Phylogenetic study and taxonomic revision of the Xanthoparmeliamexicana group, including the description of a new species (Parmeliaceae, Ascomycota)

Abstract Xanthoparmelia (Parmeliaceae, Ascomycota) is the most species-rich genus of lichen-forming fungi. Species boundaries are based on morphological and chemical features, varying reproductive strategies and, more recently, molecular sequence data. The isidiate Xanthoparmeliamexicana group is common in arid regions of North and Central America and includes a range of morphological variation and variable secondary metabolites – salazinic or stictic acids mainly. In order to better understand the evolutionary history of this group and potential taxonomic implications, a molecular phylogeny representing 58 ingroup samples was reconstructed using four loci, including ITS, mtSSU, nuLSU rDNA and MCM7. Results indicate the existence of multiple, distinct lineages phenotypically agreeing with X.mexicana. One of these isidiate, salazinic acid-containing lineages is described here as a new species, X.pedregalensis sp. nov., including populations from xerophytic scrub vegetation in Pedregal de San Angel, Mexico City. X.mexicana s. str. is less isidiate than X.pedregalensis and has salazinic and consalazinic acid, occasionally with norstictic acid; whereas X.pedregalensis contains salazinic and norstictic acids and an unknown substance. Samples from the Old World, morphologically agreeing with X.mexicana, are only distantly related to X.mexicana s. str. Our results indicate that X.mexicana is likely less common than previously assumed and ongoing taxonomic revisions are required for isidiate Xanthoparmelia species.


Introduction
The family Parmeliaceae is the largest family of lichenised fungi (Jaklitsch et al. 2016) currently classified in approximately 70 genera with almost 2,800 species (Lumbsch andHuhndorf 2010, Divakar et al. 2017). Xanthoparmelia, with about 800 described species, is the largest genus of lichen-forming fungi , with two centres of distribution in Australia and southern Africa; a smaller number of species occur in the Holarctic (Blanco et al. 2004, Eriksson et al. 2004, Thell et al. 2012, Leavitt et al. 2018. To date, 74 species have been reported from Mexico, amongst these species, 27 are isidiate .
Isidiate Xanthoparmelia species are distributed in boreal, temperate and tropical regions. However, they commonly occur in semi-arid to arid regions worldwide especially on siliceous rocks, such as granite and sandstone. In North and Central America, Xanthoparmelia mexicana (Gyelnik) Hale ranks amongst the most common isidiate species. This taxon is widely distributed and has been reported from western USA, Mexico, Dominican Republic, Argentina, Kenya, Australia, New Zealand, Japan, China and Nepal (Hale 1990, Elix 1994, Nash and Elix 2004. X. mexicana is part of a complex of morphologically similar species, with adnate to slightly attached thalli, cylindrical isidia and a brown lower side of the thalli, which are primarily separated by their secondary metabolites. The species complex also includes X. ajoensis (T. H. Nash) Egan (diffractaic acid), X. dierythra (Hale) Hale (norstictic acid), X. joranadia (T. H. Nash) Hale (lecanoric acid), X. maricopensis T. H. Nash & Elix (norstictic and hyposalazinic acids), X. moctezumensis T. H. Nash (3-α-hydroxybarbatic acid), X. plittii (Gyelnik) Hale (stictic acid), X. schmidtii Hale (barbatic, norstictic and salazinic acids), X. subramigera (Gyelnik) Hale (fumarprotocetraric acid) and X. weberi (Hale) Hale (hypoprotocetraric acid) (Hale 1990. However, previous studies indicate that current interpretations of morphological features and secondary metabolites likely fail to accurately characterise species-level diversity in isidiate Xanthoparmelia species (Leavitt et al. , 2013. To better understand the evolutionary history of the Xanthoparmelia mexicana complex and potential taxonomic implications, isidiate Xanthoparmelia specimens were collected from different locations throughout arid regions of Mexico and supplemented with previously available sequence data. The new samples came from xerophytic scrublands in the states Puebla, Oaxaca, San Luis Potosí, Querétaro, Estado de México, Mexico City, Guanajuato, Zacatecas and Hidalgo, all in the central part of Mexico. We focused on sampling Xanthoparmelia populations that were phenotypically similar to X. mexicana, e.g. isidiate specimens containing salazinic acid. X. mexicana was originally described by Gyelnik (1931) as Parmelia mexicana and was later combined into Xanthoparmelia by Hale (1974). The type specimen was collected from San Jerónimo, in Pedregal de San Angel, Mexico City. The syntype in the Bouly de Lesdain herbarium was destroyed during World War II, whereas the lectotype in the Budapest herbarium (BP) was not available for molecular study. Therefore, we attempted to recollect material at the type locality of X. mexicana and other regions throughout Mexico. Based on the results of this study, we formally describe a previously unrecognised species-level lineage comprised of isidiate specimens as new to science.

