Uncovering Chemical Variability: Molecular

North American Fungi

Volume 8, Number 12, Pages 1-14 Published September 16, 2013

Uncovering chemical variability: molecular data reveal the identity of a sterile crustose lichen from the Yukon and affirm an expanded circumscription for Buellia griseovirens

Jessica L. Allen and James C. Lendemer Institute of Systematic Botany, The New York Botanical Garden, Bronx, NY 10458-5126, U.S.A.

Allen, J. L., and J. C. Lendemer. 2013. Uncovering chemical variability: molecular data reveal the identity of a sterile crustose lichen from the Yukon and affirm an expanded circumscription for Buellia griseovirens. North American Fungi 8(12): 1-14. doi: http://dx.doi:10.2509/naf2013.008.012

Abstract: Here we present the results of a study using molecular data (nrITS and mtSSU sequences) to aid in the identification of a sterile, sorediate crustose lichen from the Yukon Territory of Canada. BLASTn and megaBLAST indicated an affinity to the family Caliciaceae, specifically the Buellioideae. Molecular phylogenetic analyses of mtSSU sequences from the Caliciaceae affirmed this placement, and recovered a strongly supported clade composed of the unidentified populations and members of Buellia s.l. Internal transcribed spacer sequences of the taxon were >99% similar to two sequences of B. griseovirens, a relationship supported by additional molecular phylogenetic analyses. This is the first report of the occurrence of norstictic acid deficient populations of B. griseovirens from North America. It 2 Allen & Lendemer. Lichen chemical variability. North American Fungi 8(12): 1-14

is also the first study to use molecular data to examine, and confirm, some of the chemical variability in B. griseovirens proposed in previous revisions.

Key words: chemotaxonomy, biogeography, boreal forest.

Introduction: To identify this material we sequenced two loci: Accurate identification of sterile asexually nrITS, which is widely used to identify and reproducing crustose lichens can be difficult, delineate fungal species (Schoch et al. 2012), and even when a combination of geographical, mtSSU, which aids in placement of taxa in a morphological and chemical data is used (Fryday higher-level evolutionary framework (Hodkinson & Coppins 1997). Although such lichens can be & Lendemer 2012). readily identified when found in a fertile state, the frequent absence of sexual characters Materials and Methods typically requires one to use correlated suites of Microscopy and chemistry non-molecular characters to elucidate the This study is based on material collected by the taxonomic affinities of a given unidentified second author in June 2011 and deposited in the specimen. Often these data alone are insufficient herbarium of The New York Botanical Garden to identify sterile specimens to species or even (NY). Reference material of Buellia griseovirens family. Thus it is not surprising that molecular from that institution was also consulted. All sequence data have begun to be used to specimens examined in this study are listed investigate these problematical taxa. This under collections examined. Light micrograph approach has proven to be effective in several and illustration preparation followed the studies (Arup & Grube 1999, Hodkinson & methods of Allen et al. (2012). Thin Layer Lendemer 2012, Lendemer & Lumbsch 2008, Chromatography (TLC) was performed in Nelsen et al. 2012, Vondrák et al. 2010, Solvents A and C following Culberson and Lendemer, in press), however the use of Kristinsson (1970). molecular data in this manner is not yet widespread. DNA extraction, PCR amplification, and sequencing Here we report the results of an effort to employ DNA extraction, PCR amplification, sequencing, this method to identify material of a sterile and sequence assembly follow the methods of sorediate crustose lichen from northern Canada Hodkinson and Lendemer (2012) for both mtSSU that could not be identified using traditional and nrITS. Sequences for both loci were obtained characters alone. This material is characterized from five specimens collected in the Yukon by dark gray soredia, immersed soralia, Territory (Table 1). Four of these represented production of atranorin, and occurrence on large, populations of the unknown sorediate crustose old, dry logs in boreal habitats. Although the lichen discussed in the introduction, and the fifth taxon could easily be recognized by its gestalt, it represented a typical population of Buellia proved impossible to identify with certainty griseovirens that produced both norstictic acid owing to the number of taxa with similar and atranorin. The newly generated nrITS and morphological and chemical features occurring in mtSSU sequences were >99% identical between different, unrelated genera. For instance, the samples, and as such only those from one sample taxon could easily have been accommodated in (KC681818) were used to query GenBank and either Lecanora Ach. or Buellia Ach. on the basis inform taxon sampling for the datasets outlined of vegetative morphology and chemistry alone. below.

