Orbilia jesu-laurae (Ascomycota, Orbiliomycetes), a new species of neotropical nematode-trapping fungus from Puerto Rico, supported by morphology and molecular phylogenetics

Orbilia jesu-laurae is a new species of nematode-trapping fungus found on decorticated angiosperm wood in a tropical rainforest in Puerto Rico. The single specimen was studied from fresh apothecia and cultures. Morphology was studied and phylogenetic analysis (rDNA: ITS and LSU) was conducted using both sexual and asexual morphs. Nematodes were added to cultures to verify the formation and morphology of the trapping structures. Our results show that the species is in the Arthrobotrys clade, the phylogenetically closest relative being a possibly Mexican genotype with unknown morphology, erroneously referred to as Arthrobotrys musiformis in GenBank. Macroand micromorphological, ecological and biogeographic data are provided along with a discussion of closely related species.


Introduction
Fungi are recognized worldwide as saprobes, parasites and mutualists, but a small percentage (0.5 %) of Mucoromycota, Basidiomycota and Ascomycota are carnivorous (Yang & al. 2012;Spatafora & al. 2016). Fungi with this lifestyle often occur in microhabitats with low concentrations of available nitrogen. In such situations they capture and consume small animals and thus can colonize available substrates (Yang & al. 2012). Carnivorous Ascomycota produce complex trapping structures such as constricting rings, sessile adhesive knobs and adhesive networks (Yang & al. 2012). Examples of all these trapping structures are found in association with the asexual states of the genus Orbilia Fr. Based on their conidial and trapping states, all those taxa that produce adhesive networks were previously referred to Arthrobotrys Corda, originally described in 1839, three years after Fries published the genus Orbilia. Baral & al. (2018) treated these fungi as an unpublished series within the genus Orbilia. Because life cycles in some fungi are polymorphic and include both meiosporic and mitosporic structures (teleomorphs and anamorphs), occurring separately in space and time, the same species may have been studied and named independently. In the present case of Arthrobotrys, the first teleomorphic state was elucidated by Pfister (1997). The genus Arthrobotrys, originally characterized by 1-septate conidia produced in single or superposed 242 Quijada & al.: Orbilia jesu-laurae from Puerto Rico clusters on mononematous conidiophores, is a good example of the problem of dual names. If one is recognizing the genus Orbilia in a wide sense, then under the onefungus-one-name rule the name Orbilia takes precedence over Arthrobotrys. Members of Orbilia with predacious capabilities together with some non-predacious taxa form a monophyletic group, here referred to as the Arthrobotrys clade. A majority of species in the Arthrobotrys clade capture and consume copepods, mites, collembola, dipterans, but primarily nematodes (Baral & al. 2018). They have been tested for possible use in biocontrol of pests (Niu & Zhang 2011). The Arthrobotrys clade as treated by Baral & al. (2018) includes five unpublished infrageneric taxa, which are basically equivalent to the five previously described genera Arthrobotrys, Dactylella Grove, Dactylellina Morelet, Drechslerella Subram. and Gamsylella Scholler & al. Of these genera, four are predacious and monophyletic, whereas Dactylella, which comprises all non-predacious members of the group, is paraphyletic according to Baral & al. (2018). More than 100 species have been published in the genus Arthrobotrys. However, Scholler & al. (1999) recognized only 46 species in the genus when newly circumscribed to include those species with three-dimensional adhesive networks. Teleomorphs of the Arthrobotrys clade are mostly characterized by narrow, subulate and curved, falcate ascospores with a small, apical spore body (Baral & al. in press).
Fungal diversity has only rudimentary documentation for the Caribbean Islands, although the region is one of the world's biodiversity hotspots (Myers & al. 2000). Fungal species richness in Puerto Rico has been better explored in some localities compared to other islands. According to the locality index (http://www.cybertruffle .org.uk/), no orbiliaceous fungi have been reported in a preserved area near San Juan National Park (Julio Enrique Monagas National Park). Little is known about the diversity of Orbiliomycetes from Puerto Rico. Cantrell & Lodge (2008) Baral & al. (in press) in their monographic work on the class. The names listed by Cantrell & Lodge (2008) are to be considered as doubtful due to the lack of any morphological or specimen data. Some occurrences were missed in their list, for example O. pilosa (Dennis) Baral [Pfister 1997, as -11), ascospore isolates from which were produced Arthrobotrys-like anamorphs. However, these collections were not studied in detail (except for the teleomorph of PR-11). PR-98-20 was identified as Arthroborys musiformis Drechsler by Pfister, but cultures are now unavailable. These three species represent the only known Puerto Rican reports of species belonging to the nematode-trapping Arthrobotrys clade.
During a field trip of the ascomycete workshop arranged in connection with the 11 th International Mycological Congress held in Puerto Rico in 2018, several collections of Orbiliomycetes were made in the Julio Enrique Monagas National Park. One of them caught the attention of the first author due to its distinctive morphological features. This species was obviously not reported by Cantrell & Lodge (2008) or Pfister (1997). The morphology of the asci and ascospores clearly indicates a relationship to Orbilia auricolor (A. Bloxam) Sacc. and related species of the Arthrobotrys clade. The aim of this investigation is to describe this apparently new species and to provide morphological, biogeographical and phylogenetic evidence for its distinction.

