13 C aromatic signals. HMBC correlation (Figures S7 and S8) between signals at δ 7.99 and δ 170.9 was suggestive of a benzoate moiety. Signals corresponding to a glycosidic structure appeared in the central region of the 1 H NMR spectrum. A typical doublet (δ 5.08, 3 J = 7.5 Hz) corresponded to the H-1′ anomeric hydrogen. A pair of double-doublet signals at δ 4.71 ( 2 J = 12.0, 3 J = 2.4 Hz) and δ 4.52 ( 2 J = 12.0, 3 J = 8.1 Hz) were coupled to each other and corresponded to the glycosidic 6′ hydrogens. Their downfield position indicated a substitution at this position of the hexose ring in addition to substitution at 1′, suggesting that 5 was a 1′,6′-disubstituted glycoside. H-5′ (δ 3.95, ddd, 3 J = 9.9, 7.9, 2.4 Hz) could be identified via coupling to H-6′ β and the remaining three glucosyl signals appeared between δ 3.72 and δ 3.55. The large coupling constants of the corresponding 3 J coupling values indicated axial configuration of H-1′-H-5′, thus assigning the sugar unit as β-glucose. 13 C chemical shifts were obtained from HSQC and HMBC spectra and confirmed the presence of the 1′, 6′-disubstituted β-glucopyranosyl unit. Additionally, HMBC correlations between H 2 -6´ and the carbonyl signal at δ 170.9 placed the free benzoate group at the C-6′ position of the glucose unit. An additional AMX spin system ( 3 J = 8.0, 8.1 Hz; 4 J = 1.4 Hz) was evident in the 1 H NMR spectrum and was confirmed via correlations present in the 1 H-1 H COSY spectrum. A series of two double-doublets and a triplet, all having ortho-coupling constants, appearing at δ 7.43, δ 7.16 and δ 6.55 indicated a 1,2,3-trisubstituted aromatic ring, and these signals corresponded to H-6, H-4 and H-5, respectively. 13 C data from the HMBC spectrum indicated an additional carbonyl signal at δ 177.8 consistent with carboxylic acid or ester function, and a correlation between this signal and the double-doublet at δ 7.43 suggested that the carboxyl moiety was attached to C-1 of the tri-substituted aromatic ring. Further HMBC correlations ( Figures S7 and S8) confirmed the presence of the 1,2,3-trisubstituted aromatic ring and allowed placement of the remaining functionality. Key correlations were between H-5 and C-3 (δ 147.3), the latter having a further correlation to H-1' (δ 5.08). This, therefore, provided evidence that the glucose unit was attached via the H-1′ position to C-3 of the tri-substituted aromatic ring. A remaining hydroxyl group at C-2 completed the structural unit and was confirmed via HMBC correlations of both H-6 and H-4 to the peak at δ 153.3 (C-2). NMR data for this portion of the molecule agreed with those reported (Sakushima et al. 1995;Rashid et al. 1996) for 2-hydroxy-3-O-β-d-glucosyl-benzoic acid which possessed the same aromatic ring substitution pattern. Confirmation of the presence of a free carboxylic acid was confirmed via esterification of 5 with an ethereal solution of diazomethane. The product of this reaction now contained a new 3H singlet at δ 3.93 ( Figure S9) confirming the presence of the methyl ester. Additional changes included a movement downfield of the H-4 and H-6 signals to δ 7.29 and δ 7.50, respectively. The structure of 5 was further confirmed by analysis of the MS 2 fragmentation pattern. MS-MS of the m/z 419 ion gave four confirmatory fragments ( Figure S3). The major fragment at m/z 153 with a formula of C 7 H 5 O 4 was consistent with dihydroxybenzoate. A smaller fragment at m/z 109 (C 6 H 5 O 2 ) corresponded to further loss of CO 2 from this fragment. A fragment with molecular formula C 13 H 13 O 6 (m/z 297) confirmed the attachment of the trisubstituted aromatic ring to the glucose moiety directly. Finally, the free benzoate group was confirmed by a small ion at m/z 121 (C 7 H 5 O 2 ). These ions are consistent with the MS-MS pattern obtained for 2,3-dihydroxybenzoic 3-O-β-d-xyloside which has similar functionality and has been isolated from Arabidopsis thaliana leaves (Bartsch et al. 2010). Finally, methylation of the free carboxyl group at C-1 resulted in a new peak in the UHPLC trace ( Figure  S10), which appeared at 31.89 min with an ion at m/z 433 corresponding to a formula of C 21 H 21 O 10. MS 2 analysis of m/z 433 showed a fragment at m/z 167, corresponding to a methyl, 2,3-dihydroxy carboxylate ion.
Compound 5 has not been previously reported. There is, however, a precedent for lower molecular weight, 1,2,3-substituted phenolic glycosides in other plant species such as Geniostoma antherotrichum (Rashid et al. 1996) and Boreava orientalis (Sakushima et al. 1995). In S. acutifolia, compound 5 represents a new 6′-O-benzoylated derivative which we have named acutifoliside. This molecule is a departure from the common salicinoid molecules, e.g. salicin and salicortin, typically found in the Salicaceae, and which have been shown to be biosynthesised from phenylalanine via cinnamic acid (Babst et al. 2010). Acutifoliside does not possess the salicyl alcohol unit in its structure and, therefore, represents a different class of molecule, presumably arising from different biosynthetic origins. Bartsch et al. (2010) identified 2,3-dihydroxybenzoic acid xyloside in Arabidopsis and suggested that this molecule had arisen from isochorismate and that its accumulation depended on eDS1 during ageing or resistance responses to pathogens. We, therefore, presume that acutifoliside has also arisen from isochorismate and thus is produced from the shikimate pathway from a branch that precedes phenyalanine and the salicinoids.

