Cadophora meredithiae and C. interclivum, new species from roots of sedge and spruce in a western Canada subalpine forest

ABSTRACT Two new species of Cadophora are described based on multigene phylogenetic analyses and phenotypic and ecological characters. The species delimitation was based on concordance of gene genealogies. The cultures of the Cadophora species were isolated from the roots of long-beaked sedge and white spruce from a subalpine forest in western Canada; however, they probably have a broader distribution because their internal transcribed spacer (ITS) sequences have high similarity with a number of GenBank sequences from ecological studies of plant roots. The taxonomy of Cadophora in Leotiomycetes is discussed based on the phylogeny generated in this study. Results from this work will facilitate ecological and evolutionary studies on root-associated fungi.


INTRODUCTION
In 1927, Lagerberg and Melin described Cadophora with C. fastigiata Lagerb. & Melin as the type species of this dematiaceous hyphomycetous genus that produces solitary phialides with distinct hyaline, flared collarettes (Lagerberg et al. 1927). In 1937, based on the similarities of phialide morphology to that of Phialophora verrucosa Medlar (Conant 1937), Conant transferred eight Cadophora species to Phialophora Medlar (Medlar 1915). Reclassification of phialophora-like anamorphs based on morphological characters (Gams 2000) and subsequent DNA sequence-based analyses (Harrington and McNew 2003) demonstrated that Cadophora is distinct from Phialophora and that the latter belongs to Chaetothyriales (Eurotiomycetes) and the former to the Helotiales (Leotiomycetes).
Cadophora species are primarily isolated from living plants, as pathogens or root colonizers, and produce melanized, septate hyphae that aptly include them among the fungi labeled dark septate endophytes (DSEs) (Jumpponen and Trappe 1998;Zijlstra et al. 2005). The best-studied DSEs are the Phialocephala fortinii-Acephala applanata complex (PAC), a group of asexual fungi in Helotiales that are closely related to Cadophora (Wang et al. 2006b). Phialocephala species are similarly characterized by dark-pigmented hyphae and produce conidiophores and hyaline to pale brown phialides with collarettes (Yu et al. 2001;Grünig et al. 2008aGrünig et al. , 2008b. Despite their morphological similarities, the two genera are distinguished based on the complexity of conidiophore branching. Cadophora phialides often occur singly or at most in groups of 2-3, with conidiophores lacking a distinct stipe and a loosely divergent branching pattern, whereas Phialocephala species often have a distinct, dark stipe, terminating in a complexly penicillately branched apex, with many densely packed phialides.
In this study, we propose two new species, Cadophora meredithiae and C. interclivum, isolated from apparently healthy sedge and white spruce roots in a subalpine forest in western Canada. They can be distinguished from PAC species and other related fungi based on phylogenetic analyses, ecology, and morphological characters. We conducted plant-fungal interaction experiments and enzymatic tests to further understand their roles in the ecosystem.

MATERIALS AND METHODS
Fungal isolation.-Roots of long-beaked sedge (Carex sprengelii) and white spruce (Picea glauca) were collected from a subalpine forest (51.122541, −115.382972) in the Harvie Heights of Canmore, Alberta, Canada, in 2015, close to but outside the boundaries of Banff National Park. The soil of the sampling location was an alkaline (pH 8.1), calcareous (30260 Ca ppm, 63.2 Ca:Mg ratio ppm) blue-gray sandy clay. Ten individual plants were sampled for each species. Samples were kept on ice during expedited shipping to the laboratory where they were processed upon arrival. Root samples were rinsed thoroughly to remove soil from the surface, cut into 10-20 mm lengths, then surface disinfected with sequential washes of 95% ethanol for 30 s, 0.5% NaOCl for 2 min, and 70% ethanol for 2 min. After several rinses with sterile water, root samples were dried and cut into 5-mm pieces and then plated on acidified malt extract agar (AMEA; 1.5 mL 85% lactic acid per liter of 2% malt extract agar). Plates were incubated at room temperature (RT) with 12 h light/12 h dark cycles. Fungal cultures were transferred to fresh AMEA and purified by subculturing from emergent hyphal tips.
