Integrative taxonomy confirms three species of Coniocarpon (Arthoniaceae) in Norway

Abstract We have studied the highly oceanic genus Coniocarpon in Norway. Our aim has been to delimit species of Coniocarpon in Norway based on an integrative taxonomic approach. The material studied comprises 120 specimens of Coniocarpon, obtained through recent collecting efforts (2017 and 2018) or received from major fungaria in Denmark, Finland, Norway and Sweden, as well as from private collectors. We have assessed (1) species delimitations and relationships based on Bayesian and maximum likelihood phylogenetic analyses of three genetic markers (mtSSU, nucITS and RPB2), (2) morphology and anatomy using standard light microscopy, and (3) secondary lichen chemistry using high-performance thin-layer chromatography. The results show three genetically distinct lineages of Coniocarpon, representing C. cinnabarinum, C. fallax and C. cuspidans comb. nov. The latter was originally described as Arthonia cinnabarina f. cuspidans and is herein raised to species level. All three species are supported by morphological, anatomical and chemical data.


Introduction
Species recognition is crucial for improved natural resource management and biodiversity conservation (Lumbsch and Leavitt 2011). Species recognizing can be challenging in lichenized fungi due to unclear species boundaries and/or cryptic diversity

Collection methods and handling of fresh specimens
New specimens of Coniocarpon for this study were collected in boreo-nemoral rainforests on the west coast of Norway from Vest-Agder to Møre og Romsdal in 2017 and 2018. Specimens were placed in paper bags, allowed to air-dry and later stored at -20 °C due to prior knowledge of fast DNA degradation (e.g., Frisch et al. 2014). After DNA extraction, specimens were incorporated for long-term storage and access in the fungarium in Trondheim (TRH). Additional specimens were made available from fungaria in Bergen (BG), London (BM), Copenhagen (C), Edinburgh (E), Helsinki (H), Oslo (O), Paris (PC), Prag (PRA), Stockholm (S), Trondheim (TRH) and Uppsala (UPS).

Taxon sampling
In total, 26 specimens were used for the molecular data production; 25 specimens from Norway and 1 specimen from Great Britain. Outgroup taxa and additional sequences of Coniocarpon were downloaded from GenBank, in total 32 sequences. Eighteen of these (nine mtSSU and nine RPB2) represent the nine outgroup taxa, whereas 14 (eight mtSSU and six RPB2) were from eight specimens of Coniocarpon originating from Great Britain, Japan, Norway, Rwanda and Uganda. Outgroup taxa were selected based on their phylogenetic position in Frisch et al. (2014) and Van den Broeck et al. (2018).

DNA extraction and sequencing
DNA was isolated from specimens up to one year old. Genomic DNA was extracted following one of three methods. (1) Five to eight ascomata were sampled in 2 ml microcentrifuge tubes with two 3 mm diam. tungsten carbide beads each and crushed into a fine powder using a Retsch TissueLyser II. Subsequently, genomic DNA was extracted using the E.Z.N.A. SP Plant DNA Kit (Omega BIO-TEK, USA) following the manufacturer's instructions. (2) Three to five ascomata were sampled directly in 0.2 ml Eppendorf PCR Tubes with 30 µl Dilution Buffer (Phire Plant Direct PCR Kit, ThermoFisher Scientific, Lithuania) and crushed with tweezers. (3) Small cuttings (ca. 50-100 µm × 50-100 µm in size) of the hymenium were sampled in 0.2 ml Eppendorf PCR Tubes and directly used for PCR amplification. The Phire Plant Direct PCR Kit (ThermoFisher Scientific, Lithuania) was used for PCR amplification in all three methods. Each PCR reaction contained 10 µl 2× Phire Plant PCR Buffer, 0.4 µl Phire Hot Start II DNA Polymerase, 1 µl of each primer for all genetic markers except RPB2 (1.5 µl of each primer), 1 µl genomic DNA (1:1) or the lichen sample, and was filled with H 2 O to the final volume of 20 µl. If PCR amplification resulted in weak products, 2 µl genomic DNA (1:10) was added. PCR amplification was done for the mtSSU, nucITS and the protein-coding gene RPB2 with the following primers: mtSSU1 + mtSSU3R (Zoller et al. 1999), ITS-1F + ITS4 (Larena et al. 1999, White et al. 1990) and RPB2-7cF + RPB2-11aR (Liu et al. 1999), respectively. PCR cycling conditions for mtSSU and nucITS started with an initial denaturation at 98 °C for 5 min, followed by 40 cycles of 98 °C for 5 s, 59 °C for 5 s, and 72 °C for 30 s, followed by a final extension of 72 °C for 1 min. For RPB2, annealing temperature was set to 57 °C. The PCR products were visualized on a 1% agarose gel stained with SYBR Safe DNA gel stain (ThermoFisher Scientific, USA) under UV light. Clean PCR products (i.e., those lacking visible contamination) were purified by adding 5 µl ExoSAP-IT Express PCR Cleanup (1:3 concentration; ThermoFisher Scientific, United Kingdom) to the PCR reactions. PCR reactions resulting in more than a single product were purified using the E.Z.N.A. Gel Extraction Kit (Omega BIO-TEK, USA) following the manufacturer's instructions, except that we performed an additional wash buffer step. The PCR products were sent to Eurofins Genomics (Germany) for Sanger sequencing using the same primers as for the PCR reactions.

