Two new species of Endocarpon (Verrucariaceae, Ascomycota) from China

Endocarpon species are key components of biological soil crusts. Phenotypic and systematic molecular analyses were carried out to identify samples of Endocarpon collected from the southeast edge of the Tengger Desert in China. These morphological and molecular analyses revealed two previously undescribed species that form highly supported independent monophyletic clades within Endocarpon. The new taxa were named Endocarpon deserticola sp. nov. and E. unifoliatum sp. nov. Furthermore, our results indicated that the newly developed protein coding markers adenylate kinase (ADK) and ubiquitin-conjugating enzyme h (UCEH) are useful for assessing species boundaries in phylogenic analyses.

In order to seek for new species resource with sand-stabilisation potential besides Endocarpon pusillum, we carried out taxonomic study on Chinese Endocarpon and two specific species drew our attention because they were not assignable to any known species based on phenotypic characteristics. In view of insufficiency of phylogenetic data in the genus Endocarpon, we aim to study the two species based on both phenotypic traits and phylogeny and accumulate more DNA data for the further study.
Nowadays, besides nrDNA ITS region, which is often used in fungal species delimitation 46 , some protein-coding genes (e.g., RPB1, MCM7) have also been successfully used 47 . In this study, two protein-coding genes (ADK and UCEH) were developed and used for the first time for phylogenetic analyses. ADK is a phosphotransferase enzyme that catalyses the interconversion of adenine nucleotides, thus playing an important role in cellular energy homeostasis 48 . UCEH is a subunit of ubiquitin-conjugating enzymes and performs the second step of the ubiquitination reaction that targets a protein for degradation by the proteasome 49 . We newly designed the primers for ADK and UCEH based on the cDNA library of E. pusillum 50 in this study. Our major goals are (1) to describe the phenotypic characteristics of these two species, (2) understand the phylogenetic relationships in the genus Endocarpon and (3) as well as develop two new protein coding markers (ADK, UCEH) to strengthen the phylogenetic information.

Results
Molecular phylogeny. The aligned matrix contained 476 unambiguous nucleotide position characters for the internal transcribed spacer (ITS), 679 for the adenylate kinase (ADK), and 279 for the ubiquitin-conjugating enzyme h (UCEH). The final alignment of the concatenated data set was 1434 positions in length. Seventy-six sequences were newly generated for this study ( Table 1).
The single-locus gene trees for the three markers are illustrated in Figures S1-S3. The topologies of the single-locus phylogenies did not exhibit obviously supported conflicts (i.e. bootstrap values ≥75%), and thus they were analysed in a concatenated data matrix. The best-fitting models corresponding to the three single-locus markers are listed in Table 2.  The concatenated three-locus data sets contained 40 sequences (1434 nucleotides sites), comprising 11 Endocarpon species. The maximum likelihood (ML) tree for the concatenated data sets constructed using partitioned models are presented in Fig. 2. The maximum likelihood analyses (ML, RAxML) shows the same highly supported clades as the Bayesian analyses. Both analyses were merged in one phylogenic tree, and the respective values (bootstrap values ≥75, posterior probability values ≥95) were plotted directly on the branches (Fig. 2). Based on the phylogenetic results the genus Endocarpon forms a strongly supported monophyletic clade and is obviously separated from the other genera within Verrucaricaceae, i.e., Dermatocarpon spp., Staurothele spp., Verrucaria spp. and Willeya spp. Within the Endocarpon clade, all 11 studied species formed highly supported lineages.

Scanning electron microscope (SEM) images of rhizines in the two new species. Both of the new
Endocarpon species may fulfil potentially important roles by stabilising soils via sand particles consolidation with their rhizines, as inferred from SEM observations (Fig. 3). The sand particle surface is covered by the squamose thalli of the Endocarpon species (Fig. 1C), and sand crystals are wrapped in their branched rhizines (Fig. 3).

