A new depsidone from the neotricone-rich chemotype of the lichenised fungus Usnea fulvoreagens

Abstract Individuals of Usnea fulvoreagens (Parmeliaceae, lichenised Ascomycota), a shrubby corticolous species that is widespread in Europe, East Asia and North America, produce medullary lichen acids in several distinct chemotypic patterns. One such chemotype reportedly contains an unidentified substance as the major secondary metabolite. We isolated this compound from Californian specimens of U. fulvoreagens and identified it as the rare depsidone neotricone. A co-occurring compound, conneotricone, was identified as 4,10-dihydroxy-5-(hydroxymethyl)-8-methyl-3,7-dioxo-1,3-dihydro-7H-isobenzofuro[4,5-b][1,4]benzodioxepine-11-carboxylic acid by NMR and HPLC-UV-MSn comparison with the material synthesised from salazinic acid. Graphical Abstract


Introduction
The lichen genus Usnea is easily recognised by the beard-like fruticose thallus with radially symmetric branches consisting of an elastic cartilaginous central axis surrounded by a medulla of fungal hyphae, a subcortical layer of algal cells, and a dense translucent cortex of agglutinated hyphae. As with most lichen genera, the thallus of Usnea accumulates lichen acids, secondary metabolites of great variety that occur almost exclusively in lichens. One of these, the dibenzofuran usnic acid, is invariably found in the cortex of Usnea and gives it a yellowish cast. This compound has antibiotic properties and also presumably helps shield the underlying tissues from harmful UV radiation (Ara ujo et al. 2015).
The medulla of Usnea also usually contains lichen acids, and their composition is often characteristic of a species. These metabolites are thought to serve as a defence against herbivory, an important function in these extremely slow-growing organisms. This has been experimentally demonstrated in the case of lichen-grazing snails (Gauslaa 2005;Solhaug et al. 2009).
Among the most common medullary lichen acids are so-called b-orcinol depsidones derived from two molecules of 2,4-dihydroxy-3,6-dimethylbenzoic acid connected by an ester and a diaryl ether linkage. The structurally simplest depsidone in this class, hypoprotocetraric acid (Figure 1), occurs in a number of lichens, though it is not particularly common. Much more widespread are some of the vast number of b-orcinol depsidones that vary from this basic structure by modifications of the eight substituent positions in rings A and C ( Figure 1). In particular, some of the methyl groups are typically oxidised to some degree (and often further functionalised), and O-methylation of the phenolic hydroxyls is common.
A sub-group of b-orcinol depsidones consists of those with a c-lactone ring bearing an -OH group, as depicted in ring D of structures 3-5 in Figure 2. The origin of this substructure can be attributed to ring-chain isomerism between -CHO and -COOH groups at positions C-12b and C-3b of ring C of the corresponding depsidone tautomer. Dozens of depsidones with this substructure have been identified, and many of them are common medullary lichen acids in Usnea species.
In general, the lichen acid composition is a helpful diagnostic character that is widely employed by lichenologists for separating species. However, it is sometimes the case that an Usnea species exhibits chemotypes, where the lichen acid composition is qualitatively different between morphologically indistinguishable specimens. One such species is U. fulvoreagens, which is widely distributed in northern Europe, East Asia and North America. Although typification, and thus denomination, of the species is problematic (Tavares 2002;Halonen and Ahti 2002), the taxon to which this species epithet is applied is distinctive and generally well known. T€ orra and Randlane (2007) and Halonen et al. (1999) describe several chemotypes of U. fulvoreagens in Europe, one of which produces a major unidentified substance. Since lichen acids can have both biological and chemotaxonomic significance, correct identification of these metabolites is of value in lichenological research. This led us to undertake the identification of the major U. fulvoreagens compound and of a cooccurring minor unknown that we detected in survey analyses of Californian specimens of the species.

