Striatiguttulaceae, a new pleosporalean family to accommodate Longicorpus and Striatiguttula gen. nov. from palms

Abstract Palms represent the most morphological diverse monocotyledonous plants and support a vast array of fungi. Recent examinations of palmicolous fungi in Thailand led to the discovery of a group of morphologically similar and interesting taxa. A polyphasic approach based on morphology, multi-gene phylogenetic analyses and divergence time estimates supports the establishment of a novel pleosporalean family Striatiguttulaceae, which diversified approximately 39 (20–63) MYA (crown age) and 60 (35–91) MYA (stem age). Striatiguttulaceae is characterized by stromata or ascomata with a short to long neck, trabeculate pseudoparaphyses and fusiform to ellipsoidal, 1–3-septate ascospores, with longitudinal striations and paler end cells, surrounded by a mucilaginous sheath. Multi-gene phylogenetic analysis showed that taxa of Striatiguttulaceae form a well-supported and distinct monophyletic clade in Pleosporales, and related to Ligninsphaeriaceae and Pseudoastrosphaeriellaceae. However, these families can be morphologically demarcated by the slit-like ascomata and extremely large ascospores in Ligninsphaeriaceae and the rather narrow fusiform ascospores in Pseudoastrosphaeriellaceae. Eight strains of Striatiguttulaceae formed two monophyletic sub-clades, which can be recognized as Longicorpusgen. nov. and Striatiguttulagen. nov. Morphologically, the genus Longicorpus can be differentiated from Striatiguttula by its elongated immersed ascomata and fusiform ascospores with relatively larger middle cells and paler end cells. Two new species Striatiguttulanypae and S.phoenicis, and one new combination, Longicorpusstriataspora are introduced with morphological details, and phylogenetic relationships are discussed based on DNA sequence data.


Introduction
Fungi associated with palms have been intensively investigated by Hyde and his coworkers (Goh and Hyde 1996, Hyde and Alias 2000, Yanna et al. 2001a,b,c, Taylor and Hyde 2003, and provided a significant contribution to their diversity and taxonomy. There have been a number of interesting studies on palm fungi. For example, Fröhlich and Hyde (1999) reviewed the biodiversity of palm fungi in the tropics, and proposed the ratio of host specific fungi to palm species as 33 to 1 rather than the general ratio of 6 to 1 for all plants proposed by Hawksworth (1991). Taylor et al. (2000) investigated biogeographical distribution of microfungi from temperate and tropical palms, and found different fungal assemblages from these two regions, and also revealed that the difference was more related to climatic influences than hosts sampled. Subsequently, Yanna et al. (2001bYanna et al. ( , 2002 studied fungal communities and succession of palms, and pointed out that fungal species compositions were distinct on different hosts and at different sites, and even differed from different palm tissues. In addition, some studies were dedicated to endophytic palmicolous fungi (Rodrigues and Samuels 1990, Taylor et al. 1999, Hyde and Soytong 2008, Pinruan et al. 2010a, Mahmoud et al. 2017) and pathogens (Fröhlich et al. 1997, Hyde and Cannon 1999, Elliott et al. 2010, Mohammadi 2014. Other studies have focused on fungi on peat swamp palms (Pinruan et al. 2002, 2010b, Pinnoi et al. 2003) and from mangrove palms (Suetrong et al. 2009, Loilong et al. 2012. All these examples indicate that species are diverse and palms harbour numerous undescribed microfungi.
Nypa fruticans is an ancient palm that grows in brackish water, while Phoenix paludosa is found in the upper parts of mangroves and tolerates salt water, with both occurring in Thailand mangrove sites. In an ongoing study on the taxonomy of fungi occurring on palms, we collected fungi colonizing these two palm hosts from different mangrove sites in Thailand. Interestingly, a group of ascomycetes recovered appears to be new to science based on morphology and multi-gene phylogenetic evidence. The aim of this study was to characterize the novel taxa and investigate their phylogenetic relationships in the order Pleosporales, as well as apply the divergence times as additional evidence, especially in higher taxa ranking, for the establishment of new family Striatiguttulaceae.

