**1. Introduction**

Natural history collections are invaluable resources for assessing biodiversity and studying the evolution, biology, ecology, morphology, anatomy, chemistry, and genetics of species [1–3]. They have primarily been used for biodiversity, taxonomy, and evolutionary research, but recent efforts, including those employing machine learning, have substantially broadened their use, allowing, for example, the evaluation of shifts associated with climate change [1,4–6]. In mycology, historical records have been used, for instance, to analyze

**Citation:** Dal Forno, M.; Lawrey, J.D.; Moncada, B.; Bungartz, F.; Grube, M.; Schuettpelz, E.; Lücking, R. DNA Barcoding of Fresh and Historical Collections of Lichen-Forming Basidiomycetes in the Genera *Cora* and *Corella* (Agaricales: Hygrophoraceae): A Success Story? . *Diversity* **2022**, *14*, 284. https://doi.org/10.3390/ d14040284

Academic Editors: W. John Kress and Morgan Gostel

Received: 24 February 2022 Accepted: 6 April 2022 Published: 10 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

changes in the fruiting date of mushrooms and how that relates to climate change [7], and to track phytopathogenic fungi (review by [8]). For lichenology in particular, a unique use of herbarium specimens has been to compare concentrations of different minerals in contemporary versus historical samples to assess how these levels may have changed over time [9–11].

High-throughput sequencing (HTS) has revolutionized collections-based research, as the commonly fragmented nature of DNA from historical (archival) specimens generally poses a challenge in Sanger sequencing [12,13]. This problem can be overcome by HTS approaches, not only allowing amplicon sequencing from less-degraded samples but also the assembly of partial or whole genomes in many cases [14,15].

Historical lichen collections, a term utilized here to describe material of at least two decades, have only been used in a limited number of genetic studies. Sohrabi et al. [16] were the first to successfully obtain a sequence acquired from a 75-year-old herbarium specimen of *Aspicilia aschabadensis* (J. Steiner) Mereschk. Subsequently, Redchenko et al. [17] sequenced multiple archival *Caloplaca* specimens, including the current record for oldest lichen specimen sequenced (a material from 1859) and Bendiksby et al. [18] sequenced specimens of *Staurolemma omphalarioides* (Anzi) P.M. Jørg. and Henssen of up to 100 years old. More recently, studies have started to utilize HTS to obtain DNA sequences from historical lichen collections. With two-step PCR and multiple primer combinations, Kistenich et al. [19] aimed to amplify a 900 base-pair mtSSU fragment from 56 specimens of eight different species, varying in age up to 125 years-old. They showed that consensus Ion Torrent sequences presented overall better quality than those generated by Sanger sequencing. Gueidan et al. [20] utilized PacBio amplicon sequencing and successfully recovered target sequence data from 89 of 96 samples (88.5%), with the studied samples being up to 25 years old. More recently, Gueidan and Li [21] furthered their studies to include 384 specimens collected between 1966 and 2020, quadrupling their sampling and including older samples with equivalent success rates (86.5%).

The first molecular study involving a historical basidiolichen collection was that of Schmull et al. [22], who successfully sequence the ITS1 region of a 30-year-old specimen of *Dictyonema*, describing a new, potentially hallucinogenic lichen, *Dictyonema huaorani* Dal Forno, Schmull, Lücking, and Lawrey. The first application of HTS for basidiolichen collections has been relatively recent: Lücking et al. [23] described the new species *Cora timucua* Dal Forno, Kaminsky, and Lücking, known only from archival specimens collected in Florida. The authors were able to generate sequences with both Sanger and Illumina sequencing from a sample from 1901, demonstrating that even Sanger sequencing can still be a valuable tool in attempting to acquire DNA sequences from archival specimens, as also shown by other studies [16,18,19]. Indeed, Sanger sequencing has been used successfully to obtain sequence data from historical collections of non-lichenized Basidiomycota [24,25].

The lichenized basidiomycete genus *Cora* Fr. presents a notable example of the usefulness of ITS in fungal barcoding for delimiting species and assessing biodiversity, as proposed by Schoch et al. [26]. *Cora* forms part of the *Dictyonematinae* subtribe, a lichenized lineage of five genera with different morphologies related to the non-lichenized mushroom genus *Arrhenia* Fr. [27–29]. Within this subtribe, *Cora* and the related genus *Corella* Vain. are distinguished by a foliose thallus with a compact surface, whereas the other genera (*Acantholichen* P.M. Jørg., *Cyphellostereum* D.A. Reid, *Dictyonema* C. Agardh ex Kunth) are filamentous or microsquamulose [28]. The taxonomy of foliose basidiolichens has suffered from the typical problems of traditional revisions based largely on herbarium collections that fail to recognize important field characters, such as consistency, lobe arrangement, color, and substrate (Figure 1). In addition, emphasis was historically placed on mycological characters, such as hymenophore anatomy and basidiospores, to establish species boundaries [30]. As a result, for a long time, only a single foliose species was recognized, named *Dictyonema pavonium* (Sw.) Parmasto and subsequently *D. glabratum* (Spreng.) D. Hawksw. [31], with all other previously proposed taxa, including those in the genus *Corella*, as synonyms [30].

