LC-DAD–ESI-MS/MS and NMR Analysis of Conifer Wood Specialized Metabolites

Many species from the Pinaceae family have been recognized as a rich source of lignans, flavonoids, and other polyphenolics. The great common occurrence of conifers in Europe, as well as their use in the wood industry, makes both plant material and industrial waste material easily accessible and inexpensive. This is a promising prognosis for both discovery of new active compounds as well as for finding new applications for wood and its industry waste products. This study aimed to analyze and phytochemically profile 13 wood extracts of the Pinaceae family species, endemic or introduced in Polish flora, using the LC-DAD–ESI-MS/MS method and compare their respective metabolite profiles. Branch wood methanolic extracts were phytochemically profiled. Lignans, stilbenes, flavonoids, diterpenes, procyanidins, and other compounds were detected, with a considerable variety of chemical content among distinct species. Norway spruce (Picea abies (L.) H.Karst.) branch wood was the most abundant source of stilbenes, European larch (Larix decidua Mill.) mostly contained flavonoids, while silver fir (Abies alba Mill.) was rich in lignans. Furthermore, 10 lignans were isolated from the studied material. Our findings confirm that wood industry waste materials, such as conifer branches, can be a potent source of different phytochemicals, with the plant matrix being relatively simple, facilitating future isolation of target compounds.


Introduction
The Pinaceae family encompasses 11 genera and about 230 species, making it the largest family of conifers as well as the Gymnospermae. Genera include Abies (47 spp.), Cathaya ). The members of this family are monoecious, resinous trees distributed widely in the Northern Hemisphere, with just one equatorial crossover species [1]. Although trees from the Pinaceae family occupy an area which is 68.4% of all forests in Poland [2], species diversity is not very large, with only a few species: Pinus sylvestris L.  [3]. Most of the genera present in Poland are potential differences caused by geographical location, studying the wood of most common Polish conifers may bring some insight into their polyphenolic composition as well as their activity and further possibility of use in the food or pharmaceutical industry.
This study aimed to analyze and phytochemically profile wood extracts of the Pinaceae family species, endemic or introduced in Polish flora, using the LC-DAD-ESI-MS/MS (liquid chromatography with diode array detector coupled with electrospray ionization tandem mass spectrometry) method and compare their respective metabolite profiles. We also aimed to extract some of the more uncommon compounds to elucidate their structure under NMR (nuclear magnetic resonance) and use them as reference compounds in MS/MS (tandem mass spectrometry) analysis.

Plant Material and Extract Preparation
Branch wood of Pinus cembra L., Pinus mugo Turra, Pinus strobus L., Pinus ×rhaetica Brügger, Abies alba Mill., Picea abies L., Picea glauca (Moench) Voss, Pseudotsuga menziesii Mirb., Tsuga canadensis Carrière, and Larix kaempferi Lamb. was collected from the Polish Academy of Sciences Botanical Garden. Branch wood of Larix decidua Mill. and Larix polonica Rac. was collected from the University of Warsaw Botanical Garden. Pinus sylvestris L. branch wood was collected from Chojnów Landscape Park. The plant material was authenticated according to Flora Europaea by botanical gardens botanists. Voucher specimens have been deposited in the Plant Collection, Department of Pharmacognosy and Molecular Basis of Phytotherapy, Medical University of Warsaw, Poland. Plant material was dried at room temperature and shredded. Afterward, each sample was weighed (5.0 g), defatted with 50 mL n-hexane, and extracted with 50 mL of 90% (v/v) aqueous methanol under reflux for 2 h. The extract was then filtered through a 0.45 µm membrane, reduced in a rotary evaporator and lyophilized, resulting in dried methanolic extracts.

Isolation of Lignans
Pure lignans were isolated from the branch wood of Abies alba Mill. (660 g) and Pinus sylvestris L. (250 g). Briefly, the plant material was degreased with 1 L of n-hexane and extracted 3 times in a SONIC-5 ultrasonic bath (POLSONIC, Poznań, Poland) with 100% methanol at 60 • C for 2 h and evaporated to dryness on a Rotavator rotary evaporator R-100 (Buchi, Flawil, Switzerland) yielding 43.7 g and 11.5 g of dry silver fir and Scots pine extracts, respectively, which was phytochemically characterized.
These compounds were used as reference compounds for ESI-MS/MS analysis.