Discussion
Phenotype. According to our morphological assessment, some samples collected from the Tengger Desert in China were not able to be categorised as any previously described Endocarpon species 9, 15, 31-45 . Samples corresponding to the newly described Endocarpon deserticola are characterised by abundant perithecia dispersed throughout nearly all squamules, and the perithecia reach 15-60 (up to 100) in number (Fig. 4A). This species is most similar to E. helmsianum Müll. Arg. of Australia, which is also characterised by abundant perithecia 41 ; however, E. helmsianum exhibits wider squamules (5-25 mm), a more contiguous to overlapping thallus, and much larger ascospores 41 .
Within the Endocarpon clade was no explicit interspecific relationship reflected in the phylogenetic tree (Fig. 2), which may be the results of the limited number of species and gene loci included in the analyses. Nevertheless, each species was highly supported and obviously separated from others, e.g. the two new putative species Endocarpon deserticola and Endocarpon unifoliatum, formed two strongly supported clades separated from other species. The two newly developed protein coding markers (ADK and UCEH) show a quite high phylogenetic informative signal on species level and can be useful for future phylogenetic analyses, not only for Endocarpon, but also for other fungal genera ( Figures S1-S3). This finding confirms the reliability of gene trees for phylogenetic analyses based on concatenated data sets.
The species Endocarpon tenuissimum is nested within the monophyletic Willeya diffractella (Nyl.) Müll. Arg. and has been placed in synonymy 30 , which is also supported by the ML tree based on ITS sequences in the present study ( Figure S1). However, within the monophyletic Willeya diffractella clade, neither morphology nor geography was found to be corresponding to the main infraspecific groupings except the nature of the substrate (calcareous vs non-calcareous). As known Endocarpon tenuissimum shared the same character, i.e. non-calcareous substrate, with one group of Willeya diffractella, but there are some subtle differences in thallus color, ecology and ascospore size between Endocarpon tenuissimum and Willeya diffractella 30 . More samples and genes are required to further explore the consistency between phenotype and phylogeny within Willeya diffractella.
Role of rhizines in soil stabilisation. Both new species of Endocarpon may fulfil important roles by stabilising soils by sand particle consolidation with their rhizines, which were inferred from SEM observations (Fig. 3). This finding is consistent with previous studies investigating lichenised BSC fungal communities in desert ecosystems 15,51 .  Table 1. Specimen information and GenBank accession numbers for the taxa used in this study. *DNA sequences were downloaded from GenBank. Othe specimens were sequenced by the authors; all sequences were deposited in HMAS-L; missing sequences are indicated by dashes.
Lichens are more effective than cyanobacteria at reducing soil erosion because the fungal hyphae of the lichen thallus penetrate more deeply and the tissue extends above the soil surface 52 .
In summary, based on morphological and molecular phylogenetic data, two new putative species, Endocarpon sp. 1 and Endocarpon sp. 2, have been described under the names Endocarpon deserticola and Endocarpon unifoliatum, respectively. In previous studies, the lichen species Endocarpon pusillum exhibited drought resistance 15 and sand and carbon fixation 9,11 , and thus served as an important species for desert bio-carpet engineering and the study of stress tolerance mechanisms in lichens in China [12][13][14] . The two new species, E. deserticola and E. unifoliatum, exhibit different morphological and phylogenetic characteristics from those of E. pusillum and may also play important roles in desert sand stabilisation. Further characterisation of features such as stress tolerance and   Etymology: The epithet of the new species 'deserticola' is a Latin compound consisting of the Latin noun 'desertum: desert' and the Latin adjective suffix '-colus: inhabiting' , meaning that the new species grows in the desert.
Diagnosis: This species is characterised by its abundant perithecia, which is dispersed throughout almost all squamules, up to 100 or more in number.
Morphology: Thallus terricolous, squamulose; squamules solitary or contiguous, with slightly upturned margins, rounded, elongate or irregular, 1-3 (−4) mm in width; upper surface pale brownish to brownish; lower cortex well-developed, dark brown to black, with brown to black rhizines, 4-6 mm long, irregularly branching in the terminal region.   Comments: This species is most similar to E. helmsianum found in Australia, which is also characterised by abundant perithecia. However, E. helmsianum has wider squamules (5-25 mm), a more contiguous-to-overlapping thallus, and much larger ascospores. Etymology: The epithet of the new species 'unifoliatum' is the nominative singular neuter of the Latin adjective 'unifoliatus: with one thallus'.
Diagnosis: This species is characterised by its unifoliate, concave, lobate thallus with slightly upturned margins.
Morphology: Thallus terricolous, squamulose, concave, and lobate, sometimes greyish-white to white at the central part of thallus, brown at the thallus edges; squamules mostly solitary, not contiguous, with slightly upturned margins, rounded, elongate or irregular, 1-2 (−4) mm wide; upper surface pale to yellowish brown; lower cortex well developed, dark brown to black, with black rhizines 2-3 mm long, irregularly branching in the terminal region. Comments: This species is similar to E. pusillum, but E. pusillum is delimited by its plane, fully adnate thallus, and tightly aggregated, nearly inseparable squamules.

