How well do ITS rDNA sequences differentiatespecies of true morels (Morchella)?

Xi-Hui DuQi ZhaoZhu L. Yang

Key Laboratory of Biodiversity and Biogeography,Kunming Institute of Botany, Chinese Academy ofSciences, Lanhei Road No. 132, Kunming 650201,Yunnan Province, P.R. China

Graduate University of Chinese Academy of Sciences,Beijing 100049, P.R. China

Karen HansenCryptogamic Botany, Swedish Museum of NaturalHistory, P.O. Box 50007, SE-10405 Stockholm, Sweden

Hatıra TaskınSaadet Buyukalaca

Department of Horticulture, Faculty of Agriculture,Cukurova University, Adana 01330, Turkey

Damon DewsburyJean-Marc Moncalvo

Department of Ecology and Evolutionary Biology,University of Toronto, Toronto, Ontario M5S2C6,Canada

Department of Natural History, Royal OntarioMuseum, and Department of Ecology and EvolutionaryBiology, University of Toronto, Toronto,Ontario M5S2C6, Canada

Greg W. DouhanDepartment of Plant Pathology and Microbiology,University of California Riverside, Riverside,California 92521

Vincent A.R.G. RobertPedro W. Crous

CBS-KNAW Fungal Biodiversity Center, Utrecht,the Netherlands

Stephen A. RehnerSystematic Mycology and Microbiology Laboratory,United States Department of Agriculture, AgriculturalResearch Service, Beltsville, Maryland 20705

Alejandro P. RooneyCrop Bioprotection Research Unit, National Center forAgricultural Utilization Research, United StatesDepartment of Agriculture, Agricultural ResearchService, 1815 North University Street, Peoria,Illinois 61604

Stacy SinkKerry O’Donnell1

Bacterial Foodborne Pathogens and Mycology ResearchUnit, National Center for Agricultural Utilization

Research, United States Department of Agriculture,Agricultural Research Service, 1815 North UniversityStreet, Peoria, Illinois 61604

Abstract: Arguably more mycophiles hunt truemorels (Morchella) during their brief fruiting seasoneach spring in the northern hemisphere than anyother wild edible fungus. Concerns about overhar-vesting by individual collectors and commercialenterprises make it essential that science-basedmanagement practices and conservation policies aredeveloped to ensure the sustainability of commercialharvests and to protect and preserve morel speciesdiversity. Therefore, the primary objectives of thepresent study were to: (i) investigate the utility of theITS rDNA locus for identifying Morchella species,using phylogenetic species previously inferred frommultilocus DNA sequence data as a reference; and (ii)clarify insufficiently identified sequences and deter-mine whether the named sequences in GenBank wereidentified correctly. To this end, we generated 553Morchella ITS rDNA sequences and downloaded 312additional ones generated by other researchers fromGenBank using emerencia and analyzed them phylo-genetically. Three major findings emerged: (i) ITSrDNA sequences were useful in identifying 48/62(77.4%) of the known phylospecies; however, theyfailed to identify 12 of the 22 species within thespecies-rich Elata Subclade and two closely relatedspecies in the Esculenta Clade; (ii) at least 66% of thenamed Morchella sequences in GenBank are misiden-tified; and (iii) ITS rDNA sequences of up to sixputatively novel Morchella species were represented inGenBank. Recognizing the need for a dedicated Web-accessible reference database to facilitate the rapididentification of known and novel species, weconstructed Morchella MLST (http://www.cbs.knaw.nl/morchella/), which can be queried with ITS rDNAsequences and those of the four other genes used inour prior multilocus molecular systematic studies ofthis charismatic genus.

Key words: Ascomycota, biodiversity, biogeogra-phy, emerencia, GCPSR, GenBank, phylogeny, specieslimits

INTRODUCTION

With the notable exception of true truffles in thegenus Tuber (Bonito et al. 2010), few fungal genera are

Submitted 17 Feb 2012; accepted for publication 16 May 2012.1 Corresponding author. E-mail: kerry.odonnell@ars.usda.gov

Mycologia, 104(6), 2012, pp. 1351–1368. DOI: 10.3852/12-056# 2012 by The Mycological Society of America, Lawrence, KS 66044-8897

1351

as synonymous with epicurean cuisine as true morels(Morchella, phylum Ascomycota). Literally thousandsof mycophiles scour mixed hardwood and coniferousforests throughout the northern hemisphere in searchof these highly prized edible fungi during their shortand often sporadic fruiting season each spring. Addingto their huge popularity is the dramatic growth ofcommercial harvesting of morels into highly lucrativecottage industries in several morel-rich countries,including China, Turkey and the United States (Pilzet al. 2007), together with significant advances towardtheir cultivation on a commercial scale (Ower et al.1986, Zhao et al. 2009).

Due to the increasing commercial demand, it isimperative that comprehensive studies are conductedto develop informed science-based conservationpolicies and management practices that will ensurethat species diversity is maximally preserved andcommercial harvests are sustainable (see Pilz et al.2007 and references therein). To this end, ourmultilocus molecular phylogenetic analyses haveidentified at least 62 phylogenetically distinct speciesworldwide based on their genealogical exclusivity (Duet al. 2012; O’Donnell et al. 2011; Taskın et al. 2010,2012). Robust hypotheses of their species limits,employing genealogical concordance/discordancephylogenetic species recognition (GCPSR; Dettmanet al. 2003, Taylor et al. 2000), provides the requisiteframework for assessing the utility of ITS rDNA foridentifying species so that their geographic distribu-tion and area of endemism can be critically assessedand so that unknown taxa can be detected andidentified for further taxonomic evaluation.

Even though the ITS rDNA region has been used asthe sole locus in more studies than any other forassessing Morchella genetic diversity (Pagliaccia et al.2011, Stefani et al. 2010, Taskın et al. 2012 andreferences therein), a search of GenBank usingemerencia (Ryberg et al. 2009) revealed that mostsequences were insufficiently identified (n 5 190),and in the near absence of type studies, binomialsappear to have been arbitrarily applied to the vastmajority of those designated fully identified (n 5 127)sensu Nilsson et al. (2005). Thus, the present study ofMorchella was conducted to: (i) determine how wellITS rDNA sequence data resolve species we previouslydelimited via GCPSR, (ii) assess whether the fullyidentified sequences in GenBank were identifiedcorrectly and identify ones classified as insufficientlyidentified, (iii) mine metadata to better understandspecies diversity and their geographic distributionsand (iv) develop a dedicated, Web-accessible infor-mational database, including references to voucherspecimens and/or cultures (Agerer et al. 2000), thataids DNA sequence-based identifications of true

morels via the Internet. We have accomplished thelatter objective by constructing Morchella MLST(multilocus sequence typing, http://www.cbs.knaw.nl/morchella/) and populated it with the data wegenerated both in the current and our prior multi-locus molecular systematic studies of this charismaticgenus.

MATERIALS AND METHODS

Source of sequence data.—We generated ITS sequence datafrom 553 collections during extensive surveys of China (Duet al. 2012), Turkey (Taskın et al. 2010, 2012) and a globalsurvey of the genus (O’Donnell et al. 2011). In addition,312 sequences were retrieved from GenBank with the genussearch tool in emerencia (Nilsson et al. 2005, Ryberg et al.2009). Five additional sequences were retrieved by emer-encia, including three of M. rufobrunnea, although one ofthese (GQ304946) was misidentified as M. crassipes, possi-bly because the pileus of this species turns yellow dur-ing ascosporogenesis. The other two GenBank accessions(EU480151.1, U61390.1) were unidentifiable sequencesunrelated to Morchella based on BLAST queries ofGenBank. Sequences of M. rufobrunnea were not analyzedfurther because they represent a highly divergent basalsister to the rest of Morchella. Thus, a total of 865 MorchellaITS rDNA nucleotide sequences were analyzed.

Molecular methods.—Total genomic DNA was extractedfrom herbarium specimens or freeze-dried myceliumcultivated from pure cultures using a CTAB protocol(Gardes and Bruns 1993, O’Donnell et al. 2011). PCRamplification and sequencing of the nuclear ribosomaltranscribed spacer region (ITS rDNA) was conducted usingthe ITS5 3 ITS4, ITS1 3 ITS4 (White et al. 1990) or ITS5 3

NL4 (O’Donnell et al. 2011) primer pairs. To completelysequence the latter amplicon ITS4 and NL1 also were usedas internal sequencing primers (O’Donnell et al. 2011). AllPCR amplifications were performed with Platinum TaqDNA Polymerase High Fidelity (Invitrogen Life Technolo-gies, Carlsbad, California), after which amplicons werepurified with Montage96 filter plates (Millipore Corp.,Bellerica, Massachusetts), and then sequenced with ABIBigDye chemistry 3.1 (Applied Biosystems). ABI XTermi-nator was used to purify the sequencing reactions beforerunning them on an ABI 3730 genetic analyzer as describedin O’Donnell et al. (2011).

