*3.2. Fungal Isolation and Identification*

In June 2008, a healthy individual of *Equisetum arvense* was collected from Beringian tundra in the Seward Peninsula of Western Alaska (64◦30- 04" N, 165◦24- 23" W; 6 m.a.s.l.) [23]. The photosynthetic stem was washed in tap water and cut into ca. 2 mm2 segments that were surface-sterilized by agitating sequentially in 95% EtOH for 30 s, 0.5% NaOCl for 2 min, and 70% EtOH for 2 min [23]. Forty-eight tissue segments were surface-dried under sterile conditions and then placed individually onto 2% malt extract agar (MEA) in sterile 1.5 mL micro centrifuge tubes. Tubes were sealed with ParafilmTM and incubated under ambient light/dark conditions at room temperature (ca. 21.5 ◦C) for up to one year. Emergent fungi were isolated into pure culture on 2% MEA, vouchered in sterile water, and deposited as living vouchers at the Robert L. Gilbertson Mycological Herbarium at the University of Arizona. One fungus of interest, isolate AK1128, was used for the present study. This fungus has been deposited at the University of Arizona Robert L. Gilbertson Mycological Herbarium (accession number AK1128).

Total genomic DNA was isolated from fresh mycelium of the isolate AK1128 and the nuclear ribosomal internal transcribed spacers and 5.8s gene (ITS rDNA; ca. 600 base pairs (bp)) and an adjacent portion of the nuclear ribosomal large subunit (LSU rDNA; ca. 500 bp) was amplified as a single fragment by PCR. Positive amplicons were sequenced bidirectionally as described previously [23]. A consensus sequence was assembled and basecalls were made by *phred* [24] and *phrap* [25] with orchestration by Mesquite [26], followed by manual editing in Sequencher (Gene Codes Corp.). The resulting sequence was deposited in GenBank (accession JQ759476).

Because the isolate did not produce diagnostic fruiting structures in culture, two methods were used to tentatively identify isolate AK1128 via molecular sequence data. First, the LSU rDNA portion of the sequence was evaluated using the naïve Bayesian classifier for fungi [27] available through the Ribosomal Database Project (http://rdp.cme.msu.edu/). The Bayesian classifier estimated placement within the Capnodiales (Dothideomycetes) with high support, but placement at finer taxonomic levels was not possible. Therefore, the entire sequence was compared against the GenBank database using BLAST [28]. The top ten BLAST matches were to unidentified dothideomycetous endophytes or uncultured ascomycetous clones, except for two strains of *Colletogloeopsis dimorpha* (strains CBS 120085 and CBS 120086). The matches to *Colletogloeopsis dimorpha* had 97% coverage and a maximum identity of only 92%, and similar levels of match precision to other taxa restricted our taxonomic inference.

Therefore, to clarify the phylogenetic placement and taxonomic assignment of AK1128, two phylogenetic analyses were conducted. First, the top 99 BLAST matches were downloaded from GenBank, four problematic sequences for which quality was suspect were removed, and AK1128 and the resulting dataset was aligned automatically via MUSCLE (http://www.ebi.ac.uk/Tools/msa/ muscle/) with default parameters. The alignment consisted of dothideomycetous endophytes as well as described species of Dothideomycetes that mostly comprised taxa affiliated with some lineages recognized within *Mycosphaerella* (e.g., *Teratosphaeria, Colletogloeopsis, Catulenostroma,* etc.). The alignment was trimmed so that starting and ending points were generally consistent with the sequence length for AK1128 and adjusted manually in Mesquite [26] prior to analysis. The final dataset consisted of 96 sequences and 1084 characters. This dataset was analyzed by maximum likelihood in GARLI (Zwickl, D. J. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion (Ph.D. Dissertation, The University of Texas at Austin, Austin, TX, USA, 2008) using the GTR + I + G model of evolution as determined by ModelTest [29]. The resulting topology indicated that AK1128 had an affinity for *Teratosphaeria, Catenulostroma*, and relatives but was not phylogenetically affiliated with *Colletogloeopsis* (data not shown). Because this analysis included many unknown taxa (endophytes) and we could not root the tree with certainty, the taxon sampling was found to be insufficient to infer with confidence the placement of the AK1128.

A second phylogenetic analysis was therefore conducted using taxa affiliated with *Teratosphaeria, Catulenostroma,* and relatives, as analyzed previously [30]. The alignment of ITS rDNA sequences from [30] was obtained from TreeBase and pruned to include only those taxa of interest based on our analysis above (i.e., the lower half of Figure 1 in [30] in a preliminary analysis, and then only those taxa most closely related within *Teratosphaeria* in the final analysis). The sequence for AK1128 was integrated into the pruned dataset and the data were realigned, adjusted, and analyzed as described above. The resulting dataset consisted of 79 sequences and 582 characters in the preliminary analysis. A bootstrap analysis with 1000 replicates was conducted to assess topological support and the topology was rooted with *Devriesia strelitziae* [30]. The resulting tree suggested that AK1128 is affiliated with *Teratosphaeria* species (data not shown).

Therefore, the dataset was pruned further, focusing only on those *Teratosphaeria* species suggested to be close relatives of AK1128. The final analysis included 42 terminal taxa and 554 characters and was rooted with *Catulenostroma macowanii* based on the topology of the preliminary analysis. The final tree with maximum likelihood bootstrap values is shown in Supplemental Figure S13. The final analysis placed AK1128 with certainty within the genus *Teratosphaeria* (Teratosphaeriaceae), likely in affiliation with *T. associata (*known from *Eucalyptus* [31] and *Protea* [30] from Australia). Given the distinctive geographical origin and host of AK1128, and the sister relationship of AK1128 to known strains of *T. associata* (Supplemental Figure S13), we designate the strain as *Teratosphaeria* sp. AK1128, affiliated with but distinct from known variants of *T. associata*.