Phytochemical Characterization by LC-DAD-ESI-MS/MS Method
LC-DAD-ESI-MS/MS analysis was performed on a UHPLC-3000 RS system (Dionex, Dreieich, Germany) with DAD detection (Dionex, Dreieich, Germany) and an AmaZon SL ion trap mass spectrometer with an ESI interface (Bruker Daltonik GmbH, Bremen, Germany). Separation was performed on a Zorbax SB-C18 column (150 mm × 2.1 mm, 1.9 µm) (Agilent, Santa Clara, CA, USA). The mobile phase consisted of 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B) using the following gradient: 0-60 min, 15-100% B, then 10 min of equilibration. Samples for LC-DAD-ESI-MS/MS analysis were prepared by dissolving dried extracts in 0.1% formic acid in methanol at the concentration of 10 mg/mL. Standards were prepared in the same way at the concentration of 1 mg/mL). The flow rate was 0.2 mL/min, injection volume was 5 µL, column temperature was set at 25 • C. The LC eluate was introduced into the ESI interface without splitting, and compounds were analyzed in both positive and negative ion mode with the following settings: nebulizer pressure of 40 psi, drying gas flow rate of 9 L/min, nitrogen gas temperature of 300 • C, and a capillary voltage of 4.5 kV. The mass scan ranged from 100 to 2200 m/z. Low-energy collision-induced dissociation (less than 100 eV) was used to obtain MS/MS spectra, with collision energies chosen automatically based on precursor ion masses by the integrated SmartFrag mode.
UV spectra were recorded in the range 190-400 nm. Compounds were identified by comparing their retention time and UV-visible and mass spectra with those obtained from reference compounds and/or tentatively identified by comparison with literature information.

Results
Through comprehensive LC-DAD-ESI-MS/MS analysis of Pinus sylvestris L., Pinus cembra L., Pinus mugo Turra, Pinus strobus L., Pinus ×rhaetica Brügger, Abies alba Mill., Picea abies L., Picea glauca (Moench) Voss, Pseudotsuga menziesii Mirb., Tsuga canadensis Carrière, Larix decidua Mill., Larix polonica Rac., and Larix kaempferi Lamb. branch wood methanolic extracts, 40 compounds were identified or partially identified based on the elution order, UV maxima, pseudomolecular and fragmentation ions in both positive and negative ion modes and comparison to literature data ( Table 1). Presence of some flavan-3ols, flavonoids and abietic acid was confirmed by comparison with commercial standards. In the case of lignans, most of the compounds were isolated and their structures were confirmed by NMR analysis. Then, their presence in selected wood extracts was confirmed by LC-DAD-ESI-MS/MS analysis by comparison with pure compounds. The investigated plant material contained specialized metabolites belonging to flavan-3-ols, flavonoids, lignans, sesquilignans, stilbenes, sesquiterpenes, and diterpenes. The presence of identified compounds in each studied species is presented in ( Table 2). The UV chromatograms of each wood extract are provided in Supplementary Materials (Figures S1-S13).

Flavan-3-ols
Eight compounds belonging to the group of flavan-3-ols were identified in selected wood extracts. Compounds 5 (t r = 4.5 min) and compound 7 (t r = 6.6 min) exhibited a pseudomolecular ion at m/z 289 and had identical fragmentation patterns. Based on their elution order and comparison to reference standards, these compounds were identified as catechin and epi-catechin, respectively. Whilst catechin was present in all studied plant samples, we were not able to detect epi-catechin in P. sylvestris, P. mugo, P. ×rhaetica, and both spruces ( Table 2). -product of retro-Diels-Alder reaction followed by the loss of water and m/z 289, which corresponds to (epi)catechin. Based on fragmentation patterns, they were all characterized as type B dimeric procyanidins [41]. Previously, procyanidins B1-B4 have been isolated from the bark of Larix gmelinii (Rupr.) Kuzen. [42] and Pseudotsuga menziesii Mirb. [31]. In each plant sample, at least two dimeric procyanidins B were present. Larch wood was the richest in procyanidins, containing all four compounds in comparison to all studied genera (Table 2).
Additionally, one more procyanidin was detected in all wood extracts-compound 4 (t r = 4.2 min). The main pseudomolecular ion was at m/z 865 and fragmented into m/z 695, m/z 577, m/z 407 and m/z 289. Based on the fragmentation pattern [43], this compound was assigned as type B trimeric procyanidin (procyanidin C). Furthermore, compound 1 (t r = 3.0 min) gave a pseudomolecular ion at m/z 305, which fragmented into m/z 219, m/z 179 and m/z 137. This fragmentation pattern was typical for gallocatechin [41]. The presence of this flavan-3-ol were found in A. alba, all studies Pinus spp., and Larix spp. (Table 2).

