News from the Sea: A New Genus and Seven New Species in the Pleosporalean Families Roussoellaceae and Thyridariaceae

: Nineteen fungal strains associated with the seagrass Posidonia oceanica , with the green alga Flabellia petiolata , and the brown alga Padina pavonica were collected in the Mediterranean Sea. These strains were previously identified at the family level and hypothesised to be undescribed species. Strains were examined by deep multi-loci phylogenetic and morphological analyses. Maximum-likelihood and Bayesian phylogenies proved that Parathyridariella gen. nov. is a distinct genus in the family Thyriadriaceae. Analyses based on five genetic markers revealed seven new species: Neoroussoella lignicola sp. nov., Roussoella margidorensis sp. nov . , R. mediterranea sp. nov ., and R. padinae sp. nov. within the family Roussellaceae, and Parathyridaria flabelliae sp. nov., P. tyrrhenica sp. nov., and Parathyridariella dematiacea gen. nov. et sp. nov. within the family Thyridariaceae


Introduction
Marine fungi are a relevant and active component of the microbial communities that inhabit the oceans [1]. Fungi in the marine environment live as mutualists, parasites, pathogens and saprobes, and are pivotal to marine food webs because of the recycling of recalcitrant substrata [2]; besides, these widely dispersed organisms are a source of novel bioactive compounds [3].
Marine fungi have been recovered worldwide from a broad range of biotic and abiotic substrata, such as driftwood algae, sponges, corals, sediments, etc. [4,5]. Following the definition of Pang et al [6] that considered "a marine fungus" to be any fungus retrieved repeatedly from marine environment and that reproduces in the marine environment, Jones et al. [7] listed 1680 fungal species belonging to 693 genera, 223 families, 87 orders, 21 classes and six phyla. However, considering that the total number of marine fungi has been estimated to exceed 10,000 taxa [8], fungal diversity remains largely undescribed. With more than 900 species [9], the Ascomycota are the dominant fungal phylum in the sea; the most represented lineages include the order Pleosporales (class Dothideomycetes) with 36 families, 95 genera and 194 species described to date (www.marinefungi.org).
Many new species of Roussoellaceae and Thyridariaceae have recently been described on terrestrial plants including bamboo, palms and mangroves [14,17,20,21]. This paper provides a more precise phylogenetic placement of the 19 strains isolated from marine substrata together with morphological insights of those strains that represent new species within these two families.

Fungal Isolates
The fungal isolates analyzed in this paper were retrieved in the Mediterranean Sea from P. oceanica (2), collected in Riva Trigoso bay and Elba island, P. pavonica (12), and F. petiolata (3) from the coastal waters of Elba island [10][11][12]. A single isolate was previously retrieved in association with D. fragilis in the Atlantic Ocean [13] ( Table 1).
The strains investigated were originally isolated on Corn Meal Agar medium supplemented with sea salts (CMASS; 3.5% w/v sea salt mix, Sigma-Aldrich, Saint Louis, USA, in ddH2O) and are preserved at the Mycotheca Universitatis Taurinensis (MUT), Italy.

Morphological Analysis
All isolates were pre-grown on Malt Extract Agar-sea water (MEASW; 20 g malt extract, 20 g glucose, 2 g peptone, 20 g agar in 1 L of sea water) for one month at 24 °C prior to inoculation in triplicate onto new Petri dishes (9 cm Ø) containing i) MEASW, ii) Oatmeal Agar-sea water (OASW; 30 g oatmeal, 20 g agar in 1 L of sea water), or iii) Potato Dextrose Agar-sea water (PDASW; 4 g potato extract, 20 g dextrose, 20 g agar in 1 L of sea water). Petri dishes were incubated at 15 and/or 24 °C. The colony growth was monitored periodically for 28 days. Macroscopic and microscopic traits, were assessed for strains grown on MEASW at the end of the incubation period.
In an attempt to induce sporulation, sterile pieces of Quercus ruber cork and Pinus pinaster wood (species autochthonous to the Mediterranean area) were placed on 3 week old fungal colonies grown on MEASW ( [22], modified). Petri dishes were further incubated for 4 weeks at 24 °C. Subsequently, cork and wood pieces were transferred to 50 mL tubes containing 20 mL of sterile sea water. Samples were incubated at 24 °C for one month. In parallel, the strains were also plated on Syntetic Nutrient Agar-sea water (SNASW; 1 g KH2PO4, 1 g KNO3, 0.5 g MgSO4 • 7H2O, 0.5 g KCl, 0.2 g glucose, 0.2 g sucrose, 20 g agar in 1 L of sea water) supplemented with sterile pine needles. Petri dishes were incubated at 24 °C for one month.
Morphological structures were observed, and images captured using an optical microscope (Leica DM4500B, Leica microsystems GmbH, Germany) equipped with a camera (Leica DFC320, Leica microsystems GmbH, Germany). Macro-and microscopic features were compared with the available description of Roussoellaceae and Thyridariaceae [14,15,17,18,20].

