Notes on Trochila (Ascomycota, Leotiomycetes), with new species and combinations

Abstract Studies of Trochila (Leotiomycetes, Helotiales, Cenangiaceae) are scarce. Here, we describe two new species based on molecular phylogenetic data and morphology. Trochilabostonensis was collected at the Boston Harbor Islands National Recreation Area, Massachusetts. It was found on the stem of Asclepiassyriaca, representing the first report of any Trochila species from a plant host in the family Apocynaceae. Trochilaurediniophila is associated with the uredinia of the rust fungus Ceroteliumfici. It was discovered during a survey for rust hyperparasites conducted at the Arthur Fungarium, in a single sample from 1912 collected in Trinidad. Macro- and micromorphological descriptions, illustrations, and molecular phylogenetic analyses are presented. The two new species are placed in Trochila with high support in both our six-locus (SSU, ITS, LSU, rpb1, rpb2, tef1) and two-locus (ITS, LSU) phylogenetic reconstructions. In addition, two species are combined in Trochila: Trochilacolensoi (formerly placed in Pseudopeziza) and T.xishuangbanna (originally described as the only species in Calycellinopsis). This study reveals new host plant families, a new ecological strategy, and a new country record for the genus Trochila. Finally, our work emphasizes the importance of specimens deposited in biological collections such as fungaria.


Introduction
The genus Trochila Fr. (Ascomycota, Leotiomycetes) was erected by Fries (1849) to accommodate four species previously placed in Phacidium Fr., Sphaeria Haller, and Xyloma Pers. Trochila craterium (DC) Fr. was the first species listed by Fries, based on Sphaeria craterium DC., which was later selected by Clements and Shear (1931) as the type species of Trochila. The other three species included by Fries (1849)  Only the genus and one species (T. laurocerasi) were briefly described by Fries (1849). However, the type species, T. craterium, was well described macromorphologically by Lamarck and de Candolle (1805). The description can be translated loosely from French as "a fungus growing on the lower surface of ivy leaves, initially forming a flat white disc, then turning blackish and concave opening by a split along radial lines, the disc usually surrounded by a whitish membrane" (Lamarck and de Candolle 1805). Later, the generic concept was expanded to include other types of apothecial opening. Rehm (1896) remarked that the covering layer of the apothecia could also open completely like a lid depending on host characters such as cuticle thickness. After the inclusion of this new character describing the genus, Stegia ilicis (Chevall.) Gillet was transferred as Trochila ilicina (Nees ex Fr.) Courtec (Crouan and Crouan 1867;Rehm 1896).
In our current circumscription of the genus Trochila, apothecia are sunken in the host tissues and hymenia are exposed either by splitting along radial lines or by splitting into a number of lobes that roll outward exposing the hymenium. The excipulum is composed of dark, globose-angular cells; asci contain eight ellipsoid, hyaline ascospores with oil guttules (except T. substictica Rehm and T. tetraspora E. Müll. & Gamundí, which both have asci containing four ascospores); and paraphyses possess yellowish guttules (Dennis 1978;Baral and Marson 2005). Thirty-three names have been applied in the genus (Index Fungorum 2021). Jaklitsch et al. (2016) suggest that only ca. 10 names should be accepted.
Most species of Trochila have been described from their sexual morph. The asexual morph has the characteristics of the form-genus Cryptocline Petr. (Morgan-Jones 1973;Kiffer and Morelet 2000;Hyde et al. 2011). Two species of Trochila have been linked to their asexual morphs: T. craterium to C. paradoxa (De Not.) Arx and T. laurocerasi to C. phacidiella (Grove) Arx (von Arx 1957). The paucity of culture and molecular data of both Cryptocline and Trochila species has hindered the linkage of sexual and asexual morphs for most species. Trochila viburnicola Crous & Denman was the first species of the genus to be described based on the combination of morphology and molecular data, but only its asexual morph is known (Crous et al. 2018). The species was named referring to its host, Viburnum sp. (Dipsacales, Adoxaceae). In addition to T. viburnicola, two other species have been reported on this host genus, but only from their sexual morph, T. ramulorum Feltgen and T. tini (Duby) Quél. [currently Pyrenopeziza tini (Duby) Nannf.]. Due to the lack of sequences or cultures of these two species, a comparison with T. viburnicola is impossible (Feltgen 1903;Crous et al. 2018).
Most Trochila members have a restricted record of geographical distribution and ecological strategy. Trochila records typically originate from the Northern Hemisphere limited to temperate regions in Europe and North America (Ziolo et al. 2005;Stoykov and Assyov 2009;Crous et al. 2018;Stoykov 2019;Global Biodiversity Information Facility 2020). Nonetheless, a number of putative Trochila reports are known from the Southern hemisphere (Spegazzini 1888(Spegazzini , 1910(Spegazzini , 1921Rehm 1909;Gamundí et al. 1978). In addition, species of Trochila are typically recorded as saprotrophs on dead leaves and branches of both herbaceous plants and trees. However, a few species have been found infecting living plant tissues. Trochila ilicina is reported as both a weak parasite and a saprotroph because of its presence on living, decaying, and fallen leaves of Ilex aquifolium (Aquifoliales, Aquifoliaceae) (Ziolo et al. 2005), T. laurocerasi as a parasite of living leaves of Prunus laurocerasus (Rosales, Rosaceae) (Gregor 1936), and T. symploci as a pathogen of living leaves of Symplocos japonica (Ericales, Symplocaceae) (Hennings 1900;Stevenson 1926).
Here, we describe two new species, T. bostonensis and T. urediniophila, collected at the Boston Harbor Islands National Recreation Area, Massachusetts and at Port of Spain, Trinidad, respectively. We also make two new combinations in Trochila based on morphological studies and phylogenetic analyses. We reveal two new host plant families (Apocynaceae and Asparagaceae) and a new ecological strategy (fungicolous symbiont) for the genus. Finally, we provide a comparative table of characters, based on literature review, for all currently accepted species of Trochila (sensu Index Fungorum 2021).