**Figure 1.** Comparison of fresh samples in their natural habitat (**<sup>a</sup>**,**b**) versus dried and deposited in the herbarium (**<sup>c</sup>**,**d**)—(**<sup>a</sup>**,**<sup>c</sup>**): *Cora* spec-023 (Dal Forno 2042), (**b**,**d**): *Cora cyphellifera* (Dal Forno 1808).

Molecular approaches led to the realization that these foliose lichens represented more than one species and also supported the separation of the genera *Cora* and *Corella* [27,28]. A first broad sampling using the ITS barcoding locus resulted in an estimated 116 species of *Cora* and ten of *Corella* [32]. Although constituting a dramatic increase in species count, this estimate was still considered conservative: quantitative species recognition methods suggested up to 170 species based on the same data, and a novel prediction method, which takes into account unsampled regions and habitat suitability, estimated more than 450 species [32]. Soon after, with much increased sampling, Lücking et al. [33] distinguished 189 species. It is possible to challenge these findings based on potential problems in ITS barcoding, such as improper assessment of intragenomic variation or the occurrence of multiple ITS copies in the genome as a result of hybridization and introgression or gene duplication, possibly leading to taxonomic inflation or the recognition of artifactual taxa [34,35]. On the other hand, ITS has also been shown to lack resolution in recently evolving species complexes, including both non-lichenized and lichenized fungi [35–39], potentially counterbalancing issues with wrongly assessed ITS variation in terms of species counts but introducing additional inaccuracy.

In the genera *Cora* and *Corella* and in subtribe *Dictyonematinae* in general, the topology of ITS-based phylogenies has been found to be highly congruen<sup>t</sup> with those of other markers, such as nuLSU and RPB2 [28,40], suggesting that ITS resolves these lineages accurately. Analysis of intragenomic variation of the ITS in *Cora inversa* Lücking and B. Moncada using a 454 pyrosequencing approach, did not demonstrate potential gene duplication or hybrid ITS arrays; instead, the variation detected stemmed almost entirely from sequencing errors and had no effect on accurate species delimitation when using a phylogenetic approach [41]. A similar result has been reported for the *Rhizoplaca melanophthalma* species complex, in which the observed ITS variation did not interfere with species discrimination in that group [42].

Phenotypes in phylogenetically delimited species of *Cora* are highly consistent with the underlying molecular data [33] and even photobiont haplotypes showed a high level of congruence with the ITS-based phylogeny of the associated mycobionts [43]. Therefore, in these lichenized Basidiomycota, ITS barcoding appears to provide highly accurate assessments of species richness.

Unfortunately, these earlier studies were biased towards geographic areas from which fresh material could be readily sampled. Natural history collections provide access not only to a much broader geographic range, but also to specimens that may have been sampled in regions that are now heavily altered or with their original habitats destroyed, as shown by the example of the possibly extinct *Cora timucua* [23,44]. Inserted within a broad phylogenetic and phenotypic framework, herbarium collections of *Cora* and *Corella* can thus provide unique opportunities to expand taxon and specimen sampling for these genera, to extend our understanding of their biology, and to test assumptions about geographical distributions of species currently only documented by recently collected material. The utility of historical collections in this regard is, however, dependent on the quality of the sequences obtained from these samples.

The United States National Herbarium (USA) at the Smithsonian National Museum of Natural History (SI-NMNH) is home to one of the largest lichen collections worldwide, with over 250,000 specimens. It contains one of the largest and most diverse collections globally of the subtribe *Dictyonematinae*, with over 400 specimens having broad geographical and temporal representation and unique morphologies. From these collections, we analyzed all *Cora* and *Corella* specimens, for a total of 274 samples with collection dates ranging from 1888 to 1998, complementing our already large dataset of 856 recently collected specimens from 18 countries (Figure 2) and representing the largest study of historical lichen collections using molecular approaches including HTS to date focusing on a single genus (most samples belong to *Cora*).

Our objectives for this study included: (1) testing the success rate of ITS barcoding from historical samples of *Cora* and *Corella* and comparing Sanger versus Illumina sequencing success; (2) expanding the existing ITS-based phylogeny with newly generated sequences and assessing the number of phylogenetically delimited species using various quantitative approaches (GMYC, bPTP, ABGD, ASAP) to test our earlier prediction of more than 450 species in this group; (3) further assessing potential intragenomic ITS variation by comparing Sanger and Illumina data from the same specimens; (4) testing the performance of ITS relative to a six-marker dataset using a subset of terminals; and (5) exploring the level of potential cryptic speciation in *Cora* and *Corella* by testing for consistency between molecular and morphological data.

**Figure 2.** Map showing country-based availability of sequenced contemporaneous (fresh) and historical (herbarium) samples of *Cora* (and *Corella*) throughout the Americas (generated with mapchart.net).