Morphological study and growth rates.-Isolates were grown on 2% MEA (BD Difco, Franklin Lakes, Maryland, USA) and 2% water agar (WA). Cultures were incubated at 20 C with three replicates. Colony diam was measured after 14 d. The following media also were used in attempts to induce sporulation: cornmeal agar (CMA; BD Difco), oatmeal agar (OA; BD Difco), potato dextrose agar (PDA; BD Difco), filtered ground pine needle agar (Luchi et al. 2007), WA amended with autoclaved Carex sprengelii seeds, and WA amended with autoclaved switchgrass (Panicum virgatum) shoots. Cultures were incubated at 25 C in the dark with three replicates and were checked weekly for 12 mo. Color names used in colony descriptions follow Ridgway (1912).
DNA extraction, amplification, and sequencing.-Genomic DNA was extracted from fungal mycelium using the UltraClean Soil DNA isolation kit (MoBio, Carlsbad, California, USA) following the manufacturer's instructions. Polymerase chain reaction (PCR) was performed with Taq 2× Master Mix (New England BioLabs, Ipswich, Maine, USA), following the manufacturer's instructions. Primers used were ITS1 and ITS4 for the internal transcribed spacer (ITS1-5.8S-ITS2 = ITS) region, NS1 and NS4 for partial nuc 18S rRNA genes (18S; White et al. 1990), ITS1 and LR5 for the D1/D2 region of the nuc 28S rRNA genes (28S; Rehner and Samuels 1995), and RPB1 Af (Stiller and Hall 1997) and RPB1 CrRev (Matheny et al. 2002) for the largest subunit of RNA polymerase II (RPB1) gene, EF1-728F and EF1-986R (Carbone and Kohn 1999) for the translation elongation factor 1-alpha (TEF1-α) gene, and BTCadF/R (Travadon et al. 2015) for β-tubulin (β-TUB) gene. PCR conditions for the ITS, 18S, and the 28S consisted of an initial denaturation step at 95 C for 2 min, 35 cycles of 95 C for 45 s, 54 C for 45 s, and 72 C for 1.5 min, and a final extension at 72 C for 5 min. For RPB1, the PCR conditions included an initial denaturation step at 95 C for 2 min, 35 cycles of 95 C for 60 s, 55 C for 1.5 min, and 72 C for 2 min, and a final extension at 72 C for 10 min. For TEF1-α, the PCR conditions included an initial denaturation step at 95 C for 2 min, 35 cycles of 95 C for 60 s, 55 C for 1.5 min, and 72 C for 2 min, and a final extension at 72 C for 10 min. For β-TUB, the PCR conditions included an initial denaturation step at 95 C for 2 min, 35 cycles of 95 C for 60 s, 56 C for 30 s, and 72 C for 60 s, and a final extension at 72 C for 10 min. PCR products were purified with ExoSAP-IT (Affymetrix, California) and sequenced with the same primers used for PCR.
Sequence alignment and phylogenetic analyses.-Six representative isolates of the new species (BAG2, BAG4, BAP6, BAP13, BAP33, and BAP37) were included in the phylogenetic analyses along with reference sequences for other Leotiomycetes species (TABLE 1). A multigene (ITS, 28S, 18S, TEF1-α, and RPB1) alignment that included the six new sequences and 14 reference sequences of Helotiales also was assembled. Sequences were aligned with MUSCLE (Edgar 2004) and manually adjusted. Genealogical concordance was evaluated with the nonparametric Templeton Wilcoxon signed-rank test in PAUP* 4.0b10 (Swofford 2002), with 95% bootstrap consensus trees as constraints. No significant conflicts were found between the individual gene data sets, so we constructed the combined phylogenetic tree. Because many Leotiomycetes species that are related to the new taxa lack sequences for some of these loci, we assembled single-locus data sets to maximize taxon coverage in our phylogenetic analyses. Maximum likelihood (ML) tree was generated with MEGA6 (Tamura et al. 2013). Models with the lowest Bayesian information criterion (BIC) scores were considered to describe the substitution pattern the best. The best models for 28S and multigene data sets were Kimura 2-parameter and Tamura 3-parameter, respectively. Initial tree(s) for the heuristic search were obtained automatically by applying neighbor-joining (NJ) and BioNJ algorithms to a matrix of pairwise distances estimated using the maximum composite likelihood approach and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites for all data sets. In addition, the rate variation model allowed for some sites to be evolutionarily invariable for the five- min, and 70% ethanol for 1 min, rinsed with sterile distilled H 2 O, and allowed to germinate in the dark at 25 C for 3 d. Plates of Agargel (Sigma-Aldrich, St. Louis, Missouri, USA), a medium suitable for plant cell culture, were made following the manufacturer's instructions. The Agargel in the plate was cut in half, and one half of the Agargel was removed. On the cut surface of the remaining half of the Agargel in the plate, three 10 mm × 10 mm × 5 mm plugs from a 1-wk-old fungal culture grown on MEA were placed equidistant from one another. Germinated switchgrass seeds with visible radicle were then placed on the plugs. Sterile MEA plugs were used as a negative control. Cultures were incubated at 25 C under 12 h light/12 h dark cycle with nine replicates. The same protocol was used for surface sterilized leek seeds. Root length was measured 7 d after inoculation.