Collected samples
Samples were collected in the field and from fungaria. One collection of Trochila was discovered during the Boston Harbor Islands (BHI) National Recreation Area fungal ATBI (Haelewaters et al. 2018a). In this project, above-ground, ephemeral fruiting bodies of non-lichenized fungi were collected. In the field, specimens were placed in plastic containers or brown paper bags. BHI-F collection numbers were assigned. Date, specific locality when applicable, GPS coordinates, substrate, and habitat notes were recorded. Specimens were dried using a Presto Dehydro food dehydrator (National Presto Industries, Eau Claire, Wisconsin) set at 35 °C for 7-9 hours. Collections were packaged, labeled, and deposited at FH. A second Trochila collection came to our attention during a survey for hyperparasites of rust fungi at PUR. The specimen was found on the uredinia of the rust fungus Cerotelium fici on the underside of Ficus maxima leaves. Fungarium acronyms follow Thiers (continuously updated).

Morphological studies
Methods to study the morphological characteristics of the Trochila specimens followed the process given in Baral (1992). Macro-and micromorphological features were examined on both fresh and dried apothecia for the specimen collected at the BHI and on dried apothecia for the specimen found at PUR. Apothecia from the BHI specimen were observed under an EZ4 stereomicroscope (Leica, Wetzlar, Germany) and studied under a B1 compound microscope (Motic, Barcelona, Spain). Apothecia from the PUR specimen were examined on an SZ2-ILTS dissecting microscope (Olympus, Center Valley, Pennsylvania) and studied using a BH2-RFCA compound microscope (Olympus). Sections of apothecia were cut free-hand and mounted in water or pretreated in 5% KOH. Sections were also mounted in Melzer's reagent with and without KOH-pretreatment to determine dextrinoid or amyloid reactions. At least 10 measurements were made for each structure at 400-1000× magnification. Measurements for each character are given as (a-)b-c(-d), with b-c indicating the 95% confidence interval and a and d representing the smallest and large single measurement, respectively. Macro-and microphotographs were taken with a USB Moticam 2500 camera (Motic) (BHI specimen) or an Olympus SC30 camera (PUR specimen). Measurements were made using the following software suites: Motic Images Plus 2.0 and cellSens Standard 1.18 Imaging Software (Olympus). Color coding refers to Kelly (1965). Abbreviations were adopted from Baral (1992) and Baral and Marson (2005) Haelewaters et al. (2018a). We amplified the following loci: nuclear small and large ribosomal subunits (SSU and LSU), internal transcribed spacer region of the ribosomal DNA (ITS), RNA polymerase II second largest subunit (rpb2), and translation elongation factor 1-α (tef1). Primer combinations were as follows: NS1/NS2 and NS1/ NS4 for SSU (White et al. 1990); LR0R/LR5 for LSU (Vilgalys and Hester 1990;Hopple 1994); ITS1F/ITS4, ITS9mun/ITS4A, and ITS5/ITS2 for ITS (White et al. 1990; Gardes and Bruns 1993;Egger 1995); RPB2-5F2/fRPB2-7cR for rpb2 (Liu et al. 1999;Sung et al. 2007); and EF1-983F/EF1-1567R and EF1-983F/EF1-2218R for tef1 (Rehner and Buckley 2005 (Don et al. 1991;Haelewaters et al. 2018b). PCR products were visualized by gel electrophoresis. Purification of successful PCR products and subsequent sequencing in both directions were outsourced to Genewiz (South Plainfield, New Jersey). Raw sequence reads were assembled and edited using Sequencher version 5.2.3 (Gene Codes Co., Ann Arbor, Michigan).