Material and methods
The specimen was collected in Bayamón, Puerto Rico, on 15 July 2015 in the Julio Enrique Monagas National Park. The collection was air-dried and subsequently deposited in the Farlow Herbarium, Harvard University (FH). Months later, pieces of wood with the apothecia were placed on a black matboard and rehydrated with a spray bottle for macrophotography. This was done with a Canon EOS 60d digital SLR camera using a Canon EF-S 60 mm macro lens. An Olympus SZX9 stereomicroscope was employed to observe and characterize macromorphology and to perform hand-sectioning. For micromorphological observations a Motic B1 light microscope was used. Digital images were taken with a USB Moticam 2500 camera and biometry was done with the software Motic images Plus 2.0. For each informative morphological feature, 10 -30 photographs were taken prior to biometric analysis. The living or dead state of the cells was determined based on the findings of Baral (1992). Mounting media employed were tap water (H 2 O) for observing living cells, Congo red (CR) to raise wall contrast, particularly of dead cells, potassium hydroxide 5 % (KOH) for killing cells or rehydrating dead specimens, and Melzer's reagent (MLZ) for exploring amyloid or dextrinoid reactions. The symbols and abbreviations were adopted from Baral (1992): * = living state; † = dead state; * † = living and dead state (no difference noted); SCBs = KOH-soluble cytoplasmic bodies; SBs = spore bodies; VBs = refractive vacuolar bodies; LBs = lipid bodies. Colour coding refers to Anonymous (1976).
Characters of the asexual morphs are presented in this study as well as those of the sexual morph. Difco potato dextrose agar (PDA) and corn meal agar (CMA) were prepared according to manufacturer's instructions and used to culture the spores from our specimen and obtain the asexual morph. The substratum bearing one apothecium was placed on the inner surface of the lid of a Petri Willdenowia 50 -2020 dish on a small piece of dampened fi lter paper. The bottom of the Petri dish containing the medium was inverted, so the surface of the PDA or CMA was facing down. The progress of ascospore discharge was checked under the stereomicroscope after several hours. Once spores where deposited on the medium, the lid was replaced with a new sp. nov.

0.08
Hypothetical distribution of species morphologically related to Orbilia jesu-laurae Fig. 1. Bayesian majority-rule consensus tree of the Arthrobotrys clade sensu Baral & al. (in press) based on ITS1-5.8S-ITS2 + LSU. Thickened branches are those with support above 0.95. Each taxon included in the phylogeny contains the following notation: species name + isolate number + geographical origin + substrate/host + GenBank accession numbers of sequences. Rectangles with colours to right of taxon information represent biomes (legend at upper right). Hypothetical distribution of taxa morphologically most similar to Orbilia jesu-laurae in N and S hemisphere represented with colours (legend at bottom right). one and the apothecia from which the spores were ejected were used for DNA extraction. Once mycelia covered the media, subcultures were made. DNA was also extracted from cultures. Nematodes (Caenorhabditis elegans Maupas) were added to some of the subcultures to observe the formation of traps. A block of PDA or CMA covered with hyphae and one from a nematode-infested culture were placed facing down and opposing each other at some distance in a new Petri dish of CMA to observe trap formation. The culture was monitored over 24 hours to observe the formation of the traps and capture of nematodes and to observe the different stages of the process. A series of images was made over two days at 40× magnification using the same tools explained above.
DNA was extracted using Qiagen QIAamp DNA Micro Kit from mycelia from cultures and from apothe cia according to the manufacturer's instructions adding a period of 24 hours of incubation in the lysis buffer at 56°C. DNA extract dilutions of 1/10 were used to perform PCR amplification of the entire internal transcribed spacer region (ITS1-5.8S-ITS2) and partial nuclear large subunit (LSU). Primers, TAQ and information on quantities for PCR followed Karekehian & al. (2019). PCR products were sequenced by GENEWIZ (Cambridge, Massachusetts) using the same primer pairs. Quality of the sequences was checked in Geneious v.6.1.7. and they were checked for contamination using BLASTn (Alt schul & al. 1990). Phylogenetic analyses were done using two rDNA regions (ITS and LSU) for species of the Arthrobotrys clade. Sequences were taken from GenBank (https:// www.ncbi.nlm.nih.gov/genbank/). For details about species and GenBank accession numbers see Fig. 1. The constructed alignment (see Supplemental content online) included sequences of 71 isolates in the Arthrobotrys clade. The sequences were aligned using Mafft v7.017 (Katoh & al. 2002). Gblocks v. 0.91 was used to identify and eliminate ambiguously aligned regions (Castresana 2000). The analyses were performed using the optimal model of nucleotide substitution (Jmodeltest; Posada 2008). Bayesian inference (BI) was done using Geneious v. 6.1.7. For more details about phylogenetic methods see Quijada & al. (2014Quijada & al. ( , 2017. Adobe Illustrator CS5 was used to assemble figures. Names of the biomes for the biogeographic analyses follow Olson & al. (2001).
Distribution and ecology -So far known from a single collection in the Neotropics, in a lowland tropical rain forest of northeastern Puerto Rico on unidentified angiosperm wood on the ground.
Etymology -The specific epithet refers to Jesús Laura Rodríguez Armas in recognition of her many contributions to mycology, of her work in education and of our friendship.