Specimen collection, examination and single spore isolation
Decayed rachides or petioles of Nypa fruticans and Phoenix paludosa were collected from Chanthaburi, Krabi and Ranong provinces in Thailand. The collected specimens were washed under running water and examined via laboratory procedures as outlined by Jones and Hyde (1988). Morphological characters were observed using a Carl Zeiss stereo microscope fitted with an AxioCam ERC 5S camera and photographed by a Nikon ECLIPSE 80i compound microscope fitted with a Canon EOS 600D digital camera. Free hand sections of fruiting bodies were made into slides within water mounts and observed under Motic SMZ 168 stereo microscope. Measurements were taken by Tarosoft Image Frame Work program v. 0.9.7 and images used for figures processed with Adobe Photoshop CS6 Extended v. 13.0 software. Isolations were obtained from single spores as described in Choi et al. (1999). New taxa were established based on recommendations outlined by Jeewon and Hyde (2016). The strains isolated in this study were deposited in Mae Fah Luang University Culture Collection (MFLUCC) and Guizhou Culture Collection (GZCC). Herbarium specimens were deposited at the herbaria of Mae Fah Luang University (MFLU), Chiang Rai, Thailand and Kunming Institute of Botany Academia Sinica (HKAS), Kunming, China. MycoBank numbers (Crous et al. 2004) and Facesoffungi numbers (Jayasiri et al. 2015) are provided.

DNA extraction, PCR amplification and sequencing
Fungal genomic DNA was extracted from fresh mycelia scraped from the margin of a colony on PDA that was incubated at 25 °C-28 °C for 30 days, followed by the Ezup Column Fungi Genomic DNA Purification Kit (Sangon Biotech (Shanghai) Co., Ltd, China) following the manufacturer's instructions. Two partial rDNA genes and two protein coding genes were used in this study: the large subunit of the nuclear ribosomal RNA genes (LSU), the small subunit of the nuclear ribosomal RNA (SSU), the translation elongation factor 1-alpha (TEF1α) and the second largest subunit of RNA polymerase II (RPB2). The primers used were LR0R and LR5 for LSU (Vilgalys and Hester 1990), NS1/NS4 for SSU (White et al. 1990), EF1-983F/EF1-2218R for TEF1α (Rehner and Buckley 2005) and fRPB2-5F/fRPB2-7cR for RPB2 (Liu et al. 1999). The amplification reactions were performed in 25μL of PCR mixtures containing 9.5μL ddH 2 O, 12.5μL 2× PCR MasterMix (TIANGEN Co., China), 1μL DNA temple and 1μL of each primer. The PCR thermal cycle program for LSU, SSU and TEF1α amplification were as follows: initial denaturing step of 94 °C for 3 min, followed by 40 cycles of denaturation at 94 °C for 45 seconds, annealing at 56 °C for 50 seconds, elongation at 72 °C for 1 min, and final extension at 72 °C for 10 min. The PCR thermal cycle program for the partial RNA polymerase second largest subunit (RPB2) was followed as initially 95 °C for 5 min, followed by 40 cycles of denaturation at 95 °C for 1 min, annealing at 52 °C for 2 min, elongation at 72 °C for 90 seconds, and final extension at 72 °C for 10 min. Purification and sequencing of PCR products were carried out with primers mentioned above at Sangon Biotech (Shanghai) Co., Ltd, China.