Sequence alignment and phylogenetic analysis
Edited sequences were blasted against the NCBI GenBank nucleotide database (http:// ncbi.nlm.nih.gov/blast/Blast.cgi) to search for closest relatives. For phylogenetic placement of our isolates, we downloaded SSU, ITS, LSU, rpb1, rpb2, and tef1 sequences of Trochila from GenBank. We also downloaded sequence data of selected clades of Helotiales, mainly from Pärtel et al. (2017) but also other sources (details in Table 1), as a basis for our six-locus phylogenetic analysis. We selected representative taxa of Cenangiaceae, Cordieritidaceae, Rutstroemiaceae, and Sclerotiniaceae, with taxa in the family Chlorociboriaceae serving as outgroups (Johnston et al. 2019). Alignment of DNA sequences was done for each locus separately using MUSCLE version 3.7 (Edgar 2004), available on the Cipres Science Gateway 3.3 (Miller et al. 2010). The aligned sequences for each locus were concatenated in MEGA7 (Kumar et al. 2016). Maximum likelihood (ML) inference was performed using IQ-TREE from the command line (Nguyen et al. 2015) under partitioned models (Chernomor et al. 2016). Nucleotide substitution models were selected under Akaike's information criterion corrected for small sample size (AICc) with the help of the built-in program ModelFinder (Kalyaanamoorthy et al. 2017). Ultrafast bootstrap analysis was implemented with 1000 replicates (Hoang et al. 2017).
For the purpose of species delimitation, we constructed a second dataset of ITS-LSU consisting of isolates of Trochila and closely related taxa in the family Cenangiaceae. We included Trochila spp., Calycellinopsis xishuangbanna, and Pseudopeziza colensoi, with Cenangiopsis spp. serving as outgroup. In this analysis, we included T. ilicina, for Table 1. Sequences used in phylogenetic analyses. Accession numbers in boldface indicate sequences that were generated during the course of this study.  which only a single ITS sequence is available. The same methods as above were applied: alignment using MUSCLE (Edgar 2004), selection of nucleotide substitution models with the help of ModelFinder (Kalyaanamoorthy et al. 2017), ML using IQ-TREE (Nguyen et al. 2015;Chernomor et al. 2016;Hoang et al. 2017
Etymology. bostonensis -referring to Boston, Massachusetts, the locality of the type collection.
Notes. Trochila bostonensis is the only species of the genus found on a member of Apocynaceae (Table 2). It was growing in the outer layer of a dead stem of Asclepias syriaca, which had fallen on the ground. The host was close to the shore in a shrubby thicket of Rhus. There are two similar species. Trochila laurocerasi has wider asci (6.0-8.0 µm vs. 4.5-6.0 µm) and larger ascospores (6.3-10 × 2.5-4.6 µm vs. 5.8-7.2 × 2.6-2.8 µm) compared to T. bostonensis. Ascus and ascospore length are similar in T. bostonensis and T. craterium, although ascospores are slightly larger in T. craterium. The two species mostly differ in the width of their asci (7-12 µm in T. craterium vs. 4.5-6.0 µm in T. bostonensis). We used the measurements in dead state to compare T. bostonensis with other species in the genus (see Table 2). Diagnosis. Differs from Trochila ilicina in ecological strategy (fungicolous symbiont); sizes of asci (102.4-111.2 × 10.5-11.6 µm), ascospores (9.0-9.7 × 6.3-7.1 µm), paraphyses (3.2-3.6 µm wide); and the inamyloidity of its ascus apex. Etymology. Referring to the intimate association of the fungus with the uredinia of Cerotelium fici.
