2.1. Qualitative determination of secondary metabolites in lichen thalli

Ten compounds (hp1–hp10) were isolated from H. physodes thalli during the UPLC–MS/MS procedure (Fig. 1). The first three compounds (hp1–hp3) showed absorption maxima in the range of 211–216 nm, 234–242 nm and 310–321 nm (Table 1), which are characteristic ranges for depsidones of the β -orcinol-type with an aldehyde or a hydroxymethyl group at position R 1 (Fig. 2) in the aromatic ring on the left-hand side (A-ring) of the molecule (Huneck and Yoshimura, 1996). The mass spectra in negative ion mode displayed deprotonated molecules [M — H] — that were detected at m / z 417, 373 and 415 for compounds hp1, hp2 and hp3, respectively (Table 1). A detailed analysis of the MS/MS spectra revealed that the most intense peaks of the product ions were formed due to the loss of small molecules (H 2 O, CO 2, CH 3 COOH or OCH 2) from the parent and daughter ions (Suppl. Fig. S1A–C). The proposed scheme of the fragmentation pathways is presented in Fig. 2. The loss of CH 3 COOH groups from the parent ions of compounds hp1 and hp3 formed the product ions observed at m / z 357 and 355, respectively. Further loss of one and two CO 2 groups from the ion at m / z 357 (hp1) led to ions at m / z 313 and 269. The corresponding ion fragments for compound hp3 were observed at m / z 311 and 267, respectively. Additional elimination of the H 2 O molecule from the product ion at m / z 269 (hp1) yielded the ion fragment at m / z 251. The loss of all of the above-mentioned groups in an alternative sequence resulted in the following fragments in the mass spectrum of compound hp1: m / z 399, 355, 339, 311, 295 and 251. It is worth noting that both compounds hp1 and hp3 have a partially coincident fragmentation pattern (Fig. 2). In addition, the product ions at m / z 311 and 267 are likely structurally equivalent with hp2 and hp3 (Suppl. Fig. S1B–C). In the lower m / z range of the MS/MS spectra of compounds hp2 and hp3, daughter ions at m / z 151 and 177 derived from typical A-ring fragmentation were observed (Suppl. Fig. S1). Most of the obtained fragments from hp2 (Suppl. Fig. S1B) corresponded to those presented previously by Huneck et al. (2004) and Parrot et al. (2013) for protocetraric acid. The partial compatibility between compounds hp1–hp3 supported their structure similarities (Huneck and Yoshimura, 1996) and biosynthetic relationship (Elix, 2014). Based on the present data (Table 1, Suppl. Fig. S1A–C) and proposed scheme of fragmentation (Fig. 2), they were identified as conphysodalic acid (hp1), protocetraric acid (hp2) and physodalic acid (hp3); all of these acids belong to the β -orcinol depsidones. The identification of hp3 as physodalic acid was further confirmed by comparing its Rf value of the TLC analysis (Table 1) with that described by Orange et al. (2001). No spots were detected for conphysodalic acid and protocetraric acid because of their very low concentrations in the extract.

Hp4–hp8 compounds had UV absorption maxima in the range of 208–224 nm, 250–270 nm and 300–333 nm (Table 1), which are typical for orcinol depsidones (Huneck and Yoshimura, 1996). The major compounds hp4 and hp7 that revealed the deprotonated molecules [M — H] — at m / z 485 and 469 (Table 1) as well as similarities in the fragmentation pathways (Fig. 3) were considered to be 3-hydroxyphysodic acid and physodic acid, respectively. Their fragmentation spectra revealed product ions corresponding to the loss of H 2 O or one or two CO 2 groups (Fig. 3). Additionally, the product ions at m / z 423 (hp4) and 407 (hp7) corresponding to the loss of H 2 O molecule and successively one CO 2 group were detected. Furthermore, the fragments at m / z 399 (hp4) and 383 (hp7) were formed by the loss of CO 2 and (CH 2 –CO) groups. The authenticity of 3-hydroxyphysodic acid and physodic acid was confirmed by comparing the results of the TLC analysis (Table 1) with those reported previously (Orange et al., 2001). The minor compound hp8 that displayed the deprotonated molecular ion [M — H] — at m / z 511 has been identified as OE -alectoronic acid. Its fragmentation pattern was partially consistent with the fragmentation route described above for 3-hydroxyphysodic acid and physodic acid (Fig. 3). The MS/MS spectrum of OE -alectoronic acid (Suppl. Fig. S1H) revealed product ions corresponding to the loss of H 2 O (at m / z 493) or a CO 2 group (at m / z 467). Moreover, the daughter ions at m / z 449 and 423 resulted from the loss of a CO 2 group from the above-mentioned product ions. In addition, trace quantities were detected of the other two compounds hp5 and hp6, which did not form distinct peaks on the UPLC chromatogram (Fig. 1). Based on the conformity of their UV absorption maxima (Table 1) with literature (Huneck and Yoshimura, 1996) and similarities of the mass spectra (Suppl. Fig. S1E–F), they were identified as 4- O - methylphysodic acid and 2 0 -O -methylphysodic acid, respectively; all of which belong to the orcinol depsidones. They were recognised by their characteristic daughter ions derived from the typical depsidone A-ring fragmentation (Elix, 1975). In the negative ion mass spectrum of 4- O -methylphysodic acid, the daughter ion was observed at m / z 261 (Suppl. Fig. S1E), in contrast to the fragmentation spectrum of 2 0 -O -methylphysodic acid in which the corresponding product ion was detected at m / z 247 (Suppl. Fig. S1F). The ion at m / z 247 was also present in the mass spectra of physodic acid and the related depsidone OE -alectoronic acid (Suppl. Fig. S1G– H). The presence of this product ion resulted from the structural similarities of these molecules; they all have the same aromatic ring structure on the left-hand side of the molecule (Fig. 3).