Allen & Lendemer. Lichen chemical variability North American Fungi 8(12): 1-14 3

Assembling the datasets and aligning the Phylogenetic analyses sequences The NEXUS file for each locus was assembled We used BLASTn (Altschul et al. 1997) to search separately and prepared for weighted maximum the NCBI non-redundant nucleotide database parsimony analysis in PAUP* according to against the mtSSU sequence from KC681818 to Hodkinson and Lendemer (2011). Ambiguously determine the general placement of the aligned regions were recoded using INAASE 3.0 specimens in a higher level taxonomic framework with a transition:transversion:gap ratio of 1:1:1. (the search was limited to this sequence because Step matrices were also created for the the other newly generated mtSSU sequences were unambiguously aligned regions using STMatrix identical). The search revealed that KC681818 3.0 (written by S. Zoller, Miadlikowska et al showed a high affinity to the family Caliciaceae, 2002). For the mtSSU a single step matrix was and specifically members of the Buellioideae generated for the entire alignment excluding the (sensu Gaya et al. 2012). Although some studies ambiguous regions. For ITS the matrix was (e.g., Tibell 2003; Williams and Tibell 2008) partitioned into the ITS1, 5.8S and ITS2 and step superficially refute the circumscription of matrices were created for each region separately. Buellioideae and Calicioideae established by Gaya PAUP* 4.0d125 (Swofford 2012) was used to et al. (2012), closer inspection reveals that in all perform maximum parsimony analyses (MP). A cases sampling is too narrow and/or there is no first round of searches was made with 1000 support for the recovered higher level random-addition-sequence (RAS) replicates and relationships. Thus these studies do not conflict tree bisection reconnection (TBR) branch with those obtained by Gaya et al. (2012) and we swapping. The MULTREES option was in effect refer to Buellioideae throughout the remainder of and zero-length branches were collapsed. The the text. To place our newly generated sequences mtSSU analysis resulted in a single most in a higher-level phylogenetic framework we parsimonious tree and the ITS analysis resulted assembled a datasets including our newly in two equally parsimonious trees. The latter generated sequences and all mtSSU sequences differed only in the arrangement of sequences from the Caliciaceae in Gaya et al. (2012). To within the B. griseovirens clade, all of which are assemble the ITS dataset we used megaBLAST to >99% similar. Both analyses reached only one search the NCBI non-redundant nucleotide tree island which was hit in >99% of the database for sequences that were highly similar replicates. Branch support for the MP trees was to KC681818. We downloaded the 40 sequences estimated through bootstrap analyses with the highest total scores in relation to (Felsenstein 1985) by performing 1000 bootstrap KC681818, pruned all sequences representing replicates with 10 RAS per bootstrap replicate, species that did not have at least two accessions with all other settings as above. and only included the top two representatives from each species with over two accessions, thus For maximum likelihood (ML) and Bayesian eliminating further duplicates. Both regions were Inference analyses (BI) ambiguously aligned aligned separately with the online version of regions were recoded using PICS-Ord (Lücking et MAFFT v. 5 (Katoh et al. 2004) using the default al. 2011). Maximum likelihood analyses were settings. The alignments were then examined and performed using RAxML 7.2.8 (Stamatakis et al. manually adjusted in Mesquite 2.75 (Maddison 2008). The –f a option was used to conduct rapid and Maddison 2011). Using the same program bootstrap analysis while simultaneously ambiguously aligned regions were manually searching for the best-scoring ML tree. 1000 excluded and defined as part of an exclusion set, bootstraps were performed and the GTRGAMMA ambiguous bases and terminal gaps were model of rate heterogeneity was used. For converted to missing. Bayesian analysis MrModeltest (Nylander 2004) 4 Allen & Lendemer. Lichen chemical variability. North American Fungi 8(12): 1-14