Sequence analyses.—The Morchella ITS rDNA sequences wegenerated (n 5 553) and those downloaded from GenBank(n 5 312) were too divergent to align across the breadth ofthe genus. Therefore, the following three clade-specificalignments were constructed: (i) Esculenta Clade (n 5 222:164 new + 58 from GenBank, SUPPLEMENTARY TABLE I), (ii)Elata Clade (n 5 200: 110 new + 90 from GenBank,SUPPLEMENTARY TABLE I), and (iii) Elata Subclade (n 5 443:348 + 164 from GenBank by others, SUPPLEMENTARY TABLE

III). A fourth clade of highly divergent M. rufobrunneasequences (n 516: 13 new + three from GenBank,SUPPLEMENTARY TABLE IV) also was identified. However,

1352 MYCOLOGIA

because these sequences were highly divergent from theaforementioned three clades, they were excluded fromfurther analysis. Sequence chromatograms generated inthis study were imported into Sequencher 4.9 (Gene CodesCorp., Ann Arbor, Michigan), where they were edited,aligned and exported as NEXUS files. Assignment of theMorchella sequences downloaded from GenBank usingemerencia (Nilsson et al. 2005) to one of the fouraforementioned clades was based on an initial maximumparsimony analysis conducted in PAUP* 4.0b10 (Swofford2002). Sequences in each of the three clade-specific datasetswere aligned with MUSCLE 3.8 (Edgar 2004) in SeaView4.3.0 (http://pbil.univ-lyon1.fr/software/seaview) and thenimprovements in the alignments were made visually withTextPad 5.1.0 for Windows (http://textpad.com/). Toimprove phylogenetic resolution sequences of M. tomentosa(Mel-1) and M. steppicola (Mes-1), the earliest divergingspecies within the Elata and Esculenta clades respectively,were used to root these phylogenies. Sequences of Mel-11and Mel-25, the sister group of the Elata Subclade (Du et al.2012; O’Donnell et al. 2011; Taskın et al. 2010, 2012), wereused to root this phylogeny.

In addition to excluding two unidentifiable highlydivergent sequences from the emerencia search of GenBank(EU480151.1, U61390.1), the 59 or 39 ends of threesequences obtained from GenBank were trimmed toremove poor sequence (93 bp from 59 end of FJ237237,194 bp from 39 end of GQ223457, 101 bp from 59 end ofEU834821). To construct partitions that only containedunique haplotypes for the final analyses, COLLAPSE 1.2(http://darwin.uvigo.es/software/collapse) was used toidentify haplotypes within each dataset. Instead of usingpercent identity to place unidentified sequences withinsimilarity clusters (Bonito et al. 2010, Horton and Bruns2001), haplotypes were identified to species, where possible,by including ITS rDNA sequences of the phylogeneticspecies identified with multilocus DNA sequence data (Duet al. 2012; O’Donnell et al. 2011; Taskın et al. 2010, 2012).Each partition was analyzed via maximum parsimony(MP) in PAUP* (Swofford 2002) to obtain tree statistics,including number of parsimony informative characters(PIC); clade support was assessed based on 1000 MPpseudoreplicates of the data (O’Donnell et al. 2011). Dueto the large number of MPTs, single phylograms werechosen randomly to present the metadata and cladesupport for the Esculenta (FIG. 1) and Elata Clades(FIG. 2). However, due to the large number of ITShaplotypes, a 70% majority rule bootstrap consensus waschosen to present the Elata Subclade phylogeny (FIG. 3).

All DNA sequence data generated in this study wasdeposited in GenBank (JQ723016–JQ723151), TreeBASE(S12489, Tr51345-Tr51347) and Morchella MLST at CBS-KNAW (http://www.cbs.knaw.nl/morchella/). In addition,accession numbers identifying voucher specimens deposit-ed in herbaria and cultures are listed (SUPPLEMENTARY

TABLES I–IV).

Construction of Morchella MLST.—This database wascreated with the BioloMICS software (Robert et al. 2011).All available molecular, specimen and/or culture data were

imported into the database. The database can be queried bykeywords or using one of the two available identificationtools. The first one consists of a modified version of BLAST(Altschul et al. 1990). Single sequences can be alignedagainst the reference database and ranked according toa combination of parameters including similarity, BLASTscore, overlap percentage between query and referencesequences and number of fragments. The second tool iscalled polyphasic identification because it allows aligning(pairwise alignments) one or several sequences fromdifferent loci at the same time. A global similarity coefficientis computed with this formula (Gower 1971):

Sjk~1

Pn

i~1Wi

Xn

i~1

Si Wi

Where Sjk is the global similarity coefficient for the

comparison of records j and k; Si is the local similarity

coefficient for the pairwise alignment of locus I; Wi is the

weight of locus I (no differential weighting is applied in the

Morchella database); n is the number of loci in the similarity

comparison. Results are ranked according to the combina-

tion of this global similarity coefficient, the number of loci

and the overlap percentage between query and reference

sequences. Phenetic trees based on one of the following

methods can be produced dynamically to help end users

locate the clade to which the unknown specimen belongs:

UPGMA, WPGMA, single linkage, complete linkage,

UPGMC, WPGMC, ward or neighbor joining using a fixed

model (Saitou and Nei 1987).

RESULTS

Esculenta clade.—The 222-sequence Esculenta Cladedataset included 58 sequences we deposited inGenBank (O’Donnell et al. 2011; Taskın et al. 2010,2012) and 164 collected in the present study(SUPPLEMENTARY TABLE I). COLLAPSE 1.2 analysis ofthe 222-sequence dataset (1349 bp alignment) iden-tified 125 unique haplotypes. The latter dataset wasused for all subsequent analyses, after 253 nucleotidepositions within the ITS1 were excluded as ambigu-ously aligned. Maximum parsimony (MP) analysis ofthe 125-sequence Esculenta Clade dataset recovered. 10 000 equally most parsimonious trees (MPTs)1045 steps long (FIG. 1, consistency index 5 0.67,retention index 5 0.96). The Esculenta Cladephylogeny was rooted on sequences of M. steppicolaMes-1 based on more inclusive analyses (O’Donnellet al. 2011). Of the 27 species previously identifiedwith multilocus GCPSR, we were able to apply nameswith confidence only to M. steppicola and five speciesendemic to North America (FIG. 1, Kuo et al. 2012).Parsimony analysis of the Esculenta Clade datasetindicated that 212/222 (95.5%) of ITS sequences

DU ET AL.: MORCHELLA ITS RDNA PHYLOGENETICS 1353

FIG. 1. One of . 10000 equally most parsimonious trees (MPTs) inferred from Esculenta Clade (yellow morels) ITS sequences,rooted on sequences of Morchella steppicola (Mes-1), the basal-most member of this clade (O’Donnell et al. 2011). The phylogram containsthe 125 unique haplotypes identified within the 222-sequence dataset (SUPPLEMENTARY TABLE I). Each sequence is identified by GenBankor herbarium accession number or laboratory code (SUPPLEMENTARY TABLE I), geographic origin, name used as deposited in GenBankand number of sequences in parentheses that comprise each haplotype represented by two or more sequences. ITS sequences identifiedby an asterisk were used as a reference for each phylospecies based on prior multilocus GCPSR analysis (Du et al. 2012; O’Donnell et al.2011; Taskın et al. 2010, 2012). Because binomials can be applied with confidence to only six of the species, each phylogenetic species (i.e.phylospecies) is identified by Mes followed by a unique Arabic number. Numbers to the right of the phylospecies identify the size andnumber of sequences (in subscript) of the (ITS1) (5.8SrDNA) (ITS2) (complete ITS region) determined by DNA sequence analysis. Notethat ‘‘s’’ indicates the sequence was short and ‘‘?’’ indicates that the species and/or ITS length is unknown. The number above internodesrepresents maximum parsimony bootstrap support $ 70% based on 1000 pseudoreplicates of the data. Phylospecies Mes-23 and Mes-24cannot be distinguished with ITS data. Mes-27 from Sichuan Province, China, is featured.

1354 MYCOLOGIA

could be identified to species. The exception involvedGenBank sequences of nine collections from Chinaand one from India, none of which were full length(SUPPLEMENTARY TABLE I, FIG. 1). In analyses of theEsculenta Clade dataset, five collections of Mes-23 orMes-24 from China could not be distinguished whenambiguously aligned regions were included or exclud-ed. In addition, three collections from Xingiang andone from Jilin, which we listed as Mes-?, could be

Mes-21. As with the aforementioned unidentifiedcollections, additional data is needed to determinethe identity of one collection from India listed as Mes-?.Six of the species (SUPPLEMENTARY TABLE I; i.e. M.diminutiva Mes-2, M. esculentoides Mes-4, Mes-6, Mes-8,Mes-9 and Mes-16) were represented by 15–42 sequenc-es and comprised 132/222 (59.5%) of Esculenta Cladesequences. Of the 27 Esculenta Clade species, M.esculentoides, the most common yellow morel endemic

FIG. 1. Continued.

DU ET AL.: MORCHELLA ITS RDNA PHYLOGENETICS 1355

FIG. 2. Phylogenetic relationships within Elata Clade (black morels) inferred by MP analysis of 60 unique ITS haplotypesidentified within the 200-sequence dataset (SUPPLEMENTARY TABLE II). The phylogram is one of . 10 000 MPTs 373 steps long,rooted on sequences of Morchella tomentosa (Mel-1). Haplotype sequences are identified by a herbarium or GenBank accessionnumber or laboratory number (SUPPLEMENTARY TABLE II), geographic origin, the name used for the deposit in GenBank andthe number of sequences (in parentheses) that comprise each non-singleton haplotype. An asterisk identifies ITS sequencesthat were used as a reference for each phylogenetic species (i.e. phylospecies) previously resolved by multilocus GCPSR data.MP bootstrap support is identified above internodes when it was $ 70%. Nine of the species can be identified by binomials; the10 known phylospecies (O’Donnell et al. 2011) are identified by Mel followed by a unique Arabic number. Two putativelynovel species represented by ITS sequences in GenBank are identified by NEW-1? and NEW-2? The size and number ofsequences of the (ITS1) (5.8S rDNA) (ITS2) (entire ITS region) analyzed is indicated to the right of each species. Note thatITS sequences for M. populiphila (Mel-5) and NEW-1? from China are not full length. Mel-10 from Yunnan Province, China,is featured.