Sequence alignment and phylogenetic analyses
The sequences were edited and aligned using BioEdit v.7.0.5.3 (Hall 1999). The identity of the sequences was verified using the nucleotide BLAST search in GenBank. For the examination of topological incongruence among gene trees, maximum likelihood (ML) bootstrapping analyses was carried out on each of the data sets using the RAxML-HPC Blackbox ver. 8.2.10 (Stamatakis 2014). The standard settings without estimation of the proportion of invariable sites (GTRGAMMA) were used. Topological incongruence was assumed if conflicting tree topologies were supported by ≥ 70% bootstrap support. Since topological incongruence could not be observed, maximum likelihood (ML) bootstrapping analysis was carried out on the concatenated three-locus dataset of 43 accessions for Coniocarpon using default settings and adding single genes as parameter partition. The RAxML analysis was stopped automatically after 402 bootstrap replicates using the MRE-based bootstopping criterion (Pattengale et al. 2010).

Morphological and chemical investigations
The morphology of 120 specimens of Coniocarpon (Norway 87, Sweden 8, Denmark 16, Great Britain 7, Austria 1, Turkey 1) was studied. The morphology of all specimens was examined using a Leica M80 stereomicroscope and a Zeiss Standard Binocular microscope. Macroscopic photographs were taken with a Leica MZ16A stereomicroscope fitted with a Leica DFC420 camera. Microscopic photographs were taken with a Leica CTR6000 microscope fitted with a Leica DFC365 camera. Sections of ascomata and lichen thalli were cut by hand and mounted in water or lactic acid cotton blue (LCB). Length and width were measured for single ascomata as well as for aggregations composed of several ascomata. For the epithecium, exciple, hymenium and hypothecium, measurements were performed in LCB. Measurements of asci and ascospores were performed in water using squashed preparations. Only fully developed ascospores and commonly asci containing mature ascospores (sometimes asci without mature ascospores) were measured. Ascospore measurements are presented as (min.-)mean ± SD(-max.). The amyloid reaction of the apothecia was tested using 0.2% (Iodine diluted ) and 1% (I), and 1% (I) solution after pretreatment with 10% potassium hydroxide (KOH) in water (KI). The quinoid pigments and Ca-oxalate crystals were measured in water and their shape studied. Later, the crystals reaction with KOH was observed. The quinoid pigments were identified by HPTLC (Arup et al. 1993) in solvent C. Quionid pigments are named according to Frisch et al. (2018) except for the newly identified A4.

Distribution maps
Based on occurrence information of all revised specimens, the distribution of Coniocarpon species in Scandinavia was illustrated by adding a delimited text layer to a Wikimedia map from QuickMap services in QGIS ver. 3.6.2. (QGIS Development Team 2019). The younger specimens (i.e., collected from the mid-80s and onwards) were placed on the maps by their geographical coordinates, while older specimens (collected 1870-1983) lacking geographical coordinates were placed on the maps based on locality information. The Earth Point coordinate converter (http://www.earthpoint.us/ Convert.aspx) was used to convert coordinates.