asci, inamyloid ascus apex, and wider apex of the paraphyses. The uredinia of the host fungus, C. fici, become a solidified mass that changes in color from dark orange yellow (72.d.OY) without apothecia of Trochila to brownish black (65.brBlack) where apothecia are present.
A second duplicate of the Reliquiae Farlowiana No. 723 is also deposited at PUR (accession PUR F1098). However, no apothecia were present on this specimen, nor could additional specimens of T. urediniophila be found on any of the other specimens of C. fici housed at PUR. At least eight other duplicates are housed at BPI, CINC, CUP, F, ISC, MICH, and UC (MyCoPortal 2020). It is unknown whether any of them may host T. urediniophila.

Mycobank No: 836591
≡ Cenangium colensoi Berk., Hooker, Bot. Antarct. Voy. Erebus Terror 1839-1843, II, Fl. Nov.-Zeal.: 201 (1855. [Basionym] = Pseudopeziza colensoi (Berk.) Massee, J. Linn. Soc., Bot. 31: 468 (1896) Notes. Cenangium colensoi is described from dead leaves of Cordyline sp. (Asparagales, Asparagaceae) in New Zealand (Hooker 1855). The host had been mistakenly reported as Phormium (Asparagales, Asphodelaceae) by Berkeley in Hooker (1855) and only recently corrected after re-study of the type collection (Landcare Research 2020). Cenangium colensoi was later combined in Pseudopeziza and described in more detail by Massee (1896). Both authors commented on the watery-grey disc and brownish receptacle of the apothecia. The apothecia develop among the rigid vascular bundles of the epidermis, first covered by the cuticle, then erumpent and opening by a narrow slit, becoming discoid when mature (Hooker 1855;Massee 1896). The habit of this fungus fits well with typical macromorphological features of the genus Trochila -a dark brown to black receptacle, which develops beneath the host tissues and eventually becomes erumpent to expose the hymenium by splitting along radial lines or by its splitting into lobes (von Höhnel 1917;Greenhalgh and Morgan-Jones 1964;Dennis 1978;Baral and Marson 2005). Microscopically, P. colensoi was described with a parenchymatous excipulum (angular-globose or isodiametric cells), hyaline under the hymenium and dark brown at the cortex (Berkeley in Hooker 1855; Massee 1896), which is also in agreement with the excipular features of Trochila species. Finally, the hymenium of P. colensoi was described as composed of inamyloid, 8-spored asci with elliptical hyaline ascospores and slender paraphyses (op. cit.).
In 2018, P.R. Johnston collected two specimens (PDD:112240, PDD:112242, Landcare Research 2020) on leaves of Cordyline australis (Asparagaceae). The morphology, ecology (host), and locality of these new collections agree with P. colensoi. The photographs of both specimens reveal features such as guttules in ascospores and paraphyses, protruding hyaline cells in the cortical layer of the upper flank and margin, and hyaline gelatinized hyphae covering the dark globose-angular cells of the ectal excipulum at the base and lower flanks. The latter excipular feature of the receptacle is reminiscent of Zhuang's (1990) description of Calycellinopsis xishuangbanna. An ITS sequence of this species was generated from the recent material (PDD:112240) and included in the Leotiomycetes-wide ITS phylogeny of Johnston et al. (2019). Their results and those in this study (Figs 1, 2) show that P. colensoi is placed among species of Trochila.  Zhuang, Mycotaxon 38: 121 (1990).

