Metacapnodiaceae is one of several sooty mold families in Ascomycota. With the goal of better characterizing species diversity, we determined complete or partial DNA sequence barcodes for ribosomal internal transcribed spacer (ITS) regions for 16 collections of Metacapnodium using a Metacapnodium-specific primer, followed by phylogenetic and morphological analyses. Tapering, moniliform hyphae, cells wider than long, and a distinctive phialidic asexual state were good predictors of membership in a well-supported Metacapnodium clade. Sequences from the 16 collections represent 9 named species of Metacapnodium. Barcoded species include: M. stanhughesii sp. nov., M. vancouverensis, sp. nov., and M. australis comb. nov. Based on morphological characters, we propose M. atro-olivaceus comb. nov., M. novae-zelandiae comb. nov., and M. pacificus comb. nov. We provide a key to the identification of 15 species studied. To investigate the deeper phylogenetic relationships of Metacapnodiaceae, we sequenced partial nuclear ribosomal large subunit (LSU) gene regions from five specimens and elongation factor 1-alpha (ef1-α) gene regions from two specimens. In our analysis of concatenated sequences from ribosomal DNA, ef1-a, and from rpb2, the gene encoding the 2nd largest subunit of RNA polymerase II protein, Metacapnodium appeared within the subclass Chaetothyriomycetidae, class Eurotiomycetes with strong support, and as the sister group to Pleostigmataceae but without strong statistical support. Our study adds Metacapnodiaceae to the clades of enigmatic, slow-growing fungi of harsh environments with lichenized, lichenicolous, resinicolous, and rock-inhabiting niches. Resolving family relationships is relevant to age estimates of Ascomycota, as fossilized Metacapnodium specimens in amber potentially contribute to the calibration of divergence times.

Biological material – Specimens examined, culturing method, and separation of individual species from multi-species tangles

 We selected specimens of Metacapnodiaceae, collected more recently than 1990, where possible, using the Mycology Collections Portal (Miller and Bates 2017). Table 1 contains accession information for the 20 specimens sequenced or appearing in figures. We collected three species of Metacapnodium, one from leaves of Pieris japonica, one from the stem of cultivated Taxus sp. from plants cultivated at the University of British Columbia campus, Vancouver, BC, and one species from Kauaii, Hawaii, from leaves of Metrosideros cf. polymorpha (Table 1). Of the fresh collections, only the specimen from Hawaii grew in pure culture. Successful isolation of the specimen began by pipetting sterile potato dextrose broth over sterile Kimwipe tissues (Kimberly-Clark, Texas, USA) in glass Petri dishes (Keith Seifert 2015, Botanical Society of America Annual Meeting, personal communication). We quickly scraped conspicuous Metacapnodium thalli from fresh leaves of Metrosideros cf. polymorpha over the open Petri dish with a sterile needle, spreading fragments of hyphae. After 7-10 days, we could observe signs of hyphal growth under a dissecting microscope in uncontaminated areas of the wet tissue.

In all work with sooty molds, it was essential to be attentive to the possibility that multiple fungal species were intermingled on the same leaf or twig surface. For morphological study, we carefully untangled small pieces of the sooty mold subiculum, searching for shiny, tapering, moniliform hyphae that characterize Metacapnodium as a genus, and paying close attention to any patterns of cell size, shape, and ornamentation that distinguish somatic hyphae of different species. We teased apart a small piece of sooty mold subiculum from each fresh or herbarium specimen in distilled water on a slide, then mounted the mycelia in 85% lactic acid on slides. Repeated observations from the same specimen were usually necessary to find reproductive structures among the dark colored, densely branching hyphae. We observed and measured any asexual or sexual structures using a Leitz DMRB DIC Microscope (Germany) and photographed structures using a Leica DFC420 Digital Color Camera (Germany).

DNA extraction, amplification, and sequencing

For a phylogenetic analysis of Metacapnodium, we extracted DNA from herbarium specimens, from fresh collections, and from the culture of the unidentified Metacapnodium sp. from Hawaii (DAOMC 252865, CCCM F128, corresponding to dried voucher UBC F33050). To harvest the target species from each specimen, we carefully selected small patches of mycelium under the dissecting microscope, moistening the subicula with a drop of water. We teased hyphae of Metacapnodiaceae apart from other sooty molds or other fungi using a dissecting needle, and placed them on a microscopic slide. Under the compound microscope, more than one fungal taxon was, in some cases, es still evident. When this happened, we re-examined the target specimen under the dissecting microscope and selected several small patches from different parts of the subiculum. From each small patch, we teased hyphae with Metacapnodium spp. morphology apart and transferred them onto at least five slides, to be examined under the compound microscope. If a patch had an individual Metacapnodiaceae species or was at least dominantly occupied by one species, we harvested hyphae from the part of the subiculum where the patch had originated for DNA extraction.