Phylogeny
The phylogeny based on Bayesian and maximum likelihood analyses present Coniocarpon as monophyletic using the selected taxon and outgroup sampling. Three discrete, well-supported lineages are recovered within Coniocarpon. These are separated from each other by branches clearly exceeding the observed infraspecific branch-lengths. The three lineages represent C. cinnabarinum, C. cuspidans (Nyl.) Moen, Frisch and Grube and C. fallax (Fig. 1). Coniocarpon cinnabarinum is the supported sister taxon to C. cuspidans, while C. fallax is sister to these two taxa.
Coniocarpon cinnabarinum from Rwanda and Uganda form a well-supported clade and are sisters to C. cinnabarinum in Norway, while C. cinnabarinum from Japan is genetically distinct and sister to the remaining taxa of Coniocarpon. The two sampled specimens from Great Britain are genetically close to C. cuspidans and C. fallax in Norway, respectively.

Morphology and chemical characters
Forty-four specimens were identified as C. cinnabarinum, as C. cuspidans and 42 as C. fallax. Ascospore size (Fig. 2), ascospore septation ( Fig. 3), ascoma shape and the distribution of pruina ( Fig. 4) were identified as useful characters for species distinction. Moreover, differences were observed in the quinoid pigment patterns revealed by HPTLC (Fig. 5) and in the amyloidity of the ascomatal gels. Four quinoid pigments were identified showing different colors on the HPTLC plates prior to sulphuric acid treatment and charring; reddish (A1, A4), purple (A2), yellow (A3) (Fig. 5A). The color of the spots under UV 365 light are deep purple (A1, A2), buff (A3) and dark salmon (A4) (Fig. 5B). The chemical results for all species are summarized in the Taxonomy section at the end of the Discussion.

Distribution and ecology
Distribution maps based on all revised specimens of Coniocarpon from Scandinavia confirm C. cinnabarinum for Denmark, Norway and Sweden, C. cuspidans for Norway, and C. fallax for Norway and Sweden (Fig. 6). Coniocarpon cinnabarinum has been collected in Norway in the boreo-nemoral rainforests in Rogaland and Hordaland, while the species occurs in other humid forests in Sweden (Skåne and Gotland) and Denmark (Sjaelland and Jylland). Specimens of C. cuspidans have been seen in Norway from the boreo-nemoral rainforests in Vest-Agder, Rogaland and Hordaland. Coniocarpon fallax has been collected in Norway in the boreo-nemoral rainforests in Vest-Agder, Rogaland, Hordaland and Møre og Romsdal. This study further reports C. fallax from Sweden (Gotland) for the first time. Outside Scandinavia, C. cuspidans is confirmed for Great Britain and C. fallax for Austria, Great Britain, Switzerland and Turkey (not shown). Coniocarpon preferably grows on trees with smooth bark and the selected host tree species slightly follow a geographical pattern (see Specimens examined below Taxonomic conclusions). Most collections of C. cinnabarinum from Norway have been made from Corylus avellana L., including few from Fraxinus excelsior L. and Sorbus aucuparia L. The species is collected in Denmark from C. avellana, F. excelsior and Fagus sylvatica L., and in Sweden from F. excelsior. Most specimens of C. cuspidans have been collected from C. avellana, but the species has been seen from a rather wide range of trees including F. excelsior, Ilex aquifolium L., Quercus robur L. and S. aucuparia. Coniocarpon fallax has mainly been collected from F. excelsior from Vest-Agder to Hordaland (more rarely from C. avellana), while all specimens from Møre og Romsdal are from C. avellana. The species is further collected from F. excelsior on Gotland and Austria, and from Picea orientalis (L.) Link in Turkey.