*

Trace amount.

The UV spectra of compounds hp9 and hp10 displayed a high similarity (Table 1) and corresponded to the β -orcinol para -depsides atranorin and chloroatranorin, respectively (Huneck and Yoshimura, 1996). Their ion mass spectra contained the deprotonated molecules [M — H] — at m / z 373 and 407, respectively (Table 1). The major fragment ions, which were detected at m / z 195, 177 and 163 for atranorin (Suppl. Fig. S1I) and at m / z 211, 195 and 163 for chloroatranorin (Suppl. Fig. S1J), were consistent with the data reported by Hiserodt et al. (2000) and Huneck and Schmidt (2006). The product ions at m / z 177 (atranorin) or 211 (chloroatranorin) and 195 originated from the cleavage of the ester bond of the molecule (Fig. 4). The product ion at m / z 163 present for both compounds was produced by the loss of methanol from the ion at m / z 195. The relative retention factor (Rf) obtained from the TLC analysis for atranorin (Table 1) was consistent with that described by Orange et al. (2001). Chloroatranorin was not detected.

In these analyses, we used a liquid chromatography–tandem mass spectrometry (LC–MS/MS) method in negative ionisation mode, in which data for lichen metabolites were hardly available compared to data gained by positive ionisation mode (Huneck and Yoshimura, 1996). The obtained results for some of the H. physodes secondary compounds (protocetraric acid, atranorin and chloroatranorin) were similar to previous reports (Hiserodt et al., 2000; Huneck and Schmidt, 2006; Huneck et al., 2004; Parrot et al., 2013). For the others acids (conphysodalic, physodalic, 3-hydroxyphysodic, 4- O -methylphysodic, 2 0 -O -methylphysodic, physodic and OE -alectoronic), we propose their fragmentation patterns in negative ionisation mode for the first time. The use of a modern sensitive analytical method allowed us to identify three additional metabolites of the lichen, i.e., conphysodalic, OE -alectoronic, and 4- O -methylphysodic acids, which accumulated in thalli in minor or trace amounts (Fig. 1). To the best of our knowledge, OE -alectoronic acid has been previously described only in some species of the genus Hypogymnia growing in Australasia (Elix and James, 1992). However, to date, none of these compounds has been identified in the lichen H. physodes . All seven well-known secondary products of H. physodes were detected in the present study (Table 1). The major compounds were physodalic, 3-hydroxyphysodic and physodic acids (Fig. 1). The present results confirm that this lichen species is qualitatively uniform and has only one chemotype (Molnár and Farkas, 2011). The newly identified compounds belong to characteristic suite of H. physodes secondary metabolites, but until now, they have remained undetected because they are present in minor or trace amounts and/or as a consequence of equipment limitations. As previously reported for other H. physodes metabolites, the content of these new substances in thalli may be modified by environmental factors such as light, UV radiation, temperature, air pollution, seasonal changes and geographical localisation (Białońska and Dayan, 2005; Hauck et al., 2013; Solhaug et al., 2009). Therefore, in some cases, these can be present in amounts that are below the limit of detection. Alternatively, these may occur as accessory substances, which are not consistently present in this species.

2.2. Detection of H. physodes secondary metabolites in spruce bark

The UPLC chromatogram of spruce bark extract at 254 nm revealed many peaks of unknown substances in the range of retention times from 1.4 to 3.7 min and from 9.6 to 10.8 min (Fig. 5A). In these two parts of the chromatogram, there were several intense peaks (e.g., Rt at 2.6, 3.2, 9.8, and 10.3 min), which seemed to correspond with the characteristic components of the bark. They also appeared in the control bark of spruce branches (but in relatively lower amounts) (Fig. 5B), however they were absent in the extract of the lichen thalli (Fig. 1). These substances likely belong to the plant phenolic compounds, which have been shown to be accumulated into a higher extent in the spruce bark covered with the lichen H. physodes (Latkowska et al., 2015) as a response of the tree to lichen infection.