Results and discussion
In addition to usnic acid, analysis of specimens of U. fulvoreagens by HPLC-UV-MS n indicated two unknown medullary metabolites in ca. 8:1 ratio. Neither of them had UV or MS n spectra that matched available literature data for common b-orcinol depsidones. The major unknown, which had an apparent M r ¼ 372, was isolated, and its 1 D and 2 D 1 H and 13 C NMR spectra in (CD 3 ) 2 SO were obtained (see Supplementary Material). The NMR data suggested a structure similar to the common depsidone norstictic acid (Figure 2, structure 5) but without a C-1 hydroxyl group on the c-lactone ring. The other difference from norstictic acid (5) was the presence of a carboxyl at C-11 instead of a formyl group. 1 H NMR assignments for the four proton-bearing carbon atoms was by 1 H-13 C heteronuclear single-quantum coherence (HSQC), and the longrange 1 H-13 C heteronuclear multiple-bond correlations (HMBC) were consistent with (though not conclusive for) the suspected structure ( Figure 2, structure 1).
A depsidone with this structure was reported from the crustose lichens Phaeographis neotricosa and P. syngraphizans (Elix et al. 2003) and named neotricone. However, the 1 H and 13 C NMR data of Elix et al. (2003) for this compound, reportedly run in (CD 3 ) 2 SO, differed from our data on the U. fulvoreagens compound (see Supplementary Table S1). Since Elix et al. (2003) also report 1 H shifts in (CD 3 ) 2 CO for the C-H n protons of neotricone, we measured these for the U. fulvoreagens material in that solvent as well (see Supplementary Material). Our 1 H shift values for the C-5 methyl and C-1 protons were very close to the published data, but those for the C-8 methyl and C-9 protons were again somewhat different.
Confronted with this conflicting NMR data, we obtained samples of P. syngraphizans and P. neotricosa and analysed them by HPLC-UV-MS n under the same conditions as with our U. fulvoreagens material. We found the retention time and UV, MS 2 and MS 3 spectra of the main Phaeographis metabolite to be indistinguishable from those of the compound from U. fulvoreagens. Unfortunately, there was not enough Phaeographis material to allow us to also purify the main metabolite and measure its NMR spectra. Support for our identification of the main metabolite in U. fulvoreagens as neotricone (1) has recently been reported by Recchia et al. (2020), and their NMR data measured in (CD 3 ) 2 SO agree with ours. In that study, the theoretical chemical shifts for structure 1, calculated by density functional theory (DFT), more closely match our NMR data than that of Elix et al. (2003). Residual chemical shift anisotropy (RCSA) measurements by Recchia et al. (2020) on the U. fulvoreagens material also are in excellent agreement with theoretical RCSA calculations for structure 1.
A co-occurring compound from U. fulvoreagens that eluted earlier than 1 on reverse-phase HPLC had an apparent M r ¼ 388. The HPLC behaviour and molecular mass suggested a candidate structure (Figure 2, structure 2) related to 1 by oxidation of the C-5 methyl to a hydroxymethyl group. Since Elix et al. (2003) successfully synthesised neotricone (1) from norstictic acid (5), we undertook the synthesis of 2 from salazinic acid (Figure 2, structure 3) in the same manner. Oxidation of 3 gave perisalazinic acid (Figure 2, structure 4) (see Sestile et al. 2021 for details on the synthesis and complete NMR characterisation of 4). Reduction of the C-1 hydroxyl group of 4 gave a major product that was identical in all respects to the natural product 2, for which we adopted the trivial name conneotricone. Its 1 D and 2 D NMR data were fully consistent with the assigned structure (see Supplementary Material).
Although structure 2 has not been previously reported, the name conneotricone was given by Elix et al. (2007) to an unknown minor metabolite found in neotriconecontaining samples of U. trachycarpa ('chemotype 6 0 ). The U. trachycarpa unknown eluted earlier (i.e., was more polar) than neotricone (1) in HPLC, as was the case for 2 (J.A. Elix, personal communication). Our HPLC-UV-MS n analysis of a sample of this chemotype of U. trachycarpa found an early-eluting minor compound identical in all respects to 2 from U. fulvoreagens, thus confirming the identity of the minor unknowns from U. trachycarpa and U. fulvoreagens as conneotricone (2).

General experimental methods
Details of materials and methods are given in the Supplementary Material.

Lichen material
Specimens of U. fulvoreagens were collected from shrubs and tree branches along the Skyline Trail in Sibley Volcanic Regional Preserve, Alameda County, California, U.S.A.
(ca. N37.85712, W122.20817). Each specimen was tested by HPLC-UV-MS to verify presence of the metabolites being studied. One such specimen (UC2082255) has been deposited in the University of California Herbarium.

Compound isolation and identification
Experimental details on the isolation and identification of 1 and 2, and the partial synthesis of 2, are presented in the Supplementary Material.

Conclusion
In addition to neotricone (1) we detected conneotricone (2) in all of the U. fulvoreagens samples we examined, as well as in the neotricone-containing chemotype of a distantly-related Patagonian species, U. trachycarpa (Elix et al. 2007). Thus, in Usnea, neotricone (1) and conneotricone (2) apparently comprise a chemosyndrome, where two or more structurally related compounds consistently occur together, indicating a shared biosynthetic pathway (Culberson and Culberson 1976). Since we confirmed the structure of conneotricone (2) by partial synthesis from salazinic acid (3), the co-occurrence of 1 and 2 gives indirect support for the assigned structure of 1. The previouslymentioned DFT and RCSA NMR analyses of the U. fulvoreagens material by Recchia et al. (2020) provide further confirmation that our structural assignment for neotricone (1) is correct.
Given the discrepancy between the published NMR data for neotricone from Phaeographis syngraphizans (Elix et al. 2003) and that of our U. fulvoreagens material, we believe that a re-examination of the major P. syngraphizans depsidone is warranted.
The lack of an -OH group at C-1 in the c-lactone ring of neotricone (1) and conneotricone (2) is a feature that has very rarely been observed in lichen depsides or depsidones (Papadopoulou et al., 2007). This structure implies that the precursor bears a -CH 2 OH group at position C-12b (Figure 1) rather than the much more typical formyl group. Lactonisation between the vicinal -CH 2 OH and -COOH groups in ring C leads to the unsubstituted c-lactone ring seen in 1 and 2. Furthermore, the presence of a -COOH group at C-11 is a feature of 1 and 2 that is also quite rare, typically occurring only in trace constituents that accompany major components of a chemosyndrome (see, for example, Elix and Wardlaw 2000). The combination of these structural features in 1 and 2 appears to be unique.