Flavonoids
A total of 10 flavonoids were found in studied wood extracts. Most of them were present in aglycone form, with just one glycoside form detected. Compound 18 (t r = 12.4 min) was the most abundant flavonoid in these extracts, as it was not present only in A. alba and T. canadensis (Table 2), and had the highest peak of all flavonoid aglycones. Compound 18 had its UV maximum at 280 nm and exhibited a pseudomolecular ion at m/z 303, which fragmented into m/z 285 and m/z 177. The compound was then identified by comparison with the reference standard and comparison to the literature data [44][45][46] as taxifolin (syn. dihydroquercetin). Another major flavonoid compound 13 (t r = 10.2 min) had its UV maximum at 287 nm and its pseudomolecular ion at m/z 465 and gave a primary fragment ion at m/z 303 [M−H−162] − corresponding to the cleavage of hexose (probably glucose). Thus, compound 13 was assigned as taxifolin hexoside, which was consistent with the fragmentation pattern reported previously [44].
Four more flavanonols were detected in Pinaceae wood, with their UV maxima at around 290 nm.  [48], as pinobanksin, and compound 36 (t r = 30.5 min), which exhibited a similar fragmentation pattern, was assigned as pinobanksin 3-O-acetate, also consistent with the data found in the literature [49]. The presence of pinobanksin was not genus-specific, while pinobanksin 3-O-acetate was only found in P. strobus wood ( Table 2).
We have found presence of two flavanones, compound 24 (t r = 15.4 min) and 35 (t r = 30.3 min), which had their UV maxima at around 290 nm. Compound 24 presented a similar fragmentation pattern to that of compound 11. Based on the elution profile and referring to the literature [50], it was tentatively assigned as eriodictyol. It was present in all larch species, Douglas fir, mountain pine, Swiss pine, and Weymouth pine. Pseudomolecular ion of compounds 35 was at m/z 255 and fragmented into m/z 213 and m/z 211. By comparison with the literature data [51] it was assigned as pinocembrin. We have found its presence in most pine species (apart from P. cembra) and in Douglas fir ( Table 2).
We report presence of two flavonols in studied wood extracts, which were characterized by their UV maxima at around 370 nm. Compounds 25 (t r = 15.7 min) and 31 (t r = 19.8 min) had typical fragmentation patterns of myricetin and quercetin, respectively [52], which was confirmed by comparison with reference standards. Quercetin and myricetin were found in larches and Douglas fir. Additionally, Weymouth pine contained quercetin, while myricetin was also present in mountain pine ( Table 2).
The 1D NMR and ESI-MS/MS spectra along with NMR chemical shifts of all isolated item are presented in Supplementary Materials (Figures S94−S103; Table S1).
Based on the LC-DAD-ESI-MS/MS analysis, three compounds, 9 (t r = 7.8 min), 12 (t r = 9.9min), and 16 (t r = 11.6 min), had their pseudomolecular ion at m/z 375 [M−H] − and UV maxima at around 225 and 280 nm. There have been previously reported three lignan isomers in conifer wood, that could give pseudomolecular ion at m/z 375: liovil, todolactol and 7-hydroxylariciresinol [57]. Unfortunately, no ESI-MS/MS fragmentation pattern data could be found for liovil and 7-hydrolariciresinol., with just one source for todolactol [57]. Through NMR analysis, we were able to assign compounds 9 and 16 as 