1356 MYCOLOGIA

to North America, was represented by the mostsequences (n 5 42). By contrast, we were unable toobtain ITS sequence data from collections of theEuropean endemic Mes-5, the only known phylogenet-ic species missing from the Morchella ITS dataset.

With ITS rDNA sequences of species identifiedpreviously via multilocus GCPSR as a reference,separate analyses using maximum parsimony andCOLLAPSE 1.2 revealed that 58 Esculenta Cladesequences deposited in GenBank by other researcherscomprised at least 10 species (SUPPLEMENTARY TABLE

I). Three of the 58 accessions in GenBank were listedas Morchella sp., with the remaining 55 reported as M.crassipes (n 5 29), M. esculenta (n 5 16), M. spongiola(n 5 9) and M. vulgaris (n 5 1). Our results,however, indicate the former three names have beenmisapplied broadly given that sequences of eightspecies were deposited in GenBank as M. esculentaand M. crassipes and four as M. spongiola (SUPPLE-

MENTARY TABLE I). All species represented by sequenc-es deposited under the latter three names are nestedwithin the Esculenta Clade, whereas those listed as M.esculenta included three species within the EsculentaClade, two within the Elata Clade and three within theElata Subclade (SUPPLEMENTARY TABLES I–III).

Furthermore, our analyses indicate that, at best,only 18/55 (32.7%) of the Esculenta Clade sequencesin GenBank are correctly identified to species. Wefound that 12 of the 16 accessions deposited as M.esculenta are the putative North American endemicM. esculentoides (Mes-4). Moreover, it is plausiblethat some of the 18 sequences that we assumed to becorrect are actually misidentified. For these sequencesto be identified correctly the sequences of Mes-16(n 5 10), Mes-17 (n 5 7) and M. esculentoides (Mes-4,n 5 1) respectively would have to be authentic for M.crassipes, M. spongiola and M. vulgaris. Studies oforiginal material, epi- or neo-typification of recentmaterial and additional sampling of Morchella inEurope are needed to test whether this assumption iscorrect. We were able to obtain complete sequencesof the ITS rDNA region for 24 of the 27 species withinthe Esculenta Clade in the present study. In additionto missing ITS rDNA data from the Europeanendemic Mes-5, the ITS1 and ITS2 were sequencedonly partially in GenBank accession GQ228465 fromIndia, the ITS1 was missing in Mes-18 from Turkeyand it was represented by a partial sequence in Mes-20from Yunnan Province, China. Although six sequenc-es downloaded from GenBank possessed insertions(1–11 bp) and deletions (1–7 bp) within the 5.8SrDNA (SUPPLEMENTARY TABLE V), these indels mayrepresent sequencing errors because the length ofthis gene appears to be fixed at 156 bp in Morchellabased on the sequences we generated from 663

collections. However, without the ability to rese-quence the individuals from which these sequenceswere generated, we cannot discount the possibilitythat they represent rare variants within their popula-tions of origin.

Analyses of the present dataset have highlightedseveral noteworthy features of the ITS1 and ITS2spacers within the Esculenta Clade. Mapping thesize of the ITS regions on the molecular phylogeny(FIG. 1) revealed that the ITS1 more than doubled inlength during the evolutionary diversification of thisclade; the two basal-most lineages possessed theshortest ITS1 spacer (285 bp in M. steppicola Mes-1and 285–286 bp in M. virginiana Mes-3), whereas itwas largest in the two most derived species (610–611 bp in Mes-8 and 617 bp in Mes-9). By contrast, theITS2 was remarkably conserved in length within theEsculenta Clade in that the shortest (371 bp in Mes-20) and longest (387 bp in Mes-25) spacers differedby only 16 bp (FIG. 1). Intraspecific variation wasdetected only within the ITS1 (1 bp) in five speciesand ITS2 (1–2 bp) in four (FIG. 1). Analyses of theITS regions revealed that the ITS1 (313 PIC/501 bp5 0.624 PIC/bp) was 2.77 times more informativephylogenetically than the ITS2 (96 PIC/426 bp 5

0.225 PIC/bp). By comparison, the 5.8S rDNA,excluding 13 putatively erroneous indel containingnucleotide positions, was highly conserved in that itpossessed only 8 PIC/156 bp (5 0.051 PIC/bp).

To assess whether sequences of the ITS1 or ITS2, orportions of these two spacers could be used to identifyEsculenta Clade species, we coded the 59, middle and39 thirds of the ITS1 (59 5 251 bp, middle 5 251 bp,39 5 252 bp) and 59 and 39 halves of the ITS2 (59 5

213 bp, 39 5 213 bp) as separate character sets andanalyzed them and the separate spacers individuallywith maximum parsimony. These analyses revealedthat the ITS1 and 59 third of the ITS1 distinguishedthe most species (22/26), whereas the 39 third of theITS1 (16/26) and 59 (16/26) and 39 (14/26) halves ofthe ITS2 were the least discriminating (SUPPLEMENTA-

RY TABLE VI).

Elata clade.—The 200 sequences in the Elata Cladedataset included 90 that we downloaded fromGenBank and 110 generated in the present study(SUPPLEMENTARY TABLE II). Validly published binomi-als can be applied with confidence to only three ofthe species (i.e. 5 M. tomentosa Mel-1, M. semiliberaMel-3 and M. punctipes Mel-4); however, names for sixspecies endemic to North America have beenproposed (FIG. 2, Kuo et al. 2012). Using COLLAPSE1.2, we identified 60 unique haplotypes within the 200sequence Elata Clade dataset (996 bp alignment)for subsequent MP phylogenetic analysis. Before

DU ET AL.: MORCHELLA ITS RDNA PHYLOGENETICS 1357

FIG. 3. Bootstrapped 70% majority rule cladogram inferred from maximum parsimony analysis of 106 unique ITShaplotypes in the 443-sequence Elata Subclade (black morels) dataset (SUPPLEMENTARY TABLE III). Sequences of the sister taxaMel-11 and Mel-25 were used to root the phylogeny. Each haplotype is identified by GenBank or herbarium accession numberor laboratory code (SUPPLEMENTARY TABLE III), geographic origin and number of sequences (in parentheses) that compriseeach non-singleton haplotype. Haplotypes identified by an asterisk represent ITS sequences used as a reference for eachphylogenetically distinct species based on multilocus support for their genealogical exclusivity. Because binomials can be

1358 MYCOLOGIA

r

applied with confidence to only four of the species, each phylogenetic species (i.e. phylospecies) is identified by Mel followedby a unique Arabic number. Note that ITS sequences of four putatively novel Elata Subclade species (NEW-3?-NEW-6?) werediscovered in GenBank. The size and number of sequences of the (ITS1) (5.8SrDNA) (ITS2) (complete ITS region)determined by DNA sequence analysis is indicated to the right of each phylospecies, except for five sequences where the ITS1is short (s). Internodes that were supported by $ 70% maximum parsimony bootstrap intervals are indicated. Note that severalgroups of phylospecies cannot be distinguished using ITS sequence data, with Mel-17/Mel-19/Mel-20/Mel-34, indicated bygray highlight being the most notable. Mel-20 from Yunnan Province, China, is featured.

FIG. 3. Continued.

DU ET AL.: MORCHELLA ITS RDNA PHYLOGENETICS 1359

conducting phylogenetic analyses, 102 ambiguouslyaligned nucleotide positions within the ITS1 and 27positions within the ITS2 were excluded from thedataset. To minimize the number of ambiguouslyaligned nucleotide characters, Elata Clade phylogenywas rooted on sequences of Mel-1 based on moreinclusive analyses (O’Donnell et al. 2011). MP analysisof the Elata Clade dataset revealed that all 200sequences could be identified to species with theITS rDNA data (SUPPLEMENTARY TABLE II). However,the genealogical exclusivity of two putatively novelspecies designated NEW-1? and NEW-2? needs to beaccessed via GCPSR. Four species composed 79%

(158/200) of the sequences, with M. capitata (Mel-9,n 5 53) and M. importuna (Mel-10, n 5 56) being thebest represented, followed by M. tomentosa (Mel-1, n5 25) and M. septimelata (Mel-7, n 5 24). By contrast,seven of the Elata Clade species each were represent-ed by fewer than eight sequences (SUPPLEMENTARY

TABLE II), including two putatively novel speciesdesignated NEW-1? from China and NEW-2? fromChina and Germany (SUPPLEMENTARY TABLE II).Moreover, only one-quarter (22/90) of the ElataClade sequences in GenBank were identified byvalidly published binomials. Analysis of these acces-sions revealed that M. importuna (Mel-10) wasdeposited under four names, one sequence of M.sextelata (Mel-6) and one of M. importuna (Mel-10)were deposited as M. esculenta and three sequencesof M. semilibera (Mel-3) were deposited as M. gigas.Although the six sequences of M. tomentosa (Mel-1)are authentic for this species, type studies are neededto determine whether the eight sequences of Mel-10that we obtained from GenBank as M. conica areidentified correctly. Even if we assume that they are,at least 8/22 (36.4%) of the Elata Clade sequencesin GenBank can be inferred to be misidentified.However, if M. conica is shown to represent a speciesother than Mel-10, then 14–16 (63.6–72.7%) of thesequences from this clade can be inferred to bemisidentified in GenBank.