DNA Extraction, PCR Amplification, and Data Assembling
Genomic DNA was extracted from about 100 mg of fresh mycelium grown on MEASW plates. Mycelium was disrupted by the mean of a MM400 tissue lyzer (Retsch GmbH, Haan, Germany) and DNA extracted using a NucleoSpin kit (Macherey Nagel GmbH, Duren, DE, USA) following the manufacturer's instructions. The quality and quantity of DNA were measured spectrophotometrically (Infinite 200 PRO NanoQuant; TECAN, Switzerland); DNA was stored at −20 °C.
Amplicons, together with a GelPilot 1 kb plus DNA Ladder, were visualized on a 1.5% agarose gel stained with 5 mL 100 mL −1 ethidium bromide; PCR products were purified and sequenced at the Macrogen Europe Laboratory (Madrid, Span). The resulting Applied Biosystem (ABI) chromatograms were inspected, trimmed and assembled to obtain consensus sequences using Sequencer 5.0 (GeneCodes Corporation, Ann Arbor, Michigan, USA http://www.genecodes.com). Newly generated sequences were deposited in GenBank (Table 1).

Sequence Alignment and Phylogenetic Analysis
A dataset consisting of nrSSU, nrITS, nrLSU, TEF1α and RPB2 was assembled on the basis of BLASTn results and of recent phylogenetic studies focused on Roussoellaceae and Thyridariaceae [18,20]. Reference sequences were retrieved from GenBank (Table 1).
Sequences were aligned using MUSCLE (default conditions for gap openings and gap extension penalties), implemented in MEGA v. 7.0 (Molecular Evolutionary Genetics Analysis), visually inspected and trimmed by TrimAl v. 1.2 (http://trimal.cgenomics.org) to delimit and discard ambiguously aligned regions. Since no incongruence was observed among single-loci phylogenetic trees, alignments were concatenated into a single data matrix with SequenceMatrix [27]. The best evolutionary model under the Akaike Information Criterion (AIC) was determined with jModelTest 2 [28].
Phylogenetic inference was estimated using Maximum Likehood (ML) and Bayesian Inference (BI) criteria. The ML analysis was generated using RAxML v. 8.1.2 [29] under GTR + I + G evolutionary model and 1000 bootstrap replicates. Support values from bootstrapping runs (MLB) were mapped on the globally best tree using the "-f a" option of RAxML and "-x 12345" as a random seed to invoke the novel rapid bootstrapping algorithm. BI was performed with MrBayes 3.2.2 [30] with the same substitution model (GTR + I + G). The alignment was run for 10 million generations with two independent runs each containing four Markov Chains Monte Carlo (MCMC) and sampling every 100 iterations. The first 25% of generated trees were discarded as "burn-in". A consensus tree was generated using the "sumt" function of MrBayes and Bayesian posterior probabilities (BPP) were calculated. Consensus trees were visualized in FigTree v. 1.4.2 (http://tree.bio.ed.ac.uk/software/figtree).
Two strains of Occutibambusa bambusae (Occultibambusaceae) were used to root the tree. Due to topological similarity of the two resulting trees, only ML analysis with MLB and BPP values was reported ( Figure 1). DNA diagnostic characters were visually identified by the presence of heterozygous bases. For each locus, aligned sequences of the individual clusters containing new species, were inspected. Nucleotide diversities of the novel species were annotated when occurred (Tables S1-S18).
Following phylogenetic tree inspection, isolates that clustered in the same group and that derived from the same substrate were subjected to PCR-fingerprinting by using the micro-and mini-satellite primers (GTG)5 and M13 [31,32] to exclude duplicates from further analysis. DNA fingerprints were visualized with 1.5% agarose gel stained with 5 mL 100 mL −1 ethidium bromide while a GelPilot 1 kb plusDNA Ladder was used as a reference. Images were acquired with a Gel Doc1000 System (Bio-Rad, Hercules, CA, USA) and fingerprints analyzed using Bionumerics v 7.6 (http://www.applied-maths.com).

Phylogenetic Inference
Preliminary analyses carried out individually with nrITS, nrSSU, nrLSU, TEF1α and RPB2 denoted no incongruence in the topology of the single-locus trees. The combined five-markers dataset-built on the basis of BLASTn results and of recent phylogenetic studies [18,20]-consisted of 81 taxa (including MUT isolates) that represented 16 genera and 56 species (Table 1)  Nucleotide divergence between each novel species and members of the same clusters were annotated for each locus, when occurred (Tables S1-S18).