General experimental procedures
UHPLC-MS were recorded with an Ultimate 3000 RS UHPLC system, equipped with a DAD-3000 photodiode array detector, coupled to an LTQ-Orbitrap elite mass spectrometer (Thermo Scientific, Germany). UHPLC separation was carried out using a reversed-phase C 18 Hypersil gold column (1.9 μm, 30 × 2.1 mm i.d. Thermo, Hemel Hempstead) which was maintained at 35 °C. The solvent system consisted of water/0.1% formic acid (A) and acetonitrile/0.1% formic acid (B). Separation was carried out for 30 min under the following conditions: 0 min, 0% B; 27 min, 70% B; 28 min, 100% B. The flow rate was 0.3 mL/min, and the injection volume was 10 μL.
Mass spectra were collected using an LTQ-Orbitrap elite with a heated eSI source (Thermo Scientific, Germany). Mass spectra were acquired in negative mode with a resolution of 120,000 over m/z 50-1500. The source voltage, sheath gas, auxiliary gas, sweep gas and capillary temperature were set to 2.5 kV, 35 (arbitrary units), 10 (arbitrary units), 0.0 (arbitrary units) and 350 °C, respectively. Default values were used for other acquisition parameters. Automatic MS-MS was performed on the three most abundant ions and an isolation width of m/z 2 was used. Ions were fragmented using high-energy C-trap dissociation with a normalised collision energy of 65 and an activation time of 0.1 ms. Data analysis was carried out using Xcalibur v. 2.2 (Thermo Scientific, Germany).
For compound isolation, 12 repeated injections (100 μL each) were made into an analytical HPLC using an Agilent 1100 HPLC system equipped with a quaternary pump, diode array detector, column oven and auto sampler (Anatune, UK). Samples were separated using reversed-phase chromatography (Column: Ascentis C18, 5 μm, 5 × 250 mm (Supelco, UK)). The operating solvents were as follows: Solvent A: H 2 O with 0.1% formic acid and solvent B: acetonitrile with 0.1% formic acid. The operating gradient for peak isolation was from 5% B (10 min) to 33.8% B (70 min) at a constant flow of 1 mL/min and using an injection volume of 100 μL and a total chromatographic run of 72 min. Peaks were identified and monitored using a wavelength of 254 nm and were collected manually into glass tubes. equivalent fractions from multiple runs were combined and evaporated using a Speedvac concentrator (Genevac, Suffolk, UK). 1 H-1-D and 1 H-1 H and 1 H-13 C 2-D NMR spectra were acquired on a Bruker Avance 600 MHz NMR spectrometer (Bruker Biospin, Germany), operating at 600.05 MHz for 1 H and 150.9 MHz for 13 C NMR spectra, using a 5 mm selective inverse probe. 1-D 1 H spectra were collected using 128 scans and using the zgpr pulse sequence with a 90° angle. The residual HOD signal was suppressed by pre-saturation during a 5 s delay. Spectra consisted of 64,000 data points with a spectral width of 12 ppm. FIDs were automatically Fourier transformed using an exponential window function with a line broadening of 0.5 Hz. Phasing and baseline correction were carried out within the instrument software. COSY, HSQC and HMBC spectra were collected using standard Bruker parameter sets and acquisition details are given in Supporting Information. 1-D 13 C NMR spectra were recorded on a Bruker 400 MHz NMR spectrometer (Bruker Biospin, Germany), operating at 100.61 MHz, equipped with a 5 mm broadband probe. All spectra were collected at 300 K in D 2 O:CD 3 OD (8:2) and chemical shifts are given in δ, relative to d 4 -TSP [(trimethylsilyl) propionic acid, 0.01% w/v] added as a chemical shift reference standard. NMR data were processed using TOPSPIN v. 2.1 (Bruker Biospin, Germany) and ACD NMR Processor (ACD Labs, Toronto, Canada)