[Basionym]
Notes. The genus Calycellinopsis was proposed with a single species, C. xishuangbanna, which is a petiole-inhabiting fungus (Zhuang 1990). The genus was placed in Dermateaceae because of its isodiametric dark brownish excipular cells (Zhuang 1990). In 2002, a second collection of the same species was sampled (HMAS:187063), which was sequenced (Zhuang et al. 2010). Additional morphological details were provided, and the genus was placed in Helotiaceae (Zhuang et al. 2010). Trochila was treated in Dermateaceae until recently because of its excipular features (Fuckel 1869; Karsten 1869; Saccardo 1884; Lambotte 1888; Lumbsch and Huhndorf 2010). Collections of Calycellinopsis have a well-developed excipulum, with an outer layer of angular to isodiametric cells with brownish walls and cortical cells at flanks and margin with protruding hyaline cells. The medullary excipulum is subhyaline and composed of textura angularis to textura intricata (Zhuang 1990;Zhuang et al. 2010).
Species in Trochila usually have a poorly developed excipulum. For example, T. bostonensis and T. craterium produce only a thin layer of globose to angular dark excipular cells (von Höhnel 1917;Greenhalgh and Morgan-Jones 1964;Baral and Marson 2005). However, other species, such as T. laurocerasi and T. urediniophila, have a well-developed excipulum (op. cit.). The excipulum of Calycellinopsis is very similar to those species of Trochila with a well-developed excipulum, composed of an outer layer of dark textura globulosa-angularis and an inner layer of hyaline medulla made of textura angularis-porrecta-intricata. At the flanks and margin of the excipulum, Calycellinopsis has protruding hyaline cells similar to Trochila species with a welldeveloped excipulum (Fig. 4). Although limited details about the living features can be obtained from the original description of Calycellinopsis, its hymenial features are consistent with Trochila. The ascospores of Calycellinopsis are described with several guttules, a feature that is also observed in species of Trochila.

Taxonomy of Trochila
This study represents the first attempt to investigate the systematics of Trochila using both morphological features and DNA sequences. We have added four species to Trochila, bringing the total number of species described in the genus to 37. Most Trochila species have been delimited based on the size of asci and ascospores, but we find that amyloidity of ascus apex, excipular features, details of the paraphyses, and presence vs. absence of guttules are also diagnostic ( Table 2). For this study, we also applied a two-dataset approach for phylogenetic analyses (e.g., Aime and Phillips-Mora 2005;Haelewaters et al. 2019). Our phylogenetic reconstruction of a six-locus dataset resolved Trochila as polyphyletic with respect to C. xishuangbanna and P. colensoi (Fig.  1). Because morphological data of these two taxa agree with Trochila, we recombined them in this genus. The second, two-locus dataset was used for species delimitation, which showed T. bostonensis and T. urediniophila as distinct from the other Trochila species. Our molecular phylogenetic results (Figs 1, 2) and morphological comparisons of Trochila species (Table 2) will facilitate future taxonomic studies in the genus.