Taxon sampling
Specimens were studied from the herbaria ASU, BRY, F, MAF and new collections from different localities throughout arid regions from the central part of Mexico (Table 1, Fig. 1). A total of 83 specimens, representing 43 species were included, with an emphasis on isidiate species/populations from Central and North America (all epithets are validly published, with the exception of X. isidiomontana nom prov assigned to the clade 'D2' from Leavitt et al. 2013). New sequences were generated from 25 specimens and supplemented with 34 sequences from a previous analysis (Leavitt et al. 2018) and 24 additional sequences from GenBank (Table 1). Four species in the genus Xanthoparmelia that have previously been shown to be distantly related to X. mexicana were used as outgroup -X. beatricea, X. austroafricana, X. subramigera and X. aff. subramigera (Leavitt et al. 2018).

Morphology and chemistry
Morphological characters were observed using a Zeiss Stemi 2000-C stereoscope. Ascomatal anatomy, ascospore in addition to conidia shape and size were observed using a Zeiss Axioscope. Secondary metabolites were identified using spot test KOH 10%, KC, C, PD and high-performance thin layer chromatography (HPTLC), using solvent systems C following established methods (Culberson and Johnson 1982, Arup et al. 1993, Lumbsch 2002, Orange et al. 2010).

Molecular methods
Total genomic DNA was extracted from thallus fragments following the manufacturers' instructions using the ZR Fungal/Bacterial DNA Miniprep Kit (Zymo Research Corp., Irvine, CA). DNA sequences were generated for four markers using polymerase chain reaction (PCR): the nuclear ribosomal internal transcribed spacer region (ITS), a fragment of nuclear large subunit rDNA (nuLSU), the nuclear protein-coding marker minichromosome maintenance complex component 7 (MCM7) and a fragment of the mitochondrial small subunit rDNA (mtSSU). PCR reactions contained 6.25 ml of MyTaq Mix, 25 ml H 2 O, 0.25 ml forward and reverse primer and 0.5 ml template DNA, for a total reaction volume of 12.5 ml. The ITS region was amplified using primers ITS1F (Gardes and Bruns 1993) and ITS4 (White et al. 1990); MCM7 using primers MCM7-709f and Mcm7-1348r (Schmitt et al. 2009), mtSSU using primers mrSSU1 and mrSSU3R (Zoller et al. 1999) and nuLSU rDNA using primers AL2R (Mangold et al. 2008) and LR6 (Vilgalys and Hester 1990). PCR products were sequenced using an ABI PRISM 3730 DNA Analyser (Applied Biosystems) at the Pritzker Laboratory for Molecular Systematics and Evolution at The Field Museum, Chicago, Illinois, USA. Nine specimens were obtained previously for a global phylogenetic study of the genus and sequenced using next generation sequencing technology as described in Leavitt et al. (2018) (Table 1). In short, metagenomic Nextera libraries (prepared from total DNA extraction) were sequenced on the Nextseq platform at the Core Genomics Facility at the University of Illinois at Chicago, USA. Illumina reads of each specimen were mapped to reference marker sequences downloaded from Genbank (ITS AY581063, nuLSU HM125760, MCM7 HM579689, mtSSU KR995373) using the mapping feature implemented in Geneious v11.0.3 (http://www.geneious. com, Kearse et al. 2012). The consensus sequence of each locus was extracted and added to the data set of Sanger sequences to build a combined alignment.

Sequence alignment and phylogenetic analysis
Sanger sequences, consensus Illumina reads and sequences available on GenBank were added to an alignment published in Leavitt et al. (2018) using Mafft v7 with the option 'add sequence' (Table 1). ITS, MCM7, mtSSU and nuLSU sequences were aligned independently using the 'automatic' option in Mafft v7, with the remaining parameters set to default values. Ambiguous positions of each one-locus alignment were removed using options for a "less stringent" selection on Gblocks 0.91b (Castresana 2000). SequenceMatrix software (Vaidya et al. 2011) was used for the alignment concatenation. Phylogenetic analyses were performed using Maximum Likelihood (ML) and Bayesian Analysis (BA). ML trees were calculated with RAxML-HPC2 on XSEDE 8.2.10 (Stamatakis 2014) on the Cipres Science Gateway (Miller et al. 2010) using GTR+G+I substitution model with 1000 bootstrap pseudoreplicates. For the BA, substitution models for each locus were estimated using jModelTest-2.1.9 (Guindon andGascuel 2003, Darriba et al. 2012): in ITS the TIM2ef+I+G, in MCM7 the K80+G, in mtSSU the TPM2uf+I and in nuLSU the TrN+I were used. Two parallel Markov chain Monte Carlo (MCMC) runs were performed in MrBayes 3.2.6 (Huelsenbeck andRonquist 2001, Ronquist andHuelsenbeck 2003), each using 10,000,000 generations which were sampled every 100 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. Tree files were visualised with FigTree 1.4.2 (Rambaut 2014). The ITS, MCM7, mtSSU and nuLSU sequences are deposited in GenBank (Table 1).