Plant material
Multiple juvenile shoots were harvested in May 2014 from the new growth of S. acutifolia (plot 1155 of the National Willow Collection, Rothamsted Research, UK) that had been coppiced at the end of the previous growing season. each plot of the collection contains 10 plants that were generated from separate stem cuttings. Plant tissue from all 10 plants was combined to give a single sample. Tissue was kept at −80 °C prior to freeze-drying to remove residual water. Stem and leaf tissues were separated after lyophilisation prior to milling to a fine power (Retsch Ultra Centrifugal Mill ZM200, Retsch, UK). Milled tissue was maintained at −80 °C until analysis. A voucher specimen has been retained and is available on request.

Metabolite extraction and isolation
For initial metabolite profiling by UHPLC-MS triplicate, aliquots of milled freeze-dried S. acutifolia stem or leaf tissue (15 mg) were extracted as previously described (Corol et al. 2014). After extraction, aliquots (120 μL) were removed to a clean glass vial and diluted with H 2 O:MeOH (80:20, 880 μL) for UHPLC-MS analysis which was carried out immediately. Relative standard deviations were calculated from the peak areas of individual peaks in the TIC ( Figure S11) and ranged from 2.62% (salicin) to 8.78% (acutifoliside).
For compound isolation, freeze-dried, milled, S. acutifolia leaf tissue (150 mg) was extracted at 50 °C (10 min) in H 2 O: MeOH (80:20, 2 mL). The sample was centrifuged (5 min) and the supernatant transferred to a new tube and heated at 90 °C (2 min). After cooling and centrifugation, the supernatant (1.6 mL) was removed to a glass HPLC vial for purification by HPLC peak collection.

Conclusions
This study has examined, by UHPLC-MS, the polar metabolite extract from juvenile stem tissue of S. acutifolia and shown that the major component was salicortin 1. From the other minor components, we identified acutifoliside, 5, a novel tri-substituted benzoic acid glycoside which also contained a benzoate group at the 6′-position of the glucose moiety. This metabolite provided evidence that phenolic glycosides in the Salicaceae can arise via branches of the shikimate pathway that precede phenylalanine.

Supplementary material
1 H NMR and 2-D 1 H-1 H and 1 H-13 C spectra are available as supporting information.