Microscopy.-Fungal strains were visualized directly on the Agargel growth plates for infection structures such as hyphopodia or appresoria. Micromorphology was studied on slide mounts in lactophenol. All images were obtained by a Nikon DS-Fi1 camera mounted on a Nikon Eclipse 80i compound microscope using the 40× or 60× objectives (Nikon, Melville, New York, USA).
Images were measured and analyzed using the Nikon NIS-Elements D3.0 software.
Enzyme experiments.-The methods of Rice and Currah (2005) were followed to test amylase, gelatinase, and lipase activities. Phosphatase and cellulase activity tests followed Pikovskaya (1948) and Gupta et al. (2012), respectively. Cultures of BAG2, BAG4, Acidomelania panicicola (CBS 137156), Barrenia panicia (WSF 1R37), Barrenia taeda (WSF 14P22), Acephala macrosclerotium (PP16P100), Cadophora luteo-olivacea (CBS 141.41), and Cadophora malorum (CBS 165.42) were grown at RT on modified Melin-Norkrans (MMN) medium agar plates containing the target macromolecule with or without an indicator (Rice and Currah 2005). Sterile MEA plugs were used as a negative control. Amylase activity was scored as follows: after isolates had grown for 3 wk on plates of MMN containing 2 g/L potato starch, plates were flooded with iodine solution and the solution decanted after several minutes to reveal a clear zone around the mycelium in strains positive for this enzyme. Phosphatase activity was scored at 7 d on Pikovskaya medium supplemented with 0.025 g/L bromophenol blue (Pikovskaya 1948). Cellulase activity was scored on MMN medium amended with 2 g/L carboxymethylcellulose and 0.2 g/L Congo red at 5 wk Figure 2. Maximum likelihood phylogenetic tree inferred from combined ITS, 28S, 18S, TEF1-α, and RPB1 gene sequence data sets. Bootstraps higher than 70% have thickened branches. (Gupta et al. 2012). Gelatin medium had 12% gelatin added to MMN medium instead of agar; liquefaction after 3 wk was considered a positive reading for this enzyme. Lipase synthesis on MMN containing 0.1 g/L CaCl 2 and 10 mL/L Tween 20 (polyoxyethylene sorbitan monolaurate; Sigma-Aldrich) was determined by the formation of visible crystals beneath the mycelium after 16 wk.

RESULTS
Culture morphology and growth rates.  Figure 3. Maximum likelihood phylogenetic tree inferred from the ITS sequence data set that includes environmental sequences from GenBank that are closely related to the new Cadophora species. Bootstraps higher than 70% have thickened branches.
nor did they induce the isolates to produce infection structures.
Enzyme experiments.-Cadophora meredithiae BAG2 and C. interclivum BAG4 were both negative for amylase and phosphatase activities. Although A. panicicola and A. macrosclerotium were negative for amylase activity, they were positive for phosphatase activity. Cadophora luteo-olivacea, C. malorum, B. panicia, and B. taeda were positive for amylase and phosphatase activities. Cadophora orchidicola and C. orientoamericana were strongly positive for amylase activity, whereas C. gregata and C. novi-eboraci scored weakly positive with only a faint halo. All were negative for phosphatase activity. All tested strains demonstrated cellulase activity except for C. interclivum BAG4. Gelatinase activity was scored positive for all the tested fungi except for C. gregata and B. panicia. Lipase synthesis was scored positive for C. meredithiae BAG2, A. macrosclerotium, and B. taeda, while all other tested strains scored negative (TABLE 2).