The Elata Clade dataset contained complete ITSsequences for all of the species except M. populiphila(Mel-5), where approximately one-half of the ITS1spacer region was missing (FIG. 1). In contrast to whatwas observed within the Esculenta Clade, the ITSregion was largest in the two basal-most taxa, M.tomentosa (Mel-1, 834–835 bp) and M. frustrata (Mel-2, 903–904 bp), and smallest in the five most derivedtaxa, where it was 626–655 bp. By mapping the size ofthe ITS1/5.8S rDNA/ITS2 on the ITS phylogeny(FIG. 2) and the topologically concordant multilocusphylogeny (O’Donnell et al. 2011 FIG. 3), our resultsindicate that the ITS1 and ITS2 spacers both havecontracted during the evolutionary diversification of

the Elata Clade. Comparison of the two species withthe largest (M. tomentosa Mel-1, 834–835 bp and M.frustrata Mel-2, 903–904 bp) and smallest (M.septimelata Mel-7, 626–628 bp and Morchella sp. Mel-8, 624 bp) spacers revealed that the ITS1 and ITS2contracted respectively by a factor of 1.7–1.9 and 1.2–1.3 during the evolution of the Elata Clade. Minorlength variation was observed within the ITS1 (2–5 bp)in five species and within the ITS2 (2–3 bp) in three(FIG. 2). Assessment of phylogenetic signal within theITS regions revealed that, in contrast to the EsculentaClade, the ITS2 (138 PIC/308 bp 5 0.448 PIC/bp)was 1.08 times more phylogenetically informativethan the ITS1 (166 PIC/402 bp 5 0.413 PIC/bp)within the Elata Clade. The 5.8S rDNA, by compar-ison, was highly conserved, possessing only 5 PIC/156 bp (5 0.032 PIC/bp).

Finally, to determine how well the ITS1 and ITS2,and equally divided portions of these two spacers,performed in distinguishing species, we coded the 59

and 39 halves of the ITS1 (59 5 252 bp, 39 5 252 bp)and ITS2 (59 5 168 bp, 39 5 167 bp) as separatecharacter sets and analyzed them and alignments ofeach spacer separately via maximum parsimony. Theresults of these analyses indicated that sequences fromeach of these partitions can be used to identify all thespecies or all but two closely related species pairs withinthe Elata Clade (SUPPLEMENTARY TABLE VI).

Elata subclade.—The 443-sequence Elata Subcladedataset included 164 sequences downloaded fromGenBank and 279 sequences generated in the presentstudy. We were able to apply binomials with confi-dence to only four of the 22 phylogenetically distinctspecies previously identified within this subcladeusing multilocus DNA sequence data (Du et al.2012; O’Donnell et al. 2011; Taskın et al. 2010,2012). One of these was for M. angusticeps (Mel-15,Kuo et al. 2012), a species restricted to eastern NorthAmerica, and three others were names recentlyproposed for North American endemics (Kuo et al.2012). Analysis using COLLAPSE 1.2 identifed 106unique haplotypes (FIG. 3, SUPPLEMENTARY TABLE V).The MP and COLLAPSE 1.2 analyses revealed thatonly 10/22 species within this clade could beidentified with ITS rDNA sequence data (FIG. 3). Nosequences were excluded as ambiguously alignedfrom the Elata Subclade dataset. Therefore, phyloge-netic analyses and BLAST queries using ITS rDNAsequences cannot be used to definitively identify 10species because they share an identical allele with atleast one other species (FIG. 3).

The MP analysis also suggested that 21 of thesequences obtained from GenBank might representas many as four putatively novel species that we

1360 MYCOLOGIA

designated NEW-3?–NEW-6? (FIG. 3); however,GCPSR-based analyses are needed to determinewhether these lineages are genealogically exclusive.One of these species (NEW-4?) appears to be apreviously unknown species on the basis of phyloge-netic analyses of ITS + LSU rDNA sequence data(Pagliaccia et al. 2011). Forty of the Elata Subcladesequences downloaded from GenBank were deposit-ed as Mel-12 (Pagliaccia et al. 2011); the remaining 55sequences could not be identified to species (SUPPLE-

MENTARY TABLE III). Although 95 Elata Subcladesequences were represented in GenBank, only 50 ofthese were identified to species. However, ouranalyses indicated that at least 35 of the 50 identifi-cations are incorrect. Seven species were listed as M.elata, four as M. angusticeps and three as M. conicaand M. esculenta (SUPPLEMENTARY TABLE III).

Complete sequences of the ITS rDNA region wereobtained for all species within the Elata Subclade,except for an unassigned species (Morchella sp.,GenBank accession number AF000970) and theaforementioned putative species NEW-4? (GenBankaccession numbers JF319903–JF319906), which en-compassed ITS sequences that were missing approx-imately 17 bp from the 59 end of the ITS1(SUPPLEMENTARY TABLE III). Although the ITS1 spacerwithin this subclade showed slight variation (213–221 bp) in three species (FIG. 3), the ITS2 was fixed at277 bp. The latter spacer is identical in length to theITS2 in Mel-11 and Mel-25, the sisters of the ElataSubclade (O’Donnell et al. 2011), which was used toroot the Elata Subclade phylogeny. Intraspecificvariation was observed only within the ITS1 of Mel-13 (213 or 217 bp), Mel-14 (217 or 221 bp) and aputatively novel species (NEW-5?, 215–216 bp) repre-sented by two collections incorrectly deposited inGenBank as M. elata from India (SUPPLEMENTARY

TABLE III, Kanwal et al. 2011). A comparison ofphylogenetic signal within the ITS regions showedthat the ITS1 (65 PIC/231 bp 5 0.281 PIC/bp) was2.3 times more informative than the ITS2 (34 PIC/279 bp 5 0.122 PIC/bp). By contrast, only twosynapomorphic positions were found within thehighly conserved 5.8S rDNA (2 PIC/156 bp 5 0.013PIC/bp).

Morchella MLST.—The Web-accessible MorchellaMLST database hosted at the CBS-KNAW wasconstructed to aid molecular identification of morelsvia the Internet with BLAST queries or phylogeneticanalysis. In contrast to GenBank, the Morchelladatabase will include only accession sequences forwhich voucher specimens or cultures are availablefrom publically accessible herbaria or culture collec-tions. This criterion is the same as that employed by

the FUSARIUM-ID (Geiser et al. 2004) and FusariumMLST (http://www.cbs.knaw.nl/fusarium) databases.In addition, the Morchella MLST database housesmetadata that can be downloaded for further study.The current version of Morchella MLST containsbroad taxonomic sampling from four loci (ITS andLSU rDNA, RPB1, RPB2 and EF-1a). Two options areavailable for identifying sequences via BLAST queriesof the database: (i) a single sequence from one of thefour loci can be compared with those housed inMorchella MLST or Morchella MLST and GenBankdatabases or (ii) multiple sequences can be used,which is recommended especially to obtain a defin-itive identification for species within the ElataSubclade. In addition to being able to conductBLAST queries, neighbor joining phylogenetic anal-ysis can be conducted to infer evolutionary relation-ships of an unknown to species represented in thedatabase. The Website contains a portal by whichsequences can be added to the database under theconditions that vouchers, including accession num-bers, are made available from publically accessibleherbaria and/or culture collections.

DISCUSSION

Identification of Morchella phylogenetic species usingITS rDNA sequence data.—The present study wasmodeled after one conducted on a grass-associatedclade of Colletotrichum (Crouch et al. 2009) in whichspecies limits were inferred from multilocus GCPSRto provide a framework for determining the utility ofITS data for differentiating species from each othervia molecular phylogenetics and by BLAST queries ofa dedicated database for morels, Morchella MLST.The primary objective of the present study was togenerate an inclusive set of taxonomically validatedITS sequences for morels and to evaluate the accuracyof this single, widely used locus for the dual purposeof phylogenetic reconstruction and species identifi-cation (Du et al. 2012; O’Donnell et al. 2011; Taskınet al. 2010, 2012). To assess their utility withinMorchella we generated 553 ITS rDNA sequencesrepresenting 61 of the 62 known phylogeneticallydistinct species and downloaded another 312 se-quences from GenBank with emerencia. Althoughefforts to obtain ITS rDNA sequence data from theEuropean endemic Mes-5 were unsuccessful, wepredict that the ITS length will match that of theclosely related species Mes-6/Mes-7/Mes-21, where itis conserved at 1089 bp.

Our analyses elucidated the strengths and limita-tions of ITS as the sole locus for distinguishing specieswithin Morchella. The present study revealed that theITS data could be used to distinguish 77.4% (48/62)

DU ET AL.: MORCHELLA ITS RDNA PHYLOGENETICS 1361

of the known phylogenetic species within Morchella.However, only 68.4% (592/865) of the sequencesanalyzed could be assigned to one of those species.ITS sequences resolved most species within the Elataand Esculenta clades, but fewer than half of the specieswithin the Elata Subclade. Moreover, as a result ofmining ITS data in GenBank, the existence of sixputatively novel species was revealed, although theirgenealogical exclusivity needs to be critically accessedvia GCPSR. Two of the potentially novel species werenested within the Elata Clade, where ITS sequencesdifferentiated the 10 known phylogenetic species (Mel-1 through Mel-10) and all 200 sequences analyzed fromthis clade. Another noteworthy finding was thatsequences of the ITS1 (504 bp) or ITS2 (335 bp)and even the 59 or 39 halves of either spacer resolved allof the species, or all but two closely related sisterspecies within the Elata Clade (SUPPLEMENTARY TABLE

VI). We also discovered that ITS data performed wellwithin the Esculenta Clade in that 92.3% (24/26) ofthe phylogenetic species were diagnosable and 95.5%

(212/222) of the sequences analyzed could beassigned to one of these species. In addition, ouranalyses revealed that sequences of the ITS1 (754 bp),or only a 251 bp fragment comprising the 39 third ofthe ITS1, could be used to identify 22/26 of the specieswithin this clade (SUPPLEMENTARY TABLE VI). While itwas possible to differentiate close to three-quarters ofthe Esculenta Clade species (19/26) with completeITS2 sequences, the 59 and 39 halves of the ITS2differentiated only 50–60% of the species.