was first used to select nucleotide substitution (MP-BP/ML-BP/BI-PP:98/83/0.85) (Fig. 1). model. We selected the SYM+I model based on Based on the results obtained from the analyses the Akaike Information Criterion (Akaike 1973) outlined above, the similarity of the sequences, for both the ITS and mtSSU. Bayesian Inference and morphological similarities of the unidentified was implemented in MrBayes 3.1.2 (Ronquist & specimens to reference material of B. Huelsenbeck 2003) with the model parameters griseovirens, we consider the unidentified for SYM+I specified via MrModeltest and populations from the Yukon to be referable to B. inserted into a MrBayes block that was produced griseovirens, representing the first norstictic acid through the online automated form found at the deficient populations reported from North Santos Lab website America. http://131.204.120.103/srsantos/mrbayes_form/ index.html. The Markov Chain Monte Carlo Discussion: parameters consisted of 10000000 generations, Here we show for the first time that molecular with four chains, and a tree sampled every 100 data support the inclusion of norstictic acid generations. The first 10000 trees were discarded deficient populations in the circumscription of as burn-in, and results were summarized in the Buellia griseovirens. Previous authors have form of a 50% majority- rule consensus tree. The proposed that such chemical variability be input files for all analyses reported here have included within B. griseovirens (Tønsberg 1992, been deposited in http://datadryad.org/ Nordin 2001), but this hypothesis had not been examined with an independent molecular Results: dataset. Nordin (2001) also noted the existence of Morphological and chemical study of the populations that produced stictic acid instead of unidentified material from the Yukon revealed norstictic acid. Although we were unable to that it was characterized by dark gray soredia, generate sequence data for such populations, we immersed soralia, occurrence on large, dry, old can report the existence of this chemical variant logs, and the production of only atranorin. The in North America (New York: J.C. Lendemer four mtSSU sequences generated from these 3086, NY1054078). Additional chemical specimens were identical to the reference variability recognized to fall within the sequence of B. griseovirens generated for this circumscription of this taxon are a suite of study. Furthermore, all ITS sequences from the unidentified pigments (Tønsberg 1992, Elix and unidentified material, as well as those from Tønsberg 1999), the variability of which were not typical B. griseovirens generated for this study examined here. and available in GenBank were >99% identical (there were only three nucleotide differences As has already been stated, this is also the first among all six sequences, none of which were report of norstictic acid deficient populations of synapomorphies for the sequences derived from Buellia griseovirens from North America. the four unidentified populations). Phylogenetic Identification of the specimens from the Yukon analyses of the mtSSU dataset recovered the as B. griseovirens is supported by multiple newly generated sequences as members of independent lines of evidence including: (1) they Buellia s.l. in a strongly supported Buellioideae were found within the known range of the clade (ML-BP:99, BI-PP:1.0) in the Caliciaceae species, and even sympatric with, typical (Appendices 1 and 2). Analyses of the ITS dataset populations of B. griseovirens (Fig. 2), (2) they recovered the newly generated sequences in a are morphologically identical to reference strongly supported clade with the two B. specimens of B. griseovirens that were examined griseovirens reference sequences, one of which (Fig. 3) and (3) they are essentially genetically was from the Yukon and the other from Sweden identical to known reference sequences of B. Allen & Lendemer. Lichen chemical variability North American Fungi 8(12): 1-14 5 griseovirens s. str. using two independent loci have been able to confidently identify these (Fig. 1 and Appendices 1 and 2). The only specimens nor been alerted to the presence of a apparent difference between these populations norstictic acid deficient chemotype of B. and those of B. griseovirens s. str. is the lack of griseovirens in North America. In the field, norstictic acid. norstictic deficient specimens of B. griseovirens, like those with norstictic acid, are most likely to Buellia griseovirens is morphologically quite be confused with other sorediate crustose lichens variable (Nordin 2001, Tønsberg 1992). In the which occur on lignum and can have dark colored material that we have examined the thallus soralia. All of the morphologically similar species ranges from endosubstratal to episubstratal, such as Caloplaca species with thalli lacking patchy to continuous and, when episubstratal, it anthraquinones (Wetmore 2004), Thelomma is rimose to areolate. One of its most distinctive ocellatum (Tibell 1976) and Xyloborus vitiligo features is its round, immersed, concave, gray- (Ach.) J.R. Laundon (Holien & Tønsberg 2008) white to yellowish soralia. About 25% of samples differ chemically in lacking atranorin. Although are fertile (Tønsberg 1992), and the apothecia are there are some species with overlapping ranges black, mostly flat, with a thick outer rim and that may be confused with B. griseovirens when contain submuriform spores. When sterile the identifications are based solely on morphology, primary diagnostic characters for B. griseovirens the chemistry of this species is distinctive. In the are the presence of both norstictic acid and future, more detailed documentation of the atranorin together with the production of dark variability of the chemistry of this species, and colored soredia in immersed soralia. Without the the range of its chemotypes would improve the aforementioned compounds it would be difficult, ease with which it can be identified. if not impossible, to distinguish sterile B. griseovirens from other sterile crustose lichens Collections examined: Canada: British with either atranorin or norstictic acid (Nordin Columbia: Greater Vancouver Regional District, 2001). Lynn Headwaters Regional Park, E side Lynn Creek, ~4 km N of terminus of Lynn Creek Rd., J. The presence or absence of norstictic acid has C. Lendemer 22325 & A. Moroz (NY); Ontario: been widely used to delimit species in the Bruce Co., Bruce Peninsula National Park, Caliciaceae (Marbach 2000) and the results of Shingle Marsh along road to Halfway Log Dump, this study suggest this approach may not always 2008, J. C. Lendemer 14322A (NY); Bruce result in accurate taxonomic circumscription of Peninsula National Park, Singing Sands, The monophyletic entities. Rather, our data affirm the Dorcas Alvar, J. C. Lendemer 14330 (NY); circumscriptions proposed by previous authors Fathom Five National Marine Park, Flowerpot that have allowed for chemical variability, and Island, Flowerpot Loop Trail, 2008, J. C. specifically the presence or absence of norstictic Lendemer 14587 (NY); Bruce Alvar Nature acid, within a given taxon in Buellia s.l. (Harris Reserve, NW of Hwy 6, 0.4 km N of Dyer's Bay 1995, Nordin 2001, Scheidegger 1993, Tønsberg Road, 2010, W. R. Buck 56613 (NY); Bruce 1992). This study is a further illustration that Peninsula National Park, Halfway Log Dump on molecular data can be a powerful tool for those Georgian Bay, 2008, J. C. Lendemer 14132A(NY); attempting to elucidate the higher-level Yukon Territory: W-slope of Grey Mountain, placement and species-level identification of Money Shot Trail, 2 km below otherwise unidentifiable lichens (Hodkinson & telecommunication towers, 9 km SE of Lendemer 2012, Lendemer & Lumbsch 2008, Whitehorse in Grey Mountain Rd., 2011, J. C. Nelsen et al. 2012, Vondrák et al. 2010). In this Lendemer 28474, 28500 (NY); E side of Klondike case, without molecular data, we would neither Hwy. (Yukon 2), 3.5 km S of Carcross, 2011, J. C. 6 Allen & Lendemer. Lichen chemical variability. North American Fungi 8(12): 1-14