Discussion
Species designations are hypotheses to be tested as new evidence becomes available. Recent molecular systematic studies have repeatedly revealed evolutionary lineages within phenotypically delimited lichenized fungi (e.g., Steinová et al. 2013;Lücking et al. 2014;Bendiksby et al. 2015;Alors et al. 2016;Hawksworth and Lücking 2017;Boluda et al. 2019). Whether or not such morphologically indistinguishable, or "cryptic", evolutionary lineages should be recognized at species level may have critical implications for conservation biology and other fields of biology that use species as a fundamental unit (Lumbsch and Leavitt 2011).
In general, several factors should be considered in the process of assessing species status: proper selection of genetic markers (multiple, unlinked loci and from different genomic compartments), presence of statistically supported phylogenetic lineages, sufficiently large sample size, corroborating non-molecular character variation, and thorough review of the taxonomic and nomenclatural history (Grube and Kroken 2000;Printzen 2010;Lumbsch and Leavitt 2011). Our phylogenetic analyses of Coniocarpon are based on unlinked, multilocus DNA sequence data (mtSSU, nucITS, RPB2) showing high statistical support for three distinct genetic lineages (Fig. 1). Re-examination of morphology against the molecular phylogeny of 26 specimens revealed that the three lineages are further supported by differences in ascospore size (Fig. 2), ascospore septation (Fig. 3), distribution of pruina and ascoma shape (Fig. 4). Furthermore, the three lineages differ in the amyloidity of the ascomatal gels and pigment patterns revealed by HPTLC (Fig. 5). Finally, an additional 94 specimens, for which molecular data are not available, were revised using the same morphological and/or chemical characters. As such, this study fulfills the factors recommended for assessing species status.
Evolutionary lineages that remain intact when living in sympatry with close relatives might deserve species status (Coates et al. 2018). The distribution maps in Fig. 6 show sympatry for the three distinct genetic lineages of Coniocarpon in Norway, providing strong indirect evidence that there are mechanisms prohibiting exchange of genetic material among them, supporting their acceptance at species level. Hence, the integrated data gathered in this study jointly support the hypothesis of three distinct species of Coniocarpon in Norway, viz. C. cinnabarinum, C. fallax and C. cuspidans. The latter species was hidden in the extensive synonymy of C. cinnabarinum as Arthonia cinnabarina f. cuspidans Nyl. and is herein resurrected (see Taxonomic conclusions below).
The present data indicate a narrower distribution in Norway for C. cinnabarinum and C. cuspidans as compared to C. fallax (Fig. 6). All available collections from Møre og Romsdal were identified as C. fallax. Coniocarpon cinnabarinum, the only species reported from that county in the Norwegian Red List of Species 2015 could not be confirmed. However, not all collections of Coniocarpon in Norway were available for this study, and previous distributions were partly based on human observations as well. The distribution of the three Coniocarpon species in Norway needs further evaluation in light of the present investigation. Moreover, this study confirms C. cinnabarinum for Denmark and Sweden. Coniocarpon fallax is reported from Sweden (Gotland) for the first time (Fig. 6), based on specimens from S and UPS previously identified as C. cinnabarinum.
Species diversity and abundance generally correlate with habitat preference. Most collections of Coniocarpon in Norway were made in Hordaland (C. cinnabarinum 22, C. cuspidans 24, C. fallax 20), in the core area of the boreo-nemoral rainforests, having the highest levels of humidity and low average winter temperatures (Blom et al. 2015;Moen 1999). In comparison, the number of studied specimens from Vest-Agder (C. cinnabarinum 0, C. cuspidans 2, C. fallax 3) at the southern limit of the boreo-nemoral rainforests is distinctly lower, which might be partly explained by fewer rainforest localities (Blom et al. 2015). The occurrence of C. fallax in Møre og Romsdal indicates a wider ecological amplitude as compared to C. cinnabarinum and C. cuspidans in terms of oceanity and temperature. Coniocarpon cuspidans is currently only confirmed for Norway and Great Britain, but as the species has not been distinguished from C. fallax in the past, it might have a wider distribution in Western Europe.
The protologue of Arthonia cinnabarina f. cuspidans (Nylander 1876) cites specimens from Ireland and Cuba as original material: "Ilicicola in Hibernia (Larbalestier)" and "Exotica eadem datur in C. Wright. Cub. no. 123 a et b". Nylander obviously considered the material from Ireland as the factual type collection. We have selected a specimen from the Nylander herbarium in Helsinki as the lectotype, which is the only specimen from Ireland that undoubtedly has been seen by Nylander. A possible syntype exists in BM: "Derryclare, Connemara, ilicicola, 1876 [BM000974345]". Another specimen [BM000974347] from Larbalestier's herbarium has a printed later label that only states "On young trees. Doughruagh Mountain and other places in Connemara". The type status of this specimen is unclear. Both specimens have been seen by us as high-resolution pictures obtained from the data portal of BM.