Morphological and phylogenetic characterisation of novel Cytospora species associated with mangroves

Abstract Mangroves are relatively unexplored habitats and have been shown to harbour a number of novel species of fungi. In this study, samples of microfungi were collected from symptomatic branches, stem and leaves of the mangrove species Xylocarpusgranatum, X.moluccensis and Lumnitzeraracemosa and examined morphologically. The phylogeny recovered supports our morphological data to introduce three new species, Cytosporalumnitzericola, C.thailandica and C.xylocarpi. In addition, a combined multi-gene DNA sequence dataset (ITS, LSU, ACT and RPB2) was analysed to investigate phylogenetic relationships of isolates and help in a more reliable species identification.


Introduction
Mangroves are forests established in tropical and subtropical backwaters, estuaries, deltas and lagoons. These forests play a major role in the ecology of coastal tropical/ subtropical waters, as they serve as hatchery and nursery habitats for marine organisms and protect coastlines from catastrophic events such as storms and tidal surges (Hyde and Jones 1988, Fisher and Spalding 1993, Hyde and Lee 1995, Hyde et al. 1998). The greatest diversity of mangrove species occurs in the mangroves of Indonesia, Malaysia and Thailand Jones 2009, Alias et al. 2010).
Reports of fungi associated with mangroves are relatively few and data on diseases of mangroves are uncommon (Cribb and Cribb 1955, Kohlmeyer and Kohlmeyer 1979, Hyde and Jones 1988. So far, a number of fungi collected from mangroves are either saprobes (e.g Swe et al. 2008a, b, Devadatha et al. 2018, Li et al. 2018 or endophytes (e.g Liu et al. 2012. One early species documented from mangroves is that of Stevens (1920) who reported a species of Anthostomella that was found from a leaf spot in red mangroves (Rhizophora mangle) in Puerto Rico. Later, McMillan (1964) reported Cercospora which caused leaf spot on red mangroves in Florida and Kohlmeyer (1969) documented an undescribed Cytospora species on R. mangle in Hawaii. Cytospora rhizophorae has also been reported as a marine fungus from Rhizophora mangle in southwest Puerto Rico (Wier et al. 2000). Later, Shivas et al. (2009) reported a serious disease, caused by Pseudocercospora avicenniae, on leaves of Avicennia marina in Cape Tribulation, Queensland.
Cytospora was introduced by Ehrenberg (1818) and belongs to the family Cytosporaceae in Diaporthales (Wijayawardene et al. 2018). Cytospora species are phytopathogens or saprobes (Wehmeyer 1975, Barr 1978, Eriksson 2001, Castlebury et al. 2002, Wijayawardene et al. 2018. Cytospora has a worldwide distribution and is an important pathogenic genus, causing canker and dieback disease on branches of a wide range of plants (Adams et al. 2005, 2006. Currently, there are 614 epithets for Cytospora (Index Fungorum 2018, 14 June 2018 with an estimated 110 species in Kirk et al. (2008). Recently, fourteen new species were introduced to this genus by Norphanphoun et al. (2017). In this study, we report on three novel species of Cytospora associated with mangroves in Thailand. Detailed descriptions and illustrations of all the species identified are provided in this paper.