Lendemer 28952 (NY); N side of Wheaton River, Lendemer 27586 & M. Sundue (NY). Sweden: S of Annie Lake Rd., at km post 24, 24 km W of Närke Province: Svennevad, Södra Berg., 1952, G. jct w/ Klondike Hwy. (Yukon 2), 2011, J. C. Kjellmert; Svennevad, Zettermanstorp, 1951, G. Lendemer 28771 (NY); bluffs above N shore of Kjellmert (NY). Värmland Province: Örebro Co., Fish Creek, 1.4 km W of jct. of Fish Creek Rod. Karlskoga par., Lerängen, 1981, L. E. Muhr 4039 And Alaska Highway (Yukon 1), ~4 km NW of (NY). Whitehorse, 2011, J. C. Lendemer 28530 (NY); N side of Alaska Highway (Yukon 1) 0.5 km N of Acknowledgements: turnoff for Swan Haven Dr./McClintock We thank Brendan Hodkinson for helpful Subdivision and crossing over the McClintock discussion, especially with regards to River, 2011, J. C. Lendemer 28586, 28628 (NY) – phylogenetic analyses and Richard Harris for USA: California: Mariposa Co., Yosemite discussion and comments on the manuscript. We National Park, 1 km E of Hodgdon Meadow on are also grateful to Tør Tonsberg and Bruce Old Big Oak Flat Rd. where it crosses Hazel McCune for helpful reviews. This study is a result Green Creek, 2009, J. C. Lendemer 19550 (NY); of work conducted by the authors while Yosemite National Park, bottom of Tuolumne supported by NSF DEB-1145511. Additional Grove, North Crane Creek, 2009, J. C. Lendemer support for the first author was provided by The 19575A (NY); Sierra Nevada Mountains, City University of New York and The New York Yosemite National Park, on slope above Merced Botanical Garden. Lendemer’s fieldwork in the River near Pohono Bridge, El Capitan Quad., Yukon Territory was funded by NSF DEB- 2000, K. Knudsen 11601 (NY). Massachusetts: 1110433 and conducted under the auspices of Berkshire Co., Town of Adams, Mt. Greylock Yukon-Canada Scientists and Explorers Act State Reservation, W slope of Mt. Fitch along License #11-16S&E. The second author extends Notch Road, W. R. Buck 27773 (NY). New York: special thanks to Jennifer Staniforth of the Yukon Essex Co., Low east-facing slopes of Mt. Jo, Ministry of the Environment for her logistical northeast of Heart Lake, J. C. Lendemer 3086 help and hospitality. (NY). Oregon: Douglas Co., The Dell. Rocky slopes of conglomerate and sandstone, with Literature cited: Pseudotsuga and Arbutus menziesii, 2009, B. Akaike, H. 1973. Information theory and an McCune 29912 (OSC); On BLM road 29-7-3.0, extension of the maximum likelihood principle. Coast Range SW of Winston, 2010, B. McCune in Procreeding of the 2nd International 30976 (OSC); Jefferson Co., Above Simtustus Symposium on Information Theory (B.N. Petrov Reservoir on Deschutes River, 2000, B. McCune & F. Csaki, eds): Akademiai Kiado: Budapest, 25287 (OSC); Lincoln Co., north of Drift Creek, Hungary. on Forest Road 1928, Coast Range, Siuslaw National Forest, 2011, B. McCune 31566 (OSC); Allen, J. L., B. P. Hodkinson, and C. R. Björk. Wallowa Co., Pseudotsuga forest, trale to Ice 2012. A major range expansion for Platismatia Lake, Wallowa Mountains, Eagle Cap Wilderness, wheeleri. North American Fungi 7(10): 1-12. 2005, B. McCune 27992, 27989 (OSC); Wheeler http://dx.doi: 10.2509/naf2012.007.010 Co., Abies grandis forest on bench in valley bottom, Honeymoon Creek Camp, Black Canyon Altschul, S., T. L. Madden, A. A. Schaffer, J. Wilderness, tributary of South Fork John Day Zhang, Z. Zhang, W. Miller and D. J. Lipman. River, Ochocho Mountains, 2006, B. McCune 1997. Gapped BLAST and PSI-BLAST: a new 28286 (OSC). Vermont: Caledonia Co., Town of generation of protein database search programs. Wheelock, Wheelock Farm, NW shore of Flagg Nucleic Acids Research 25: 3389–3402. Pond, S of W Wheelock Rd., 2010, J. C. http://dx.doi.org/10.1093/nar/25.17.3389 Allen & Lendemer. Lichen chemical variability North American Fungi 8(12): 1-14 7