One important finding is ITS sequence data haslimited utility in differentiating species within thespecies-rich Elata Subclade. Counting sequences oftwo sister species used to root the phylogeny (i.e. Mel-11 and Mel-25), our results show that only 50% (12/24) of species in the Elata Subclade + Mel-11/Mel-25could be identified with the ITS data and only 44.7%

(198/443) of the sequences analyzed could beassigned to one of these species. We attribute thefailure of the ITS to resolve species within the ElataSubclade to its relatively recent origin in the middleMiocene, about 12.15 Mya (95% HPD: 9.46–15.48)(Du et al. 2012) and to its subsequent rapid radiationinto a species-rich lineage comprising one-third of thetrue morels (O’Donnell et al. 2011). Rapid radiationsresult in short internal branch lengths in many genesequence phylogenies; these short branches have littlephylogenetic information (Kraus and Miyamoto1991) and may cluster randomly due to short-branchattraction (Nei 1996). Thus, understanding speciesdiversity within the Elata Subclade will requireidentification of additional marker loci with higherphylogenetic signal (Lopez-Giraldez and Townsend2011) to resolve closely related species and the

backbone of the phylogeny; it is likely that onlyrapidly evolving loci will be informative due to theshort-branch phenomena described previously.

Our results also should serve as a cautionary note tonot assume a priori 3% ITS interspecific variabilitywithin Morchella and other fungi (Nilsson et al. 2006).While it is convenient to define species or phylotypesphenetically in fungal environmental surveys (Peayet al. 2008), or in systematic studies focused onmining metadata from GenBank accessions (Bonitoet al. 2010, Ryberg et al. 2008), our results and thoseof others (Crouch et al. 2009, Roe et al. 2010, Willet al. 2005) strongly indicate that this approach islikely to significantly underestimate species diversity.Fortunately, in instances where researchers queryingMorchella MLST are unable to identify unknowns withITS data, they can conduct additional searches of thisreference database using sequences of one or more ofthe three single-copy protein-coding genes (i.e. RPB1,RPB2, EF-1a), which were developed initially withinthe framework of the AFTOL project (James et al.2006, Matheny 2005, Reeb et al. 2004).

Evolution of the ITS regions.—The current study addsto our growing knowledge of ITS evolution andvariability within the Fungi. In contrast to thesignificant ITS divergence reported for diverse speciesinferred from GenBank data (Nilsson et al. 2008), wedetected surprisingly little intraspecific variabilitywithin the ITS region in Morchella; the highest weobserved was only 1.3% in M. esculentoides (Mes-4).However, had we grouped sequences by the namesused to identify them in GenBank, this estimatewould have been considerably higher because multi-ple species are accessioned under the same name inthis database. Our analyses of GenBank accessions,for example, revealed that those listed as M. esculenta,M. crassipes and M. elata each comprised eightdifferent phylogenetic species (SUPPLEMENTARY TABLES

I–III). To further illustrate problems caused bymisidentified sequences in GenBank (Bridge et al.2003, Vilgalys 2003), Nilsson et al. (2008) reported3.1% ITS divergence in Fusarium solani; however, thisestimate is flawed because our GCPSR studiesrevealed that this morphospecies comprises morethan 50 distinct phylogenetic species (O’Donnell2000, O’Donnell et al. 2008, Zhang et al. 2006). Incontrast to the aforementioned survey of ITS varia-tion within diverse fungal species, our analysesshowed that the highest ITS divergence for anyspecies within the F. solani species complex was only0.1% in F. falciforme. Estimates of ITS variation can beaffected significantly by putative paralogs and intra-genomic variation, as reported in diverse fungi(Lindner and Banik 2011, O’Donnell et al. 1997,

1362 MYCOLOGIA

Simon and Weiß 2008), but these were not detectedwithin the ITS of Morchella. In addition, we did notdetect any gene tree-species tree discordance betweenthe morel ITS rDNA and the RPB1, RPB2 and EF-1aphylogenies.

Results of the present study fit experimental data inbudding and fission yeast that indicate the ITS1 isgenerally less constrained evolutionarily than theITS2 (Abeyrathne and Nazar 2005, van Nues et al.1995). Early in our studies we discovered that ITSlength variation was primarily due to length muta-tions within the ITS1 (O’Donnell et al. 1993), so weused variation at this locus to screen cultures andherbarium specimens via PCR to sample for thegreatest global genetic diversity (O’Donnell et al.2011), using the universal PCR primers ITS5 3 ITS4(White et al. 1990). In the present study wediscovered that the ITS1 more than doubled inlength during the evolutionary diversification of theEsculenta Clade whereas its length contracted by 50%

within the Elata Clade. Significant length variationhas been documented within the ITS region in otherfungi (reviewed in Hausner and Wang 2005), with thegreater than threefold increase reported within theITS1 of Cantharellus being one of the most dramatic(Feibelman et al. 1994). One of the more noteworthyfindings of the present study was that ITS2 length washighly conserved during the estimated 68.7 millionyear (95% HPD interval: 49.19–93.03] evolutionaryhistory of the Esculenta Clade (Du et al. 2012), withthe shortest and longest spacers differing in length byonly 16 bp. Although not as highly conserved, theITS2 only varied by 53 bp across the phylogeneticbreadth of the Elata Clade. Although MUSCLE(Edgar 2004) produced fewer gaps in the finalalignment when compared to Clustal W (Thompsonet al. 1994) and MAAFT 6 (Katoh and Toh 2008)(data not shown), considerable manual adjustmentswere required and even then one-third and one-fifthof the ITS1 nucleotide positions were excludedrespectively from the Esculenta and Elata Cladedatasets due to ambiguity in the alignment. Consis-tent with uniform reports of the 5.8S rDNA beinghighly conserved in the Fungi (Nilsson et al. 2008),our analyses indicate the 5.8S rDNA was fixed at156 bp within Morchella and in the hypogeous(Trappe et al. 2010) and other epigeous membersof the Morchellaceae (Hansen and Pfiser 2006,O’Donnell et al. 1997). Conservation of the 5.8SrDNA also was reported in an extensive survey of thecore group of Peziza, where it was fixed at 157 bp(Hansen et al. 2002). Based on our findings that six ofthe Morchella sequences we downloaded from Gen-Bank had length mutations within the 5.8S rDNA thatnecessitated inserting indels within the alignments

(SUPPLEMENTARY TABLE V), and three other ITSsequences could be used only once, bad sequencedata at the 59 or 39 end were removed, it would behighly desirable if automated processes at GenBankwere in place to identify sequences with probablesequencing errors before they end up in the database.Of course, this necessitates the adoption of asequence standard for each species or strain to whicha deposited sequence could be compared, in whichcase rare variant alleles that encompass insertions ordeletions also must be taken into account. While sucha method could be based on models of nucleotidesubstitution, they do not incorporate the probabilitythat an insertion or deletion occurs due to theintractability of that problem (i.e. While the likeli-hood of an ARG transition can be predicted basedon observations from the sequence data, how can theprobability be calculated that a nucleotide will beinserted between any two nucleotides?). This problemrequires that an alternative solution be found.

Can GenBank and emerencia be used to identifyMorchella species?— With 66% or more of thesequences misidentified, Morchella may have thedubious distinction of having the highest percentageof poorly annotated named fungal ITS sequences inGenBank. This estimate significantly eclipses theoften cited ‘‘upwards of 20%’’ reported as the worstcase scenario for certain groups of the Fungi (Bridgeet al. 2003, Nilsson et al. 2006, Vilgalys 2003). Ourstartling discovery prompts three obvious questions:(i) Why are Morchella sequences so poorly annotatedin GenBank?; (ii) Can these databases be used forspecies identifications, and if so, how? And (iii) is thepoor state of Morchella annotation in GenBank closerto the rule rather than the exception?

Lack of knowledge of species boundaries andthe absence of type studies both have contributed tothe large number of misidentifications. Contrary tothe general assumption, our studies indicate thatMorchella species are not cosmopolitan in distribution(Du et al. 2012; O’Donnell et al. 2011; Taskın et al.2010, 2012). However, because Kanwal et al. (2011)incorrectly assumed M. crassipes, M. elata and M.angusticeps were present in India, and did not identifytheir accession, GQ228474, as M. tomentosa (Mel-2),75% (12/16) of the ITS sequences they deposited inGenBank were misidentified. Once misidentifiedsequences such as these are accessioned, they thencan serve as the reference for subsequent BLASTqueries of GenBank and identifications using emer-encia (Nilsson et al. 2006). Another potential sourceof error is the inclination of individual researchers todeposit sequences under a validly published binomialregardless of whether the sequence is actually

DU ET AL.: MORCHELLA ITS RDNA PHYLOGENETICS 1363

assignable to that species. However, because almostevery GCPSR-based study, including our own, hasdiscovered widespread cryptic speciation, consistentwith the prediction that only , 7% of fungal diversityhas been discovered (Hawksworth 2004), it is essentialthat sequences of phylogenetically distinct species beidentified by some standardized informal nomen-clature when they are deposited in GenBank andemerencia, especially when applying a binomial isproblematic. To this end, we distinguished sequencesof phylogenetic species within the Esculenta and ElataClades respectively by Mes and Mel followed by aunique Arabic number, thereby facilitating DNAsequence-based identifications of Morchella in Gen-Bank. A BLAST query of GenBank matching one ofthese sequences identifies the organism (e.g. ‘‘Morch-ella sp. Mel-14’’ represented by GU551435) andprovides a clarifying note (e.g. note 5 ‘‘phylogeneticspecies Mel-14’’). However, because at least 66% ofthe named sequences in GenBank are misidentified,interpreting BLAST results can be complicated.

