The colorful fruiting bodies (FBs) of several dermocyboid Cortinarii were collected in three forest sites in Tyrol (Austria) and identified according to their macroscopic and microscopic characteristics. Due to the wide morphological variety and overlapping morphological characters, a reliable identification beyond species complexes (i.e., species with predominantly yellow, red, and orange lamellae) was not always possible. In consequence, the fungi were initially sorted according to their pigmentation type. The latter was defined by Gruber [27] based on the appearance of the gills and the detected anthraquinones. It was later extended by Keller [28] as well as Høiland [17]. Five types are known, whereof three are relevant for this study (i.e., Cinnamomea, Sanguinea, and Malicoria type, which have yellow, red, and orange gills) and are distinguished by their major pigment/s (see Table S2).

A trustworthy assignment was impeded by the lack of a recent taxonomic key containing all dermocyboid Cortinarii. Thus, after processing and air-drying, DNA extraction was performed, and the internal transcribed spacer (ITS) sequences were analyzed. As a result, many unexpected species were observed: especially the collections of fruiting bodies belonging to the Cinnamomea (yellow) and the Sanguinea (red) pigmentation type [28] showed a great species diversity; the rarely confirmed species C. holoxanthus, C. huronensis, and C. hadrocroceus were, for example, identified. In contrast, the frequently reported, morphologically similar species C. croceus and C. bataillei were not confirmed, even though macroscopical features and the utilized identification keys ‘Kleine Kryptogamenflora’ [29] or Funga Nordica [30] initially hinted toward those species.

Another challenging, morphologically very similar pair of species are C. malicorius and C. rubrophyllus. On a macroscopic level, they differ in the orange color of the veil remnants on the stipe and the brim of the pileus [30]. However, phylogenetic analysis allowed a precise identification.

Most collections with red pilei and red lamellae, and thus belonging to the Sanguinea pigmentation type, were identified as C. sanguineus var. aurantiovaginatus (12 of 21) or C. ominosus (5 of 21). Nevertheless, also some rare species were discovered like C. fervidus, C. purpureus, and C. vitiosus. The typical habitat of all collected dermocyboid Cortinarii was Sphagnum bog co-dwelled with young conifers (e.g., Picea abies or Pinus sylvestris). Interestingly, several dermocyboid Cortinarii species are often occurring intermixed with each other or intermixed with other species within one square meter. This made the identification of closely related species especially challenging.

Altogether, 49 collections belonging to ten different dermocyboid Cortinarius species (see Table S1) were gathered and processed (i.e., identification, logging, separating into different tissues, and air-drying). This high diversity of different species, especially the rare species, prompted our interest in testing the hypothesis that the photoactivity of dermocyboid Cortinarii is concentrated in their reproductive organs. Thus, we used collections (representing biological replicates) belonging to ten different species (see Figure S1-S10) for our extensive HPLC-DAD(-MS) and photoactivity analysis. At least five fruiting bodies were analyzed of each collection. For C. hadrocroceus, C. malicorius, C. ominosus, and C. sanguineus var. aurantiovaginatus, even ten fruiting bodies of one collection were examined. All individually collected fruiting bodies were processed prior to air-drying. In detail, the sporocarp was cut lengthwise into two halves, whereby one half was further separated into the pileus, stipe, and gills. This labor-intensive approach was chosen to reduce the effect of abiotic and biotic factors (e.g., temperature or age), which affect the content of the secondary metabolites. If the samples had been pooled, an undesired level of uncertainty would have been gained. Where the biomaterial allowed it, three technical replicates were investigated. In sum, 533 extracts were analyzed and statistically processed.

