Reassessment of Allantonectria, phylogenetic position of Thyronectroidea, and Thyronectria caraganae sp. nov

The genus Allantonectria is synonymised with Thyronectria, based on morphological and molecular phylogenetic considerations. Investigations of types and fresh collections revealed that Tubercularia concentrica is an earlier name for Sphaeria (syn. Allantonectria) miltina and is thus combined in Thyronectria. Allantonectria yuccae is recognised as a distinct species and transferred to Thyronectria, as well as the recently described A. zangii. Descriptions and illustrations are provided for T. concentrica and T. yuccae. A recent collection of the North American Thyronectroidea chrysogramma, the generic type of Thyronectroidea, was studied with respect to morphology of its sexual morph in fresh condition and of its asexual morph produced in pure culture. Molecular phylogenies based on six loci (ITS and LSU regions of nuc rDNA, act1, rpb1, rpb2, tef1 and tub2 genes) place T. chrysogramma within Thyronectria, confirming synonymy of Thyronectroidea with Thyronectria, but remarkably a relationship to European species with green to brown spores (former genus Mattirolia) receives no support, and its closest relatives remain unclear. Thyronectria caraganae is described as a new species from herbarium specimens of Caragana arborescens collected in the Ukraine. It is characterised by morphology of the sexual morph and by DNA sequence data, which place it within the T. austroamericana - T. rhodochlora clade with high support. This is also supported by its morphology, specifically ascomata partly embedded within a stroma, muriform ascospores not budding within the ascus, becoming yellowish to rosy at maturity.


Introduction
The genus Thyronectria was recently re-instated by , as it takes precedence over the younger genus Pleonectria. Based on detailed morphological as well as molecular phylogenetic analyses, they synonymised Mattirolia, Pleonectria and Thyronectroidea with Thyronectria and combined all respective epithets in Thyronectria. This was also implemented by Lombard et al. (2015) in their overview of genera of the Nectriaceae.
as a separate monotypic genus, characterized by minute, one-celled allantoid to rod-shaped ascospores and growth on monocotyledonous hosts as diagnostic characters. In previous works (Höhnel and Weese 1910;Lowen 1991;Hirooka et al. 2012), the generic type of Allantonectria, A. yuccae from Yucca spp., was regarded as a synonym of A. miltina, which was described from Agave americana. However, none of these authors mentioned Tubercularia concentrica, although this species was considered to be the asexual morph of A. miltina in the older literature (e.g., Saccardo 1878Saccardo , 1883 and which, in case of conspecificity, would take precedence due to priority. The description of A. zangii from Populus sp. Zhuang 2012, 2013), a nonmonocotyledonous host, and the lack of significant support for a sister group relationship of Thyronectria to Allantonectria in recent molecular phylogenies Checa et al. 2015) called for a re-evaluation of the generic status of Allantonectria. We performed detailed morphological studies of Allantonectria specimens from Agave and Yucca spp. for comparison with Thyronectria and to evaluate the species status of A. miltina and A. yuccae. In addition, we examined the types and fresh collections of Tubercularia concentrica and A. miltina to investigate whether they are conspecific. Based on these investigations, we conclude that T. concentrica and A. miltina are synonymous, A. yuccae represents a distinct species and Allantonectria should be merged with Thyronectria.

Morphological observations
Microscopic preparations were mounted in water, 3 % potassium hydroxide (KOH) or lactic acid (LA). Methods of microscopy included stereomicroscopy using a Nikon SMZ 1500 and Nomarski differential interference contrast (DIC) using the Zeiss Axio Imager.A1 compound microscope. Images and data were gathered using the Nikon DS-U2 or Zeiss Axiocam 506 color digital cameras and measured by using the NIS-Elements D v.3.0 or Zeiss ZEN Blue Edition softwares. For certain images of ascomata the stacking software Zerene Stacker v. 1.04 (Zerene Systems, Richland, WA, USA) was used. Measurements are reported as maxima and minima in parentheses and the range representing the mean plus and minus the standard deviation of a number of measurements given in parentheses.

