Secondary chemistry

Poelt (Reference Poelt1974, p. 190) suggested that among Candelariaceae ‘…Their chemistry is mostly uniform…’ (‘…Ihr Chemismus ist weitgehend einheitlich…’) and that the presence of ‘stictaurin’ (i.e. pulvinic dilacetone and calycin occurring together; Zopf Reference Zopf1907) was generally characteristic. In their revision of Placomaronea, Westberg et al. (Reference Westberg, Frödén and Wedin2009) list only secondary metabolites generally characteristic for the genus, not for individual species, reporting calycin, pulvinic dilacetone (= pulvic acid lactone), vulpinic acid and pulvinic acid. This agrees largely with our own analyses, although vulpinic acid could not be confirmed and, for the first time, 4-hydroxypulvinic acid is reported, present in virtually all specimens examined.

Overall, chemical variation in the material examined is low. Based on quantitative analyses using high performance liquid chromatography (HPLC), Westberg (Reference Westberg2005) distinguished only two chemosyndromes, ChA (with calycin as the dominant metabolite; characteristic for most species of Candelariella, Placomaronea, Candelina, and the 8-spored species of Candelaria) and ChB (with the secondary metabolites in approximately equal quantities; documented for polyspored species of Candelaria, Candelariella subdeflexa, and possibly Candelariella lutella).

We did not have access to HPLC and the chemotypes distinguished here are based on secondary metabolites detected by TLC. Here we document variation in whether terpenoids are present or not and if pulvinic dilacetone and/or calycin are absent, occur in minor concentration, or as a trace (Table 1). Whereas a much larger number of specimens and taxa needs to be analyzed to assess if any of this variation correlates with particular genera or species. This may seem surprising. In many groups of lichenized fungi, secondary chemistry is taxonomically relevant and chemotype patterns are often species-specific. Only recently, however, Díaz-Escandón et al. (Reference Díaz-Escandón, Tagirdzhanova, Vanderpool, Allen, Aptroot, Češka, Hawksworth, Huereca, Knudsen and Kocourková2022) suggested including Candelariaceae within Lichinomycetes, based on their much smaller genomes, arguing that these fungi are characterized by ‘reduced arsenals of carbohydrate-degrading enzymes and secondary metabolite gene clusters’ (Díaz-Escandón et al. Reference Díaz-Escandón, Tagirdzhanova, Vanderpool, Allen, Aptroot, Češka, Hawksworth, Huereca, Knudsen and Kocourková2022, p. 5209).

Most surprising in the material analyzed is the inconsistency of reddish spot test reactions in K. Zopf (Reference Zopf1895) reported that pure calycin reacts K+ red, but in the presence of pulvinic dilacetone (= pulvinic acid lactone) forms a complex called ‘stictaurin’, which no longer reacts with K. This observation cannot be confirmed. Virtually all specimens that contain both calycin and pulvinic dilacetone react distinctly K+ red, or at least K± weakly reddish (chemotypes A & B; Table 1), and specimens that contain only calycin and lack pulvinic dilacetone instead mostly react K− (chemotype C; Table 1). Curiously, even some specimens that contain neither pulvinic dilacetone nor calycin react distinctly K+ red, or K± weakly reddish.

Poelt (Reference Poelt1974) suggested, based on observations by Santesson, that polyporic acid might be responsible for a K+ weakly reddish spot test reaction in Candelina mexicana. For the material that we examined, this cannot be confirmed. Polyporic acid is a deep red pigment, whereas the medulla of C. mexicana is bright yellow, no different in colour than the cortex of that species (Fig. 7B). It seems unlikely that terphenylquinones would be present in a family of lichens generally characterized by pulvinic acid derivatives. Instead, in much of the material that we examined, we discovered a substance that most likely refers to 4-hydroxypulvinic acid, a pulvinic acid derivative (in solvent C it occurs at R _f 3 as a distinctly pale yellow spot in λ 365 nm UV light before H₂SO₄ treatment; the same spot is pale yellow in visible light after H₂SO₄ treatment, and it turns dull greenish yellow in λ 365 nm UV light after H₂SO₄ treatment; Figs 1 & 2). This secondary metabolite was present in all chemotypes and for all chemotypes at least some specimens displayed a positive spot test K reaction. If 4-hydroxypulvinic acid is indeed responsible for this reaction, the strength of the reaction may be correlated with its concentration in the cortex. Comparing whether secondary metabolites show up on a thin-layer chromatography plate as saturated, pale or barely visible (major, minor or trace) depends on many factors and it is, at least to some extent, subject to experimental variation. More sensitive, quantitative analytical methods are therefore necessary to assess whether differences in concentration of 4-hydroxypulvinic acid might explain the variation in K reactions documented here.

