Psoroma capense and P. esterhuyseniae (Pannariaceae), two new alpine species from South Africa

Abstract The new species Psoroma capense and P. esterhuyseniae are described from four alpine localities in the Western Cape Province of South Africa and are the only known Psoroma species from Africa. The specimens were all collected from moist sites near watercourses, on cool and mostly south-facing cliffs. Psoroma capense resembles P. tenue in gross morphology but differs in the ascending thallus squamules, lack of secondary compounds and short-ellipsoid to ovoid ascospores. However, a phylogenetic analysis involving the markers ITS, nucLSU, mtSSU and Mcm7, comparing the only recent collection of P. capense with previously published sequences, shows that it belongs to the P. hypnorum lineage, with no known, closely related species. Psoroma esterhuyseniae resembles P. hypnorum but has subglobose to short-ellipsoid ascospores without apical perispore extensions. The two species are thought to have evolved from one or two long-distance dispersal events during the Pleistocene.


Introduction
Previously, the genus Psoroma was interpreted widely, and included almost all tripartite Pannariaceae species (Jørgensen & Galloway 1992), with green algae as major photobionts, and cyanobacteria located in smaller cephalodia. However, many foliose tripartite species were later transferred to the genera Pannaria (see review in Elvebakk & Elix (2017)) and the newly described Gibbosporina . Some squamulose species were transferred to the new genus Joergensenia (Passo et al. 2008) and to Psorophorus and Xanthopsoroma (Elvebakk et al. 2010), whereas six bipartite, squamulose species were transferred from Pannaria and Santessoniella to Psoroma (Ekman et al. 2014). After these revisions, the genus Psoroma became much more homogeneous. Its species have squamules, densely distributed or more scattered, and connected by a distinct or indistinct hypothallus/prothallus. In most species, the squamules are brown from melanins and lack secondary compounds that can be detected by TLC, with the presence of pannaric acid and porphyrilic acid or porphyrilic acid methyl ester as the most common exceptions.
The lichen genus Psoroma is mostly terricolous, occasionally corticolous in austral forests, rarely saxicolous, and is concentrated in the Southern Hemisphere, with most of its species distributed in southern South America, south-eastern Australia, and New Zealand (Galloway 2007). Park et al. (2018) showed that no less than 10 species are known from Antarctica and also listed the four species reaching the Northern Hemisphere. Since then, two rare species have been reported from Alaska and arctic Canada (Elvebakk & Tønsberg 2018;Fryday et al. 2019).
When Jørgensen (2003) reviewed the Pannariaceae flora of the African continent, he included Psoroma asperellum Nyl. and P. fruticulosum P. James & Henssen, both from South Africa. The report of the former from 'Promontorio Bonae Spei ' by Nylander (1863) was later shown to be a misinterpretation of 'Montis Tabularis', an old name for Mt Wellington in Tasmania (Galloway 2007). An unpublished record from South Africa of Psoroma hypnorum (Vahl) S.F. Gray has also been posted in the GBIF database. A report from South Africa of the austral species 'Psoroma sphinctrinum Nyl.' by van der Byl (1931: 9) and cited by Doidge (1950), possibly refers to material of the tropical genus Gibbosporina.
During studies of Pannariaceae in various herbaria, no African material of Psoroma has been discovered, except the two Esterhuysen collections from BG determined as P. fruticulosum and P. hypnorum, and a third Esterhuysen sample borrowed from BOL. The three Esterhuysen specimens were collected between 1943 and 1951, and a fourth specimen was collected by T. Rämä in 2018. The aim of the present study is therefore to describe the species, but also to search for related species through phylogenetic analyses. This was carried out by comparing the recent collection with other Psoroma species using four phylogenetic markers.

Lichen material
Herbarium materials used for this study are housed at BG and BOL, and the species was not found during extensive studies of the Pannariaceae collections in herbaria such as B, BM, C, CANB, O, S, SGO, UPS and W. In microscope sections, iodine reactions were tested by adding IKI to mounts pretreated with KOH (Orange et al. 2001). Perispore structures were studied in water mounts and restricted to spores liberated from the asci. Ascospore morphology was studied in detail by drawing detailed sketches of c. 80 ascospores and copies of all original drawings have been included with the specimens. Several specimens of other species were studied specifically for comparison. Thin-layer chromatography of acetone extracts followed standardized procedures and used solvents A and C (Culberson 1972;Orange et al. 2001). Nomenclature of ascospore structures follows Nordin (1997).