Discussion
The Arthrobotrys clade was proposed in Baral & al. (2018) for a subgroup of Orbilia with predominantly nematode-trapping capabilities using diverse types of trapping organs. Species in this clade (Baral & al. 2018) are characterized by narrowly sickle-shaped (Fig. 3, E1 -3), rod-shaped or ellipsoid ascospores and apothecia being either tolerant of or sensitive to desiccation (Baral & al. in press;Quijada & Baral 2017). The connection between an Orbilia teleomorph and an Arthrobotrys state was reported for the first time by Pfister (1994). Only three years later, ITS phylogenetic analysis showed that trapping organs reflect evolutionary relationships in this group of fungi (Liou & Tzean 1997). Here, we use cultures obtained from ascospores to link an anamorph (Fig. 4) to the Puerto Rican apothecial collection and demonstrate the type of trapping organs formed (Fig.  5). Morphology of both teleomorph and anamorph was studied, and sequence comparisons supported the connection (Fig. 1). None of the species reported from Puerto Rico in the Arthrobotrys clade (Pfister 1997 (Baral & al. in press). Orbilia blumenaviensis has a pantropical distribution (Qiao & al. 2012) which includes neotropical (Caribbean belt, subtropical humid eastern South America), afrotropical (central Africa, Comoros), Indo-Malayan (subtropical humid SE Asia) and European (thermo-temperate NW Spain) areas (Baral & al. in press). The isolates of O. blumenaviensis included in our phylogeny (Fig. 1) reflect this distribution. The three collections clustered unsupported and are distant with respect to the placement of O. jesu-laurae. Morphologically, O. blumenaviensis differs in its lanceolate paraphyses and 1 -3-septate conidia (Qiao & al. 2012). Orbilia terrestris lacks sequence and anamorph data. It is only known from the type collection in mountainous central Asia. Its ascospores are wider than in O. jesu-laurae ( †1.4 -1.7 µm vs. †0.9 -1.1 µm) and have a distinct tail-like base (Baral & al in press;Raitviir & Faizova 1983). Arthrobotrys javanicus and A. musiformis are the morphologically most similar species to O. jesu-laurae. Arthrobotrys javanicus has an Indo-Malay distribution (Fig. 1), appearing in tropical or subtropical forests of Java, Yunnan and Taiwan (Baral & al. in press). Arthrobotrys musiformis is widespread with a Palearctic and Indo-Malayan distribution (Fig. 1); specimens have been reported in temperate (Europe, North America) and tropical (China) ecosystems (Baral & al. in press;Drechsler 1937;van Oorschot 1985). The new species is phylogenetically distant from these two species. The clade that includes the two collections from Mexico and the new species from Puerto Rico have a neotropical distribution (Fig. 1). In addition, we found biometric differences among them (measurements for A. javanicus and A. musiformis taken from Baral & al. in press). First, the ascospore length of O. jesu-laurae (*10 -11.8 µm) is larger than in the other two species (A. javanicus = *7 -9 µm; A. musiformis = *7.7 -10 µm). Second, the conidial length of O. jesu-laurae is shorter than in the other two species and the width is in between (O. jesu-laurae = *14 -26 × 7.5 -13 µm; A. javanicus = *25 -42.5 × 9.5 -15.5 µm; A. musiformis = *18.5 -32.5 × 6 -8.5 µm).
Our phylogenetic and morphological analysis showed that Orbilia jesu-laurae is not related to any of the taxa described so far in the Arthrobotrys clade (Baral & al. 2018). The new species has a neotropical distribution, at least so far as can be said based on a single collection. The closest phylogenetically related strains are two Mexican isolates identified in GenBank as A. musiformis (EVLL02, EVLL-2), here called "Arthrobotrys sp." because the ex-type culture of A. musiformis is in a different clade (Fig. 1). Our work contributes to the knowledge of this group of nematode-trapping fungi in a little-explored geographic area of the world.