We found emerencia (Nilsson et al. 2005) to beextraordinarily useful in retrieving Morchella ITSsequences represented in GenBank because it alsofound several from environmental studies that werelisted ‘‘uncultured fungus’’ (Taylor et al. 2008) andone as ‘‘Pezizomycetes sp.’’ (SUPPLEMENTARY TABLE

III). Results of the present study, however, show thatemerencia cannot be used to obtain reliable speciesidentifications for Morchella because two-thirds of the‘‘fully identified sequences’’ (i.e. those with binomi-als) it retrieved from GenBank as a reference weremisidentified. To provide an example of the problemswe encountered when we queried emerencia withseparate sequences of phylogenetic species Morchellasp. Mel-20 (GU551406, O’Donnell et al. 2011;HM056383, Taskın et al. 2010), the best BLASTmatches were identified respectively as M. esculenta(EU600241) and M. elata (EF081000). This heuristicshould serve to alert users of emerencia, as cau-tioned by its authors (Nilsson et al. 2005), that‘‘fully identified’’ and correctly identified are notsynonymous.

Because the factors that contribute to poorlyannotated sequences in GenBank, such as widespreadcryptic speciation and failure to include type speci-mens as a reference, are hardly unique to Morchella,it seems likely that the percentage of misidentifiedsequences for certain groups of fungi in GenBankcould easily exceed the current estimate of 20%

(Bridge et al. 2003). Looking to the future, it is clearthat multilocus GCPSR-based studies will play anever-increasing role in assessing the genetic diversityof fungi, increasing our ability to optimally minesequence data in GenBank. By constructing an ITS

dataset that comprised all but one of the Morchellaspecies inferred via GCPSR, we were able to identifyendophytes of Bromus (Baynes et al. 2012) andJuniperus scopulorum (http://www.ncbi.nlm.nih.gov/nuccore/GQ153010) in western North America, aputative ectomycorrhizal associate of the terrestrialorchid Gymnadenia conopsea in Germany (Stark et al.2009), hypothesize the existence of six putativelynovel species from ITS sequences in GenBank(SUPPLEMENTARY TABLES II–III) and identify three-quarters of the Morchella accessions in GenBank tophylospecies. The latter revealed: (i) several specieswith intercontinental distributions, including fourfire-adapted species native to western North America(Du et al. 2012, O’Donnell et al. 2011, Taskın et al.2012); (ii) the putative North American endemic M.esculentoides (Mes-4) in central Europe (Kellner et al.2005) and (iii) that Morchella sp. (Mes-16) appears tobe the most widely distributed species, based oncollections from China (Du et al. 2012), Israel(Masaphy et al. 2010), Turkey (Taskın et al. 2012),New Zealand (GenBank JF423317), India (Kanwalet al. 2011), three African countries (Degreef et al.2009), and Hawaii and Java (O’Donnell et al. 2011).As proposed for true truffles in the genus Tuber(Bonito et al. 2010) and diverse ectomycorrhizal fungi(Vellinga et al. 2009), our working hypothesis is thatmost if not all disjunct distributions in Morchella aredue to human introductions of species into nonin-digenous areas. Examples include the North Ameri-can endemics M. esculentoides (Mes-4) and M.importuna (Mel-10), which were discovered respec-tively in Europe and Turkey (Taskin et al. 2012) andEurope, Turkey and China (Du et al. 2012, Taskin etal. 2012). We theorize that their introduction intoEurope and China might have been mediated byrelatively recent anthropogenic activities, whichshould minimize the likelihood that they representlater synonyms of taxa described from European andChinese collections (Kuo et al. 2012).

The relatively recent discoveries that morels(Baynes et al. 2012) and Verpa (Stark et al. 2009) cangrow endophytically in plants and form mycorrhizal-like associations with angiosperms (Stark et al. 2009)and gymnosperms (Dahlstrom et al. 2000) suggest thateconomically important plants likely serve as vectorsfor the vertical transmission of morels across oceans.With the whole genome sequencing of one specieswithin the Elata Clade nearing completion (J. Spata-fora pers comm) and others likely to follow in the nearfuture, rich genetic resources such as these shouldprovide a wealth of molecular markers for phylogeo-graphic and phylogenetic studies needed to distin-guish between human-mediated introductions andlong distance dispersal. Phylogenomic analyses should

1364 MYCOLOGIA

help elucidate the underlying genetic basis of nicheadaption of the Esculenta and Elata Clades respectivelyto temperate deciduous and coniferous forests, helpguide genetic studies focused on strain improvementfor commercial cultivation and advance our under-standing of species diversity and distribution so thatthis invaluable genetic resource can be managed toprevent overharvesting, thereby promoting long termspecies conservation.

Morchella MLST.—We were motivated to constructMorchella MLST based on our discovery that at leasttwo-thirds of the named ITS sequences of Morchella inGenBank were misidentified and because this prob-lem is exacerbated when emerencia is applied (Nilssonet al. 2006), given that the latter database uses poorlyannotated named sequences in GenBank as thereference to re-identify sequences, including thoseof distinct phylogenetic species inferred to begenealogically exclusive from multilocus DNA se-quence data (O’Donnell et al. 2011; Taskın et al.2010, 2012). Because many people who depositsequences use the names in GenBank as a reference,the system unfortunately is designed to perpetuateitself (Bidartondo et al. 2008, Vilgalys 2003). Toobviate potential problems created by poorly anno-tated sequences in GenBank we constructed theInternet-accessible Morchella MLST to provide ameans by which the scientific community cancontribute to and access high quality electrophero-grams of voucher specimens and/or cultures that areaccessible to interested researchers (Kang et al. 2010).An added advantage of a dedicated database is thatthe informal nomenclatural system that we developedfor distinguishing the phylogenetically distinct speciesby clade (i.e. Mes or Esculenta and Mel for Elata/ElataSubclade) followed by a unique Arabic number canbe implemented. This informal system was adoptedout of necessity because our studies surprisinglyrevealed that most of the species in North America,Turkey and Asia are novel, negating the assumptionthat European taxa are cosmopolitan, and because ofthe broad uncertainty on how to apply validlypublished names. The utility of Morchella MLST, asfor any reference library dedicated to species identi-fication and discovery, is the dense taxon sampling inwhich species were resolved a priori by multilocusGCPSR. One of the overarching goals of MorchellaMLST is to stimulate integrative taxonomic studies sothat all of the phylospecies receive specific epithets,even in instances where they cannot be reliablydiagnosed phenotypically (Kuo et al. 2012).

A comprehensive nomenclatural and comparativemorphotaxonomic study of these species is needed tobe able to apply European names of Morchella to

species delimited via GCPSR. Roughly 82 validEuropean species names of Morchella are available,excluding many subspecific epithets (see IndexFungorum http://www.indexfungorum.org/), themajority of which are based on material from France,Czechoslovakia, Italy, the Netherlands and Sweden.

Because many of these names date back to 1800 andearly 1900, it is likely that original material either doesnot exist or that it will be much too old to yield usefulDNA sequence data. These names should be lectotypi-fied as needed, or if neither material nor an illustrationexists a neotype should be selected (the latter withnewly collected material, including cultures and multi-locus DNA sequence data). If an original type is too oldto yield useful DNA or is an illustration, it will benecessary to designate a recent collection (with culturesand DNA sequence data) as an epitype to be able tostabilize the name, considering the high morphologicalplasticity of macroscopic characters within species ofMorchella and stasis in several microscopic charactersacross the genus. The neotype or epitype preferablyshould be from the type locality or country. Europe isthe poorest sampled area in our surveys, and to fullyexplore the diversity and boundaries of these species amore thorough molecular and morphological study offresh collections from for example France, CzechRepublic, Slovakia, Italy, the Netherlands, Sweden,Germany and Spain will be essential. Circumscriptionsand synonymies of iconic European species, such as M.esculenta, M. crassipes, M. deliciosa, M. conica, M. elata,M. distans and M. hiemalis, need to be clarified. Typesof a number of recently published taxa from Europe(e.g. Jacquetant 1984, Jacquetant and Bon 1984) needto be located and subjected to similar morphologicaland molecular phylogenetic analyses.

ACKNOWLEDGMENTS

Special thanks are due Nathane Orwig for processing DNAsequences in the NCAUR DNA core facility. The research visitof X.-H.D. to NCAUR in 2011 was supported by the NationalBasic Research Program of China (No. 2009CB522300), theHundred Talents Program of the Chinese Academy ofSciences, Joint Funds of the National Natural ScienceFoundation of China and the Yunnan provincial government(No. U0836604). The mention of trade names or commercialproducts in this publication is solely for the purpose ofproviding specific information and does not imply recom-mendation or endorsement by the U.S. Department ofAgriculture. The USDA is an equal opportunity providerand employer.