Sequence alignment and phylogeny analyses
A concatenated data set of LSU, SSU, TEF1α and RPB2 sequences was used for phylogenetic analyses with the inclusion of reference taxa from GenBank (Table 1). Sequences were aligned using MAFFT v.7 (http://mafft.cbrc.jp/alignment/server/) (Katoh and Standley 2013) and then checked visually and manually optimized using BioEdit v.7.0.9 (Hall 1999). Representative families in Pleosporales and several major groups in Dothideomycetes were included in our analyses, and taxa in Arthoniomycetes were selected as outgroup. A maximum likelihood (ML) analysis was performed at the CIPRES web portal (Miller et al. 2010) using RAxML v.7.2.8 as part of the "RAxML-HPC Blackbox (8.2.10)" tool (Stamatakis 2006, Stamatakis et al. 2008. A general time-reversible model (GTR) was applied with a discrete GAMMA distribution and four rate classes. Fifty thorough ML tree searches were carried out in RAxML v.7.2.7 under the same model. One thousand non-parametric bootstrap iterations were run with the GTR model and a discrete gamma distribution. The resulting replicates were plotted on to the best scoring tree obtained previously.   Maximum parsimony (MP) analyses were performed using the heuristic search option with 1000 random taxa additions and tree bisection and reconnection (TBR) as the branch-swapping algorithm. All characters were unordered and of equally weight; gaps were treated as missing data. Maxtrees setting was 1000, and zero-length branches were collapsed, and all parsimonious trees were saved. Clade stability was assessed using a bootstrap (BT) analysis with 1000 replicates, each with 10 replicates of random stepwise addition of taxa (Hillis and Bull 1993 The Bayesian analysis was performed using PAUP v.4.0b10 (Swofford 2002) and MrBayes v.3.1.2 (Ronquist and Huelsenbeck 2003). The best model for different genes partition in the concatenated data set was determined by MrModeltest 2.3 (Nylander 2004). Posterior probabilities (Rannala and Yang 1996) were determined by Markov Chain Monte Carlo sampling (MCMC) (Larget and Simon 1999) in MrBayes v.3.1.2. Four simultaneous Markov chains were run for 10 million generations and trees were sampled every 1000 th generation, thus 10,000 trees were obtained. The suitable burn-in phases were determined by inspecting likelihoods and parameters in Tracer version 1.6 (Rambaut et al. 2013). Based on the tracer analysis, the first 1,000 trees representing 10% were discarded as the burn-in phase in the analysis. The remaining trees were used to calculate posterior probabilities in the majority rule consensus tree (critical value for the topological convergence diagnostic set to 0.01). Phylogenetic tree was visualized by FigTree v.1.4.2 (Rambaut 2014), and the alignment is deposited in TreeBASE under the accession number TB2: S23392 (http://purl.org/phylo/treebase/phylows/study/TB2:S23392).

Divergence times estimates
One secondary data and two fungal fossil calibrations were used in this study. The split between Arthoniomycetes and Dothideomycetes was selected as a secondary calibration point referring to previous evolutionary molecular studies (Gueidan et al. 2011, Prieto and Wedin 2013, Pérez-Ortega et al. 2016, Phukhamsakda et al. 2016, with a mean of 300 MYA and standard deviation (SD) of 50 MYA in a normal posterior distribution. Simultaneously, one ascomycete fossil Metacapnodiaceae , was used as the common ancestor of Capnodiales, with constraint of mean 100 MYA and SD 150 MYA in a normal posterior distribution (Pérez-Ortega et al. 2016, Phukhamsakda et al. 2016. Whereas the fossil Margaretbarromyces dictyosporus (Mindell et al. 2007, Berbee and Taylor 2010 was used to calibrate the Aigialus (Aigialaceae) crown, with an offset of 35 MYA in a gamma distribution (Phukhamsakda et al. 2016). Divergence time estimates were carried out by BEAST v 1.8.0 (Drummond et al. 2012). Aligned sequence data were partitioned separately for LSU, SSU, TEF1α and RPB2 data set, and loaded to prepare an XML file constructed with BEAUti v1.8.0. The substitution models, clock models and the tree prior parameters were set to be linked. The nucleotide substitution model was set to GTR (Generalized Time Reversible) + Gamma + Invariant sites, with estimated base frequencies, four gamma categories and without partitions. An uncorrelated relaxed clock model (Drummond et al. 2007) with a lognormal distribution of rates for each gene estimate was used for the analyses. We used a Yule tree prior, which assumes a constant speciation rate per lineage, and a randomly generated starting tree. The analysis was run for 100 million generations and parameters were sampled every 10,000 generations. Tracer v.1.6 (Rambaut et al. 2013) was used to analyze the trace files, and the acceptable effective sample sizes (ESS) values were greater than 200. Maximum clade creditability (MCC) trees were annotated using TreeAnnotator v1.8.0 and then visualized in FigTree v.1.4.2 (Rambaut 2014).