Host associations
Thus far, members of Trochila have been reported from 31 families of both monocots and dicots (Table 2). In this study, we add two plant family hosts, Apocynaceae (for T. bostonensis) and Asparagaceae (for T. colensoi). In addition, we reveal a new ecological niche (for T. urediniophila) -a species that associates with uredinia of the rust species Cerotelium fici. This sample was collected in 1912 as a rust specimen and deposited in the Arthur Fungarium (PUR) at Purdue University. More than a century later, the exsiccatae sample was scanned for the presence of hyperparasites of rust fungi from South America. Apothecia of T. urediniophila were found exclusively on uredinia without any direct contact with the host plant. Due to the age and limited available material, ultrastructural examinations of the interaction between these two fungi could not be made. However, T. urediniophila is the first species in the genus that fruits exclusively from another fungus, hinting at more complex associations among Trochila species and other fungi on which they might act as mycoparasites.

Trochila in the Neotropics
South America is known to be one of the most biodiverse continents in the world (Dourojeanni 1990;Hawksworth 2001). However, its fungal communities are thought to be severely understudied (Mueller and Schmit 2007). Members of Trochila are no exception to this. Six species of Trochila have been described from South America. These are T. chilensis Speg., T. jaffuelii Speg., and T. perseae Speg. from Chile; T. leopoldina Rehm from Brazil; and T. tetraspora, and T. winteri Speg. from Argentina (Spegazzini 1888(Spegazzini , 1910(Spegazzini , 1921Rehm 1909;Gamundí et al. 1978). Their type collections need to be re-examined to determine if these species are in fact members of Trochila. One of our new species, T. urediniophila, was collected in Port of Spain, Trinidad. Little data are available regarding the Funga (sensu Kuhar et al. 2018) of Trinidad and Tobago (Baker and Dale 1951;Dennis 1954a, b). The most recent work on the fungal diversity from this country was published online (Jodhan and Minter 2006) derived from reference collections and data from scientific literature. Based on the available literature, no records of Trochila are known in Trinidad. As a result, T. urediniophila represents the first published report of the genus from Trinidad, and by extension from the Caribbean (Minter et al. 2001).
Trochila species are likely more broadly distributed than generally thought, and certainly not limited to the Northern Hemisphere. This is often the case for many fungi that are based on limited regional collecting and thus may not represent the full extent of their distributional ranges due to, for example, the lack of studies in subtropical and tropical ecosystems (Groombridge 1992;Hawksworth and Mueller 2005;Mueller and Schmit 2007;Aime and Brearley 2012;Cheek et al. 2020).

The importance of biological collections
Our work emphasizes the importance of specimens preserved in biological collections -such as fungaria and herbaria -for studies of biodiversity and applied biological sciences, and for climate change research (Hawksworth and Lücking 2017;Andrew et al. 2019;Lang et al. 2019;Ristaino 2020;Wijayawardene et al. 2020). Because of the well-preserved specimens deposited at PUR, the genus Trochila is now known to be present in Trinidad and to form fungicolous associations. Another interesting example of the use of collections is Trochila colensoi. Known only from the type specimen for more than 100 years, additional specimens were only reported following the correction of the host substrate (as Cordyline rather than Phormium), which was based on re-examination of the type specimen preserved at K. Biological collections are not only important for morphological studies, but also as sources of genetic and genomic information Brock et al. 2009;Redchenko et al. 2012;Dentinger et al. 2016; this study). The single-oldest fungal specimen used for DNA extraction and sequencing was the type of Hygrophorus cossus (Sowerby) Fr. (Agaricales, Hygrophoraceae), collected in 1794 and deposited at K (Larsson and Jacobsson 2004). Our material of T. urediniophila gathered by Roland Thaxter in 1912 proves again that old samples can be used successfully for modern molecular phylogenetic analyses.