The extracts of the fruiting bodies and the respective tissues (pileus, stipe, and gills) were submitted to the DMA-assay (9,10-dimethylanthracene, a chemical probe for singlet oxygen [5]) to test whether photoactivity is a common phenomenon in the AQ-producing species of Cortinarius subgenus Dermocybe. All species produced singlet oxygen, as depicted in Supplementary Figure S11 and Table S3, with a relative photoactivity of at least 33 ± 15% (for C. ominosus) compared to the natural photosensitizer berberine. Extracts of the root of Berberis ilicifolia containing berberine [31] yielded a relative photoactivity value of 29% [6]. Evaluating just the complete basidiocarps, extracts from fungi of the Malicoria pigmentation type (C. malicorius and C. rubrophyllus) were the most active, followed by those from the Sanguinea type and the Cinnamomea type. In the present set of fruiting bodies, no clear ranking between the latter two pigmentation types could be drawn, which seemingly contradicted our previous results [6], where species of the Cinnamomea pigmentation type were found to be highly active (average photoactivity 180–260%). Nevertheless, the species analyzed in this study were different ones (i.e., C. olivaceofuscus, C. uliginosus, and C. cinnamomeoluteus vs C. holoxanthus, C. huronensis, and C. hadrocroceus) and the extracts were prepared in another manner (i.e., sequential extraction with petroleum ether and methanol vs methanol only). The most prominent difference—and probably the reason behind these diverging results—was, however, the age; the basidiocarps of the initial study [6] were sourced from an herbarium and were up to 45 years old. As a consequence, e.g., the air and light unstable pre-anthraquinones flavomannin-6,6′-di-O-methyl ether (FDM) (11) and anhydroflavomannin-6,6′-di-O-methyl ether (AFDM) (14) [32, 33] were nearby completely oxidized to 7,7’-biphyscion (15), the most active photosensitizer isolated from C. uliginosus. In the present study, however, freshly collected and air-dried fruiting bodies were investigated, thus the precursor FDM (11) was still present as major compound. In another recent study, the photoyield of FDM (11) was determined to be φΔ = 2% [14]. 7,7’-Biphyscion (15), in contrast, yields ten times more singlet oxygen (φΔ = 20%), corroborating this explanation.

Strikingly, independent of the ability to produce singlet oxygen (1O2), our analysis showed that the extracts of all tested lamellae produce 1O2 the most efficiently (Fig. 1). This also holds true for C. huronensis and C. vitiosus if one processes the data in such a way that the results of the individual investigated tissues are correlated to the corresponding individual whole half fruiting body, verifying the necessity of our tedious experimental design. A clear tendency was shown for all species, though a significant effect was just calculated for C. hadrocroceus, C. holoxanthus, and C. purpureus. A larger dataset with more individual samples and consequently more data points as part of future research projects will increase the degree of freedom and thus yield true significance. Compared to the fleshy stipes, the filigree nature of gills resulted in extracts of high anthraquinone content [15]. A flaw in the DMA-assay (i.e., the extracts of the gills yield a false-positive result) is debarred by an included correction factor: the results of the DMA-quenching are corrected by the probability of the absorption; thus, solely a higher concentration of anthraquinones does not lead to a higher 1O2 production value [5].

Distribution of the singlet oxygen generation across the different tissues of all ten investigated species. The results of the individual tissues were normalized to the respective whole half fruiting bodies, to allow easy recognition of the accumulation of photoactivity. Due to the limited sample size of C. fervidus only technical replicates were measured

In short, we reproduced not only the results of C. rubrophyllus collected in 2019 [15], showing that the photoactivity is steadily enhanced in the gills, even in collections of different years and sites, but also significantly extended the number of photoactive dermocyboid Cortinarii. In total, the number of Cortinarius subgenus Dermocybe species investigated for their photoactivity increased to sixteen (i.e., C. cinnabarinus [6], C. cinnamomeoluteus [6], C. croceus [5], C. hadrocroceus, C. holoxanthus [14], C. huronensis, C. malicorius, C. olivaceofuscus [6], C. ominosus, C. purpureus [6], C. fervidus, C. rubrophyllus [14], C. semisanguineus [6], C. sanguineus var. aurantiovaginatus, C. uliginosus [6], and C. vitiosus). Thus, this result substantiated the strong prospect that photoactivity is a common trait of dermocyboid Cortinarii.

HPLC-DAD-MS analysis was conducted of all extracted whole half fruiting bodies. A Synergi Max-RP column was chosen for the analysis, which is excellent for selectively separating different aromatic compounds with a wide range of polarities. Analysis of the MS signals (see Figure S12–19 for selected examples), the UV–Vis absorption, and the comparison with authentic samples (i.e., 1, 6, 7, 8, 9, 11, 12, 13, and 15) allowed the annotation of all major and most minor pigments. As depicted in Fig. 2, up to seventeen different peaks were identified. Generally, they can be classified into glycosides of the monomeric acidic and neutral anthraquinones (1–7) eluting in the front (rt = 3–5 min), followed by the acidic monomers (dermolutein (8), dermorubin (9), as well as endocrocin (10)). In the middle (rt = 6.5 min), the elution of the dimeric pre-anthraquinone FDM (11) is observed. In the apolar region (rt > 7 min), the neutral anthraquinones (emodin (12) and dermocybin (13)) eluted, followed by the oxidized derivatives of FDM (11), i.e., AFDM (14) and 7,7′-biphyscion (15). Minor pigments as, e.g., dermoglaucin or the chlorinated anthraquinones were not clearly detected in the MS-spectra. Nevertheless, that might be explained by the low concentration and non-optimized MS parameters.