Anatomy

In his revision of Candelariaceae, describing the genus Candelina, Poelt (Reference Poelt1974) was the first to recognize anatomical differences in the cortical structure of the family. He emphasized that Candelina, though appearing crustose-placodioid, is characterized by a well-defined lower cortex of few rows of closely aggregated cells, attached to its substrate with what he called ‘searching hyphae’, or ‘Suchhyphen’, whereas the lower cortex of Placomaronea, illustrating P. candelarioides as an example, has a lower cortex of several closely packed, dense cell layers (Poelt Reference Poelt1974, p. 190, fig. 1). Westberg et al. (Reference Westberg, Frödén and Wedin2009), distinguishing Placomaronea from all other genera of Candelariaceae, also emphasized cortex anatomy, which they described as ‘well-developed paraplectenchymatous…the pigments form[ing] characteristic hoods…usually also by the presence of a thin, gelatinous epicortex’ (Westberg et al. Reference Westberg, Frödén and Wedin2009, p. 526).

Closely comparing the cortical anatomy of Candelariella, Candelina and Placomaronea (Figs 4– 9), the distinction of the three genera at first seems less obvious than the assessment by Westberg et al. (Reference Westberg, Frödén and Wedin2009) suggests. All three genera show at least the tendency to form a paraplectenchymatous cortex. In all three, the yellow pigments are present as minute granules in the outer part of the cortex. A cursory comparison of, for example, the crustose-subsquamulose P. mendozae with Candelariella rosulans may not, at first, suggest conspicuous differences: both have a thin layer of unpigmented, somewhat paraplectenchymatous cells, apparently topped by yellow pigment granules. Only upon closer examination are the differences revealed. The outer cell layer of P. mendozae is distinctly paraplectenchymatous, and its cells have pigmented cell walls, the yellow granules adhering to these pigment caps. When these cells die off, they leave behind a residue of unpigmented organic material shedding off, a layer that Westberg et al. (Reference Westberg, Frödén and Wedin2009) referred to as ‘gelatinous epicortex’ (Fig. 6B & C). This is not the case for Candelariella rosulans; the species lacks this residue of organic material and the uppermost cells of the cortex are not distinctly paraplectenchymatous (Fig. 8C).

Figure 6. A–D, Placomaronea mendozae (T. H. Nash 25430 (ASUL010350)). A, squamulose thallus. B & C, paraplectenchymatous cortex with apically pigmented cells, covered by a hyaline coating, possibly the residue of dead cells. D, hooded pigment cells coated with pigment granules. E & F, Placomaronea lambii (I. M. Lamb 5413 (MSC0112896—isotype)). E, inflated erect thallus areoles. F, uppermost cortex with hooded pigment cells coated with pigment granules, covered by hyaline coating, possibly the residue of dead cells. G & H, Placomaronea candelarioides (T. H. Nash 28034 (ASUL010349)). G, upper and lower view of umbilicate thalli (some indistinctly attached by broadened multiple holdfasts). H, paraplectenchymatous cortex with apically pigmented cells, covered by a hyaline coating, possibly the residue of dead cells. Scales: A, E & G = 1 cm; B, C, D, F & H = 25 μm. In colour online.

Figure 7. A–D, Candelina mexicana (T. H. Nash 39918 (ASUL034850)). A, overview of placodioid thallus. B, central part of the thallus, parts of the cortex damaged, the bright yellow medulla exposed. C, hand-cut section of the thallus, with a thin paraplectenchymatous cortex capped by a dense layer of pigment granules, green photobiont layer and yellow medulla below. D, thin section with two layers of paraplectencymatous cortical cells, capped by a dense layer of pigment granules. E–I, Candelina submexicana (R. D. Worthington 31956 (ASUL024890)). E, overview of placodioid thallus. F, detail of thallus lobes with sessile, adnate apothecia, parts of the cortex damaged, the white medulla exposed. G & H, hand-cut sections of the thallus showing the cortex; photobiont layer transitioning into a narrow layer of hyaline paraplectenchymatous cells, capped by a dense layer of pigment granules, hyaline coating absent. I, close-up of the dense layer of yellow pigment granules (squash preparation). Scales: A & E = 1 cm; B & F = 5 mm; C & H = 50 μm; D, G & I = 25 μm. In colour online.