Phylogenetic analyses
In order to determine the phylogenetic position of the undescribed species from South Africa, the phylogenetic relationships of 12 species of the genera Psoroma, Psorophorus, Xanthoprosoma and Pannaria were reconstructed. Protopannaria pezizoides (G. H. Web.) P. M. Jørg. & S. Ekman was used as an outgroup. The reference materials were selected from those used in a previous study (Park et al. 2018). Four phylogenetic markers, 5.8S-ITS2 rRNA (ITS), the nuclear large subunit rRNA (nucLSU), the mitochondrial small subunit rRNA (mtSSU) and minichromosome maintenance component 7 (Mcm7), were used for phylogenetic reconstruction. Sequence information for ITS, nucLSU and mtSSU of the reference materials was retrieved from a previous study (Park et al. 2018). Sequence information for ITS, nucLSU and mtSSU of the new material was obtained following procedures described by Park et al. (2018). Mcm7 was amplified using the primers mcm7-709for and mcm7-1348rev (Schmitt et al. 2009). Touchdown PCR amplifications were performed in a T-gradient thermocycler (Biometra, Göttingen, Germany) with the following cycling parameters: 1 min initial denaturation at 95°C, 6 touchdown cycles of 30 s denaturation at 95°C, 50 s annealing at 60-56°C at the ramp of 1°per cycle and 1 min extension at 72°C, followed by 38 cycles of 45 s denaturation at 94°C, 50 s annealing at 56°C, and 1 min extension at 72°C, with a 5 min final extension at 72°C. The new sequences, including the holotype of Psoroma capense (cited as NK-1080 in Fig. 5 and Supplementary Material Table S1, available online) and the additional Mcm7 sequences of the samples analyzed previously were deposited in the GenBank database under the accession numbers MT316196 to MT316208 (Supplementary Material  Table S1).
Sequence alignments of ITS, nucLSU, mtSSU and Mcm7 were conducted using the software ClustalX (Larkin et al. 2007) and manually adjusted. The size variation and ambiguous alignment of the ITS1 domain resulted in it being excluded from the phylogenetic analyses, as were other ambiguously aligned sites. Phylogenetic trees were inferred from each genetic locus and the combined dataset by maximum parsimony (MP), maximum likelihood (ML), and Bayesian analyses. MP trees were obtained using the Tree-Bisection-Regrafting (TBR) algorithm of MEGA X (Kumar et al. 2018) with search level 5 in which the initial trees were obtained by the random addition of sequences (1000 replicates). ML trees were constructed using MEGA X based on the GTR + I + G evolutionary model (Lanave et al. 1984), the search options of best tree topology finding by branch swapping of NNIs and SPRs, and random addition of sequences (1000 replicates). Aligned sites with less than 95% coverage by alignment gaps, missing data, or ambiguous bases were excluded. The Bayesian tree was generated using a search approach by MrBayes ver. 3.2. (Ronquist et al. 2012) with the GTR + I + G model. Two parallel Markov chain Monte Carlo (MCMC) runs were performed for 1 000 000 cycles, each with one cold and three heated chains and the temperature parameter set to 0.1; trees were sampled every 100 generations. A consensus tree was calculated after discarding the first 25% of trees as burn-in.
Chemistry. Brownish melanins present, of a chestnut-coloured type, appearing similar to those of P. tenue Henssen. No TLC-detectable components found.
Etymology. Named after its occurrence in the Western Cape Region. Distribution and ecology. Only known as three collections from the Western Cape Province of South Africa, found at moist sites at high altitudes in cool, mostly south-facing sites, often near water.

MycoBank No.: MB 836050
Similar to Psoroma hypnorum but ascospores subglobose to shortellipsoid without apical perispore extensions, and tomentum of the lower parts of thalline excipuli less prominent.
Chemistry. Brownish melanins present, appearing similar to those of P. hypnorum. No TLC-detectable components found.
Etymology. Named after the South African botanist Elsie Elizabeth Esterhuysen , who collected three of the four samples of Psoroma known from South Africa. Distribution and ecology. Known only from the holotype collected from damp cliffs in Western Cape, South Africa.