Sample collection and examination of specimens
Samples collected were dead branches of Xylocarpus granatum K.D. Koenig, X. moluccensis (Lam.) M. Roem. and leaf spots of Lumnitzera racemosa Willd. from Phetchaburi and Ranong provinces, Thailand in 2016. Specimens were returned to the laboratory in paper bags, examined and described following Norphanphoun et al. (2017). Morphological characters of ascomata and conidiomata were examined using a Motic SMZ 168 dissecting microscope. Hand sections were mounted in water and examined for morphological details. Micro-morphology was studied using a Nikon Ni compound microscope and photographed with a Canon EOS 600D digital camera fitted to the microscope. Photo-plates were made using Adobe Photoshop CS6 Extended version 13.0 × 64 (Adobe Systems, USA), while Tarosoft (R) Image Frame Work programme v. 0.9.7 was used for measurements.
Cultures were obtained by single spore isolation method outlined in Chomnunti et al. (2014). Single germinating spores were observed and photographed using a Nikon Ni compound microscope fitted with Canon EOS 600D digital camera. Geminated spores were transferred aseptically to 2% malt extract agar (MEA, malt extract agar powder 32 g in 1000 ml water) and incubated at room temperature (18−25 °C). A tissue isolation method was used for isolation of taxa from leaf spots of Lumnitzera racemosa. Leaves with leaf spots were cut into small pieces (0.5 × 0.5 cm 2 ) using a sterilised blade and surface was sterilised using 70% ethanol for 1 minute, followed by three rinses with sterile distilled water, 1 minute in 3% sodium hypochlorite (NaOCl) and rinsed with sterile water for 1-2 minutes and dried by blotting on sterile filter paper. Four to five segments including the edge of the leaf spot were placed on water agar (WA) plates, supplemented with 100 mg/ml streptomycin. The dishes were incubated at 27 °C ± 2 °C for 7-10 days. Fungal colonies were transferred using single hyphal tips on to potato dextrose agar (PDA) plates throughout a 2-week period. Pure cultures were maintained for further studies on PDA (Bharathidasan and Panneerselvam 2011). The specimens/dried cultures and living cultures are deposited in the Herbarium Mae Fah Luang University (MFLU) and culture collection Mae Fah Luang University (MFLUCC), Chiang Rai, Thailand and duplicated in the International Collection of Micro-organisms from Plants (ICMP). Facesoffungi numbers were registered as in Jayasiri et al. (2015). New taxa are established based on recommendations as outlined by Jeewon and Hyde (2016).
The amplification reactions were carried out with the following protocol: 50 μl reaction volume containing 2 μl of DNA template, 2 μl of each forward and reverse primers, 25 μl of 2 × Bench Top TM Taq Master Mix (mixture of Taq DNA Polymerase (recombinant): 0.05 units/μl, MgCl 2 : 4 mM and dNTPs (dATP, dCTP, dGTP, dTTP): 0.4 mM) and 19 μl of double-distilled water (ddH 2 O) (sterilised water) using the thermal cycle programme in Norphanphoun et al. (2017). Purification and sequencing of PCR products with the same primers mentioned above were carried out at Life Biotechnology Co., Shanghai, China.

Phylogenetic analysis
The sequences were assembled by GENEIOUS Pro v. 11.0.5 (Biomatters) and BLAST searches were made to retrieve the closest matches in GenBank and multiple alignment also included recently published sequences . Combined analyses of ITS1, 5.8S, ITS2, LSU, RPB2 and ACT sequence data of 86 taxa were performed under different optimality criteria (MP, ML, BI). Diaporthe eres (AFTOL-ID 935) was used as the outgroup taxon. In order to obtain a better picture of the phylogenetic relationships amongst our strains and closely related strains, a separate ITS1+ITS2 phylogeny was inferred, because only ITS sequences were available for many strains in that group and because less ambiguously aligned (and excluded) positions are expected in a dataset with narrower taxonomic coverage. Nineteen strains were selected for this analysis based on preliminary analyses and results from the multigene phylogeny. All sequences were aligned separately using the MAFFT v.7.110 online programme (http://mafft.cbrc. jp/alignment/server/; Katoh and Standley 2013) and Gblocks v. 0.91b was used to exclude ambiguously aligned positions in the ITS and ACT alignments (Castresana 2000, Talavera andCastresana 2007). A partition homogeneity test (PHT) was performed with PAUP 4.0b10* (Swofford 2002) to determine whether the individual datasets were congruent and could be combined. The combined sequence alignments were obtained from MEGA7 version 7.0.14 (Kumar et al. 2015), missing data were coded as question marks (?) and further manual adjustments were made wherever necessary in BioEdit 7.2.3 (Hall 1999). The combined sequence alignment was converted to NEXUS file for maximum parsimony analysis using ClustalX v. 2 (Larkin et al. 2007). The NEXUS file was prepared for MrModeltest v. 2.2 (Nylander 2004) in PAUP v.4.0b10 (Swofford 2002).
Maximum Parsimony (MP) analysis was performed using PAUP (Phylogenetic Analysis Using Parsimony) v. 4.0b10* (Swofford 2002) with 1000 bootstrap replicates using a heuristic search with random stepwise addition and tree-bisection reconnection (TBR), as detailed by Jeewon et al. (2002) and Cai et al. (2005). Maxtrees was set to 1000, branches of zero length were collapsed. The following descriptive tree statistics were calculated: parsimony tree length [  For both Maximum Likelihood and Bayesian analyses, a partitioned analysis was performed with the following six partitions: ITS1+ITS2, 5.8S, LSU, ACT-exons, ACTintrons and RPB2. Maximum-likelihood (ML) analysis was performed with RAxML (Stamatakis 2006) implemented in the CIPRES Science Gateway web server (RAxML-HPC2 on XSEDE; Miller et al. 2010), 25 categories, 1000 rapid bootstrap replicates were run with the GTRGAMMA model of nucleotide evolution. Maximum likelihood bootstrap values (MLBS) equal or greater than 50% are given above each node.
Bayesian Inference (BI) analysis was performed using the Markov Chain Monte Carlo (MCMC) method with MrBayes 3.2.2 (Ronquist et al. 2012). The best-fit nucleotide substitution model for each dataset was separately determined using MrModeltest version 2.2 (Nylander 2004). GTR+I+G was selected as the best-fit model for the ITS1+ITS2, LSU, ACT (ACT-exons and ACT-introns) and RPB2 datasets and K80 for 5.8S. The MCMC analyses, with four chains starting from random tree topology, were run for 5,000,000 or 10,000,000 generations for the combined dataset or the ITS1+ITS2 dataset. Trees were sampled every 100 generations. Tracer v. 1.5.0 was used to check the effective sampling sizes (ESS) that should be above 200, the stable likelihood plateaus and burn-in value (Rambaut et al. 2013). The first 5000 samples were excluded as burn-in.