Phylogenetic results
The multi-gene dataset comprised 113 taxa and 4113 characters after alignment (LSU: 919 bp; SSU: 1245 bp; TEF1α: 929 bp; RPB2: 1020 bp) including gaps. RAxML, MP and Bayesian analyses were conducted and resulted in generally congruent topologies, and the familial assignments are similar to previous work (Hashimoto et al. 2017. Maximum parsimony analyses indicated that 2,302 characters were constant, 355 variable characters parsimony uninformative and 1,456 characters are parsimonyinformative. A heuristic search yield four equally most parsimonious trees (TL = 10905, CI = 0.278, RI = 0.561, RC = 0.156, HI = 0.722). The combined dataset provided higher confidence values for the familial level than those of the individual gene trees (data not shown), and RAxML analysis based on LSU, SSU, TEF1α and RPB2 yielded a best sorting tree ( Figure 1) with a final optimization likelihood value of -52455.532059.
The eight newly generated strains clustered together and positioned outside the two suborders (Massarineae and Pleosporineae) of Pleosporales, and formed a well-supported monophyletic clade and represented as a new linage of Pleosporales. The phylogeny also revealed that this clade is close to Ligninsphaeriaceae, Pseudoastrosphaeriellaceae, Testudinaceae and Tetraplosphaeriaceae, and can be recognized as a novel family (Striatiguttulaceae). Furthermore, the eight strains formed two well-supported monophyletic sub-clades, which can be identified as two new genera (Longicorpus and Striatiguttula) with three species (Longicorpus striataspora, Striatiguttula nypae and S. phoenicis).

Divergence time estimates
The maximum clade credibility (MCC) tree with divergence estimates (Figure 2) obtained through BEAST was topologically identical to those recovered by Bayesian and ML procedures with regards to the placement Pleosporales and several major lineages within Dothideomycetes. The mean dates of Pleosporales crown corroborate reported estimates (Phukhamsakda et al. 2016 are provided in Notes. The family Striatiguttulaceae is introduced to accommodate two new genera Longicorpus and Striatiguttula, characterized by the immersed, and erumpent to superficial stromata, with a papilla or a short to long neck, trabeculate pseudoparaphyses, bitunicate asci, and hyaline to brown, fusiform to ellipsoidal, striate, guttulate, 1-3-septate ascospores, with paler end cells and surrounded by a mucilaginous sheath. Members of Striatiguttulaceae are morphologically similar to the genera Leptosphaeria and Trematosphaeria, but they are phylogenetically distinct and also differ in ascospores characteristics and the latter two have coriaceous, heavily pigmented thick-walled peridium. Multi-gene phylogenetic analyses revealed a close relationship of Striatiguttulaceae to Ligninsphaeriaceae and Pseudoastrosphaeriellaceae. However, Striatiguttulaceae differs from Pseudoastrosphaeriellaceae as the latter has 1-3-septate or 2-5-septate ascospores, which are narrowly fusiform with acute ends and all cells are concolorous. The slit-like ascomata and broad fusiform, 1-septate, rather large ascospores (79-121 × 14-23 μm) in Ligninsphaeriaceae (Zhang et al. 2016) are distinct from those found in Striatiguttulaceae. Additionally, a divergence time estimate analysis indicated that the crown age 39 (20-63) MYA and stem age 60 (35-91) MYA of Striatiguttulaceae, match with the recommendations of using divergence times to recognize families in Liu et al. (2017). Attempts were made to culture the asexual morph in order to build comprehensive familial concept for Striatiguttulaceae, but it was not successful. Further morphological investigations together with more molecular data are needed. Etymology. Name refers to the striate and guttulate ascospores.
Notes. Striatiguttula nypae varies in ascomatal appearance, mostly immersed beneath the plant surface, sometimes visible as a papilla or dome-shaped area on the plant surface, and becomes erumpent to superficial, with a papilla or a short to long neck. The typical morphological characters of S. nypae are the appearance of stromata, with interwoven pale brown hyphae or setae at the apex of the neck, and the hyaline