Molecular analysis and phylogeny
The phylogeny based on the concatenated multi-locus dataset of ITS, nucLSU, mtSSU and Mcm7 indicates that Psoroma capense forms a well-supported monophyletic group (referred to here as the Psoroma hypnorum lineage) with P. antarcticum Hong & Elvebakk, P. buchananii (Knight) Nyl., P. fruticulosum, P. hypnorum and P. paleaceum (Fr.) Timdal & Tønsberg (Fig. 5). The monophyletic group was consistently recovered by MP, ML and Bayesian methods, and also based on single-locus analyses (data not shown). The group was clearly separated from the Psoroma tenue lineage, including P. cinnamomeum Malme and P. tenue, and from the genera Psorophorus, Xanthopsoroma and Pannaria. The phylogenetic position of P. capense within the Psoroma hypnorum lineage was not clearly resolved and the relationship was poorly supported by bootstrap and posterior probability. Psoroma capense was grouped with P. buchananii, P. fruticulosum and P. paleaceum in the ML tree based on the combined dataset (Fig. 5), but the relationship was not always recovered by MP, ML, and Bayesian methods with single-locus datasets. Branch lengths from the common ancestor of the group leading to terminal taxa were generally very short and statistical support for bifurcation was generally very low. Sequence similarity of the combined dataset between P. capense and the other species of the Psoroma hypnorum lineage ranged between 97 and 98%, which is close to similarity values among the other species of the group.