Arup, U., and M. Grube. 1999. Where does lichen. Systematic Botany 37: 835-844. Lecanora demissa (Ascomycota, Lecanorales) http://dx.doi.org/10.1600/036364412X656536 belong? Lichenologist 31: 419-430. http://dx.doi.org/10.1017/S0024282999000584 Holien, H. and T. Tønsberg 2008. Xylographa soralifera, a new species in the X. vitiligo Culberson, C.F. and H. Kristinsson. 1970. A complex. Graphis Scripta 20(2): 58-63. standardized method for the identification of lichen products. Journal of Chromatography 46: Katoh, K., K. Kuma, H. Toh and T. Miyata. 2004. 85–93. MAFFT version 5: improvement in accuracy of multiple sequence alignment. Nucleic Acids Elix, J. A. and T. Tønsberg. 1999. Notes on the Research 33: 511-518. chemistry of some lichens from Norway. Graphis http://dx.doi.org/10.1093/nar/gki198 Scripta 10: 4-6 Lendemer, J. C. in press. Thelomma ocellatum, a Felsenstein, J. 1985. Phylogenies and the range extension to the Yukon Territory and case comparative method. The American Naturalist study in the use of molecular data to recognize 125: 1-15. asexually reproducing crustose lichens. North American Fungi. Fryday, A. and B. Coppins. 1997. Keys to sterile, crustose saxicolous and terricolous lichens Lendemer, J. C. and H. T. Lumbsch. 2008. occurring in the British Isles. The Lichenologist Protoparmelia capitata sp. nov., and P. isidiata 29: 301-332. Diederich, Aptroot & Sérus., two species of http://dx.doi.org/10.1006/lich.1997.0080 Protoparmelia (Lecanorales, Ascomycota) from south-eastern North America. The Lichenologist Gaya, E., F. Högnabba, Á. Holguin, K. Molnar, S. 40: 329-336. Fernández-Brime, S. Stenroos, U. Arup, U. http://dx.doi.org/10.1017/S0024282908007810 Søchting, P. Van den Boom, R. Lücking, H. J. M. Sipman, and F. Lutzoni. 2012. Implementing a Lücking, R., B. P. Hodkinson, A. Stamatakis and cumulative supermatrix approach for a R. A. Cartwright. 2011. PICS-Ord: unlimited comprehensive phylogenetic study of the coding of ambiguous regions by pairwise identity Teloschistales (Pezizomycotina, Ascomycota). and cost ordination. BMC Bioinformatics 12: 10. Molecular Phylogenetics and Evolution 63: 374- http://dx.doi.org/10.1186/1471-2105-12-10 387. http://dx.doi.org/10.1016/j.ympev.2012.01.012 Maddison, W. P. and D. R. Maddison. 2011. Mesquite: a modular system for evolutionary Harris, R. C. 1995. More Florida lichens. analysis. Version 2.75 Including a 10¢ tour of the Pyrenolichens. Publ. http://mesquiteproject.org. by the Author: Bronx, NY. Marbach, B. 2000. Corticole und lignicole Arten Hodkinson, B. P. and J. C. Lendemer. 2011. der Flechtengattung Buellia sensu lato in den Molecular analyses reveal semi-cryptic species in Subtropen und Tropen. in Bibliotheca Xanthoparemlia tasmanica. Bibliotheca Lichenologica 74 J. Cramer, ed.: 384. Berlin: Lichenologica 106: 115-126. Stuttgart.