Discussion
The description of these new taxa was particularly challenging because neither asexual nor sexual reproductive structures developed in axenic conditions. Therefore, we were unable to describe the range of anatomical variations and diagnostic features among these newly recognized phylogenetic lineages. Indeed, strictly vegetative growth without sporulation is a common feature of many marine fungal strains [10,11,33]. Possibly, these organisms rely on hyphal fragmentation for their dispersal, or alternatively, the differentiation of reproductive structures may be obligatorily dependent on the peculiar environmental conditions under which they live (e.g., wet-dry cycles, high salinity, low temperature, high pressure, etc.). During the study of these fungi, we tried to mimic the saline environment by using different culture media supplemented with natural sea water or sea salts. Although these culture methods were applied to induce sporulation, we observed that only media supplemented with sea water supported a measurable growth of vegetative mycelium (data not shown). The method introduced by Panebianco et al. [22] to induce sporulation by placing wood and cork specimens on the colony surface with their subsequent transfer into sea water, was only partially successful: out of seven species, three (P. dematiacea, P. flabelliae, R. mediterranea) developed chlamydospores in the mycelium above the wood surface, two (N. lignicola, R. padinae) gave rise to resting spores inside wood vessels. Most of the strains preferred to colonize P. pinaster wood rather than Q. ruber cork. These structures were interpreted as "chlamydospores" instead of "conidia" for the following reasons: i) They were characterized by a very thick cell wall, a typical feature of resting spores; ii) conidiogenous cells were never observed. Additional efforts to force the development of reproductive structures by using SNASW and pine needles, were also unsuccessful.
Both R. padinae and N. lignicola displayed a similar lignicolous behavior, growing and producing chlamydospores inside wooden vessels, although of different size and shape. The ability to form hyphae and to grow inside the wood vessels has been reported for a number of dark septate endophyte fungi in terrestrial environment [34] and, recently, for Posidoniomyces atricolor Vohník and Réblová, a marine endophyte that lives in association with the roots of P. oceanica [35]. By definition, endophytes live inside living plant tissues. To induce sporulation, sterilized specimens of dead wood were employed, therefore R. padinae and N. lignicola were inferred to be "lignicolous fungi" rather than "endophytes". The observation of this growth characteristic in two different genera, may find its reason in an evolutionary adaptation to marine life in association with lignocellulosic matrices. Therefore, we can hypothesize their ecological role as saprobes involved in degrading organic matter.
Notwithstanding the lack of exhaustive descriptions of morphological features, the strongly supported phylogenetic and molecular analysis, conducted with five different genetic markers (nrSSU, nrITS, nrLSU, TEF1α and RPB2) undoubtedly pointed out the differences among these species and their belonging to new taxa. This is also supported by the DNA diagnostic characters identified in the individual loci (Table S1-S18). In particular, the present study introduces four new species of Roussoellaceae and three new species of Thyridariaceae. Indeed, only MUT 2452 and MUT 4893 were ascribable to the previously described P. robiniae (Figure 1). In the case of MUT 4884, the holotype of P. dematiacea, a novel genus was proposed since it formed a defined cluster with MUT 5310 and MUT 4419, well separated by the genera Cycasicola, Liua and Thyridariella.
Most of the Roussoellaceae and Thyridariaceae described to date are associated with terrestrial plants, especially bamboo and palm species [15,16]. In fact, only two species, R. mangrovei and R. nitidula have been retrieved from the marine environment (www.marinefungi.org). However, considering the present study, we can infer that these families are well represented in the sea, thus improving our knowledge on the largely unexplored fungal marine biodiversity.

Supplementary Materials:
The following are available online at www.mdpi.com/1424-2818/12/4/144/s1, Table  S1: The eight variable sites detected in the nrITS region among P. dematiacea and its neighbor species, Table S2: The single variable site detected in the nrLSU region among P. dematiacea and its neighbor species, Table S3: The five variable sites detected in the nrSSU region among P. dematiacea and its neighbor species, Table S4: The six variable sites detected in the TEF1α partial gene among P. dematiacea and its neighbor species, Table S5: The six variable sites detected in the nrITS region among P. tyrrhenica, P. flabelliae and their neighbor species, Table: S6 The eight variable sites detected in the nrLSU region among P. tyrrhenica, P. flabelliae and their neighbor species ,  Table S7: The eight variable sites detected in the TEF1α partial gene among P. tyrrhenica, P. flabelliae and their neighbor species, Table S8: The 33 variable sites detected in the RPB2 partial gene among P. tyrrhenica, P. flabelliae and their neighbor species, Table S9: The two variable sites detected in nrITS region among R. mediterranea, R. padinae and the neighbor species, Table S10: The single variable site detected in nrLSU region among R. mediterranea, R. padinae, and the neighbor species, Table S11: The six sites detected in the TEF1α partial gene among R. mediterranea, R. padinae and the neighbor species, Table S12: The six sites detected in the RPB2 partial gene among R. mediterranea, R. padinae and the neighbor species, Table S13: The eight variable sites detected in the nrITS region among N. lignicola and its neighbor species, Table S14: The three variable sites detected in the nrLSU region among N. lignicola and its neighbor species, Table S15: The eight variable sites detected in the nrSSU region among N. lignicola and its neighbor species, Table S16: The ten sites detected in the TEF1α partial gene among N. lignicola and its neighbor species, Table S17: The three variable sites detected in the nrITS region among R. margidoriensis and its neighbor species, Table S18: The 29 variable sites detected in the TEF1α partial gene among R. margidoriensis and its neighbor species