LITERATURE CITED

Abeyrathne PD, Nazar RN. 2005. Parallels in rRNAprocessing: conserved features in the processing of

DU ET AL.: MORCHELLA ITS RDNA PHYLOGENETICS 1365

the internal transcribed spacer I in the pre-rRNA fromSchizosaccharomyces pombe. Biochemistry 44:16977–16987, doi:10.1021/bi051465a

Agerer R, Ammirati J, Blanz P, Courtecuisse R, DesjardinDE, Gams W, Hallenberg N, Halling R, HawksworthDL, Horak E, Korf RP, Mueller GM, Oberwinkler F,Rambold G, Summerbell RC, Triebel D, Watling R.2000. Always deposit vouchers. Mycol Res 104:642–644.

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990.Basic local alignment search tool. J Mol Biol 215:403–410.

Baynes M, Newcombe G, Dixon L, Castlebury L, O’DonnellK. 2012. A novel plant-fungal mutualism associated withfire. Fungal Biol 116:113–144, doi:10.1016/j.funbio.2011.10.008

Bidartondo MI, Bruns TD, Blackwell M. 2008. Preservingaccuracy in GenBank. Science 319:1616, doi:10.1126/science.319.5870.1616a

Bonito GM, Gryganskyi AP, Trappe JM, Vilgalys R. 2010. Aglobal meta-analysis of Tuber ITS rDNA sequences: speciesdiversity, host association and long-distance dispersal.Mol Ecol 19:4994–5008, doi:10.1111/j.1365-294X.2010.04855.x

Bridge PD, Roberts PJ, Spooner BM, Panchal G. 2003. On theunreliability of published DNA sequences. New Phytol160:43–48, doi:10.1046/j.1469-8137.2003.00861.x

Crouch JA, Clarke BB, Hillman BI. 2009. What is the valueof ITS sequence data in Colletotrichum systematics andspecies diagnosis? A case study using the falcate-sporedgraminicolous Colletotrichum group. Mycologia 101:648–656, doi:10.3852/08-231

Dahlstrom JL, Smith JE, Weber NS. 2000. Mycorrhiza-likeinteraction by Morchella with species of the Pinaceae inpure culture synthesis. Mycorrhiza 9:279–285, doi:10.1007/PL00009992

Degreef J, Fischer E, Sharp C, Raspe O. 2009. AfricanMorchella crassipes (Ascomycota, Pezizales) revealed byanalysis of nrDNA ITS. Nova Hedwigia 88:11–22,doi:10.1127/0029-5035/2009/0088-0011

Dettman JR, Jacobson DJ, Taylor JW. 2003. A multilocusgenealogical approach to phylogenetic species recog-nition in the model eukaryote Neurospora. Evolution57:2703–2720.

Du X-H, Zhao Q, O’Donnell K, Rooney AP, Yang ZL. 2012.Multigene molecular phylogenetics reveals true morels(Morchella) are especially species-rich in China. FungalGenet Biol 49:455–469, doi:10.1016/j.fgb.2012.03.006.

Edgar RC. 2004. MUSCLE: a multiple sequence alignmentmethod with reduced time and space complexity. BMCBioinform 5:1–19, doi:10.1186/1471-2105-5-113

Feibelman T, Bayman P, Cibula WG. 1994. Length variationin the internal transcribed spacer of ribosomal DNA inchanterelles. Mycol Res 98:614–618, doi:10.1016/S0953-7562(09)80407-3

Gardes M, Bruns TD. 1993. ITS primers with enhancedspecificity for basidiomycetes—application to the iden-tification of mycorrhizae and rusts. Mol Ecol 2:113–118,doi:10.1111/j.1365-294X.1993.tb00005.x

Geiser DM, Jimenez-Gasco MdM, Kang S, Makalowska I,Veeraraghavan N, Ward TJ, Zhang N, Kuldau GA,O’Donnell K. 2004. FUSARIUM-ID v. 1.0: A DNA sequence

database for identifying Fusarium. Eur J Plant Pathol 110:473–479, doi:10.1023/B:EJPP.0000032386.75915.a0

Gower JC. 1971. A general coefficient of similarity and someof its properties. Biometrics 27:857–874, doi:10.2307/2528823

Hansen K, Laessøe T, Pfister DH. 2002. Phylogeneticdiversity in the core group of Peziza inferred fromITS sequences and morphology. Mycol Res 106:879–902, doi:10.1017/S0953756202006287

———, Pfister DH. 2006. Systematics of the Pezizomy-cetes—the operculate discomycetes. Mycologia 98:1029–1040, doi:10.3852/mycologia.98.6.1029

Hausner G, Wang X. 2005. Unusual compact rDNA genearrangements within some members of the Ascomyco-ta: evidence for molecular co-evolution between ITS1and ITS2. Genome 48:648–660, doi:10.1139/g05-037

Hawksworth DL. 2004. Fungal diversity and its implicationsfor genetic resource collections. Stud Mycol 50:9–18.

Horton TR, Bruns TD. 2001. The molecular revolution inectomycorrhizal ecology: peeking into the black box.Mol Ecol 10:1855–1871, doi:10.1046/j.0962-1083.2001.01333.x

Jacquetant E. 1984. Les Morilles. Paris: La Bibliotheque desArts. p 7–114.

———, Bon M. 1984. Typifications et mises au pointnomenclaturales dans l’ouvrage ‘‘Les Morilles’’ (de E.Jacquetant), Nature-Piantanida 1984. Doc Mycol 14:1.

James TY, Kauff F, Schoch CL, Matheny PB, Hofstetter V, CoxCJ, Celio G, Gueidan C, Fraker E, Miadlikowska J,Lumbsch HT, Rauhut A, Reeb V, Arnold AE, Amtoft A,Stajich JE, Hosaka K, Sung GH, Johnson D, O’Rourke B,Crockett M, Binder M, Curtis JM, Slot JC, Wang Z,Wilson AW, Schussler A, Longcore JE, O’Donnell K,Mozley-Standridge S, Porter D, Letcher PM, Powell MJ,Taylor JW, White MM, Griffith GW, Davies DR, HumberRA, Morton JB, Sugiyama J, Rossman AY, Rogers JD,Pfister DH, Hewitt D, Hansen K, Hambleton S,Shoemaker RA, Kohlmeyer J, Volkmann-Kohlmeyer B,Spotts RA, Serdani M, Crous PW, Hughes KW, MatsuuraK, Langer E, Langer G, Untereiner WA, Lucking R,Budel B, Geiser DM, Aptroot A, Diederich P, Schmitt I,Schultz M, Yahr R, Hibbett DS, Lutzoni F, McLaughlinDJ, Spatafora JW, Vilgalys R. 2006. Reconstructing theearly evolution of Fungi using a six-gene phylogeny.Nature 443:818–822, doi:10.1038/nature05110

Kang S, Mansfield MA, Park B, Geiser DM, Ivors KL, CoffeyMD, Grunwald NJ, Martin FN, Levesque CA, Blair JE.2010. The promise and pitfalls of sequence-basedidentification of plant-pathogenic fungi and oomy-cetes. Phytopathology 100:732–737, doi:10.1094/PHYTO-100-8-0732

Kanwal HK, Acharya K, Ramesh G, Reddy MS. 2011.Molecular characterization of Morchella species fromthe western Himalayan region of India. Curr Microbiol62:1245–1252, doi:10.1007/s00284-010-9849-1

Katoh K, Toh H. 2008. Recent developments in the MAFFTmultiple sequence alignment program. Brief Bioinform9:286–298, doi:10.1093/bib/bbn013

Kellner H, Renker C, Buscot F. 2005. Species diversity within theMorchella esculenta group (Ascomycota: Morchellaceae) in

1366 MYCOLOGIA

Germany and France. Org Diver Evol 5:101–107,doi:10.1016/j.ode.2004.07.001

Kraus F, Miyamoto MM. 1991. Cladogenesis among thepecoran ruminants: evidence from mitochondrial DNAsequences. Syst Zool 40:117–130, doi:10.2307/2992252

Kuo M, Dewsbury DR, O’Donnell K, Carter MC, Rehner SA,Moore JD, Moncalvo J-M, Canfield SA, Methven AS,Volk TJ. 2012. A taxonomic revision of true morels(Morchella) in Canada and the United States. Mycolo-gia 104:1159–1177, doi:10.3852/11-375

Lindner DL, Banik MT. 2011. Intragenomic variation in theITS rDNA region obscures phylogenetic relationshipsand inflates estimates of operational taxonomic units ingenus Laetiporus. Mycologia 103:731–740, doi:10.3852/10-331

Lopez-Giraldez F, Wang Z, Townsend JP. 2011. Assessingphylogenetic information content in fungal genomeswith PhyDesign, a Web application for smart markerselection. Inoculum 62:30.

Masaphy S, Zabari L, Goldberg D, Jander-Shgag G. 2010.The complexity of Morchella systematics: a case of theyellow morel in Israel. Fungi 3:14–18.

Matheny PB. 2005. Improving phylogenetic inference ofmushrooms with RPB1 and RPB2 nucleotide sequences(Inocybe; Agaricales). Mol Phylogenet Evol 35:1–20,doi:10.1016/j.ympev.2004.11.014

Nei M. 1996. Phylogenetic analysis in molecular evolution-ary genetics. Annu Rev Genet 30:371–403, doi:10.1146/annurev.genet.30.1.371

Nilsson RH, Kristiansson E, Ryberg M, Hallenberg N. 2008.Intraspecific ITS variability in the kingdom Fungi asexpressed in the international sequence databases andits implications for molecular species identification.Evol Bioinform 4:193–201.

