New species and a new record of Phylloporia from Benin

Species of the wood-decay genus Phylloporia (Hymenochaetaceae, Hymenochaetales, Basidiomycota) are widely distributed in the tropics. Phylloporia species are, however, morphologically and ecologically diverse, which makes morphology-based species identification challenging. In this study, we re-examined species of Phylloporia reported from Benin (West Africa). Using an integrative approach combining morphology, ecology, and phylogenetic analyses, we describe Phylloporia beninensis sp. nov. and report Phylloporia littoralis for the first time outside of its type locality. Phylloporia beninensis sp. nov. is characterized by its annual and imbricate basidiomata, duplex context with a black zone separating the upper context from the lower one, dimitic hyphal system, presence of cystidioles, basidia of 9–12 × 4–5 μm, and subglobose to ellipsoid basidiospores measuring 3–4.6 × 2.1–3.6 μm. Detailed descriptions with illustrations for the new species are provided. With the addition of the new species, 15 Phylloporia species are now known to occur in tropical Africa. Our discovery of a new Phylloporia species in Benin should stimulate further mycological investigations in tropical African ecosystems to discover other new polypore species. To facilitate further taxonomy studies on tropical African Phylloporia taxa, a key to the known tropical African species is provided.


Sequence alignment and species delimitation.
To place our newly generated sequences accurately in the phylogenetic tree, we aligned them in addition to 126 LSU sequences retrieved from GenBank and used by previous studies on Phylloporia 29 . Sequences were aligned using the online mode of MAFFT version 7 30 , with the algorithm FFT-NS-i as the most suitable. The resulting multiple sequences alignment was checked in Geneious 5.6.7 (https:// www. genei ous. com) 31 , where the ends rich in gaps were manually trimmed. Further, the multiple sequences alignment was inspected and some bases were manually adjusted using AliView 32 . Two model-based methods for species delimitation namely the Automated Barcode Gap Discovery (ABGD) 33 and the Poisson Tree Process (PTP) 34 were performed. The ABGD analysis detect potential barcode gap and use the identified barcode gap to sort the datasets into a hypothetical species. This analysis was performed on ABGD web interface using the Jukes-Cantor (JC69) and Kimura two-parameter (K2P). The relative gap width was set to 1.0 because if the gap is too large, the model will sort the dataset into a single species. We kept all other parameters as default. Like the ABGD method, the PTP is another species delimitation method that inferred putative species boundaries on a given phylogenetic input tree. To run the PTP analysis, we first built a single phylogenetic tree using IQ-tree 1.6.12 (http:// www. iqtree. org/) in command line mode. The resulted tree without annotations in Newick format was used as the input tree to run the PTP analysis on a web server (http:// speci es.h-its. org/ ptp/) for 500,000 generations and 25% were discarded as burn-in. To compare both species delimitation models to the phylogenetic analysis, Maximum likelihood (ML) analysis under the Ultrafast Bootstrap with 5000 replicates was performed on the dataset using IQ-tree 1.6.12 (http:// www. iqtree. org/) in command line mode with TM3 + F + I + G4 as the best substitution model selected using the command TESTONLY.
Phylogenetic analyses. For phylogenetic analyses, 73 sequences from the LSU region out of the 126 sequences previously used to inform species delineation in Phylloporia were selected and aligned with the 4 newly generated sequences in this study. In addition, 34 sequences from the ITS region including the type material of the genus were downloaded from GenBank and aligned together with the sequences newly generated in this study. Inonotus andersonii (Ellis & Everh.) Nikol. and I. hispidus (Bull.) P. Karst. were chosen as outgroup for both regions. Each region was aligned separately using the online mode of MAFFT version 7 30 , with the algorithm L-INS-i. The multiple sequences alignments were checked and concatenated in Geneious 5.6.7 (https:// www. genei ous. com) 31 .
Given the gap in terms of number of sequences between the ITS and LSU regions (36 vs. 77), the concatenated alignment was considered as a single region and the best-fit evolutionary model was estimated as GTR + I + G using IQ-tree 1.6.12 (http:// www. iqtree. org/) and the command TESTONLY. Following this substitution model, two phylogenetic tree inference methods, ML and Bayesian inference (BI) were performed. The ML was run using RAxML 8.2.10 35 under standard bootstrap at the Cipres Science Gateway V.3.3 36 . The BI was executed using MrBayes 3.2.7 in command line mode (https:// github. com/ NBISw eden/ MrBay es) 37 for five million generations until the standard deviation of split frequencies reached 0.01. Chain convergence was determined using Tracer. v1.7.1 (http:// tree. bio. ed. ac. uk/ softw are/ tracer/) and the first 25% (5000) trees was discarded as burn-in. The remaining trees were used to build the consensus tree using the Phylogenetic Tree Summarization (SumTrees) program within DendroPy 4.3.0. (https:// github. com/ jeets ukuma ran/ Dendr oPy) 38  www.nature.com/scientificreports/ was better resolved than that of BI, so the ML tree was targeted. To add the posterior probabilities (PP) of BI on the ML tree, the Phylogenetic Tree Summarization (SumTrees) program within DendroPy 4.3.0. (https:// github. com/ jeets ukuma ran/ Dendr oPy) 38 was used. Then, the bootstrap values were added to the ML best tree already having the posterior probabilities using IQ-tree 39 . The resulting tree with (PP/BS) is presented in Fig. 3 and the support values ≥ 80% of PP and ≥ 70% of BS are indicated on each node. Alignment and phylogenetic tree generated in the study are deposited in TreeBASE: http:// purl. org/ phylo/ treeb ase/ phylo ws/ study/ TB2: S27303.