Materials and Methods
Lichen collection and ethics statement. Lichen specimens were collected from the Shapotou region (37°32′N, 105°02′E) on the southeast fringe of the Tengger Desert (Fig. 1). The investigation areas are located at an elevation of 1339 m in the steppified desert zone, which is also a transitional zone between desert and oasis 53 . The area has a mean annual precipitation of 180.2 mm, a mean annual evaporation of 3000 mm, a mean annual   2 , and concentrated alcoholic p-phenylenediamine) and TLC (solvent system C) were used to detect lichen substances 55,56 .
DNA extraction, PCR amplification, and sequencing. Thirty-six specimens, including seven Endocarpon species, were chosen for DNA extraction, as shown in Table 1. The extraction procedure followed the modified CTAB method 57 . Three gene loci were used for PCR amplification: the nrDNA ITS region and two protein-coding genes, ADK and UCEH. The primer pairs ITS4 and ITS5 58 were used to amplify the nrITS regions, and the primers for ADK and UCEH were newly designed in this study (Table 3) based on the cDNA library of E. pusillum 50  Phylogenetic analyses. The sequences generated for this study were complemented with sequences from GenBank representing additional specimens or species, as listed in Table 1. The gene sequences of three loci, specifically nrDNA ITS, ADK and UCEH, were used for phylogenetic analyses. Sequences were aligned using ClustalW Multiple Alignment 59 in BioEdit 7.2.5 60 and introns were manual excluded. The alignment files were transformed into both phylip and nexus formats using SeaView version 4 61,62 . The best model for the three single genes used in the phylogenetic analysis was identified in advance with jModelTest-2.1.9 63,64 .

Congruence among loci.
To test the level of congruence among loci, highly supported clades (equal to or more than 75% bootstrap) from single-gene trees were compared and assessed 65,66 . Each locus was subjected to a randomised accelerated maximum likelihood (RAxML) analysis involving 1000 pseudoreplicates with RAxML-HPC BlackBox 8.2.6 (Stamatakis 2014) on the Cipres Science Gateway (http://www.phylo.org) 67 . The results were visualised with FigTree 1.4.2. When there was no conflict using a 75% bootstrap value threshold, in situations where a monophyletic group was supported with bootstrap values ≥75% at one locus and the same group of taxa was supported (≤75%) as non-monophyletic with another locus, the group was assumed to be congruent and the data set was concatenated 66 .
Phylogeny of the genus Endocarpon. Phylogenetic analyses of Endocarpon were performed using the concatenated data set, which was analysed using RAxML-HPC BlackBox 8.2.6 68 and MrBayes 3.2.6 69, 70 on the Cipres Science Gateway (http://www.phylo.org) 67 . For the ML analysis, the GTR+G+I model was used as the substitution model with 1000 pseudoreplicates. The data were partitioned according to the different genes. The best model for the three single genes used in the Bayesian analysis was obtained in advance with jModelTest-2.1.9. Data sets for the two protein-coding genes (ADK and UCEH) were also partitioned by codon position. Two parallel Markov chain Monte Carlo runs were performed, each using 8000000 generations and sampling every 1000 steps. A 50% majority rule consensus tree was generated from the combined sampled trees of both runs after discarding the first 25% as burn-in.
Scanning electron microscopy. Rhizines of the samples were observed by performing SEM. Samples were sputter-coated with gold particles using a Bio-Rad SEM coating system (Sputter Coater BALTEC SDC 005, Leica Microsystems, Liechtenstein), and SEM images were recorded using a scanning electron microscope (SEM Quanta-200, FEI, Czech Republic) with a secondary electron detector operated at 10.0 kV.

Nomenclature.
The electronic version of this article in Portable Document Format (PDF) in a work with an ISSN or ISBN will represent a published work according to the International Code of Nomenclature for algae, fungi, and plants. In addition, new names contained in this study have been submitted to Fungal Names (FN) from where they will be made available to the Global Names Index. The unique FN number can be resolved and the associated information viewed through any standard web browser by appending the FN number contained in this publication to the prefix http://www.mycobank.org/MB/.