 We extracted DNA from the mycelium using the DNeasy Plant Mini Kit (QIAGEN Inc., Canada), following the manufacturer’s protocol. To check for consistent detection of the same species and to rule out laboratory contamination, we repeated extraction, amplification, and sequencing three times, unless the amount of subiculum in the herbarium specimens was limiting. PCR amplifications targeted ITS, SSU, LSU, and ef1–α regions using universal and specific primers (Table 2). We designed Metacapnodium-specific ITS and ef1–α primers (Table 2) based on alignments including sequences from Metacapnodium neesii (Sugiyama et al. 2020) and from our cultured specimen from Hawaiian leaves (voucher UBC F33050). For PCR, each 25 μl reaction contained a master mix of 10X Taq reaction buffer (2.5 μL), MgSO4 20 mM (2.5 μl), dNTPs 2 mM (2.5 μl), Taq DNA polymerase 5U/µL (0.125 µl) (Bio Basic Inc., USA), Milli-Q water (2.375 μl), 10 μM forward primer (1.25 μl), 10 μM reverse primer (1.25 μl), and genomic DNA diluted 1:1 or 1:9 in water. We performed amplifications as follows: 40 PCR cycles, 2 min at 94 °C (initial denaturation step), 10 s at 94 °C for denaturation, 20 s at about 53°C for annealing, 30 s at 72 °C, and 7 min at 72 °C as a final elongation step. Annealing temperatures varied depending on primers. Amplifications of RPB2 using primers from Sugiyama et al. (2020) were unsuccessful, so they will not be described further.

 We purified PCR products using the QIAquick PCR Purification Kit (QIAGEN Inc., Canada) according to the manufacturer’s protocol. Each 10 μL sequencing reaction contained 2 μL of the DNA (10 ng/μL) (depends on the product length), 2 μL of 2.5 μM primer, 3 μL water, and 3 μL of diluted BigDye® Terminator v3.1 Sequencing Chemistry (Applied Biosystems, Canada). We set cycling conditions for sequencing reactions as follows: 1 min at 96 °C, followed by 25 cycles of 10 s at 96 °C, 5 s at 50°, C and 4 min at 60°C. Using the DyeEx 2.0 Spin Kit (QIAGEN Inc., Canada) according to the manufacturer’s protocol, we purified the sequencing reactions and then submitted them to the Sequencing + Bioinformatics Consortium, UBC, Vancouver, BC, Canada for electrophoresis. We assembled and edited sequences using Sequencher 4.10.1 (Gene Codes Corp., USA).

Phylogenetic taxon sampling

Sampling of the ITS regions included 16 newly determined Metacapnodium sequences, two previously determined Metacapnodium sequences, and two sequences (GenBank MH930326 and MW376663) that are highly similar to Metacapnodium species, even though they were not originally identified as such. We selected outgroups from order Verrucariales and family Pleostigmataceae, based on preliminary BLAST searches and on previously published studies (Gueidan et al. 2007; Gueidan et al. 2014; Chen et al. 2015; Crous et al. 2020; Cometto et al. 2023).

 For analysis of the SSU, LSU, ef1-a, and rpb2 regions, we selected sequences to represent the diversity of classes Dothideomycetes and Eurotiomycetes, with members of Leotiomycetes included as the outgroup based on previous publications (Gueidan et al. 2007; Gueidan et al. 2014; Chen et al. 2015; Crous et al. 2020; Cometto et al. 2023) or on BLAST search results (Suppl. material 1: Table S1). Most of the ribosomal RNA gene sequences were from GenBank. Most of the rpb2 and ef1-a sequences were from whole-genome projects at the Joint Genome Institute (JGI) (Grigoriev et al. 2011; Grigoriev et al. 2014). Where possible, we used sequences from the same fungal isolate to represent each locus.

For many of the ascomycetes closest to Metacapnodiaceae, only partial LSU sequences were publicly available. In its sequenced genomes, JGI does not include transcriptome sequences for ribosomal RNA genes, but we sometimes found LSU sequences by searching JGI EST clusters or genome assemblies. Because LSU sequences are more easily accessed through GenBank, after finding LSU sequences in JGI, we used a sequence with a GenBank accession number instead of a JGI scaffold or EST cluster name, providing that the JGI and GenBank sequences were 100% identical over the aligned region (Suppl. material 1: Table S1).

For alignments, we used MAFFT (http://mafft.cbrc.jp/alignment/server/index.html) with the L–INS–I (iterative refinement) method (Katoh et al. 2019). We edited the alignments manually using Mesquite version 3.81 (Maddison and Maddison 2023). For the two protein-coding genes, we applied minor edits to the alignment in Mesquite to maximize the similarity of aligned, predicted amino acids (Suppl. material 2). Initially, we analyzed each sequence region separately (Suppl. material 3: Table S2). The LSU sequences included unalignable regions, which we removed using GBlocks at its least stringent settings (Castresana 2000; Dereeper et al. 2008). We concatenated the SSU, LSU, ef1-, and rpb2 regions** for combined analysis, and phylogenetic analysis was carried out on data partitioned by region, and for ef1-a and rpb2 regions, also by codon position (Suppl. material 3: Table S2).

Phylogenetic analyses

To compute a matrix of uncorrected distances of the aligned ITS regions, we used the program distmat (http://emboss.open-bio.org/) (Rice et al. 2000). We analyzed individual and concatenated regions with raxmlGUI ver. 2.0.14 (Edler et al. 2021), after selecting appropriate evolutionary models (Darriba et al. 2020). We used 20 replicate searches with different starting trees and estimated support for branches using transfer bootstrap expectations (Zaharias et al. 2023) and Felsenstein branch support (Felsenstein 1985). We applied Bayesian analysis, running on MrBayes 3.2.7a, on Access, CIPRES (Miller et al. 201,0) to the ITS alignment and to the concatenated alignment. Using 10,000,000 generations with 25% burn-in was sufficient for convergence of posterior probabilities, based on standard measures (Suppl. material 3: Table S2). We considered branches in the resulting trees to be well supported when they had at least 90% Felsenstein bootstrap support, 95% support from transfer bootstrap expectations, and a Bayesian posterior probability of 1.0.