Morphological examination.
Morphological descriptions were based on dried herbarium specimens.
Macro-morphological characters were described with the aid of a stereomicroscope Leica EZ4 while microstructures were described using a Leica DM500 light microscope. For the microstructures, fine sections through the basidiomata were prepared for observation using a razor blade under a stereomicroscope and mounted in distilled water and 5% aqueous solution of potassium hydroxide (KOH) mixed with 1% aqueous solution of phloxine. Melzer's reagent (to test for dextrinoid or amyloid reactions) and cotton blue (to test for cyanophilic reaction) were used and then examined at a magnification of 1000×. Leica Application Suite EZ V.3.4 software (Leica Microsystems Ltd., Switzerland) was used to capture images from the microscope. Measurements from captured images were done with the software "Makroaufmaßprogramm" from Jens Rüdigs (https:// ruedig. de/ tmp/ messp rogra mm. htm) and analyzed with the software "Smaff " version 3.2 40 .

Results
Species delimitation. The ABGD method with parameters JC69 and K2P gave identical results and partitioned the LSU dataset into 6 partitions. The first five partitions with interspecific priority divergence ranging from P = 0.001 to P = 0.0077 contained 83 groups each while the sixth partition with interspecific priority divergence P = 0.0129 contained only one group (Fig. 1). Each group within each partition represented a hypothetical species with one or several sequences (Supplementary Table S1). Given the congruence between the first five  Table S1). The PTP species delimitation estimated that the number of species in LSU dataset was between 82 and 109, with the Mean of 97 species. The PTP species delimitation was supported by the maximum likelihood solution (PTP_Mls) and the Bayesian solution (PTP_Bs). Both solutions gave two different results in terms of the number of estimated species. The PTP_Mls yielded into 82 putative species (Supplementary Table S2) while PTP_Bs gave 100 putative species (Supplementary Table S3). Although the PTP_Mls and PTP_Bs yielded different results, the newly generated sequences formed two distinct species and are grouped identically in both outcomes (Supplementary Table S2, 3). Since species delimitation with PTP_Mls and PTP_Bs gave same results for our newly generated sequences with good support values, and considering the ML tree and ABGD results, we chose the results from PTP_Mls as the most suitable for our dataset. The Fig. 2 presents the ML tree with the putative species as found with ABGD and PTP_Mls.
Phylogenetic analyses. The combined ITS-LSU alignment contained 78 sequences with 2397 characters, of which 711 were parsimony-informative, 277 singleton sites, and 1409 constant sites. Four well supported major clades namely Fomitiporella (PP = 1.00/BS = 99), Fulvifomes (PP = 1.00/BS = 98), Inonotus (PP = 1.00/ BS = 100), and Phylloporia (PP = 1.00/BS = 88) were recovered from the phylogenetic analyses inferred from the ITS-LSU (Fig. 3). Phylloporia appeared as a well-supported monophyletic clade, which split into two well-supported groups, here named A and B (Fig. 3). Group A (PP = 1.00/BS = 76) contained the sequences of the most Phylloporia species, including the generic type (P. parasitica), while group B (PP = 1.00/BS = 96) consisted of just three species of Phylloporia. The newly generated sequences nested within group A. The sequence OAB0204 clustered together with P. littoralis as it was found in the species delimitation analyses with high support (PP = 1.00/ BS = 95). Sequences OAB0107, OAB 0142, and OAB0511 formed a distinct well-supported lineage (PP = 0.97/ BS = 93) and had as a sister lineage an unidentified Phylloporia species from Kenya with high support (PP = 1.00/ BS = 95). Since the sequences OAB0107, OAB 0142, and OAB0511 grouped together and had always formed a distinct lineage in all analyses (Figs. 2, 3; Supplementary Table S1-3), we proposed here as a new species and performed a detailed anatomical-morphological description on these specimens.