7-hydroxylariciresinol diastereoisomers, respectively. Fragmentation of m/z 375 ion corresponding to compound 12 gave m/z 327, m/z 191, and m/z 176 which was consistent with the fragmentation pattern proposed for todolactol [57], and in agreement with fragmentation of our isolated standard. The occurrence of lignans in different conifer species was as follows: 7-hydroxylariciresinol (I) was found in silver fir and both spruces, todolactol was additionally present in European and Japanese larch, Douglas fir and Canadian hemlock. The lignan 7-hydroxylariciresinol (II) was present in at least one species of each genus except spruces (Table 2) [57], and that was not observed for compound 30. Compound 30 produced a fragment ion at m/z 342 [M−H−15] − corresponding to the loss of methyl radical, and at m/z 151 (guaiacyl) and m/z 136, which were the products of α,β-cleavage in the side chain. This fragmentation pattern was characteristic for furofuran lignans [57]. According to the fragmentation, elution order in RP-LC, and comparison with reference standards, compound 30 was identified as pinoresinol, while compound 32 was identified as matairesinol. Both compounds were also isolated, and their structure was confirmed through NMR analysis. Pinoresinol was only present in Scots pine wood and matairesinol could be found in nearly every plant material except silver fir and Douglas fir ( Table 2). which corresponds to the loss of water and CO was observed,. Both fragmentation patterns and differences among them have been previously observed [58]. According to the fragmentation, elution order in RP-LC, and comparison with reference standards, compound 22 was identified as 7-hydroxymatairesinol and compound 29 was identified as nortrachelogenin (syn. 8 -hydroxymatairesinol). They were also isolated and their structure was confirmed using NMR methods. Nortrachelogenin and 7-hydroxymatairesinol were observed in most analyzed conifer wood, only excluding mountain and Rhaetic pine for the former, and additionally excluding Scots pine for the latter. tentatively identified as α-conidendric acid [57]. It was found in silver fir, mountain pine, Weymouth pine, all larch species, and Canadian hemlock ( Table 2) [57] for lariciresinol (27) and cyclolariciresinol (syn. isolariciresinol) (20), respectively, which was then confirmed by isolating these compounds and elucidating their structure through NMR analysis. Cyclolariciresinol was only found in silver fir, while lariciresinol was present in all larch species, silver fir, Douglas fir, Weymouth pine, Canadian hemlock, and white spruce ( Table 2).
Apart from cyclolariciresinol (20) , corresponding to both rearrangements. This structure was compared to fragmentation data described in the literature [57], isolated, and identified as secoisolariciresinol. This was the dominant metabolite of silver fir and was also found in all larch and spruce species, mountain and Weymouth pine, Douglas fir, and Canadian hemlock ( Table 2). The fragmentation pattern was in accordance with the data for oligolignans isolated from Norway spruce and Scots pine knots [12]. Based on fragmentation, this compound was tentatively identified as secoisolariciresinol 4-O-guaiacylglyceryl ether. It was found in silver and Douglas fir, as well as all larch species and mountain pine ( Table 2).  [12]. It was only found in silver fir wood (Table 2).