Sequence data and phylogeny.-There were 4631 characters in the five-gene alignment, including 673 from ITS, 533 from 28S, 828 from TEF1-α, 610 from RPB1, and 1987 from 18S. There were 478 characters in the 28S alignment with the expanded taxon sampling, 641 characters in the ITS alignment with additional taxa, and 342 characters in TEF1-α, 676 characters in β-TUB, and 610 characters in RPB1 alignments with additional taxa. Maximum likelihood trees based on 28S, five-gene sequences, and ITS are shown in FIGS. 1-3. The other single-gene trees, including the RPB1, TEF1-α, and the β-TUB trees, are provided in SUPPLEMENTARY FIGS. 1-3. A separate single-locus phylogenetic analysis for the fivegene alignment data set also was performed, and the trees are provided in SUPPLEMENTARY FIGS. 4-8. All phylogenies supported that the new isolates belong to the Cadophora clade in Helotiales. The five-gene tree showed that these new isolates were closely related to A. panicicola, Mollisia cinerea, Acephala applanata, and Barrenia. The five-gene tree and the individual β-TUB, TEF1-α, and RPB1 trees all indicated that Cadophora is monophyletic and recognized two groups among the new isolates: BAG4, BAP33, and BAP37 formed a well-supported group, described below as C. interclivum, whereas isolates BAG2, BAP6, and BAP13 formed another, described as C. meredithiae. The ITS tree, however, only supported the grouping of isolates BAG4, BAP33, and BAP37. Within-group phylogenetic relationships varied among different single-gene   Etymology: The epithet honors Dr. Meredith Blackwell, whose research and teaching efforts have substantially advanced the field of mycology.
Molecular description: Cadophora meredithiae differs from C. interclivum by the following fixed nucleotides in the largest subunit of RNA polymerase II gene Etymology: The epithet describes the location where the fungi were collected, "inter" meaning between and "clivum" meaning slopes.
Molecular description: Cadophora interclivum differs from C. meredithiae by the following fixed nucleotides in the largest subunit of RNA polymerase II gene Notes: Cadophora meredithiae differs from C. interclivum by having smaller, pyriform conidia (3-6 µm long), and a slower growth rate on MEA (54 mm/20 d), versus C. interclivum with elongated conidia (4-7 µm long), and MEA growth of 67 mm/20 d. In addition, the RPB1, β-TUB, and TEF1-α gene phylogenies support the separation of the two species. Based on the phylogenetic analysis, the two new species are closely related to Cadophora luteo-olivacea, C. malorum, and C. gregata. However, C. meredithiae and C. interclivum are associated with the roots of apparently healthy plants, whereas C. luteo-olivacea, C. malorum, and C. gregata are pathogenic. In addition, the two new species have solitary phialides compared with C. luteo-olivacea (1-3 phialides from a supporting cell). Cadophora meredithiae can also be differentiated from C. luteo-olivacea in having shorter conidia (3-6 vs. 4-10 μm long).

DISCUSSION
In this study, we isolated two new dark septate endophyte (DSE) species from sedge and conifer roots in alkaline, calcareous soil from a subalpine forest in Alberta, Canada. The phylogenetic analyses indicate that the two new species belong to Cadophora (Leotiomycetes) and that Cadophora is closely related to Acidomelania, Hyaloscypha, Mollisia, the Phialocephala fortinii-Acephala applanata complex (PAC), and Varicosporium. Our previous surveys of root-associated fungi in the oligotrophic pine barrens also revealed a number of novel DSEs, but the pine barrens soil is acidic and those DSEs belong in different lineages in Leotiomycetes (Luo et al. 2014a(Luo et al. , 2014bWalsh et al. 2014Walsh et al. , 2015. Cadophora differs from Varicosporium by its association with terrestrial plant roots; Varicosporium species are often saprobes on submerged, aquatic plants (Ingold 1942;Kegel 1906). Although the taxa of the PAC are also root endophytes, they exhibit more complex phialide arrangements than those in Cadophora. Mollisia also can be differentiated from Cadophora by its Phialocephala-like anamorphs. Moreover, Cadophora has 93% or less ITS sequence similarities to the above-mentioned close relatives. The family placement of Cadophora is uncertain because Leotiomycetes phylogeny is poorly resolved and several families in this class probably are polyphyletic (Wang et al. 2006a).