Phylogenetic analyses
A single accession of each Thyronectria species was included in the phylogenetic analyses.
The accessions were selected according to availability of markers and, if possible, ex-type sequences were used (marked with an asterisk in Fig. 1). In addition, four Nectria species were included, and Septofusidium berolinense, S. herbarum and Tilachlidium brachiatum (Tilachlidiaceae) were selected as outgroup according to Lombard et al. (2015). Available sequences were downloaded from GenBank; details on the sequences used in the phylogenetic analyses are provided in Table 1.
All alignments were produced with the server version of MAFFT (www.ebi.ac.uk/Tools/ mafft or http://mafft.cbrc.jp/alignment/server/), checked and refined using BioEdit v.7.0.4.1 (Hall 1999). To reveal the phylogenetic position of the newly sequenced Thyronectria species, the newly generated sequences were aligned with the GenBank sequences. The resulting combined sequence matrix contained 6292 alignment positions from six gene regions (630 from act1, 541 from ITS and 807 from LSU, 706 from rpb1, 1192 from rpb2, 1308 from tef1 and 1108 from tub2).
Maximum likelihood (ML) analyses were performed with RAxML (Stamatakis 2006) as implemented in raxmlGUI 1.3 (Silvestro and Michalak 2012) using the ML+ rapid bootstrap setting and the GTRGAMMAI substitution model with 1000 bootstrap replicates. Substitution model parameters were calculated separately for the different gene regions included in the combined analyses.
Maximum parsimony (MP) analyses were performed with PAUP v.4.0a147 (Swofford 2002), using 1000 replicates of heuristic search with random addition of sequences and subsequent TBR branch swapping (MULTREES option in effect, steepest descent option not in effect). All molecular characters were unordered and given equal weight; analyses were performed with gaps treated as missing data; the COLLAPSE command was set to MAXBRLEN. Bootstrap analysis with 1000 replicates was performed in the same way, but using 5 rounds of random sequence addition and subsequent TBR branch swapping during each bootstrap replicate, with the COLLAPSE command set to MINBRLEN.

Cultures and sequences of sexual and asexual morphs of Thyronectria concentrica
Growth characteristics of cultures of T. concentrica obtained from the presumed asexual morph matched those obtained from ascospores, and the asexual morph produced in pure cultures from ascospores and conidia was identical. The ITS-LSU sequences of cultures isolated from the sexual and asexual morphs were identical to GenBank sequences of Allantonectria miltina from Agave americana.

Molecular phylogeny
Of the 6292 characters of the combined matrix, 1857 were parsimony informative (142 in act1, 119 in ITS, 108 in LSU, 304 in rpb1, 485 in rpb2, 393 in tef1 and 306 in tub2). The phylogram of the best ML tree (lnL = −49176.5326) obtained by RAxML is shown in Fig. 1. The MP analysis revealed a single tree of length 9703 (not shown), which is similar to the ML tree except for a different position of T. chrysogramma (see below), a sister group relationship of Nectria dematiosa to the other three Nectria species, and a placement of T. quercicola following next to T. berolinensis. Except for minor differences, tree topologies agree well with those of  and Checa et al. (2015).
The Thyronectria-Allantonectria clade is highly supported in both ML and MP analyses ( Fig. 1), Allantonectria receives maximum support, but Thyronectria receives only low support. The ML and MP analyses reveal different phylogenetic positions of T. chrysogramma. In the ML analyses, Thyronectria chrysogramma is placed basal to the highly supported clade containing species with yellowish to rosy spores, but this placement does not receive ML bootstrap support. In the MP analysis, T. chrysogramma is the most basal taxon of the Allantonectria-Thyronectria clade, however, again without bootstrap support. Thyronectria caraganae is placed in the highly supported clade containing species with yellowish to rosy spores; sister group relationship to T. austroamericana receives low (MP) or high (ML) support. Voglmayr, Akulov & Jaklitsch, sp. nov. Fig. 2. MycoBank: MB 817627. Etymology: referring to its host, Caragana arborescens.