Phylogeny
Results from phylogenetic analyses presented here clearly indicate that the taxonomy in the Xanthoparmelia mexicana group requires revision because different samples assigned to the same species based on phenotypical characters may not form a monophyletic group. Specimens identified as X. mexicana from Asia (Pakistan and South Korea) were distantly related to samples of the species collected in North America and Europe (included in X. isidiomontana nom prov) (Fig. 2). The latter specimens fell into three distinct and well supported clades (clade I-III in Fig. 2). Note that the three distinct and well supported clades did not form a monophyletic group. Clade 'I' (=X. 'isidiomontana' nom prov, 'D2' in Leavitt et al. 2013) included isidiate specimens from North America and Europe and samples identified as X. dierythra, X. mexicana ( Figs. 2A and B) and X. plittii, in addition to a number of non-isidiate, fertile specimens. Additional studies will be necessary to better understand the delimitation of X. dierythra, which is also polyphyletic and is currently accommodating specimens with norstictic acid and lacking salazinic acid (Hale 1990). This clade likely represents another species-level lineage lacking formal taxonomic recognition and a formal description of this lineage will be proposed once the phylogenetic placement of X. dierythra s. str. is ascertained.
Clade 'II' included specimens collected in the Pedregal, south of Mexico City, which is also the type locality of X. mexicana. However, the new material does not correspond phenotypically with the type specimen of X. mexicana in BP (Fig. 2G). These specimens are different from X. mexicana specimens (represented by Clade III in phylogenetic analysis) in having less contiguous lobes, densely isidiate thallus, presence of salazinic acid, norstictic acid and an unknown substance. Since clade 'II' differs phylogenetically and phenotypically from clade 'III' (representing X. mexicana s. str. -see below), we describe clade 'II' as a species new to science, X. pedregalensis (Figs. 2C and D).
Clade 'III' includes the majority of samples identified as X. mexicana collected in different localities of Mexico (Oaxaca, Puebla, San Luis Potosí, Querétaro, Hidalgo). Specimens recovered in this clade were morphologically and chemically similar to the lectotype of X. mexicana in BP (Fig. 2G). Therefore, clade 'III' is here recognised as X. mexicana s. str. (Gyelnik 1931, Hale 1974) (Figs. 2E and. So far, we have only been able to confirm the presence of X. mexicana s.str. in Mexico. Specimens collected in other areas and previously identified as X. mexicana likely represent different species. For example, none of the samples from Asia or those reported in Leavitt et al. (2013) from western United States belongs to X. mexicana s. str. Further studies are needed to evaluate the occurrence of this species in other parts of the world, including North America and Europe.
Underestimates of species diversity is common amongst under-studied organismal groups (Pawar 2003, Chiarucci et al. 2011, Lücking 2012, Coleman 2015, Troia and McManamay 2016, Troudet et al. 2017, which is particularly evident in lichenised fungi (Crespo and Perez-Ortega 2009, Lumbsch and Leavitt 2011, Leavitt et al. 2013, Leavitt et al. 2018. Previous studies concluded that the species delimitation in lichenised ascomycetes with traditional morphological and chemical characters are apparently misleading with respect to species diversity. In the study of Leavitt et al. (2016), several new taxa were described primarily based on evidence from genetic data, but it does not preclude the possibility that additional studies investigating morphological and chemical characters may identify additional independent characters or combinations of characters, supporting the species circumscribed using molecular data. Our results corroborate findings from the previous studies by showing the need of an integrative approach using not only conventional (i.e. morphology and TLC data), but also new sets of data (e.g. DNA sequence data) to better understand the pattern of species diversity. Our study shows that, by incorporating molecular data, the taxonomic status of a conventionally difficult group based on morphology can be resolved: the three main clades belonging to the X. mexicana complex do not form a monophyletic group based on our newly reconstructed phylogeny (Fig. 1). In this context, the species diversity in the X. mexicana complex is likely under-estimated and morphologically cryptic species may be identified in the future. Diagnosis. Thallus moderately adnate to adnate, imbricate, upper surface yellowgreen, lower surface tan-brown, abundant isidia subglobose to cylindrical, simple to branched and medulla containing salazinic and norstictic acids as major compounds and an unknown substance. Differing from the phenotypically similar X. mexicana by nucleotide position characters in the ITS sequence as shown in Table 2.
Etymology. The taxon name is based on its occurrence in the Pedregal de San Angel region of Mexico.
Distribution and ecology. The new species was found in xerophytic scrub vegetation, in Pedregal de San Angel south of Mexico City, growing on volcanic rocks. It is currently known only from the type locality.
Notes. Xanthoparmelia pedregalensis is morphological and chemically similar to X. mexicana. However, the latter has more contiguous lobes and is less isidiate than X. pedregalensis. In addition X. mexicana has salazinic (major) and consalazinic acid (minor) and usually norstictic and protocetraric acids (trace) in the medulla, whereas X. pedregalensis contains salazinic (major) and norstictic acids (submajor) and an unknown substance. Distinguishing the two species is supported by molecular data.
Additional specimens examined.