Figure 8. A–D, Candelariella rosulans (W. C. Davis 451 (ASUL011839)). A, overview of subsquamulose areoles with abundant apothecia. B, detail of the thallus with sessile, adnate apothecia. C, hand-cut section of the thallus, showing a thin layer of proso- to indistinctly paraplectenchymatous hyaline cells above the photobiont layer, apically covered by an accumulation of pigment granules. D, close-up of hyaline cortex cells and aggregation of pigment granules. E–H, Candelariella vitellina (W. C. Davis 617 (ASUL011791)). E, overview of thallus granules with scattered apothecia. F, detail of the thallus granules with scattered apothecia. G, hand-cut and somewhat squashed section of the thallus, showing a thin layer of proso- to indistinctly paraplectenchymatous hyaline cells above the photobiont layer, apically covered by an accumulation of pigment granules. H, close-up of the hyaline cortex cells and aggregation of pigment granules (squash preparation). Scales: A = 1 cm; B & F = 2 mm; E = 5 mm; C = 50 μm; D, G & H = 25 μm. In colour online.

Originally we assumed that the organic material was a residue of dead cortical cells, perhaps better called an epinecral layer, not an epicortex. However, the thick, hyaline, layered material covering the cortex of ‘Candelariella’ kansuensis shows the same ‘shedding’ at its surface. Thus, the cortex of Placomaronea and ‘Candelariella’ kansuensis appears structurally identical (compare Figs 4– 6 with Fig. 9). Since we do not understand its origin, we prefer to call it a ‘coating of layered organic material’.

Figure 9. Candelariella kansuensis (C. M. Wetmore 55470 (ASUL011864)). A, overview of inflated, waxy thallus squamules. B, detail of squamules with few sessile, adnate apothecia. C, hand-cut section of the thallus, showing thick paraplectenchymatous layer of cortical cells, capped by swollen, hooded cells ‘peppered’ by pigment granules, covered by a thick hyaline, distinctly layered coating that apically erodes into minute, hyaline granules. D, close-up of the swollen, apical cells capped by yellow pigment hoods ‘peppered’ by pigment granules, covered in hyaline coating. Scales: A = 5 mm; B = 2 mm; C & D = 50 μm. In colour online.

Macroscopically, the genera also differ: the densely packed pigment granules on the surface of Candelariella rosulans cause a finely pruinose surface, whereas the surface of P. mendozae appears waxy, the coating of hyaline organic material smoothing the surface, covering pigment granules below. In Candelina the thallus surface appears dull, not pruinose but not smooth either, possibly a result of the much more densely packed pigment granules. The newly described Placomaronea fruticosa and P. placoidea not only appear similar macroscopically because of their placodioid growth form, but also from the outside their surface does not look different from the species of Candelina. Some material of P. fruticosa, a specimen at ASU from Argentina (T. H. Nash 27947, ASUL010376), was previously even annotated by Westberg as Candelina cf. submexicana. Entirely embedded in the substrate, the basal, fruticose growth of this specimen is easily overlooked. In addition, although the material has smaller, slightly narrower lobes, only careful examination of the cortex anatomy will reveal the distinct diagnostic differences of Placomaronea: the paraplectenchymatous cortex, apically with ‘peppered pigment hoods’, covered by a thin colourless coating sloughing off.