Discussion
In recent phylograms, the genus Psoroma has either appeared as polyphyletic (Ekman et al. 2014) or paraphyletic (Park et al. 2018; the present study), with species of the P. hypnorum and P. tenue groups forming separate lineages. The possible recognition of these two lineages as separate genera has not been proposed due to insufficient taxon sampling. The P. hypnorum and P. tenue lineages both clearly have evolutionary histories featuring adaptations to cold climates, probably initiated in or near Antarctica, where glaciation occurred at c. 34 Ma (Pollard & DeConto 2020). Data from thermophilous Psoroma species from austral forests should be incorporated in future phylogenies, since they are potential members of older lineages needed in analyses to define the genus. From its gross morphology alone, Psoroma capense resembles P. tenue. The latter species is distributed in Antarctica and subantarctic areas, but also in the Northern Hemisphere by a taxon considered to represent a separate variety (Henssen & Renner 1981;Jørgensen 2004b), a concept which has recently been challenged by Marthinsen et al. (2019). Psoroma capense and P. tenue share a related melanin colour and strongly subsessile apothecia with crenate-lobate margins, but without the excipulum squamules typical of P. hypnorum. However, P. capense differs from P. tenue by ascending thallus squamules, short-ellipsoid to ovoid ascospores and a lack of TLC-detectable compounds.
Psoroma esterhuyseniae resembles P. hypnorum, although the characteristic regular tomentum on the apothecia of P. hypnorum (see Elvebakk & Tønsberg 2018) is lacking; replaced by some mycelium-like cover in only the least exposed apothecia. The few cephalodia seen in P. esterhuyseniae are regularly coarsely coralloid, whereas they are irregular in P. hypnorum. The ascospores of P. esterhuyseniae are very different from those of both P. hypnorum and P. capense (Fig. 2C), in being subglobose to short-ellipsoid, and very rarely ovoid. The apical perispore extensions present in both these species are absent in P. esterhuyseniae.
Phylogenetically, P. capense is very distinct from P. tenue, and is instead positioned within the Psoroma hypnorum lineage based on a concatenated dataset of ITS, nucLSU, mtSSU and Mcm7 sequences. This is very well supported by all phylogenetic methods and by all datasets examined in the present study. Within the P. hypnorum lineage, P. capense is in a poorly supported sister group position to a clade including P. paleaceum, P. fruticulosum and P. buchananii. None of these have any resemblance to P. capense. The former has characteristic long scales along apothecium margins (Elvebakk & Tønsberg 2018) and the two latter species were previously considered to form a subgroup within Psoroma by Henssen et al. (1983), a conclusion confirmed by our ongoing studies, as well as by the present phylogram. The Esterhuysen 19947 specimen (erroneously cited by Jørgensen (2003) as Esterhuysen 9419) was determined as P. fruticulosum because of its 'erect, isidioid lobules, which are partly flattened' (Jørgensen 2003), a character resembling P. capense. However, both P. fruticulosum and P. buchananii have conspicuous black pycnidia, prominent apothecia almost appearing stipitate, and spores deviating from those of the remaining Psoroma species.
Subantarctic islands of the Indian Ocean are the Psoroma sites closest to the distribution area of P. capense and P. esterhuyseniae. These areas house endemic species such as Psoroma absconditum Øvstedal and P. xanthorioides (P. M. Jørg.) P. M. Jørg., and represent the major distribution area of P. dichroum (Hooker f. & Taylor) P. M. Jørg. (Jørgensen 2000(Jørgensen , 2004cØvstedal & Gremmen 2008;Ekman et al. 2014). Psoroma absconditum is the most similar to the South African species; however, it is not well understood since it was not compared to other members of the P. hypnorum lineage, but instead to the very different species P. asperellum Nyl. (Øvstedal & Gremmen 2008). Psoroma esterhuyseniae has shorter spores and apothecia with squamulose margins compared to P. absconditum, the latter named after its sunken apothecia, partly hidden by squamules.
Fresh material of P. capense appeared to have a yellowish brown melanin colour where the pigments were not strongly concentrated, which in combination with the chlorobiont cells gave the lichen a peculiar 'grass green' colour, even in a dried specimen two years after collection. A similar colour, contrasting with most other Psoroma species, has been observed in fresh specimens of the New Zealand species P. cyanosorediatum P. M. Jørg. (A. Elvebakk, unpublished data), which has very different, long and narrowly ellipsoid ascospores according to Jørgensen (2004a). Among the rather few Psoroma species described with short ascospores, P. antarcticum Elvebakk & S. G. Hong, P. saccharatum Scutari & Calvelo and P. pannarioides Henssen lack other similarities with the two new species from South Africa (Henssen 1983;Scutari & Calvelo 1995;Park et al. 2018).
All the specimens of the new species were collected from moist and S-facing sites at altitudes between 1300 and 1800 m at Sneeugatpiek, Waaihoekpiek, and Hexrivierberge in the Western Cape Province, only 140-200 km NNE of Cape Town. A search for the species by TR in Spekrivierskloof, on 18 March and 17 November in 2018, revealed no additional findings. At higher altitudes in this area the habitats were drier, and the species might be truly rare here due to a scarcity of moist, suitable habitats. There is a clear need to search in the Western Cape for more populations of these species, which appear as Red List candidates, and to determine if more species are present in South African mountains. South Africa features extreme speciation in many groups of organisms, for example in the plant genus Erica which has evolved no less than 690 endemic species in the Cape Region during the last 15 million years (Pirie et al. 2016). Chlorobiont acquisition is an important evolutionary feature in lichens but has not yet been studied in Pannariaceae. Previously, Myrmecia was the most commonly identified chlorobiont in Psoroma, but recent studies instead identify it as Trebouxia Muggia et al. 2018). In Psoroma esterhuyseniae the chlorobiont has cells with angular chloroplasts, whereas they are differently shaped and papillose in P. capense.
In conclusion, there are currently no candidates closely related to P. capense and P. esterhuyseniae, from phylogenetic analysis or by comparison of taxonomic characters. In this context, it should be added that a high proportion of specimens collected throughout the distribution area of Psoroma represent misunderstood or undescribed species (A. Elvebakk & S. G. Hong, unpublished data). For these reasons, it is difficult to hypothesize on the migration history of the ancestors of these two species into Africa, where they are the only known members of the genus Psoroma.
The bipolar element within the genus is most easily explained by migrations along American mountain chains during the Pleistocene; several Psoroma species occur in the Central Andes (Jørgensen & Palice 2010) where rapid diversification has taken place during this period in Lobariaceae lichens, as shown by Widhelm et al. (2019). The dramatic cooling during the Pleistocene probably represented a scenario of expansion of the cold-adapted groups within Psoroma, and our hypothesis is that P. esterhuyseniae and P. capense obtained their isolated and shared geographical positions as a result of one or two longdistance dispersal events during this period.
Acknowledgements. Curators of the herbaria mentioned are acknowledged for making herbarium material available for study, including T. Trinder-Smith at BOL during 2018. An anonymous referee and Associate Editor T. Randlane provided comments which greatly improved the manuscript. Photographs were taken by M. Karlstad at UiTthe Arctic University of Norway. This Black thick branches indicate those that were conserved in maximum likelihood (ML) and maximum parsimony (MP). Grey thick branches indicate those that were conserved in ML or MP. Asterisks indicate that branches were conserved but not supported by high bootstrap values. Bayesian posterior probabilities (PP ≥ 0.90) and bootstrap values in ML and MP trees (≥ 80%) are indicated above or below the nearest branches (PP/ ML/MP). The geographical origins of the specimens are shown after the voucher number. KGI = King George Island, Antarctica; P. tenue l. = P. tenue lineage.