Taxonomy. Phylloporia beninensis
Olou & Langer, sp. nov. Figures 4, 5, 6 Diagnosis. Phylloporia beninensis differs from other known species of Phylloporia by the combination of the following characteristics: basidiomata imbricate; pileus projecting up to 3 cm, 5 cm wide, and 1 cm thick at base, surface concentrically sulcate and zonate; cystidioles present, variable in size and shape; basidia 9-12 × 4-5 μm; basidiospores ellipsoid to subglobose, 3-4.6 × 2.1-3.6 μm.    Description. Basidiomata annual, pileate, sessile, imbricate with overlapping pilei, broadly attached or effused-reflexed (Fig. 4a-c), hard when dried, without odour or taste, projecting up to 3 cm, 5 cm wide, and www.nature.com/scientificreports/ 1 cm thick at the base. Pileus applanate to slightly convex, surface mustard and ferruginous brown in young or actively growing specimens and almost blackish in old specimens, velvety under stereomicroscope; surface concentrically sulcate and zonate; margin undulate, obtuse, yellowish when young or in actively growing specimens (Fig. 4c), concolorous with the pileus at maturity. Pore surface buff-yellow to honey, not shining or at least in the dried specimens, pore very small, 7-9 per mm, isodiametric to angular (Fig. 4d). Context two-layered, with a black line separating the upper context (tomentum) from the lower context, mustard brown, tomentum softer and lighter coloured than the lower context, tomentum up to 5 mm thick at the base and in the middle and thinner toward the margin, lower context up to 2 mm thick at the base and thinner at the margin (Fig. 4e). Tube layer concolorous with pore surface, up to 2 mm long.

Discussion
Phylogenetic analyses inferred from the LSU and ITS-LSU datasets, coupled with macro-and microscopic examinations and ecological analyses, support the recognition of P. beninensis as a new species. Phylloporia beninensis is morphologically distinguished from other Phylloporia species by its annual, sessile, pileate, and imbricate basidiomata, two-layered context with the layers separated by a black line, dimitic hyphal system, and presence of cystidioles that vary in size and shape.
We cannot yet confirm whether or not P. beninensis is saprotrophic even though it was mainly found on dead wood (Fig. 4a,b), because it is well evidenced, that the habit of a fungus to produce fruit body on dead wood does not necessarily indicate a saprotrophic lifestyle 42 . However, although the lifestyle of P. beninensis is not yet well known, the fact that it was mainly found on dead wood we can reasonably say that the latter is saprotroph. As saprotroph, P. beninensis is therefore ecologically different from P. minutipora and Phylloporia sp., which are mainly collected from living trees 17,41 . Like P. beninensis, P. rattanicola is also saprotrophic because it was collected from dead rattan 8 . However, knowing that Phylloporia species display a high level of host specificity 7,10,15,43 , and that P. rattanicola is only collected on rattan while P. beninensis is collected on hardwood, we can safely say that P. beninensis and P. rattanicola do not belong to the same morpho-ecological group as stated above.
We also reported here P. littoralis Decock & Yombiyeni on the basis of molecular and morphological analyses, constituting the first record of the species from Benin (Figs. 2, 3, 7). The Benin P. littoralis specimen fits well morphologically and genetically to the Central African type specimen (see Fig. 2, in Yombiyeni and Decock 2017). To our knowledge, this is the first time P. littoralis has been reported outside its type locality Gabon, and suggests that the species may be more widely distributed in sub-Saharan Africa.
The recognition of P. beninensis brings the number of described Phylloporia species to 62 worldwide. Among these 62 species, nine were described from tropical Africa 9,12,17,18,23 . Phylloporia are more diverse in tropical Africa in comparison with Europe, where only P. ribis (Schumach.) Ryvarden has been reported 4 to date. Considering that tropical Africa remains poorly explored for wood-decay fungi, it is likely that many more Phylloporia species remain to be found. We are also confident that new investigations of new still unexplored habitats and re-examination of herbarium specimens initially assigned to the genus Phellinus will reveal more new species www.nature.com/scientificreports/ of Phylloporia from tropical Africa. Aside the nine species described with type specimens, six other Phylloporia species have been reported from tropical Africa 2, 7 , which brings the number of regional Phylloporia species to 15. To facilitate future taxonomic studies in the genus, we provide a dichotomous key for identification of tropical African Phylloporia species.
Identification key to African Phylloporia species

Data availability
Alignment and phylogenetic tree from the combined ITS-LSU dataset generated in this study are available in TreeBASE under this link: http:// purl. org/ phylo/ treeb ase/ phylo ws/ study/ TB2: S27303. Newly generated sequences are available in GenBank and the accession numbers are given in Table 1. Alignment, phylogenetic tree, and accession numbers of newly generated sequences will be public after the paper is published. Collected specimens are available at the mycological herbaria of the University of Parakou (UNIPAR) in Benin and University of Kassel (KAS) in Germany. Following the new requirement of MycoBank, the new species will be registered in MycoBank and the registration number will be given in the taxonomy section of this paper as soon as the paper is accepted.