Phylogeny

Our trees are generally similar to those of Candelariaceae previously published (e.g. Westberg et al. Reference Westberg, Arup and Kärnefelt2007, Reference Westberg, Frödén and Wedin2009; Kondratyuk et al. Reference Kondratyuk, Lőkös, Jeong, Oh, Kondratyuk and Hur2020; Halıcı et al. Reference Halıcı, Kahraman Yiğit, Bölükbaşı and Güllü2023). All of the studies agree in several broad aspects: Candelariella blastidiata Yakovch. and C. subdeflexa (Nyl.) Lettau form a well-supported, monophyletic clade basal to the rest of the family (100% BS). Kondratyuk et al. (Reference Kondratyuk, Lőkös, Jeong, Oh, Kondratyuk and Hur2020) named this clade Protocandelariella. Also, Candelina and Placomaronea each form well-supported, monophyletic clades (99% and 100% BS, respectively). Polyspory must have evolved independently multiple times. Our phylogeny suggests that this occurred at least four times: three times in Candelaria and Candelariella s. lat., once when Placomaronea emerged.

Our phylogeny also agrees with Westberg et al. (Reference Westberg, Frödén and Wedin2009) in so far as the uniquely characteristic aspects of the cortex anatomy in Placomaronea must have evolved independently in Candelariella kansuensis: both are characterized by a distinctly paraplectenchymatous cortex with an outer layer of cells with granular-pigmented hoods. It remains unclear, however, whether the thick hyaline coating observed in C. kansuensis is analogous to what appears to be a residue of cells in Placomaronea.

Although several clades in our analysis received strong bootstrap support, the relationships between clades (and within clades) remains unresolved. Most clades formed a massive polytomy in the middle of the tree. The relationships between the clades in this polytomy were sensitive to the alignment and which particular ITS sequences we chose to include. For example, we were unable to determine with any support which clade might be sister to Placomaronea. Westberg et al. (Reference Westberg, Frödén and Wedin2009) placed it sister to Candelariella complanata, and Westberg et al. (Reference Westberg, Morse and Wedin2011) placed it in a polytomy similar to ours but with a few small changes: they included Candelariella aurella in the clade with Candelaria s. str., Candelariella s. str. and C. antennaria + C. aggregata (we found the position of the C. aurella + C. plumbea group to be unstable and left it in the polytomy), and they separated out a clade with C. corviniscalensis C.A. Morse & M. Westb., C. kansuensis and C. rosulans in a position basal to the polytomy (some versions of our tree had weak support for a similar basal clade which also included C. clarkii and C. spraguei (Tuck.) Zahlbr., but other versions did not, and so the conservative tree we present here leaves these three species unresolved).

Candelariella thus remains paraphyletic, and relationships between various clades of Candelaria, Candelariella, Candelina and Placomaronea remain for the most part unresolved. Notably, we were unable to confirm support for two of the three genera proposed by Kondratyuk et al. (Reference Kondratyuk, Lőkös, Jeong, Oh, Kondratyuk and Hur2020), namely Candelinella and Opeltiella. For example, one of the sequences available for Opeltiella fruticans (Poelt & Oberw.) S.Y. Kondr. (EF535207) appears to be nested within the same clade as Candelinella deppeanae (M. Westb.) S.Y. Kondr. and C. makarevichiae (S.Y. Kondr. et al.) S.Y. Kondr.

The phylogenies published so far (Westberg et al. Reference Westberg, Arup and Kärnefelt2007, Reference Westberg, Frödén and Wedin2009; Kondratyuk et al. Reference Kondratyuk, Lőkös, Jeong, Oh, Kondratyuk and Hur2020; Halıcı et al. Reference Halıcı, Kahraman Yiğit, Bölükbaşı and Güllü2023), although similar overall, nevertheless do not exactly correspond in their topologies, even though individual clades are supported by high bootstrap values. It is difficult to assess what causes these discrepancies. We had hoped that improving clade resolution and overall tree topology might be possible by adding additional loci. Unfortunately, all eight sequences newly generated by Kondratyuk et al. (Reference Kondratyuk, Lőkös, Jeong, Oh, Kondratyuk and Hur2020) remain unavailable. More importantly, of the sequences available in GenBank for different loci (303 ITS, 24 mtSSU and 21 nuLSU), not a single specimen had all three loci sampled. Constructing a multilocus ‘consensus’ tree based on such sparse and poorly overlapping data does not seem warranted, especially since some of the sequences appear to be of low quality. For example, a BLAST search with the mtSSU sequence of Candelariella terrigena (DQ986884) retrieves Porina. In addition, the mtSSU sequence for Candelaria concolor (MN508267) is missing numerous base pairs in nucleotide regions which are otherwise well conserved for other mtSSU sequences of Candelariaceae. This suggests that attempts to accurately resolve the overall phylogeny for the family remain premature without adding additional sequences from a broader array of loci. Whether any of the new genera proposed by Kondratyuk et al. (Reference Kondratyuk, Lőkös, Jeong, Oh, Kondratyuk and Hur2020) deserve recognition cannot be assessed without access to the specimens, sequences and alignments used in that study.