In addition to plant phenolics, a few lichen metabolites were detected in the samples of bark of the branch colonised by H. physodes thalli. The peaks detected at retention times of 6.1 (hp3), 6.9 (hp4), 7.6 (hp7) and 8.8 (hp9) min (Fig. 5A), which were displayed in the MS analysis as deprotonated molecules at m / z 415, 485, 469 and 373 (Suppl. Fig. S2), were recognised as physodalic, 3-hydroxyphysodic, physodic acids and atranorin, respectively. Both the retention times and molecular masses of the compounds were consistent with those obtained for the corresponding substances in the extract of the lichen thalli (Table 1). Moreover, their fragment ions (Suppl. Fig. S2) were compatible with those obtained for the corresponding compounds from lichen thalli (Suppl. Fig. S1). For example, in the MS/MS spectrum of physodalic acid, characteristic daughter ions were observed at m / z 355, 311 and 267 (Suppl. Fig. S1C and S 2A), corresponding to the sequential loss of CH 3 COOH and two CO 2 groups, respectively. The mass spectrum of 3-hydroxyphysodic acid contained product ions at m / z 467, 441, 423, 399, 397 (Suppl. Fig. S2B), all of which have been observed previously in samples of the lichen thalli (Suppl. Fig. S1D). In addition, the physodic acid isolated from bark samples displayed the formation of typical fragment ions at m / z 451, 425, 407, 383, 381 (Suppl. Fig. S2C), which were described herein for the samples of the lichen thalli (Suppl. Fig. S1G). The mass spectrum of atranorin detected in spruce bark covered by lichen thalli displayed product ions at m / z 192, 177 and 162 (Suppl. Fig. S2D). The slight divergences in mass of the fragment ions in comparison to those described in the lichen thali (Suppl. Fig. S1I) may be due to the presence of small amounts of atranorin in the sample, but they are within the limits of error of the analysis.

None of the four compounds detected in the bark of spruce branches colonised by H. physodes, or any of the other six secondary metabolites produced by this lichen, have been found in bark without lichen, which served as control (Fig. 5B). The obtained results demonstrate the presence of the lichen secondary metabolites in the spruce bark in direct contact with H. physodes . This may be the result of their translocation from the lichen thalli into the spruce bark and/or their production inside the bark by some hyphae penetrating the tree tissues. The fact, that these substances belong to three different groups of compounds, β -orcinol depsidones (physodalic acid), orcinol depsidones (3-hydroxyphysodic acid and physodic acid) and β -orcinol para -depsides (atranorin), suggests that their molecular structure has insignificant or no effect on their allocation.

The content of H. physodes depsides and depsidones may represent more than 20% of the dry weight of the thalli (Solhaug et al., 2009). Lichen substances detected in tree bark belong to major secondary metabolites that are produced and accumulated by H. physodes thalli (Solhaug et al., 2009). It seems most likely that the amount of the substances translocated into the tree bark correlates with their quantity in the lichen thalli. The other six minor metabolites of H. physodes were not detected in the tree bark samples, likely because they were present at levels below the limit of detection rather than an inability to translocate. The limited amount of the substances translocated into the bark may be also explained by their low or moderate (considering the polar groups on the molecules) solubility in water. Moreover, the lichen substances may undergo enzymatic and/or hydrolytic degradation (Lawrey et al., 1999), but in the present study, we did not investigate the degradation products. Atraric acid, the hydrolysis product of atranorin, has been identified and quantified in cork, sapwood and the outer heartwood of oak trunks colonised by Parmelia sp. (Bourgeois et al., 1999) due to the ability of this lichen to penetrate tree tissues up to the phloem and cambium (Brodo, 1973). The hyphae of some fruticose lichens such as Evernia prunastri (L.) Ach. has a compact holdfast that reaches up to the inside of xylem vessels within tree branches (Ascaso et al., 1980). Therefore, its main metabolites, evernic, everninic and usnic acids, were present in the xylem sap of oak and birch branches that were richly covered by E. prunastri (Avalos et al., 1986; Monsó et al., 1993). Some of the lichen metabolites were also translocated into the leaves (Avalos et al., 1986). The foliose lichen H. physodes, which belongs to the family Parmeliaceae, has no specific structure but is typically attached to the substratum by folds or, rarely, by most of the lower cortex (Smith et al., 2009). For this reason, it penetrates only the superficial layer of the periderm (Brodo, 1973). Our results confirmed that the secondary metabolites of this lichen species were present in the surrounding host tree tissues. These metabolites might be the cause of the long-lasting negative effects on spruce trees that we have been observed in our previous studies (Latkowska et al., 2015). However, further experiments with extracted lichen substances are necessary to justify this hypothesis.