Asexual morph in pure culture
On MEA colony up to 46 mm diam after 7 days at room temperature; surface cream, with tufts of aerial mycelium appressed to hyphal strands, turning pink from the centre due to conidial masses; reverse cream, conidiation starting on superficial substrate hyphae, later mainly on aerial hyphal strands. On CMD colony up to 59 mm diam after 7 days at room temperature; white, aerial hyphae sparse, conidiation mainly on superficial substrate hyphae. Conidiophores emerging as fasciculate side branches on strands of parallel hyaline aerial hyphae, short, 2-4.3 μm wide, simple, with few verticils of 2-3 branches. Conidiation on phialides or hyphal pegs. Phialides Rechinger (W 1914-10028); ibid., same date and collector, in Rehm, Ascomyc. 1962bAscomyc. (W 1914 Notes: This species has been commonly known as Allantonectria miltina (see Hirooka et al. 2012). In older literature (e.g., Saccardo 1878Saccardo , 1883, Tubercularia concentrica has been given as the asexual morph of A. miltina, but Lowen (1991) and Hirooka et al. (2012) did not mention the name in their synonymies. Hirooka et al. (2012) did not record an asexual morph from natural substrates. In one of our recent collections, we found a synnematous asexual morph developing below the host epidermis, and its connection with the sexual morph was subsequently proven by pure cultures and sequence data. Investigation of three isotype specimens of Tubercularia concentrica revealed that they fully match the asexual morph of our recent collection, and we here select the well-developed specimen K(M) 201845 as lectotype. As Tubercularia concentrica is older than Sphaeria miltina, the epithet concentrica has to be used, which is here combined in Thyronectria (see discussion below).
The asexual morph produced on SNA has been described as trichoderma-like by Hirooka et al. (2012). In our cultures on 2 % MEA, conidiation started on surface hyphae similar to other Thyronectria species, but was later primarily present on the hyphal strands of the aerial mycelium.
Thyronectria concentrica appears to be widespread in the Mediterranean where its main host, Agave americana, is commonly cultivated and naturalised. Like its host, it is assumed to originate from North America. Based on sequence data (Fig. 1), we do not accept synonymy of T. concentrica and T. yuccae, which mainly differ in their hosts (Agave vs. Yucca; see below) and by different ascospore sizes. In addition, T. concentrica commonly has more ascomata per stroma (up to 75 vs. up to 30 in T. yuccae). We include here a short description and illustration of T. concentrica in addition to the detailed description by Hirooka et al. (2012, under Allantonectria miltina) to facilitate comparison with T. yuccae and to describe the apical paraphyses and the asexual state on the natural substrate, which have not yet been documented for this species. As in all Thyronectria species we have investigated Checa et al. 2015), apical paraphyses are numerous and more or less persistent, but they are delicate, easily disarticulating and therefore only seen after careful ascoma preparation preferably in KOH (herbarium specimens) or water (fresh samples).
The collection data (year and collector) of the specimen given as holotype of Sphaeria miltina by Hirooka et al. (2012) do not agree with the protologue and cannot represent a type. Lowen (1991) designated a specimen from Herb. Montagne deposited in PC as lectotype, but the specimen labels of the two extant collections from Herb. Montagne do not bear the collection data she gave in the text, and the packages contain no information about lectotypification. In addition, the specimens were rearranged on a new sheet, with one fragment in PC 0723446 and four fragments in PC 0723441, whereas Lowen (1991) mentions the presence of two fragments in the specimen she examined, so it is unclear which specimen she selected. We here select specimen PC 0723442 from Herb. Durieu as lectotype of Sphaeria miltina, which is the only authentic specimen from PC on which Mustapha, the place mentioned in the protologue, is given on the original label. The two specimens from Herb. Montagne without place and date are considered to represent isotypes.
Notes: This species is the generic type of Allantonectria (Earle 1901), described from Yucca sp. collected in Colorado (USA). In combining the older Sphaeria miltina, which was described from Agave americana collected in Algeria, in Allantonectria, Höhnel and Weese (1910) synonymised A. yuccae with the older A. miltina, evidently based on similar ascospore sizes. This synonymy was also accepted by Hirooka et al. (2012), who cited specimens from both host genera, but only included sequences from a European accession from Agave americana in their phylogenies. Recently, sequences from an accession from Yucca elata collected in Arizona (USA) became available (Lombard et al. 2015), which differ substantially from the European collections of T. concentrica (see also Fig. 1). We therefore re-investigated specimens from both host genera, including an isotype of A. yuccae deposited in W, and found that A. yuccae has larger ascospores compared to T. concentrica Notes: Thyronectria zangii closely resembles T. concentrica and T. yuccae in its minute unicellular rod-shaped ascospores, but differs by thinner perithecial walls, narrower asci, and occurrence on twigs of Populus (Zeng and Zhuang 2012). Ascospore size of T. zangii (3.5-5.5(−6) × 0.9-1.2(−1.4) μm) is similar to that of T. concentrica revealed in the present study. Like for T. yuccae, no asexual morph has been recorded for T. zangii from natural substrates.