The ITS is the selected fungal DNA barcoding gene and is useful for species identification for most fungi; however, it is not an appropriate marker for species recognition for certain lineages, such as Fusarium and Alternaria (Geiser et al. 2004;Schoch et al. 2012). In this study, we found that the three sequenced protein-coding genes, especially the RPB1 gene, have a number of fixed substitutions between C. meredithiae and C. interclivum and resolved the species relationships in Cadophora better than the ITS. Therefore, we delimited the two species based on the combined analysis rather than only relying on the ITS sequences. Based on the phylogenies reconstructed from this study, we defined the genus Cadophora as a monophyletic clade in Helotiales that includes the type species C. fastigiata; the two new species C. meredithiae and C. interclivum; and C. luteo-olivacea, C. malorum, C. gregata, C. orchidicola, C. orientoamericana, C. melinii, C. viticola, C. spadicis, and C. novi-eboraci. The two Collembolispora species that were isolated from freshwater foam in the Czech Republic also are placed in the Cadophora clade based on our ITS and 28S phylogenies. The Collembolispora species are morphologically distinct from Cadophora by producing septate macroconidia except that C. barbata is dimorphic with a cadophora-like syanamorph (Marvanova et al. 2003;Crous et al. 2012). The placement of Collembolispora needs confirmation from proteincoding genes, and such sequences currently are unavailable.
BLAST searches against public records in GenBank indicated that the new Cadophora species have a broad distribution. Twenty-seven ITS sequences in GenBank had 99-100% identities with that of C. meredithiae BAG2 and that of C. interclivum BAG4, for example, JN859255 from Juniperus communis root in Hungary (Knapp 2012), KX354285 from Epipactis sp. root in Germany (Schiebold et al. 2017), KT268959 from Microthlaspi perfoliatum root in Bulgaria (Glynou et al. 2016), KP278162 from Chloraea grandiflora root in Chile (Herrera et al. 2017), KF618066 from spruce roots in Alaska (Taylor et al. 2014), and KR230080 from Vincetoxicum rossicum root in eastern Canada (Day et al. 2016). The host plants of the matched sequences in GenBank vary considerably from terrestrial orchids and herbaceous perennials to grasses and conifers, but they all come from healthy plant roots that prefer calcareous soils in subalpine regions (Wiemken et al. 2001;Selosse et al. 2004).
Currently known Cadophora species appear to be considered plant pathogens, root associates, or wood colonizers. In the TEF1-α and β-TUB trees from this study, the following groups were noticed: pathogens of trees and grapevines C. fastigiata, C. melinii, C. novieboraci, C. orientoamericana, C. spadicis, and C. viticola; saprobic, weak pathogens C. luteo-olivacea and C. malorum; and finally root-associated endophytes C. interclivum and C. meredithiae. Under certain conditions, these fungi may change from endophytes to pathogens. The hypothesis of climate-driven shifts from endophytes/saprobes to pathogens was supported by Blanchette et al. (2004Blanchette et al. ( , 2010 and Arenz and Blanchette (2009) whose studies indicated the presence of C. fastigiata, C. luteo-olivacea, and C. malorum as saprobes on wooden structures in the Antarctic Peninsula and Canadian High Arctic. However, Gramaje et al. (2011) and Navarrete et al. (2011) both found C. luteo-olivacea to be pathogenic in grapevines in Spain and Uruguay. The enzyme tests further reflected this clustering by showing that both C. luteoolivacea and C. malorum are capable of degrading a range of carbon sources and releasing soluble phosphorus, suggesting their trophic modes could vary from active decomposer to plant pathogen depending on their need to exploit nutrients from various sources. Enzyme test studies conducted by Day and Currah (2011) using the same strain of C. luteo-olivacea rendered the same results. Although C. novi-eboraci and C. orientoamericana could also degrade a range of carbon sources, they were unable to release phosphorus. Cadophora meredithiae showed some capability for cellulose and fat degradation, whereas C. interclivum showed none; however, both could liquefy gelatin to varying degrees. Because of limited sampling in this and previous studies, the trend suggested here needs further testing with more data.
In this study, we described two new species of root-colonizing fungi associated with plants living in an alkaline, nutrient-poor environment. The phylogenetic and taxonomic results, enzymatic tests, and the plant-fungal interaction results reported here will aid future ecological and evolutionary studies on root-associated fungi.