With regard to the species newly described here, our phylogenetic analysis suggests that they form a well-supported, monophyletic clade within Placomaronea. However, distinguishing the two from P. mendozae or recognizing P. fruticosa and P. placoidea as separate species is, unfortunately, not supported by the current topology. Not only are the two species part of a clade that includes P. mendozae, but P. fruticosa is also nested inside a clade that includes both P. placoidea and P. mendozae. Based on the molecular data, the hypothesis that the three taxa might be conspecific cannot therefore be rejected. Our phylogeny instead suggests that the conspicuously different growths forms could be interpreted as environmental modifications of a single, morphologically extremely plastic species: P. mendozae.

However, evolutionary processes such as introgression or incomplete lineage sorting could also explain this discrepancy, particularly in rapidly evolving lineages (Tremble et al. Reference Tremble, Suz and Dentinger2019). These processes are known to result in polyphyly within trees that are based on the analysis of a single gene only, even after speciation (Maddison & Knowles Reference Maddison and Knowles2006). Using a small number of, or even single, loci to construct a phylogeny can result in a topology where the gene tree is incongruent with the overall species tree (Górniak et al. Reference Górniak, Szlachetko, Olędrzyńska, Naczk, Mieszkowska, Boss and Ziętara2021). Such phenomena ultimately can be resolved only by expanding the scope of the molecular analysis to include additional loci and additional specimens, yielding a consensus tree which more accurately reflects the overall phylogeny of the species (Maddison & Knowles Reference Maddison and Knowles2006). Within a single gene tree, polymorphisms from an ancestor may persist in a new lineage for some time, even after species divergence (Maddison Reference Maddison1997). Relying on a single gene to construct a phylogeny, the parent population may then appear to be paraphyletic with respect to its descendants (Carstens & Knowles Reference Carstens and Knowles2007). Monophyly of the single gene tree may eventually be restored via lineage sorting and extinction, but typically only after sufficient time has passed. It has been argued that this depends on generation time and effective population size of the species involved (Knowles & Carstens Reference Knowles and Carstens2007). In such a situation, accurately inferring the species tree by molecular phylogenetic reconstruction requires increased sampling, adding both individuals and additional loci (Maddison & Knowles Reference Maddison and Knowles2006). It is not inconceivable that some of these processes are also responsible for the phylogeny inferred here.

Our analysis, where a phylogeny is based solely on ITS but is also conspicuously incongruent with distinct anatomical and morphological characters, suggests that ITS may not offer sufficient information to accurately resolve the correct phylogeny of these species. Even though nrITS is now widely accepted as the universal barcoding region for fungi (Nagy et al. Reference Nagy, Kocsubé, Csanádi, Kovács, Petkovits, Vágvölgyi and Papp2012; Schoch et al. Reference Schoch, Seifert, Huhndorf, Robert, Spouge, Levesque, Chen and Consortium2012; Xu Reference Xu2016), its effectiveness in reliably distinguishing some fungal lineages is increasingly being questioned (Badotti et al. Reference Badotti, de Oliveira, Garcia, Vaz, Fonseca, Nahum, Oliveira and Góes-Neto2017; Parks et al. Reference Parks, Kurylo, Batchelder, Vincent and Blanchard2019). Some authors suggest that this insufficient resolution can indeed be a result of incomplete lineage sorting or hybridization (Pino-Bodas et al. Reference Pino-Bodas, Martin, Burgaz and Lumbsch2013; Tremble et al. Reference Tremble, Suz and Dentinger2019).