Phylogenetic analysis of combined ITS, LSU, ACT and RPB2 sequences
The combined alignment of ITS, LSU, ACT and RPB2 sequences comprised 86 taxa, including our strains, with Diaporthe eres (CBS 183.5) as the outgroup taxon. The total length of the dataset was 2037 characters including alignment gaps (1-199, 200-357, 358-518, 519-1056, 1057-1296 and 1297-2037 corresponding to ITS1, 5.8S, ITS2, LSU, ACT and RPB2, respectively). The combined dataset contained 1426 constant, 144 parsimony uninformative and 467 parsimony informative characters. The result from the partition homogeneity test (PHT) was not significant (level 95%), indicating that the individual datasets were congruent and could be combined. The combined dataset was analysed using MP, ML and Bayesian analyses. The trees generated under different optimality criteria were essentially similar in topology and did not differ sig- Etymology. refers to the host where the fungus was isolated.
Notes. Based on the multigene phylogeny, Cytospora lumnitzericola is closely related to Cytospora thailandica (Fig. 1). Although conidial sizes of both species are similar, they have significant differences in nucleotides: ITS (26 nt), ACT (22 nt), and RPB2 (53 nt) ( Table 5). The phylogeny derived from the ITS regions depicts C. lumnitzericola as an independent lineage close to C. brevispora CBS 116829 and C. eucalyptina CMW5882 (Fig. 2). In future, more collections are needed to confirm whether C. lumnitzericola can exist as a saprobe or endophyte as well as performing tests to confirm its pathogenicity. Etymology. refers to the country where the fungus was collected.
Notes. Cytospora thailandica was collected from branches of Xylocarpus moluccensis. The new species resembles some other Cytospora species, but is characterised by unior multi-loculate ascomata/conidiomata with unicellular, subcylindrical and hyaline spores in both morphs. Cytospora species associated with Xylocarpus granatum is also reported in this study as C. xylocarpi (MFLUCC 17-0251, Fig. 5). Cytospora xylocarpi is similar to C. thailandica in its conidiomata being multi-loculate and in the length of conidia in the asexual morph (C. xylocarpi: conidia 3 × 1.1 μm versus 3.8 × 1.3 μm in C. thailandica). However, C. thailandica differs from C. xylocarpi in having shorter ostiolar necks and larger asci and ascospores (Table 2). Phylogenetic analysis of our combined gene also reveals C. thailandica is closely related to C. lumnitzericola (Fig. 1), but there are nucleotide differences as mentioned in notes of C. lumnitzericola. The individual ITS1+ITS2 phylogenetic tree also indicates that C. thailandica is distinct with good support (Fig. 2). Etymology. refers to the host genus that fungus was collected.
conidia (Kohlmeyer and Kohlmeyer 1971). However, the phylogenies, generated herein, show that C. xylocarpi is distinct from C. rhizophorae (ATCC 38475), a strain from Rhizophora mangle that was identified by Kohlmeyer, the author of the species (Fig. 2). The two species also differ by 25 substitutions in ITS1+ITS2 and were collected from different hosts. Therefore, the collection in the present study is designated as a new species. Our phylogeny also indicates a close relationship to unpublished sequences from GenBank (Figs 1, 2). Given that no morphological descriptions are available for these, the similarity in the ITS1 and ITS2 sequence between our strain and the sequences from GenBank (HAB16R13, M225, A761, MUCC302) are presented in Table 3. Those strains were collected from different hosts (Table 3) and, together with our strain, show substantial variation in ITS1 and ITS2 (Table 4). More collections are needed to further study morphological and genetic variation in this group.