HPLC–DAD fingerprint analysis of all investigated species with putative annotations of the major and most of the minor pigments. A Synergi Max RP column was used as stationary phase and H2O/ACN (+ 0.1% FA) as mobile phase. The gradient is displayed in the upper chromatogram (Cortinarius ominosus). The injection volume was 5 µL. Grey expressed numbers in the HPLC diagram indicate uncertainty in the assignment due to low resolution in the respective MS spectrum. Stereochemistry of the glycosides is shown were known from a previous study [14]. For an increased readability, the numbers of the major AQs (6, 7, 11) were only displayed once and thereafter highlighted in color (yellow … emodin-1-glycoside (6), red … dermocybin-1-glycoside (7), and orange … FDM (11))

The obtained pigment profiles of ten dermocyboid Cortinarii confirmed the pigment classification according to Keller (1982) (Table S2). Depending on the age and the extraction procedure, the major pigments FDM (11) (in the absence of emodin (12)), emodin (12), and dermocybin (13), including the glycosidic derivatives of these monomers, serve well as markers for the yellow (Cinnamomea), orange (Malicoria), and red (Sanguinea) pigmentation type, respectively. In the individual groups of pigmentation types, minor differences can help to differentiate the species (e.g., C. malicorius and C. rubrophyllus can be distinguished by the variety of acidic anthraquinones (e.g., dermorubin (9)). The yellow species seem to differ primarily in the diversity of glycosylated monomeric anthraquinones: while in C. huronensis, four different mono- and di-glycosides were annotated (i.e., dermolutein-di-glycoside (2), dermocybin-di-glycoside (3), dermolutein-glycoside (4), and endocrocin-glycoside (5)), in C. holoxanthus and C. hadrocroceus only 4 and 5 were detected. The latter two species were distinguishable by the content of 4 and 5. Nevertheless, further studies are needed to verify these annotated AQ-glycosides' chemical nature and validate if this is a robust differentiator.

Analysis of C. fervidus, being represented by a collection comprising only four fruiting bodies, led to the identification of all three major pigments (i.e., 6, 8, 11). Thus, its pigmentation type could be described as a blend of three different types. A detailed phylogenetic and mycochemical investigation is part of future projects.

The chromatograms of the red species, dominated by dermocybin-glycoside (7), can only be distinguished by their minor compounds. In our dataset, identification based on the pigment fingerprint is possible. Nevertheless, verification is needed to confirm HPLC-DAD analysis as a suitable tool for chemotaxonomy, which is urgently required to complement phylogenetic analyses with low resolution, as typical for this group of fungi. Larger datasets of unambiguously identified fruiting bodies are therefore required. Such datasets should contain different basidiocarps from different locations, collected under various abiotic conditions and at different stages of development. Furthermore, the dataset should comprise fruiting bodies of all known dermocyboid Cortinarii, i.e., also including species described from the Southern Hemisphere. All mycochemical well-described species should be included to train an algorithm with as many data points as possible. With such a dataset in hand and the means of modern data processing (i.e., machine learning), a mycochemistry-aided identification will be possible.

This report presents—to the authors' best knowledge—for all included fungi, the first mycochemical pigment profile based on combined chromatographic and mass-spectrometric means. Chromatographic investigations are known for C. holoxanthus [28], C. malicorius [28], C. purpureus [28], C. sanguineus [28, 34], and C. vitiosus [34]. They coincided with the results presented here.

After the assignment of all relevant peaks, we analyzed which pigments were significantly more enhanced in the lamellae as compared to the other parts of the fruiting body. Data processing was done in such a way that the individual components were always correlated to their content in the respective whole half fruiting body (to account for concentration differences due to age or other abiotic factors across the collection). Furthermore, the probability of photon absorption (λ = 468 nm) was included by expressing the fraction of that peak in the total extract. As depicted representatively in Fig. 3, and as a complete set in the ESI (Figure S20–27), for each species, we were able to putatively assign one or two pigments as responsible for the high phototoxicity of the gills.