Hodkinson, B. P. and J. C. Lendemer. 2012. Miadlikowska, J., B. McCune and F. Lutzoni. Phylogeny and taxonomy of an enigmatic sterile 2002. Pseudocyphellaria perpetua, a new lichen 8 Allen & Lendemer. Lichen chemical variability. North American Fungi 8(12): 1-14

from Western North America. The Bryologist Scheidegger, C. 1993. A revision of European 105: 1–10. saxicolous species of the genus Buellia De Not. And formerly included genera. The Lichenologist Nelsen, M. P., R. Lücking, C. J. Andrew, E. Rivas 25: 315-364. Plata, J. L. Chaves, M. E. S. Cáceres and N. http://dx.doi.org/10.1006/lich.1993.1001 Ventura. 2012. Dismantling Herpothallon: Herpothallon antillarum (Arthoniomycetes: Stamatakis, A., P. Hoover and J. Rougemont. Arthoniaceae) is a member of the genus 2008. A Fast Bootstrapping Algorithm for the Diorygma (Lecanoromycetes: Graphidaceae). RAxML Web-Servers. Systematic Biology 57: The Bryologist 115: 313-321. 758-771. http://dx.doi.org/10.1639/0007-2745-115.2.313 http://dx.doi.org/10.1080/10635150802429642

Nordin, A. 1996. Buellia species (Physiaceae) Swofford, D. 2012. PAUP* 4.0d125. Private with pluriseptate spores in Norden. Acta author distribution, Duke University. Universitatis Upsaliensis. Symbolae botanicae Tibell, L. 1976. The genus Thelomma. Botaniska Upsaliensis 31: 327-354. Notiser 129: 221-249.