Lack of sufficient ITS resolution has been documented for a variety of fungal lineages (O’Donnell & Cigelnik Reference O’Donnell and Cigelnik1997; Nilsson et al. Reference Nilsson, Kristiansson, Ryberg, Hallenberg and Larsson2008; Thiery et al. Reference Thiery, Vasar, Jairus, Davison, Roux, Kivistik, Metspalu, Milani, Saks and Moora2016; Hughes et al. Reference Hughes, Tulloss and Petersen2018; Kruse et al. Reference Kruse, Piątek, Lutz and Thines2018a, Reference Kruse, Dietrich, Zimmermann, Klenke, Richter, Richter and Thinesb ; Tremble et al. Reference Tremble, Suz and Dentinger2019; Stadler et al. Reference Stadler, Lambert, Wibberg, Kalinowski, Cox, Kolařík and Kuhnert2020), including some that are lichenized (Pino-Bodas et al. Reference Pino-Bodas, Martin, Burgaz and Lumbsch2013; Moncada et al. Reference Moncada, Reidy and Lücking2014; Steinová et al. Reference Steinová, Košuthová and Škaloud2022; Di Meglio & Goward Reference Di Meglio and Goward2023). It has been suggested that, even within groups of lichenized fungi where ITS generally offers good resolution (e.g. Magain & Sérusiaux Reference Magain and Sérusiaux2015; Moncada et al. Reference Moncada, Sipman and Lücking2020), augmenting the data with secondary barcodes will sometimes be necessary, particularly when dealing with rapidly evolving lineages (Lücking et al. Reference Lücking, Aime, Robbertse, Miller, Ariyawansa, Aoki, Cardinali, Crous, Druzhinina and Geiser2020).

Morphology

Despite inconclusive molecular data, accepting Placomaronea mendozae as one single, highly polymorphic species seems untenable. Both P. fruticosa and P. placoidea are characterized by a thallus morphology conspicuously different from P. mendozae.

It could be argued, of course, that the two new species are simply environmentally induced morphotypes. Countless examples of dimorphic lichens are known where a basal crustose or squamulose primary thallus gives rise to a secondary fruticose thallus. However, these transitions are easily recognized when studying specimens and no such transitional morphotypes could be observed here: Placomaronea mendozae is composed of minute, scattered subsquamulose to barely squamulose, broadly attached, gomphate areoles. Even exuberant specimens are barely squamulose, not closely aggregated, the squamules broadly attached to their substrate. The thalli are not hollow like the placodioid lobes of P. placoidea, and unlike P. fruticosa, subsquamulose areoles of P. mendozae do not show any tendency to become fruticose.

Crustose, placodioid lichens transitioning into coralloid fruticose lichens have been described from distantly related species in Caloplaca s. lat. Along the Pacific coast of Baja California and Baja California Sur in Mexico, Polycauliona thamnodes (Poelt) Arup et al. forms placodioid thalli in less exposed microhabitats, but transitions into distinctly coralloid fruticose outgrowths at more exposed sites. In the Atacama of southern Peru and northern Chile, Follmannia orthoclada (Zahlbr.) Frödén et al. also transitions from distinctly placodioid into entirely fruticose specimens. The crustose thalli are found at the most sheltered sites, and the fruticose ones at the most exposed habitat sites. Obviously, the morphology of these two not closely related species is highly plastic; it can be interpreted as fog-induced and is clearly an adaptation of the morphology to microhabitat variation at the sites where these species grow.

Interpreting the different morphologies of P. fruticosa and P. placoidea as also being environmentally induced, however, ignores how these thalli presumably arose. Both species have been found in ecologically similar areas; Placomaronea placoidea grows on the surface of rocks embedded in soil, and these surrounding soils are inhabited by thalli of P. fruticosa, which develop as an adaptation to the soil in which the species grows. At these sites the loamy soil is hardened and compact and if the lichen thalli were not ‘rooted’ deeply inside their substrate these lichens could not compete with grasses. Thalli of P. placoidea growing nearby obviously cannot penetrate the rock they grow on and any tendencies to develop a fruticose growth form would necessarily result in erect thalli emerging from the rock surface. Such thalli have, however, not been observed. Transitional forms, where lobes of P. fruticosa creep onto rocks from the surrounding soil, have also not been found. Both species thus maintain their distinct morphologies, even when growing side by side. Both also maintain distinct anatomical differences. The lobes of P. placoidea consistently have a hollow medulla, whereas the fruticose stalks of P. fruticosa as well as its placodioid lobes on the substrate surface are solid.