Sesquiterpenoids and Diterpenoids
We were able to identify one sesquiterpenoid and two diterpenes in conifer wood-all of them in positive ion mode.  [64,65] and the reference standard, we were able to identify compound 39 as neoabietic acid and compound 40 as abietic acid. Dehydrojuvabione was only found in A. alba while both abietane-type resin acids were present in all studied conifer species wood ( Table 2).

Discussion
Conifer wood and waste materials from its processing in the paper and wood industry (such as knots, bark and branches) have been identified as a good source of many polyphenols. Willfor and his team characterized wood knots of spruces, pines and firs, isolating many lignans in the process [34,39,66,67]. More recently, Gabaston et al. not only characterized wood knots of Pinus pinaster Aiton using LC-MS/MS, but also studied their antifungal activity against grapevine pathogens [53]. Another fine work offered LC-MS/MS analysis of stilbenes in Picea jezoensis Carrière bark [28]. To the best of our knowledge, most of the wood studied here was never characterized before using LC-MS/MS techniques. This study was also the first attempt at phytochemical profiling of wood from P. ×rhaetica and L. polonica, as well as the first to compare the composition of Pinaceae species growing in Poland.
Referring to previous research, our study confirms that conifers from the Pinaceae family are rich in otherwise scarce compounds, such as lignans and stilbenes, offering a large variety of these structures.
Lignans are a group of natural polyphenols (and one of the most lipophilic), located in plant cell walls. According to the recent nomenclature: lignans dimers of two coniferyl, sinapyl, 4-hydroxycinnamyl alcohol or similar alcohol monomers. Some authors restrict the term lignan only to those molecules coupled by the central carbon of the sidechain (i.e., 8,8 or β, β' dimers) [68]. They were first identified in conifer wood, which is why they often obtain their names from those species, e.g., pinoresinol from Pinus nigra Aiton and lariciresinol from Larix decidua Mill. [69]. Although lignans have been detected in numerous plants, with many being part of human diet, their content is usually very low. Apart from conifer wood knots (where they can accumulate in amounts up to 30% (w/w)), flaxseed and sesame seeds are considered rich but challenging sources of lignans. In this context, studied wood from the Pinaceae family offers a much easier-to-extract source of these polyphenols, with much fewer steps required to obtain pure compounds (directly in aglycone form with much larger structure variety). Wood from all studied species contained detectable amounts of lignans, with silver fir being the richest source of these polyphenols. Although A. alba offered large quantities of lignans, especially secoisolariciresinol and lariciresinol, P. sylvestris was far richer than all other species in nortrachelogenin, pinoresinol, and matairesinol. This is particularly interesting, as Scots pine is the most abundant and economically most important tree in Europe. Another important notion from our study is the detection of a lignan (namely matairesinol) in P. cembra, which was previously described as the only pine not containing lignans [7].
Through our analysis, we did not detect few lignans that have been previously found in some of the studied species. Willfor et al. reported presence of matairesinol, pinoresinol, and nortrachelogein in A. alba wood knots [34], but we did not observe any signal at expected retention times. In the same study, presence of secoisolariciresinol monomethyl ether and dimethyl ether was established. We cannot rule out the presence of these compounds in our samples, as pseudomolecular ions [M−H]− with expected mass for these lignans were observed. Unfortunately, no fragmentation pattern for them could be found in the literature, nor could we isolate them from the plant material. Similar problems occurred with analysis of sesquilignans, for which only few fragmentation patterns could be found, thus limiting our ability to identify them through simple LC-MS/MS analysis. Noticing that many lignans and sesquilignans did not have ESI fragmentation patterns reported previously, our study fills this knowledge gap for at least some of these compounds.
Species from the Pinaceae family are considered as one of the richest in stilbenes content. These rare polyphenols consist of two aromatic rings linked by a C 2 ethylene sidechain, formed through cyclization of 3 malonyl-CoA with cinnamoyl-CoA, coumaroyl-CoA or caffeoyl-CoA, followed by hydroxylation of aromatic rings [70]. It has been previously reported that stilbenes are abundant constituents of P. abies and P. jezoensis bark, with their levels in wood being much lower [28,32]. Hovelstad et al. reported no presence of stilbenes in P. abies heartwood and knotwood; however, their analysis was limited to pinosylvins [29]. Our study offered analysis of stilbenes from different groups and included species not screened for stilbenes before. We have shown presence of stilbenes in all studied pines, spruces, and larches wood extracts, though their structure and nature varied among species. Stilbenes present in pine wood belonged to the pinosylvins, which were more lipophilic than those found in spruce and larch wood, which mostly occurred in glycosidic form. In the case of spruces, the UV signal for trans-astringin was much stronger than that of most abundant lignan reported in these species, i.e., 7-hydroxymatairesinol, which could imply higher content of stilbenes than lignans in spruce wood.
As was the case with lignans, fragmentation patterns for stilbenes are rarely reported in the literature, thus making LC-DAD-ESI-MS/MS analysis of these compounds difficult. This study offers MS 2 fragmentations for some of these compounds, yet we see the need for further exploration of ESI fragmentation. Additional focus should be directed at differentiation of cis/trans isomers and the site of glucose substitution. In the case of our study, no differences were observed for such compounds and identification had to be based on elution order.
Our analysis also offered an insight into flavonoid, flavan-3-ol, and diterpene composition of conifer wood. These compounds were present in all studied samples, with differences concerning structures and their exact content. Larches and Douglas fir could be considered the richest in flavonoids and flavan-3-ols.

Conclusions
Overall, results obtained from this study show that conifer wood is a significant source of different phytochemicals, with many rare structures present. Silver fir wood could be a rich source of lignans, pine wood offer both lignans and stilbenes aglycones, while spruce wood had stilbenes glycosides as dominant metabolites. Larch and Douglas fir wood can be considered abundant sources of flavonoids and flavan-3-ols.
Furthermore, isolation of these metabolites from the plant matrix can be easily achieved. As these plant metabolites have already proven numerous pharmacological and dietary activities, they may find applications in pharmaceutical, cosmetic and food industries. Additionally, LC-ESI-MS has been established as quite a successful method for conifer wood metabolites detection and identification, allowing for simultaneous identification of phytoconstituents from different groups.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.