———, ———, ———, Abarenkov K, Larsson K-H, KoljalgU. 2006. Taxonomic reliability of DNA sequences inpublic sequence databases: a fungal perspective. PLoSOne 1, e59, doi:10.1371/journal.pone.0000059

———, ———, ———, Larsson K-H. 2005. Approaching thetaxonomic affiliation of unidentified sequences in publicdatabases—an example from the mycorrhizal fungi. BMCBioinform 6:178, doi:10.1186/1471-2105-6-178

O’Donnell K. 2000. Molecular phylogeny of the Nectriahaematococca-Fusarium solani species complex. Myco-logia 92:919–938, doi:10.2307/3761588

———, Cigelnik E. 1997. Two divergent intragenomicrDNA ITS2 types within a monophyletic lineage ofthe fungus Fusarium are nonorthologous. Mol Phylo-geneti Evol 7:103–116, doi:10.1006/mpev.1996.0376

———, ———, Weber NS, Trappe JM. 1997. Phylogeneticrelationships among ascomycetous truffles and the trueand false morels inferred from 18S and 28S ribosomalDNA sequence analysis. Mycologia 89:48–65, doi:10.2307/3761172

———, Rooney AP, Mills GL, Kuo M, Weber NS, RehnerSA. 2011. Phylogeny and historical biogeography oftrue morels (Morchella) reveals an early Cretaceousorigin and high continental endemism and provincial-ism in the Holarctic. Fungal Genet Biol 48:252–265,doi:10.1016/j.fgb.2010.09.006

———, Sutton DA, Fothergill A, McCarthy D, Rinaldi MG,Brandt ME, Zhang N, Geiser DM. 2008. Molecularphylogenetic diversity, multilocus haplotype nomencla-ture and in vitro antifungal resistance within theFusarium solani species complex. J Clin Microbiol 46:2477–2490, doi:10.1128/JCM.02371-07

———, Weber NS, Hains JH, Mills GL. 1993. Molecularsystematics and phylogenetics of the Morchellaceae.Inoculum 44:52.

Ower R, Mills GL, Malachowski JA. 1986. Cultivation ofMorchella. US patent 4594809.

Pagliaccia D, Douhan GW, Douhan L, Peever TL, CarrisLM, Kerrigan JL. 2011. Development of molecularmarkers and preliminary investigation of the popula-tion structure and mating system in one lineage ofblack morel (Morchella elata) in the Pacific Northwest-ern USA. Mycologia 103:969–982, doi:10.3852/10-384

Peay KG, Kennedy PG, Bruns TD. 2008. Fungal communityecology: a hybrid beast with a molecular master.BioScience 58:799–810, doi:10.1641/B580907

Pilz D, McLain R, Alexander S, Villarreal-Ruiz L, Berch S,Wurtz TL, Parks CG, McFarlane E, Baker B, Molina R,Smith JE. 2007. Ecology and Management of MorelsHarvested from the Forests of Western North America.Corvallis, OR: PNW–GTR-710. USDA Forest Service.

Reeb V, Lutzoni F, Roux C. 2004. Contribution of RPB2 tomultilocus phylogenetic studies of the euascomycetes(Pezizomycotina, Fungi) with species emphasis on thelichen-forming Acarosporaceae and evolution of poly-spory. Mol Phylogenet Evol 32:1036–1060, doi:10.1016/j.ympev.2004.04.012

Robert V, Szoke S, Jabas J, Vu D, Chouchen O, Blom E,Cardinali G. 2011. BioloMICS software, biological datamanagement, identification, classification and statistics.Open Appl Informatics J 5:87–98, doi:10.2174/1874136301105010087

Roe AD, Rice AV, Bromilow SE, Cooke JEK, Sperling FAH.2010. Multilocus species identification and fungal DNAbarcoding: insights from blue stain fungal symbionts ofthe mountain pine beetle. Mol Ecol Res 10:946–959,doi:10.1111/j.1755-0998.2010.02844.x

Ryberg M, Nilsson RH, Kristiansson E, Topel M, Jacobsson S,Larsson E. 2008. Mining metadata from unidentified ITSsequences in GenBank: a case study in Inocybe (Basidio-mycota). BMC Evol Biol, doi:10.1186/1471-2148-8-50

Saitou N, Kristiansson E, Sjokvist E, Nilsson RH. 2009. Anoutlook on the fungal internal transcribed spacer sequenc-es in GenBank and the introduction of a Web-based toolfor the exploration of fungal diversity. New Phytol 181:471–477, doi:10.1111/j.1469-8137.2008.02667.x

———, Nei N. 1987. The neighbor joining method: a newmethod for reconstructing phylogenetic trees. Mol BiolEvol 14:406–425.

Simon UK, Weiß M. 2008. Intragenomic variation of fungalribosomal genes is higher than previously thought.Mol Biol Evol 25:2251–2254, doi:10.1093/molbev/msn188

Stark C, Babik W, Durka W. 2009. Fungi from the roots of thecommon terrestrial orchid Gymnadenia conopsea. MycolRes 113:952–959, doi:10.1016/j.mycres.2009.05.002

DU ET AL.: MORCHELLA ITS RDNA PHYLOGENETICS 1367

Stefani FOP, Sokolski S, Wurtz TL, Piche Y, Hamelin RC,Fortin JA, Berube JA. 2010. Morchella tomentosa: aunique belowground structure and a new clade ofmorels. Mycologia 102:1082–1088, doi:10.3852/09-294

Swofford DL. 2002. PAUP* 4.0b10: phylogenetic analysisusing parsimony (*and other methods). Sunderland,Massachusetts: Sinauer Associates.

Taskın H, Buyukalaca S, Dogan HH, Rehner SA, O’DonnellK. 2010. A multigene molecular phylogenetic assess-ment of true morels (Morchella) in Turkey. FungalGenet Biol 47:672–682, doi:10.1016/j.fgb.2010.05.004

———, ———, Hansen K, O’Donnell K. 2012. Multilocusphylogenetic analysis of true morels (Morchella) revealshigh levels of endemics in Turkey relative to otherregions of Europe. Mycologia 104, 446–461.

Taylor DL, Booth MG, McFarland JW, Herriott IC, Lennon NJ,Nusbaum C, Marr TG. 2008. Increasing ecologicalinference from high throughput sequencing of fungi inthe environment through a tagging approach. Mol EcolResour 8:742–752, doi:10.1111/j.1755-0998.2008.02094.x

Taylor JW, Jacobson DJ, Kroken S, Kasuga T, Geiser DM,Hibbett DS, Fischer MC. 2000. Phylogenetic speciesrecognition and species concepts in fungi. FungalGenet Biol 31:21–32, doi:10.1006/fgbi.2000.1228

———, Turner E, Townsend JP, Dettman JR, Jacobson D.2006. Eukaryotic microbes, species recognition and thegeographic limits of species: examples from thekingdom Fungi. Phil Trans R Soc B 361:1947–1963,doi:10.1098/rstb.2006.1923

Thompson JD, Higgins DG, Gibson TJ. 1994. Clustal W:improving the sensitivity of progressive multiple sequencealignment through sequence weighting, position-specificgap penalties and weight matrix choice. Nucleic Acids Res22:4673–4680, doi:10.1093/nar/22.22.4673

Trappe MJ, Trappe JM, Bonito GM. 2010. Kalapuya brunneagen. and sp. nov. and its relationship to the othersequestrate genera in Morchellaceae. Mycologia 102:1058–1065, doi:10.3852/09-232

van Nues RW, Rientjes JMJ, Morre SA, Mollee E, Planta RJ,Venema J, Raue HA. 1995. Evolutionary conservedstructural elements are critical for processing oninternal transcribed spacer 2 from Saccharomycescerevisiae precursor ribosomal RNA. J Mol Biol 250:24–26, doi:10.1006/jmbi.1995.0355

Vellinga EC, Wolfe BE, Pringle A. 2009. Global patterns ofectomycorrhizal introductions. New Phytol 181:960–973, doi:10.1111/j.1469-8137.2008.02728.x

Vilgalys R. 2003. Taxonomic misidentification in publicDNA databases. New Phytol 160:4–5, doi:10.1046/j.1469-8137.2003.00894.x

White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification anddirect sequencing of fungal ribosomal RNA genes forphylogenetics. In: Innis MA, Gelfand HH, Sninsky JJ,White TJ, eds. PCR protocols: a guide to methods andapplications. San Diego: Academic Press. p 315–322.

Will KW, Mishler BD, Wheller QD. 2005. The perils of DNAbarcoding and the need for integrative taxonomy. SystBiol 56:844–851, doi:10.1080/10635150500354878

Zhang N, O’Donnell K, Sutton DA, Nalim FA, SummerbellRC, Padhye AA, Geiser DA. 2006. Members of theFusarium solani species complex that cause infectionsin both humans and plants are common in theenvironment. J Clin Microbiol 44:2186–2190,doi:10.1128/JCM.00120-06

Zhao Q, Xu ZZ, Cheng YH, Qi SW, Hou ZJ. 2009. Artificialcultivation of Morchella conica. Southwest China J AgricSci 22:1690–1693 (in Chinese).

Note. Shortly after our paper was accepted for publication we discovered the following article: Clowez P. 2012. Lesmorilles, une nouvelle approche mondiale du genre Morchella. Bull Soc Mycol Fr 126:199–376. Because severalspecies from North America were described in Clowez (2012), at least six species described in Kuo et al. (2012)and whose names were used herein (i.e. M. virginiana [Mes-3], M. esculentoides [Mes-4], M. populiphila [Mel-5],M. septimelata [Mel-7], M. importuna [Mel-10] and M. cryptic [Mel-11]) are later synonyms of taxa described inClowez (2012). Molecular phylogenetic studies are in progress to determine which phylospecies is represented byeach type included in the latter publication (Franck Richard and Mathieu Sauve pers comm).

1368 MYCOLOGIA