CC:
(Characteristics C) central cells larger, brown, end cells smaller and paler, ascospore wall covered in distinct longitudinal striations, and surrounded by a sheath.
Notes. The fusiform to ellipsoidal, 1-3-septate ascospores of Striatiguttula phoenicis is similar to those of Trematosphaeria mangrovis, associated with submerged roots of mangrove trees. However, Striatiguttula phoenicis differs from T. mangrovis (Kohlmeyer 1968) as the latter has larger ascospores and lacks striations (Table 3). Striatiguttula phoenicis is morphologically different from S. nypae as it has ellipsoidal ascospores which are broader in width. Currently, the erumpent to superficial stromata have not been found in S. phoenicis. The phylogenetic analysis also confirms that they are distinct species. There are 26 noticeable nucleotide differences across the 474 nucleotides (Suppl. material 1) of ribosomal ITS sequence data (strains: MFLUCC 18-0266 vs. MFLUCC 18-0265, MFLUCC 17-2517 andMFLUCC 17-2518). Etymology. Name refers to the elongated ascomata and ascospores.
Culture characteristics. Colonies on PDA attaining 12 mm diameter within 21 days at 25 °C under natural light, velvety, centrally raised, irregular to circular in shape, greenish grey and mixed with milky white mycelium at the edge of a colony, the reverse dull green or grey olivaceous.
Notes. Longicorpus striataspora was found on two mangrove palm species, Nypa fruticans and Phoenix paludosa. The typical characteristics of L. striataspora are the deeply immersed, carbonaceous ascomata with a long neck, and the striate, guttulate, fusiform, 1-3-septate ascospores, with larger middle cells and relatively smaller and paler end cells, surrounded by a mucilaginous sheath. However, such characteristics are similar to Trematosphaeria spp. (Table 3), and match with Trematosphaeria striataspora (Hyde 1988), the holotype collected from intertidal wood of Nypa fruticans in North Sumatra. Trematosphaeria striataspora was later accommodated in Astrosphaeriella Syd. & P. Syd. (Hyde 1992a) with proposals for recollection and further phylogenetic studies (Liu et al. 2011b. We have compared the fresh collections of Longicorpus striataspora with the type material of Trematosphaeria striataspora, and concluded that the two are identical in morphology. On the other hand, the genus Trematosphaeria Fuckel has been assigned to the family Trematosphaeriaceae K.D. Hyde, Y. Zhang ter, Suetrong & E.B.G. Jones, based on molecular data of its type species T. pertusa Fuckel. Therefore, we follow Ariyawansa et al. (2014) and designate an epitype for Longicorpus striataspora in this study.

Discussion
A novel pleosporalean family, Striatiguttulaceae is introduced herein, which has been compared to several morphologically similar genera and species recovered from mangroves. This study introduces three novel species including an epitypification. The use of divergence times as an additional evidence for ranking taxa (especially in higher taxa ranking) has become possible and several studies have been carried out across different fungal groups (Phukhamsakda et al. 2016, Samarakoon et al. 2016, Divakar et al. 2017. To better understand the placement of Striatiguttulaceae, divergence time was also estimated and this study supports taxonomic schemes proposed earlier. The recent study of ranking a family with divergence time estimates is Liu et al. (2018), who introduced Lentimurisporaceae, a new pleosporalean family. We have recovered essentially similar phylogenetic topology, and in an extensive dataset that included berkleasmium-like taxa (referred to Liu et al. 2018), phylogenies generated were also topologically identical to those recovered herein (Figure 1). The monotypic family Ligninsphaeriaceae is sister to Striatiguttulaceae, and berkleasmium-like taxa are close to Aquasubmersaceae, Hermatomycetaceae and Salsuginaceae respectively. In this study, the ages of most families in Pleosporales, especially those positioned outside the two suborders were estimated in our divergence time analysis, and the results are comparable to other studies. However, Ligninsphaeriaceae, Pseudoastrosphaeriellaceae and Testudinaceae have relatively younger stem ages than that in Liu et al. (2017), presumably due to different taxa sampling in our phylogeny.
The nature of the pseudoparaphyses (sensu Liew et al. 2000) is worth considering here and may provide evidence for separate lineages. The family Striatiguttulaceae, currently with three species, have trabeculate pseudoparaphyses, but also appearing septate. Phylogenetically closely related families of Ligninsphaeriaceae and Pseudoastrosphaeriellaceae are characterized by cellular pseudoparaphyses and trabeculate pseudoparaphyses respectively.
Considering the ecology of these Striatiguttulaceae species in relation to the mangrove ecosystem, salinity may be an important contributor to their presence. Loilong et al. (2012) have compared fungal community from Nypa fruticans at different salinities, and found freshwater species in lower salinity and marine species at higher salinity. Although no salinity was measured during our collections, Longicorpus striataspora, Striatiguttula nypae and S. phoenicis can be considered as manglicolous, because they are found from decayed rachides/petioles of palms, which are perennials submerged in soft mangrove mud and salty water, and well adapted to the varying salinity in mangroves by tidal water. On the other hand, their ascospores have mucilaginous sheaths and lack elaborate appendages, which are also typical characteristics of most mangrove fungi (Jones 2000).