Discussion
Molecular phylogenetic analyses confirm the conclusions of  that Thyronectroidea chrysogramma belongs to Thyronectria. However, it is not contained within the clade comprising European species with green to brown ascospores (formerly classified in Mattirolia), but its phylogenetic position within Thyronectria remains uncertain. In the MP analyses, it is placed basal to the Allantonectria-Thyronectria lineage, a position which does not receive support, whereas in the ML analyses it is placed in a basal position of the clade containing the species with yellowish to rosy spores, again without support.
While the Thyronectria-Allantonectria clade and Allantonectria are highly supported, the genus Thyronectria receives only low support in both MP and ML analyses (Fig. 1). It should be noted that, with the addition of species and markers, the medium to high internal support of a sister group relationship of Allantonectria to Thyronectria as revealed by Hirooka et al. (2012), has decreased significantly in subsequent publications Checa et al. 2015), which casts doubt on the status of Allantonectria as a separate genus. With the addition of A. zangii from Populus (Zeng and Zhuang 2013), monocotyledonous substrates are no longer distinctive for Allantonectria, and the only remaining difference between the two genera lies in the small, unicellular allantoid to rodshaped ascospores of Allantonectria. Considering the highly diverse ascospore sizes, shapes, colours and septations prevailing in Thyronectria, we do not regard ascospore characters alone to be suitable for generic delimitation, and it appears justified to include Allantonectria in Thyronectria. The presence of persistent apical paraphyses and the asexual morph in pure culture, which were revealed in the present study, also correspond to Thyronectria. In the light of this morphological and molecular phylogenetic evidence, we argue for inclusion of Allantonectria in Thyronectria, and we formally combine the three accepted species of Allantonectria in Thyronectria.
Based on culturing and sequencing of recent collections as well as type studies, we were able to confirm that Tubercularia concentrica is the asexual morph of A. miltina, and we here provide an extended description of the asexual morph on natural substrates for the first time. Lowen (1991) and Rossman et al. (1999) mentioned Tubercularia sp. as asexual morph; however, without giving any description. Hirooka et al. (2012) did not find an asexual morph on natural substrates, probably due to its development within decaying leaves prior to ascomata, whereas they expected it to be pycnidial. At first sight, the synnematous conidiomata of T. concentrica do not fit the concept of Thyronectria, which was considered to be characterised by pycnidial conidiomata (Hirooka et al. 2012, under Pleonectria). However, not all Thyronectria species have pycnidial conidiomata, but some like T. roseovirens show effuse sporulation on natural substrates . Also, the "trichoderma"-like conidiation on aerial hyphal strands in pure culture are not unique for T. concentrica, but similar conidiation is also observed, e.g., in T. asturiensis . Therefore, although pycnidial asexual morphs are common within Thyronectria, they cannot be considered diagnostic for the whole lineage.
Thyronectria caraganae is revealed as the closest relative of T. austroamericana within the clade comprising species with yellowish to rosy ascospores. Apart from spore colour, species within this clade are characterised by ascomata surrounded by stromatic tissue (Hirooka et al. 2012;. In addition, all species have muriform ascospores, which do not bud within the ascus or only rarely do so. All these characters are also present in T. caraganae. Remarkably, T. caraganae is so far only known from the Ukraine, although its host, Caragana arborescens, is widely planted throughout Europe at road banks and between fields as a shrub preventing soil erosion by wind. However, it appears to be much rarer than its host, and may be confined to areas of continental climate. Despite pronounced search, the fungus could not be found in eastern Austria where its host is commonly cultivated. In Voglmayr et al. Page 13 eastern Austria, Caragana arborescens is frequently colonised by another nectriaceous fungus with similar ecology, Stromatonectria caraganae (Jaklitsch and Voglmayr 2011a), which may be the reason for the absence of Thyronectria caraganae in this area. However, the latter is an example for the lack of biodiversity inventories, even in well-studied Central Europe. Despite its conspicuous bright red stromata, S. caraganae had not been recorded for more than 100 years after its description, although it not only persisted in its type locality, a botanical garden (Jaklitsch and Voglmayr 2011a), but also turned out to be very common on Colutea arborescens and Caragana arborescens, and sometimes Laburnum anagyroides (all members of the Fabaceae), in eastern Austria after more detailed investigations. The detection of T. caraganae on a widely distributed host once again demonstrates the need of biodiversity studies on corticolous ascomycetes in Europe, of which a substantial amount of species is still poorly known and undescribed (e.g., Voglmayr and Jaklitsch 2008, 2011Jaklitsch andVoglmayr 2011a, b, 2014;Voglmayr et al. 2012Voglmayr et al. , 2016Jaklitsch et al. 2013Jaklitsch et al. , 2016Checa et al. 2015). Phylogram of the best maximum likelihood tree (lnL = −49176.5326) revealed by RAxML from an analysis of the combined matrix of Thyronectria, showing the phylogenetic position of T. chrysogramma and T. caraganae (in bold). ML and MP bootstrap support above 50% are given above or below the branches. Strain or herbarium numbers are given following the taxon names; holo-, neo-or epitype strains/specimens are marked by an asterisk (*). The tree was rooted with Septofusidium berolinense, S. herbarum and Tilachlidium brachiatum (Tilachlidiaceae) according to Lombard et al. (2015)  Isolates and accession numbers used in the phylogenetic analyses; those in bold were isolated/sequenced in the present study