Nordin, A. 2001. Buellia species with pluriseptate Tibell, L. 2003. Tholurna dissimilis and generic spores and the Physciaceae (Lecanorales, delimitation in Caliciaceae inferred from nuclear Ascomycotina): taxonomic, phylogenetic, and ITS and LSU rDNA phylogenies (Lecanorales, ultrastructural studies. Ph.D. thesis, lichenized ascomycetes). Mycological Research Comprehensive Summaries of Uppsala 107: 1403-1418. Disertations from the Faculty of Science and http://dx.doi.org/10.1017/S0953756203008694 Technology 593. Tønsberg, T. 1992. The sorediate and isidiate, Nylander, J. A. A. 2004. MrModeltest v2. coritcolous, crustose lichens in Norway. Program distributed by the author. Evolutionary Sommerfeltia 14: 1-331. Biology Centre: Uppsala University. Vondrák, J., O. Redchenko, D. Himelbrant, I. Ronquist, F. and J. P. Huelsenbeck. 2003. Stepanchikova, and E. Kuznetsova. 2010. Some MrBayes 3: Bayesian Phylogenetic Inference sterile Caloplaca crusts identified by molecular under mixed models. Bioinformatics 19: 1572- data from the Leningrad region (Russia). Folia 1574. Cryptogamica Estonica 47: 97-99. http://dx.doi.org/10.1093/bioinformatics/btg18 0 Wetmore, C.M. 2004. The sorediate corticolous species of Caloplaca in North and Central Schoch, C. L., K. A. Siefert, S. Huhndorf, V. America. The Bryologist 107(4): 505-520. Robert, J. L. Spouge, C. A. Levesque, W. Chen Williams, C. B., and L. Tibell. 2008. Calicium and Fungal Barcoding Consortium. 2012. Nuclear sequoia, a new lichen species from north-western ribosomal internal transcribed spacer ITS) region California, USA. The Lichenologist 40: 185-194. as a universal DNA barcode marker for Fungi. http://dx.doi.org/10.1017/S0024282908007615 Proceedings of the National Academy of Sciences http://dx.doi.org/10.1073/pnas.1117018109

Allen & Lendemer. Lichen chemical variability North American Fungi 8(12): 1-14 9

Fig. 1. Phylogeny of Buellia griseovirens and its relatives based on ITS sequence data, inferred using ML, and presented with midpoint rooting. Branch support is shown as MP-BP/ML-BP/B-PP. Bolded branches have MP-BP and ML-BP >70 and B-PP >0.95. The shaded samples of B. griseovirens contain norstictic acid and atranorin, while the other samples contain only atranorin.

10 Allen & Lendemer. Lichen chemical variability. North American Fungi 8(12): 1-14

Fig. 2. Distribution of B. griseovirens in North America based on specimens examined from NY and McCune, as well as records from the Consortium of North American Lichen Herbaria (CNALH) website. Blue = vouchers with only atranorin, red = vouchers with both atranorin and norstictic acid, pink = vouchers with norstictic acid only, and yellow = records in the CNALH.

Allen & Lendemer. Lichen chemical variability North American Fungi 8(12): 1-14 11

Fig. 3. Light micrographs of Buellia griseovirens. A. Overall morphology, and B. soralia of a specimen with atranorin only collected from the Yukon (J. C. Lendemer 28628). C. Overall morphology and D. soralia of a specimen with norstictic acid and atranorin (J. C. Lendemer 3086). (Scale bars: 1.0 mm in A and C, 0.5 mm in B and D).

12 Allen & Lendemer. Lichen chemical variability. North American Fungi 8(12): 1-14

Table 1. Accession numbers of sequences generated for this study. All sequences are from specimens of Buellia griseovirens collected from the Yukon Territory, Canada.

ITS Isolate mtSSU Genbank Collection Chemistry No. Genbank No. No. J. C. Lendemer atranorin, norstictic NY1304 KC681816 KC681811 28530 acid J. C. Lendemer NY1305 KC686817 KC681812 atranorin 28952 J. C. Lendemer NY1314 KC681818 KC681813 atranorin 28771 J. C. Lendemer NY1315 KC681819 KC681814 atranorin 28500 J. C. Lendemer NY1316 KC681820 KC681815 atranorin 28474

Allen & Lendemer. Lichen chemical variability North American Fungi 8(12): 1-14 13

Appendix 1. Phylogeny of the Caliciaceae based on mtSSU sequence data inferred using ML with members of the Calicioideae selected as the outgroup following Gaya et al. (2012). Branch support is shown as bootstrap proportions (BP) recovered from ML (ML-BP) and posterior probabilities (PP) recovered by BI (BI-PP). Bolded branches have ML-BP >70 and B-PP >0.95. The shaded sample of B. griseovirens contains norstictic acid and atranorin while the other samples contain only atranorin. Note that the taxonomic identity of the sample used to generate the sequences of Diplotomma epipolium used here is unclear (Nordin 1996).

14 Allen & Lendemer. Lichen chemical variability. North American Fungi 8(12): 1-14

Appendix 2. Phylogeny of the Caliciaceae based on mtSSU sequence data inferred using maximum parsimony with Calicioideae selected as the outgroup following Gaya et al. (2012). Numbers above branches on the left are bootstrap proportions (BP) recovered from maximum parsimony analysis (MP- BP). Bolded branches have MP-BP >70.