Distribution of the respective major components in the different tissues of fungal fruiting bodies. Represented is the portion of the individual pigment accounting for the total absorbance at 468 nm. For C. holoxanthus and C. purpureus only those fungi tested in the DMA assay were used to allow a direct comparison. FB fruiting body; dl-6-gly Dermolutein-6-glycoside; Endo-gly endocrocin-glycoside; Em-1-gly emodin-glycoside, DC-6-gly Dermocybin-6-glycoside

The putatively responsible compounds seem to be the same for each pigmentation type (refer to supplementary part Figure S20–28). For the Malicoria type, emodin (12) was identified as significantly higher expressed in the gills, which reproduced our data from fungi collected in 2019 [15], where we investigated over twenty individual fruiting bodies of C. rubrophyllus. For C. holoxanthus, C. huronensis, C. hadrocroceus, and thus for all studied Cinnamomea type fungi, peak 5 —putatively assigned as the glycoside of endrocrocin (5)—was shown to be enhanced in the gills. Furthermore, 15 was proven to be enriched in the gills of C. huronensis and C. holoxanthus, which correlates with the high photoactivity, as 15 is described by a photoyield of 20%. Neither the glycoside of endocrocin (5) nor endocrocin (10) itself was yet photochemically characterized; thus, a precise photochemical statement is not possible. Nevertheless, from emodin and its glycosides, it is known that the photoactivity is retained, even if it is reduced [14]. For the species belonging to the Sanguinea type, dermocybin (13) and its glycoside (7), as well as emodin (12) (Figure S21, S23), seem to be responsible. Nevertheless, more detailed investigations are needed due to the complexity of the Sanguinea-type extracts (especially of the minor compounds) and the wide range in the contents of investigated pigments.

In sum, the detailed HPLC-DAD analysis of individual fruiting bodies of ten different dermocyboid Cortinarii showed that (i) the photoactivity is concentrated in the gills and that (ii) monomeric AQs, as well as their glycosides, are responsible for this action. That the gills contain the highest content of a toxin is a phenomenon well known in the genus Amanita. For Amanita phalloides [35, 36], A. exitialis [37], and A. fuliginea [38] it was shown that alpha-amanitin is concentrated in the gills followed by the cap. Thus, our results fit into the general body of knowledge.

To test for biological function, i.e., photochemical defense, a photoinsecticidal assay utilizing Chaoboridae larvae (Chaoborus crystallinus) was performed according to Preuß and colleagues [39]. In brief, ten larvae were treated with a methanolic extract of C. sanguineus var. aurantiovaginatus, one of the most abundant dermocyboid Cortinarius species in Tyrol, for 1 h. After incubation, one set of larvae was exposed to blue light (λ = 468 ± 27 nm, H = 74.16 J/cm2, t = 60 min) while the second set was stored in the dark, to evaluate the dark toxicity of the extract. The dead larvae were counted after three hours. As depicted in Fig. 4a, the irradiation alone—without extract—showed no effect. The treatment with the extract of C. sanguineus var. aurantiovaginatus induced a dose-dependent lethal effect in the dark as well as under blue light illumination. The irradiated population, however, was significantly more affected than the respective dark population.

A Larval phototoxicity of a Cortinarius sanguineus var. aurantiovaginatus extract tested at different concentrations and under blue light irradiation (IRR, λ = 468 ± 27 nm, H = 74.16 J/cm2, t = 60 min) or in the dark (CTR). The obtained data were not normally distributed. Thus, a Kruskal–Wallis test was performed with the following significance levels: *p < 0.05, **p < 0.005. B Micrographs of untreated (upper), dark treated (250 µg/mL, dark, middle), and light treated (250 µg/mL, sunlight, lower) Chaoborus crystallinus larvae

The transparency of the larvae allowed us furthermore to observe the accumulation of the colored anthraquinones (Fig. 4b). In the dark, the anthraquinone fraction accumulated in the intestine of the larvae (Fig. 4b, middle). Sunlight irradiation induced a lethal effect. Microscopic investigation showed that the larva was affected, and an extensive staining of the entire body was observed. In conclusion, the outcome of this experiment supports our hypothesis of photoactivated defense as a common trait of dermocyboid Cortinarii.